Risk Management
Vignesh Sivadass
enlightened-planning-text-book.pdf
StuDocu is not sponsored or endorsed by any college or university
Enlightened Planning Text book
Project Management (University of Southampton)
StuDocu is not sponsored or endorsed by any college or university
Enlightened Planning Text book
Project Management (University of Southampton)
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Enlightened Planning
Strategy, risk management and project management are often considered separately by
those applying their principles – but at their most effective, all are dependent on each other
for success. Enlightened Planning teaches this holistic perspective and demonstrates how a synthesis of these approaches yields far greater opportunities. A strategic, calculated risk, for
example, can be less inherently risky than chronic risk aversion over time.
Here, a respected specialist and teacher demonstrates how to become an ‘enlightened
planner’, one who is aware of project, strategy and risk concerns and their potential interplay.
Following the core principle of Keep It Simple Systematically, he shows how organised,
systematic thought processes can demystify the complexities of decision- making when
considering a huge variety of concerns at once.
Supported throughout with real- life cases from the author’s considerable experience with
commercial organisations, it is also supported by a website containing even more cases
and learning and teaching materials. This book is essential reading for any practitioner
specialising in risk management, project management or strategy; as well as those teachers
or participants in executive programmes.
Chris Chapman is an experienced consultant, the basis of his research as Professor of
Management Science at Southampton University. He has published widely on uncertainty,
opportunity, and risk management for projects, operations and corporate management. He
is a past President of the Operational Research Society.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Enlightened Planning Using Systematic Simplicity to Clarify Opportunity, Risk and Uncertainty for Much Better Management Decision Making
Chris Chapman
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
First published 2020 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN
and by Routledge 52 Vanderbilt Avenue, New York, NY 10017
Routledge is an imprint of the Taylor & Francis Group, an informa business
© 2020 Chris Chapman
The right of Chris Chapman to be identified as author of this work has been asserted by him in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988.
All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers.
Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe.
British Library Cataloguing- in- Publication Data A catalogue record for this book is available from the British Library
Library of Congress Cataloging- in- Publication Data A catalog record for this book has been requested
ISBN: 978- 1- 138- 35352- 7 (hbk) ISBN: 978- 0- 429- 42539- 4 (ebk)
Typeset in Galliard by Apex CoVantage, LLC
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Foreword by Stephen Ward ix Comments by other colleagues xi Preface xix Overview xxvii About the author xxxviii Acknowledgements xliii
PART 1
Foundations 1
1 Why planning is usually vital but often difficult and frequently inept 3
2 A ‘universal planning uncertainty and complexity management process’ (UP) 41
3 Low to high clarity approaches and the ‘estimation- efficiency spectrum’ 71
4 Confronting challenging complexities usually needing more clarity 132
PART 2
Employing planning tools in practice – five illustrative tales 161
5 Using a UP – an initially simple supply chain management example 164
6 Building ‘specific processes’ – a bidding process example 212
7 Adapting ‘generic processes’ – a project planning example 307
8 Corporate strategy formulation – an electricity utility example 399
9 Building well- founded trust about complex concerns – a railway safety example 476
Contents
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
viii Contents
PART 3
Further synthesising and reflecting 579
10 Immediate and longer term ‘what needs to be done’ priorities 581
11 Ongoing enhancement of strategic clarity and tactical clarity 601
References 628 Website information 635 Index 636
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
We all engage in planning. As individuals much of our planning is of a tacit, informal nature,
but for managers there is an expectation, even a requirement, that planning will be a more
formalised process, with clear rationale and objectives requiring the cooperation of various
other parties. However, this clarity can be difficult to achieve, not least because complexity,
uncertainty, and risk must be confronted, whatever the context. Appreciating and planning
appropriately for the implications of complexity, uncertainty, and risk associated with differ-
ent possible courses of action are major challenges.
Whatever the enterprise, some degree of planning is necessary to achieve desired out-
comes, and large projects require a great deal of planning. But common experience shows
that plans frequently require major modification, become thwarted, or fall short of achieving
desired outcomes, usually because of factors that were not considered sufficiently during the
process of planning.
In this book, Chris Chapman explains how we can do much better by adopting an ‘enlight-
ened planning’ approach that systematically addresses complexity, uncertainty and risk that
really matters, together with a creative search for opportunities. The concept of ‘enlightened
planning’ involves a coherent synthesis of a rich variety of concepts and analytical tools that
can be used directly in a wide range of contexts.
As a practical matter, all planning and attendant decision making must make simplifying
assumptions about reality. Unfortunately, managers, particularly hard- pressed ones, often
seem to be attracted to simplistic approaches that limit the scope of analysis and result in
important features of a planning context being ignored.
A very common example is a preference for estimates of key contextual variables that
consist of unrealistically narrow ranges of possible values, or even just single value (point)
estimates. Simplistic estimates of this kind can seriously compromise the credibility and
usefulness of plans.
Another example is the common practice of using probability- impact grids to characterise
sources of risk, with attendant simplistic assumptions that all sources of risk are derived only
from a set of events that may or may not occur and that events are assumed to be independ-
ent of one another. This practice severely restricts recognition of important aspects of uncer-
tainty that drive risk, including knock- on effects of one event on another.
How to keep things appropriately simple in any planning process requires careful
thought. Rather than uncritical use of simplistic approaches, enlightened planning calls
for intelligent and critical use of simplifying assumptions in scoping planning effort that
facilitates the exploration of key issues and related uncertainty that really matters. What are
these key issues? They include deciding what factors to take into account, the potential for
dynamic interactions among factors, understanding the goals of all relevant stakeholders,
Foreword by Stephen Ward
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
x Foreword by Stephen Ward
the ambiguous nature of some goals, potential trade- offs between goals, how far ahead to
plan, and how much detail to go into. Uncertainty and risk are associated not only with such
features of the planning context but also with the planning process itself.
A particularly valuable feature of this book is its focus on practice, employing detailed
case studies (tales), derived from Professor Chapman’s extensive consulting experience in
a wide range of planning contexts. Each of the tales in Part 2 illustrates aspects of enlight-
ened planning, describing how recognition and understanding of key issues and what to
do about them develops as protagonists progress their analysis and thinking. Each of these
tales is set in a particular context, but the approaches described are readily transferrable to
other organisation contexts. All the issues addressed are of strategic importance to opera-
tions management, project management, and corporate management. Most ought to be of
concern at board level to ensure effective governance of planning processes, and leadership
of enlightened planning.
This book is the culmination of decades of reflection about what best practice planning
and decision making under uncertainty should look like. Potential readers should on no
account be deterred by the length of this book, or of individual chapters. For, make no
mistake, this book represents a major opportunity for all managers, and their advisors, to
substantially enhance organisational performance through enlightened planning.
Stephen Ward is Emeritus Professor in Management, Southampton Business School, University of Southampton
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Matthew Leitch is a consultant, author, business school educator and risk standards committee member whose website is www.WorkingInUncertainty.co.uk
Enlightened Planning is not just a book about how to manage risk on projects or in busi- ness planning, but if you need a book on those topics then this is an excellent choice. It does
two things that are important steps forward. First, it includes a management process explic-
itly designed to incorporate a wide range of appropriate methods. This is not just a process
which boils down to a list of risks and a procedure for doing things with them. Instead, it
encourages consideration of a much wider range of techniques from operations research
modelling down to quick estimates. This kind of flexibility is a frequent aspiration for risk
management publications, but in this book, Chapman successfully achieves it. Second, the
book focuses on how to choose the right management technique at the right time to get the
best combination of clarity and economy. It explicitly tackles something most busy managers
can feel all the time, which is a drive to know and understand more, frustrated by a lack of
time and an opportunity to do so.
The book is extensively illustrated with detailed practical examples from real life, includ-
ing calculations and charts. The rich conceptual content and practical detail mean this is not
a book you can read over a weekend. But tackled patiently, a section at a time, over a period,
it will make most intelligent managers into better managers.
Paul Thornton was a founder and Managing Partner at The French Thornton Partnership. He is past President of the Management Consultancies Association and the Operational Research Society.
French Thornton was a programme management consultancy that led numerous well-
known initiatives in the financial, transport, and retail logistics sectors, and my earlier man-
agement consultancy career involved a wide range of clients. As a lifelong management
consultant, I focused on helping organisations to achieve lasting, beneficial change. Because
I came from an operational research background, my instincts were to adopt a modelling
approach to whatever problems we were confronted with. If standard models didn’t ade-
quately represent the real- life complexities of the situation, then develop and extend them
until they did was the basic idea. This approach risked getting bogged down in endless cycles
of model development and redevelopment. However, the main alternative (and the one
often favoured by macho managers) was to assert that the problem was really a simple one
and to solve that problem irrespective of how much genuine complexity was being brushed
under the carpet.
I first became aware of Chris Chapman’s work in the 1980s, and we got to know each
other in the early 1990s when he succeeded me as president of the Operational Research
Comments by other colleagues
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
xii Comments by other colleagues
Society. Chris was providing long- term support to major corporations and utilities on pro-
jects that flew under the ‘risk management’ banner, many of them concerned with huge
capital investment programmes. Drawing on these multiyear consulting engagements, and
his extensive research work, Chris has now codified significant subsequent evolution of his
approach, which is set out fully for the first time in this new ‘enlightened planning’ book.
A key aspect of this new book’s approach is that much of the real- world complexity is
addressed by thoughtfully tailoring the overall planning process to accommodate the cul-
tural stresses and strains within the enterprise. This frequently facilitates the use of very sim-
ple conceptual and operational tools that build on sophisticated underlying conceptual and
operational tools. ‘Systematic simplicity’ is an excellent description for this approach. For
example, as discussed in Chapter 6, simple but effective quantitative probabilistic analysis
can be used when assessing competitor behaviour in a bidding context based on work Chris
did with IBM UK and earlier work with BP International, but qualitative cultural issues also
need attention, like a corporate understanding of the difference between good luck and
good management, bad luck and bad management, and why people taking more risk and
knowing what they are doing can be crucial to organisational survival, explored in detail in
Chapter 3.
I also like the ‘systematic simplicity’ approaches to cultural concerns, including sound
ethics. For example, Figure 11.1 in the last chapter uses a ‘red face test’ notion Chris attrib-
utes to me, but what I particularly like is the way this simple ethical test is linked to more
general ‘frowns test’ and ‘smiles test’ variants, driven by a systematic approach to important
concerns which are not amenable to simple metrics but need simple systematic attention.
Examples used earlier in the book in the IBM based bidding example in Chapter 6 include
staff being needed if the bid is won who are already heavily committed to other crucial con-
tracts (a ‘frown’) and a bid in a new area of business the company doing the bidding wants
to develop expertise for with a view to future business (a ‘smile’). Railway safety as addressed
in Chapter 9 raises more complex ethical issues involving trade- offs among monetary costs,
avoided fatalities, and avoided injuries, and these ethical issues are crucial and current in a
wide range of important contexts.
Professor Jeffrey K. Pinto, Black Chair in Technology Management, Penn State University, USA. Creating and managing knowledge as it applies to project management has, at times,
taken on the nature of a Sisyphean labour. At an age when project- based work has grown
ever more common and projects of the broadest array – infrastructure, information systems,
new technology and services development – represent real opportunities to advance the
human condition, our understanding of how best to manage them for maximum value often
remains mired in misapplication, flawed thinking, and a variety of personal and organisa-
tional biases. In short, the dichotomy is real and it is growing: between, on one hand, the
increased need for organisational undertakings best supported by projects and project man-
agement and, on the other, the seemingly intractable challenges in advancing our knowl-
edge base sufficiently to gain best use from our efforts. The data support this contradiction:
the Project Management Institute reports that for every billion US dollars invested in pro-
jects, some 125 million USD is classified ‘at risk’, a figure that is actually growing at double
digits year- on year on year. Failure rates in information technology projects are high and
have remained depressingly stable for well over a decade. The data are clear: the need for
projects is at an all- time high while the manner in which practitioners and academics alike
deliver on this promise remains mired.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Comments by other colleagues xiii
It is against this backdrop: a time for projects that is both highly exciting in its pos-
sibilities and rather sterner in its realisation, that Professor Chris Chapman’s book is so
welcome. The underlying premise of his work (and it is a message that is born out again
and again in projects great and small alike) focuses on the nature of project front- end
planning and offers a deceptively simple message regarding what we keep getting wrong
with these processes – we continue to make the simple difficult while making the difficult
seem seductively simple. Professor Chapman has been one of the leading theorists and
original, holistic thinkers in the project management discipline for several decades now. In
Enlightened Planning, he brings his considerable talents to bear in a work that is compel- ling and powerful. It offers a new way of viewing the project planning process: one that
directly considers the ways in which our organisations, culture, and processes can interact
to get planning done the right way. Equally welcome is the manner in which Professor
Chapman illustrates these ideas, through a series of compelling case examples that show
in practice the principles he espouses. I have a strong admiration for this book. Terms
like ‘strategic clarity’, ‘enlightened planning’, and ‘systematic simplicity’ are certain to
become more than talking points; they offer the means to reorient our thinking. They
cut right to the heart of what we need to be doing to put into practice effective project
planning approaches. The framework provided also links this thinking directly to planning
for operations management and corporate strategy formation concerns which are directly
interrelated in important ways. The final result is a thought- provoking and important
work for practitioners and scholars alike.
Professor Terry Williams is Director of the Risk Institute, University of Hull, UK. As Professor and Head of the Management Science Department at Strathclyde University
and Professor and then Director of the Management School of the University of Southamp-
ton before coming to Hull, I have been aware of the work Chris was doing for many years.
I invited him to be a key speaker at a NATO Science Programme conference in Kiev in
1996, used Project Risk Management (Chapman and Ward, 1997) in my book Modelling Complex Projects (Williams, 2002), which shares his (and my) ‘practitioner who is also an academic’ perspective, and invited him to contribute a chapter to my book Project Govern- ance: Getting Investment Right (Williams and Samset, 2012). His new Enlightened Planning book provides a reflective synthesis of his earlier work plus a comprehensive set of important
new conceptual and operational tools with significant implications.
An irony of the risk management field is that it has become prone to the risk of stand-
ardised ways of considering and quantifying risk. The paper ‘Organising Risk: Discourse,
Power, and “Riskification” ’ (Hardy and Macquire, 2016) in the Academy of Management Review has shown how organisations and ‘experts’ in risk have developed a dominant dis- course which limits the way we think about risk. What the field urgently needs is thinking
that takes us to the basics of what risk and uncertainty are, and looks at them in a fresh way.
This book is indeed such thinking, introducing important new fundamental concepts such
as ‘clarity efficiency’ and the ‘estimation- efficiency spectrum’ (although keeping some well-
used ideas from previous books such as the ‘seven Ws’). Pleasingly, it aims to provide good
estimates rather than the simple assumption of bias and the application of standardised con-
tingency factors, often done following Flyvbjerg’s work. Having edited a couple of books
about planning projects with scant information, this book would have given us a really use-
ful structure on which to consider the ideas with which we had to grapple, and I hope its
ideas are taken up by practitioners, academics, and the various bodies of knowledge.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
xiv Comments by other colleagues
Stephen Cresswell practices as Into Risk Ltd, London UK, www.intorisk.com. I have been an independent consultant for a decade, building on an earlier decade of
experience, initially in business development in the IT and Telecoms industry, then as a pro-
ject management consultant with Bombardier Transportation and the Sweett Group. My
growing interest in risk and uncertainty led me to Chris’s publications in the early 2000s,
and I took part in his 2006 International Project Management Association ‘Managing Pro-
ject Risk & Uncertainty in New Ways’ Advanced Training Course in Copenhagen. This
course and subsequent reading of his publications have been a key influence in my develop-
ment as a ‘reflective practitioner’ – with many of the methodologies and approaches being
applied, with success and benefit, in my client assignments. Particularly important was an
appreciation of how to get beyond seeing ‘risk’ in terms of independent events, treat inter-
dependencies effectively, and meet the challenges involved in persuading clients’ personnel
conditioned by simplistic approaches to change their approaches.
The ‘enlightened planning’ perspective explored in this new book involves deep thinking
about strategy, uncertainty, complexity, and implementation, with considerable attention to
perspective and some philosophical aspects. However, this new book is also rich in simple
tools and practical concepts and mantras that help with the ‘how to do it’ and ‘craftsmanship’
associated with planning and analysis in new ways. I fully anticipate using all aspects when
faced with challenging problems. My favourite new concept is ‘closure with completeness’.
The basic underlying ideas have always been there in the approaches Chris has advocated,
but they are now brought together, made explicit, and named. Closure with completeness
gives a very concise plain English label and rationale for the inclusion of items such as an
‘allowance for unidentified risk’ in a project cost estimate. It also naturally prompts people
to question whether the analysis addresses everything relevant.
This new book builds on a consolidation of an extensive research and consulting career,
drawing on both successful applications of new ideas and some lessons learned the hard way.
Reading this book from beginning to end was a serious challenge, but I expect to get large
return on the time invested.
Mike Annon, PMP, is Owner of I&C Engineering Associates, Waterford CT, USA. My 45 years of experience with more than 50 nuclear energy and fossil- fuelled power
plant facilities in managerial and management consulting roles has convinced me that most
organisations believe they know how to plan. However, too frequently they need to ‘rescue’
their plans at considerable cost, with serious associated delays, because their initial planning
efforts were incomplete. Since the 1980s I have published, led workshops and managed
projects with a focus on the processes, other tools, and teamwork needed to ‘get started on
the right foot’ and to rescue projects which failed to do so.
After I attended a 2009 professional training course in Chicago led by Chris based on the
second edition of his book Project Risk Management (Chapman and Ward, 2003), I started to embed many of his ideas about new ways of looking at ‘uncertainty’, ‘risk’ and ‘opportu-
nity’ in my work with clients. Contributing feedback comments on his book How to Manage Project Opportunity and Risk (Chapman and Ward, 2011) helped me to extend and update these ideas. His new Enlightened Planning book has a wide range of further new ideas which will be incorporated into my planning efforts with future clients. I particularly like his new
book’s approach to visualising what I would term ‘beginning with the end in mind’, explicitly
linking the assumed project lifecycle framework plus the ‘seven Ws’ framework and a ‘goals–
plans framework’ to the project planning phase structure framework which integrates these
four key frameworks. It helps with the front end of project management, and with integrating
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Comments by other colleagues xv
operations management and corporate management team concerns, to facilitate delivering
what both the project owners and the project’s ultimate ‘customer’ actually want by asking
them the right questions at the right time and engaging in the right kind of dialogue.
Rodney Turner is now retired. Most recently he was Professor of Project Management at SKEMA Business School, in Lille, France, where he was Scientific Director for the PhD in Project and Program Management. Rodney is Vice President, Honorary Fellow and former Chairman of the UK’s Association for Project Management, and Honorary Fellow and former President and Chairman of the International Project Management Association.
Consider a trio of quotations:
No battle plan ever survives first contact with the enemy. – Field Marshal Helmuth von Moltke
In preparing for battle I have always found that plans are useless, but planning is indispensable.
– General Dwight D. Eisenhower
The perfect is the enemy of the good. By this I mean that a good plan violently executed now is better than a perfect plan next week. War is a very simple thing, and the determining char- acteristics are self- confidence, speed, and audacity. None of these things can ever be perfect, but they can be good.
– General George S. Patton
The quotation by Field Marshal von Moltke can suggest there is no point planning
because the plans will be wrong. But President Eisenhower, while agreeing that the bat-
tle will not evolve as the plans envisage, suggests that the process of planning is essential
because it creates a strategy for the battle, and though the battle will not evolve as the plan
envisages, having done the planning, we can understand what the likely scenarios are and
respond to the scenarios that occur. General Patton takes a slightly different approach. He
starts by quoting Voltaire and says we should not aim for the perfect plan because we will
already have lost the battle. But we should aim for a good plan, one of the defining charac-
teristics of what Chris Chapman refers to as ‘systematic simplicity’.
Henry Simon, in his concept of bounded rationality, agrees with these sentiments with
ideas also supportive of a systematic- simplicity approach. We can never make a perfect deci-
sion, because we do not have all the information we need, we cannot perfectly analyse all
the information we do have, and most important of all, we cannot foretell the future, so we
do not know precisely how things will evolve. Therefore, we need to make good decisions,
ones that satisfice, and not strive for perfect decisions.
Chris Chapman, in this unique book, explores how we can plan effectively in this uncertain
environment. In Chapter 2 he introduces a universal planning and complexity management
process that outlines how we can be better able to respond as plans unfold. This process is
based on ‘systematic simplicity’ with the aim of providing good plans based on a sound inter-
pretation of the data plus wider possibilities. In Chapter 3 he introduces a range of approaches
to uncertainty using this process – the plan will not evolve precisely as envisaged, and we can’t
predict the future, but we can make forecasts within sensible ranges and plan effectively for
likely scenarios. Chapters 4 to 7 further explain the use of these systematic simplicity ideas
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
xvi Comments by other colleagues
in project management areas where Chris has an established international reputation and
related operations management areas. Chapter 8 addresses strategy formation and corporate
planning, and Chapter 9 expands further when low probability but very high impact scenarios
may be involved when planning for the likely and unlikely scenarios, understanding the pos-
sible range of outcomes, and developing robust plans that are appropriately prudent.
This book will be invaluable to anyone involved in strategic planning or corporate deci-
sion making as well as those interested in project planning.
Martin Hopkinson is a project risk professional and author based in Winchester, UK. With the advent of computerised tools, our business and project planning processes have
evolved to demand ever increasing levels of breakdown and detail. This book identifies
why this approach has become less enlightened than we might think. For example, projects
often maintain schedules with thousands of activities, employing a team of planners to keep
abreast of the myriad of changes as they occur. Chris Chapman describes how these projects
could improve their estimates, make better decisions and foster a progressive planning cul-
ture by limiting the number of activities to 75 or fewer. His approach involves understand-
ing activities, interdependencies and the implications of uncertainty in greater depth. It is
underpinned by a welcomed clarity about the assumptions that we make when planning,
often without noticing.
The book is illustrated with practical examples drawn from the author’s long experience
of working with businesses and government departments in wide range of different indus-
tries and countries. If you deploy only some of the tools and techniques that are described,
it is difficult to see how your planning process cannot become more enlightened.
Jesper Schreiner, Managing Director, Danish Project Management Association. I attended the Copenhagen IPMA (International Project Management Association)
advanced training programme on Project Risk Management provided by Chris in 2012,
subsequently used his ideas as a practitioner and as a teaching consultant, and contributed a
two hour session to his 2017 IPMA programme as a Visiting Speaker, discussing my experi-
ence putting his ideas into practice with clients, so I was familiar with his overall approach
when reading his new ‘enlightened planning’ book.
What I particularly like about this new book is the way Chapter 1 clarifies key basic
concepts like the relationship between opportunity, risk, uncertainty and underlying com-
plexity, and Chapter 3 then clarifies the relationship between all the components of his
‘opportunity efficiency’ concept, using practical examples based on his work with BP, IBM
UK, the UK Ministry of Defense and other clients.
I also like the Chapter 7 ideas which are new to me – in particular, breaking down the
current practice silos within project management among risk management, estimating and
other aspects of planning.
I find the new Enlightened Planning book – and the embedded mindset of systematic simplicity – a very useful contemporary contribution to a better understanding of the fun-
damental complexity often encountered in the handling and the clarifying of risk and oppor-
tunities for better management decision making.
Dr Dale F Cooper, Director, Broadleaf Capital International, www.Broadleaf.com.au Chris Chapman has written an important but challenging book. It is important because
it addresses matters central to most organisations: how to make important decisions when
confronted by significant uncertainty. I shall return to the challenges later.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Comments by other colleagues xvii
This book describes two journeys. The first is the one we readers are invited to join, an
intellectual exploration as concepts are developed from relatively simple matters through to
ideas that seem deceptively simple at first but are embedded in a subtle matrix of nuance that
requires profound understanding and interpretation. The second is Chris Chapman’s own
journey along roughly the same route but with many of the bumps and wrinkles smoothed
out to make the lives of his readers easier. This second journey, interwoven through the first,
provides the justification for the steps along the way. It explains the practical circumstances
in which the main concepts were developed, with case studies from some of Chris’s large
clients emphasising the significant practical value and the substantial benefits that can be
obtained for modest investments of time and effort.
Here I must declare an interest. I joined Chris at the University of Southampton in 1978.
With a background in operational research and mathematical modelling, an understanding
of psychology and a fascination with decision- making, I was drawn quickly into risk man-
agement with Chris. Our first work together was with Acres, examining the reliability of
an LNG facility proposed for the high Arctic, using software Chris had developed with BP
and adapted by us for the specific reliability context, combined with semi- Markov analysis
that seemed innovative at the time. We went on to work together on other large projects:
hydroelectric developments in Alaska and northern Alberta, upstream oil and gas off New-
foundland and oil and gas pipeline transport in Alaska, some of which are described in this
book. Although our paths have diverged geographically since then, and we often use slightly
different words to describe similar things, the approaches I learned from Chris, and those
we developed together, are still central to my own international risk management practice.
In particular, my risk management work has always had a strong focus on practical value,
on how risk management can be used to support better decisions, with a clear recognition
that analysis by itself is not sufficient. This is echoed strongly in this book: uncertainty must
be analysed but only so far as is necessary to add value and make a sound decision, one that
can be explained and justified clearly to stakeholders. Enlightened planning, as described by
Chris in this book, provides a window into how this might be achieved.
Some of the core concepts in this book that I still use regularly (albeit sometimes using
different words) are risk efficiency and opportunity efficiency, diagrams like histogram and
activity trees that explore and explain important uncertainties and their interrelationships,
graphs that demonstrate key sensitivities and their practical implications, and illustrations
of the differences between options so decision makers can evaluate outcomes outside the
constraints of simplistic one- dimensional metrics.
Another important concept that resonates strongly with me and my colleagues at Broad-
leaf is that developing an understanding of the structure of uncertainty is an unequivocally
necessary precursor to quantification. We have seen far too many examples of quantitative
models where understanding was clearly lacking and the outcomes were at best misleading,
often technically incorrect and at worst fatally flawed.
A core concept that Chris develops is clarity efficiency. This reflects the notion that there
should be a balance between the amount of effort that is devoted to exploring important
decisions (with the context and uncertainty that surrounds them) and the understanding
that is generated for those who must make those decisions. Making such trade- offs is a criti-
cal part of enlightened planning, just as it is of risk management as we practise it.
This brings me back to the challenging aspect of this book – you need to read it carefully
and with close attention to detail to form the necessary understanding and to get the most
from it. There is no ‘magic formula’ that you can extract and apply in a few minutes. The
answer does not leap off the page but must be absorbed as concepts are developed along
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
xviii Comments by other colleagues
the path described here. You must follow the path, without shortcuts, to gain the enlighten-
ment that Chris offers. Your understanding will almost certainly be different in detail from
Chris’s or mine, but the effort you apply will be well worthwhile.
Michael Pidd is Professor Emeritus of Management Science in the Lancaster University Management School. He is Past President of the Operational Research Society, Fellow of the British Academy of Management, and Fellow of the Academy of Social Sciences.
Systematic simplicity, via enlightened planning, is the main theme of this welcome book
from Chris Chapman. Why is it welcome? Stories of large projects that exceed their initial
cost estimates or fail altogether are easy to find in the media. It’s easy to criticise, but much
harder to show all relevant parties how things could be done better. Introducing significant
change in an organisation is never easy. Why is it never easy? Because any major corporate
change requires a response that links and integrates operations management, project man-
agement and corporate management. Successfully managing such changes is particularly
difficult.
Chris addresses this complex challenge using a very broadly grounded ‘enlightened plan-
ning’ approach based on many years as a professor and international consultant. The scope
of the material covered, the broad intended audience, and the demonstration of important
nuances involved in practice are discussed using case studies, which explore qualitative as
well as qualitative concerns.
As Chris Chapman puts it, ‘there are no silver bullets, but some approaches are much bet-
ter than others’. Operational Research and Management Science are sometimes defined as
‘the science of better’. This fits well the book’s advocacy of a ‘systematic simplicity’ approach
based on rigorous analysis and practical insights.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
‘Enlightened planning’ provides an operational basis for an effective and efficient approach
to developing a shared understanding of all relevant uncertainty and underlying complex-
ity in order to grasp the key opportunities and avoid inappropriate risk when addressing all
aspects of an organisation’s management decision making in a holistic manner. Much better management decision making than most current ‘good practice’ is the overall goal, a step-
change improvement. Even if ‘very good practice’ is the current norm, demonstrating that
this is the case is a planned by- product of an enlightened planning approach and a good
reason for organisations formally embracing a variant of the approach explored in this book.
Enlightened planning in an ‘operations management’ context may include bottom- up
corporate strategic planning driven by operations management concerns. Enlightened plan-
ning in a ‘project management’ context may include bottom- up corporate strategic plan-
ning driven by project management and interdependent operations management concerns.
Enlightened planning in a ‘corporate management’ context should include top- down cor-
porate strategic planning, its integration with bottom- up corporate strategic planning, cor-
porate governance, and planning for all other corporate tasks which are not seen as part of
operations management or project management.
In all contexts using an enlightened planning approach implies some uncertainty usually
needs immediate attention in strategic terms, and ongoing attention to strategy and tac-
tics as operations, project and corporate plans evolve will raise new management decision-
making concerns. Interdependencies between operations, project and corporate planning
are often important. The underlying goals of all relevant parties are always central, and the
ambiguous nature of some goals plus the need to consider important trade- offs between
goals which may be complex can be crucial issues.
The focus of this book is ‘commercial organisations’ broadly defined, but both the scope
of the key tools and the origins of many of the underlying concepts are much wider.
A paper by Warren Black (Black, 2017), with the title Originals: How Non- conformists Will Ultimately Disrupt the World of Risk Management, inspired by Adam Grant’s 2016 New York Times best- selling book Originals, How Non- conformists Move the World, argues that my work with Stephen Ward on ‘uncertainty management’ makes us ‘risk originals’ who will
disrupt the world of ‘risk management’. Black’s paper places us in the same category as the
behavioural economist Daniel Kahneman, a Nobel Prize winner and the best- selling author
of Thinking, Fast and Slow (Kahneman, 2012). Black’s flattering assessment of my work with Stephen Ward involves a focus on features which we would not emphasise as much as
alternative features because our agenda is much wider than mainstream ‘risk management’,
and any comparison of this kind cannot be taken too far. However, there are features of our
‘enlightened planning’ approach to uncertainty management which have links to Black’s
Preface
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
xx Preface
paper worth exploring briefly in this Preface and the following Overview as part of the pro-
cess of explaining what this book is about and how its approach is different, with a view to
tempting you to read it whatever your background, perspective, and concerns.
Subjective biases are central to Kahneman’s work, a psychologist by background who is
widely viewed as one of the leading founders of behavioural economics. The acronym ‘KISS’
is an illustration of a common bias towards ideas which are ‘simplistic’ in a pejorative sense
when it is associated with the common interpretation ‘keep it simple, stupid’. This com-
mon interpretation of KISS implies that how to keep it simple is obvious, even if you are
stupid, requiring very little systematic and intelligent thought, demonstrably not the case.
Overcoming this common bias towards simplistic approaches to all an organisation’s man-
agement decision making is a central goal of this book. A very different reinterpretation of
KISS, ‘keep it simple systematically’, is a core mantra of the enlightened planning approach.
The key aspirations are facilitating opportunities to simplify in the ‘right way’ while avoid-
ing risks associated with simplifying in the ‘wrong way’, using systematic processes which
are well founded and make good use of intelligence and creativity to deliver effective and
efficient management decision making. Associated goals are seeking approximately correct
answers to the right questions in the right timeframe, which all those involved can trust,
based upon appropriately shared clarity about the rationale. Associated concerns are avoid-
ing precise answers to the wrong questions, answers that take too long, or answers which
generate a lack of trust. Seeking simplicity in this enlightened manner, avoiding simplistic
approaches which simplify in the wrong way, is a basic characteristic of all the approaches
discussed in this book.
Addressing all relevant planning concerns from the perspective of this enlightened plan-
ning approach is challenging. One reason is the framing assumptions must be as broad as
possible, by design, with clarity about the implications of all working assumptions. There are no ‘silver bullets’ – and no ‘golden’ or ‘magic’ variants. However, some conceptual
frameworks plus closely coupled operational toolsets are much better than others, and the implications of using the best available approaches drawing on a synthesis of ‘best prac-
tice’ from all relevant management practice and underlying useful theory can be ‘game-
changing’, not just a marginal improvement.
Successful enlightened planning depends on the capability and culture of those involved.
At the heart of success is a nuanced and well- founded corporate understanding of ‘what
needs to be done’ in all relevant contexts – ‘strategic clarity’. Organisations that have stra-
tegic clarity and can deploy dependable teams that have the tactical clarity to know which
tools to use and how to use them with an effective understanding of the interdependencies
between operations, project and corporate management decisions enjoy an important cor-
porate capability. In a competitive context this capability is a crucial competitive advantage.
The gist of what this book is about is enhancing your personal understanding of ‘what
needs to be done’ for organisations to achieve strategic clarity. However, to understand what
needs to be done, some aspects of ‘how to do it’ require an overview understanding. The
central core of this book is using illustrations drawn from practice to illustrate what needs to
be done plus the requisite overview understanding of how to do it in selected example con- texts. Strategic clarity plus the requisite tactical clarity in these example contexts will provide
you with a basis for facilitating ongoing development of operational forms of strategic and
tactical clarity for all the contexts of interest to you.
A key foundation level tool this book explores is clarity about the most general set of
framing assumptions and habitual working assumptions to employ to achieve clarity about
the best full set of working assumptions for any given context. You can integrate the
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Preface xxi
understanding this provides with your current understanding and build on it in the long-
term. But an immediate pay- off also provided is a closely coupled toolset of conceptual and
operational tools which can be used directly and immediately.
‘Systematic simplicity’ is the term adopted by this book to characterise all variants of an
enlightened planning approach which involve different people and different organisations
adopting their own particular versions of enlightened planning conceptual and operational
tools to best suit their context. The only constraint is a comparable set of shared basic
foundation framing assumptions which ensure that effective and efficient communication
is feasible. Terminology is an important component of all framing assumption sets, because
the way basic concepts and operational tools are perceived depends upon communication.
Enlightened planning is my current understanding of what systematic simplicity ought
to involve, described using terminology which provides a basis for communication which is
as simple and unambiguous but usefully nuanced as I could make it. The term ‘systematic
simplicity’ will be used sparingly in this book, and it can be viewed as a complex composite
concept which is not formally defined until the end of Chapter 1. However, a key charac-
teristic of enlightened planning is the form of adaptability highlighted by the basic nature
and purpose of the systematic simplicity concept as just explained. That is the rationale for
using the term ‘systematic simplicity’ in the subtitle of this book and explaining its basic role
in this Preface.
Target readers at an overview level share several important common attributes, but the
key one is most of them already are, or they would like to become, effective ‘reflective prac-
titioners’ who are interested in systematic approaches to keeping planning and associated
management decision making simple, and even the exceptions would like more clarity about
what being an effective reflective practitioner and using systematic approaches to keeping
it simple ought to imply. Management decision making which is effective and efficient in a
‘best practice’ sense depends on effective reflective practitioners, and this book is about best
practice in a sense experienced ‘reflective practitioners’ should be comfortable with.
A closely coupled key characteristic of the target audience is a broad range of backgrounds
in terms of education and experience – this book makes no specialist knowledge assump-
tions, and it makes minimal general background knowledge assumptions. This was deemed
essential because providing an understanding of planning which is as broadly accessible as
possible is an essential feature of an enlightened planning approach.
Background knowledge in specialist areas is not required, and general background
assumptions are minimal, but what is required is a significant interest in sustained explora-
tion of new perspectives, toolsets and mindsets, plus a tolerance for some very carefully
pruned but still extensive how to do it material which needs broad understanding in order
to appreciate what needs to be done with strategic clarity.
This book assumes that some target readers will have no mathematical or statistical back-
ground beyond their secondary education and a limited interest in mathematically based
concepts, other target readers will be highly mathematically inclined and have a very strong
formal training in mathematics and statistics, but most will occupy the middle ground. All
target readers need an overview understanding of some mathematical and statistical con-
cepts in what needs to be done terms. This common overview understanding is provided
using two frameworks via strategies introduced in Part 1 using an approach all target readers
should be comfortable with and find useful.
A central challenge which writing this book had to confront directly was avoiding unhelp-
ful detail while systematically clarifying key how to do it concepts in sufficient depth for
all target readers. As part of meeting this challenge, for some purposes target readers were
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
xxii Preface
characterised in four groups: directors, experts, students and other target readers. The pur-
pose of this decomposition was keeping it simple when considering the key implications of
different perspectives and concerns. The next few pages provide initial advice for each spe-
cific target reader group. The focus is directly relevant advice for that group which should
also be of interest to all other target reader groups. You may be in more than one group.
Directors means all those in executive and non- executive director roles in private sector
commercial organisations plus their equivalents in public sector organisations. ‘Commercial
organisations’ in this book means ‘providing services or products to customers at a price’.
This involves a very broad interpretation of ‘commercial’ if we explicitly assume that a ser-
vice receiver need not be the party paying for the service. Examples used extensively in this
book include international oil and gas energy– sector companies, an international computer
systems company, a national railway company, a provincial electricity utility, a regional water
and sewage utility, and a family- owned company manufacturing lawnmowers. But the rel-
evant spectrum is much broader, with debatable boundaries.
If you associate yourself with directors of commercial organisations in this broad sense,
and your experience base includes a reasonably broad exposure to disciplines addressing the
sciences, arts and crafts of planning and management more generally, you will already have
your own well- developed variant of the ‘enlightened planning’ frameworks developed in
this book. In some respects, your version may be significantly more enlightened than mine,
as well as being wider and deeper in the areas of relevance to you. However, this book’s
approach will probably challenge some reasonably fundamental aspects of your current
views in a useful manner. Meeting this challenge should help you to enhance your current
‘toolset’ of concepts and operational techniques for planning.
A key challenge associated with maintaining your interest is persuading you to tolerate
how to do it detail in areas which may not be of direct interest to you in order to clarify what
needs to be done. By the end of Chapter 1 you should be starting to see where this book
is taking you and why the journey may be useful, but ongoing patience is needed. The rest
of Part 1 provides much more clarity. Part 2 tales in Chapters 7, 8 and 9 are all set at board
level for major organisations, but Chapter 6 is at marketing manager level for regional sales
by an international organisation, Chapter 5 begins with what appears to be a very tactical
inventory management issue for a small family- owned manufacturing organisation, and Part
1 needs reading first. I am very aware of good practical reasons why most board level man-
agers are reluctant to read long books, but I hope you will be persuaded by the rest of this
Preface, the following Overview, and the earlier comments by my colleagues, to treat this
book as an exception.
To develop the holistic what needs to be done framework and the associated strategic
clarity which is central to my version of enlightened planning, this book uses an approach
involving reasonably close attention to selected aspects of relevant detail which is layered,
section by section, and chapter by chapter. In some areas the approaches advocated depart
from conventional frameworks in basic ways for reasons which need understanding by any-
one with your level of corporate decision- making and governance responsibilities. However,
the reasons are not easily explained in advance, and clarity about the implications requires
how to do it understanding at an overview level in some key areas.
To begin to encourage you to read further, consider one illustrative example. Risk man-
agement of many different kinds has been receiving growing attention in most organisations
over many decades. However, there are many popular approaches to risk management which
do not address all relevant risk, fail to integrate the management of risk with opportunities,
and do not link either to all relevant underlying objectives, uncertainty and complexity in
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Preface xxiii
an operationally effective manner. These shortcomings mean they are fundamentally flawed
and require radical reframing. Some key risk management concepts and operational tools
promoted by many ‘experts’ need to be scrapped. You need to be aware of the key issues,
the nature of associated controversy, and what needs to be done about it, with sufficient
clarity to deal with experts who disagree. Some of these experts may report to you, and some
may serve on the same board or report to other directors. In some cases, widely advocated
approaches are doing more harm than good. In some cases, counterproductive approaches
are being promoted by professional organisations which need to be directly challenged in
an open and public manner which, so far, they have escaped or simply failed to respond to.
Chapter 7 addresses why in the context of ‘project risk management’ in terms all board
members need to understand. Preceding chapters prepare some of the ground. Chapter 8
generalises to address some very fundamental board level ‘enterprise risk management’
(ERM) issues of general concern which are frequently not part of an ERM perspective.
Chapter 9 uses a broadly defined ERM perspective to consider approaches to ‘safety and
security management’ which can cope with very low probability but very high- impact inci-
dents, as well as a full range of less serious variants, considering an operations management
context with crucial board level corporate strategic planning implications. There are further
implications for regulators and governments, and related implications in a wide range of
different contexts.
To operate effectively at a director level in governance terms you need an overall under-
standing of how to do it for opportunity, risk and uncertainty management issues which
risk management experts in your organisation applying common practice may have seriously
misjudged. If experts have misjudged important risk management issues in your organisa-
tion, you need to be aware of what needs to be done about it, even if ‘getting it done’ is not
your responsibility.
There are further organisation- wide how to do it issues which you need to understand
at an overview level for strategic clarity about what needs to be done. One example is an
answer to the questions, ‘What discount rate is appropriate for capital investment decisions,
and why?’ explained at a level everybody who should be interested in the questions can
understand and trust. You need to understand why an ‘opportunity cost’ or a ‘risk premium’
component of the discount rate is often a driver of myopic thinking, leading to a focus
on high risk ‘quick- buck’ developments, and the neglect of ‘sure- thing’ opportunities, the
antithesis of what is needed. You may not even need to know ‘who to ask who knows how
to do it’, because that may be the responsibility of another director who will respond to your
concerns using his or her access to relevant expertise. However, you need to understand
important impacts of inappropriate discount rates in your areas of direct responsibility, as
well as the key concepts associated with what needs to be done by others, in order to play
the effective overall corporate decision- making and governance roles your position requires.
Throughout this book the treatment of how to do it issues emphasises strategic clarity
about what needs to be done using conceptual and operational tools which make framing
and working assumptions as easy to clarify and test as possible. The development of this
toolset was partly driven by a need to use and explain new decision- making tools at board
level when the common practice alternatives had proved unsatisfactory.
You and your fellow directors are key target readers. Without your support everyone
else within your organisation is likely to experience ongoing problems trying to implement
key features of an enlightened planning approach, unless your organisation experiences an
obvious crisis and those reporting to you are able to ‘lead from the middle’ or replace you.
Effective leadership from the top and from the front is always desirable, often crucial.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
xxiv Preface
If you want to apply some of the enlightened planning ideas developed in this book in
your organisation, you may need early support from at least one other board level col-
league. You may want internal or external short courses to ensure that everyone involved in
any necessary corporate change processes understands how to move in the same direction.
You may need consultancy support to help make some features operational. A crucial issue
before proceeding too far will be ensuring that your organisation has a collaborative team
which covers all relevant areas of expertise with the collective skill set and mindset to lead
the required change management. This team will have to collectively understand how to get
it done in practice as well as what needs to be done and how to do it in principle, including
understanding how to ‘learn by experience’ as progress is made. Initial clarity needs to be ‘fit
for the purpose’, not fully enlightened in terms of long- term aspirational goals.
Experts include directors who are experts, but the focus of the next page or so is those
not yet at board level who are experienced managers with line management responsibilities
in areas like marketing, supply chain management, production management, project man-
agement, safety and security management, corporate strategic management and risk man-
agement, plus experienced supporting staff with expertise in any of these areas. This includes
consultants, who may be external or internal. You will have your own well- developed views
about a variant of my enlightened planning approach which may be significantly better than
mine in some areas, as well as being broader and deeper in areas of particular interest to you.
However, this book can probably challenge some of your core beliefs in a useful manner, to
help you enhance your current toolset and mindset for planning, even if the primary focus
of your concerns is not addressed in a direct manner in this book.
Within your areas of expertise the nature of your current planning frameworks may be
much richer than most of those with director level concerns. But outside your areas of
expertise your planning skills may be much less developed than those with director level
concerns. Furthermore, both your perspective and your concerns may be very different. One key difference is that you will find the treatment of your area of expertise lacks some
of the how to do it detail you need to ‘get it done’. You will need further reading and discus-
sions with colleagues to put some of the principles and operational tools developed in this
book into practice in your areas of expertise. Chapter 11 addresses some of these how to do
it concerns on an expert area of focus basis but only at a level of detail which is of possible
relevance to experts in other areas and other target readers.
The second key difference is that you will encounter many areas of application which
may seem well beyond the scope of your current concerns, exacerbating the lack of detail
you would like in your areas of focus, and you will lack the overall corporate governance
concerns of directors to motivate your interest in these areas. However, you are encouraged
to view this potential problem as an opportunity in several important senses. You may be
seriously surprised by the usefulness of the widely applicable holistic corporate toolset of
‘clarity efficient’ approaches and operational tools which involve very little effort and cost in
a very wide range of appropriate contexts, including your own. You may also be surprised by
the impact on your mindset of an enhanced understanding of the importance of all experts
having a shared toolset, plus your potential role in collaborative approaches working across
traditional planning silos with other experts in different areas with different professional
perspectives.
Consider one example concern which is relevant to all an organisation’s experts, plus
all directors, students and other target readers – the need for a corporate- wide shared
understanding of the nature of uncertainty, including the role of subjective probabilities
and underlying complexity. Understanding uncertainty is a much broader concern than
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Preface xxv
understanding risk. For example, the creation and use of all estimates of cost, duration,
or other performance measures are directly dependent upon understanding uncertainty.
Understanding uncertainty is directly relevant to all experts, and it is also directly relevant
to everyone else they interact with. All experts who do not have an effective and well-
founded practical understanding of uncertainty need to be challenged by everyone else who
does understand uncertainty appropriately, and nobody can afford to fail to understand what needs to be done to provide unbiased estimates with an appropriate level of clarity.
Furthermore, the low to high clarity range of approaches explored in ‘estimation- efficiency
spectrum’ terms in Chapter 3 can be generalised to consider a low to high clarity range of
both processes and embedded models relevant to any area of expertise, as demonstrated in
Part 2. The overall treatment of these concerns in Parts 1, 2, and 3 should provide you with
an effective holistic framework for addressing your particular interests and areas of expertise
using the systematic approach to simplicity demonstrated by the examples.
A central assumption underlying this book’s approach is that you are prepared to work
effectively with other experts plus other key parties in a collaborative manner, whatever
your area of focus as an expert. To do this you and all those you communicate with need
compatible overall conceptual frameworks and a number of common very flexible and port-
able operational tools. Jointly you can use these shared approaches and tools to move your
organisation towards a more coherent enlightened planning perspective. Most organisations
need experts who strongly promote coordination and collaboration which involves ‘big
teams’ at several levels. These big teams should not be constrained by traditional profes-
sional perspectives, traditional corporate silos, or corporate boundaries. Holistic changes
which cannot be limited to your current spheres of interest may be big opportunities for you
and your organisation. If you are interested in the nature of the integration and interfaces
needing attention, and key aspects of what is sometimes characterised as ‘the big picture’,
you will find the challenges are significant, but successfully meeting them is a very reward-
ing experience.
You are a key target reader, and you are perhaps the most likely point of entry for enlight-
ened planning approaches in your organisation. Even if the director level managers you
report to are highly supportive of an enlightened approach to planning, perhaps because
you have initiated and then encouraged the development of your interests jointly, to get
things done you will have to ‘lead from the middle’. A management team which is led from
the top, the middle, and the bottom as and when appropriate, with everyone moving in the
same direction, is usually the best way forward.
Students, as a target reader group, includes students on university courses and partici-
pants on professional courses, but it also includes all those on an independent learning path
of their own design who see themselves as ‘a student of systematic approaches to planning
and associated management decision making’. All these students are key target readers. For
any student, it is important to appreciate that this is not a textbook in the usual sense. How-
ever, it can serve as a core text, and it can also complement conventional textbooks. This
book can help you to construct an enlightened planning framework of your own design,
to build on for the rest of your career. If your exposure to the sciences, arts, and crafts of
planning and management more generally is still in a formative stage, a crucial early acqui-
sition is a framework to help you integrate your understanding as you test the validity of
conflicting views in the literature and other inconsistent advice against actual outcomes.
A coherent framework for acquiring practical experience and testing your accumulating
expertise is essential for any effective reflective practitioner, along with a habit of critical
reflection on what has been learned so far via a synthesis of different experiences. You need
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
xxvi Preface
an ongoing interest in addressing the gaps in your understanding which might be filled, with
priorities dependent upon current and longer- term responsibilities and interests.
Avoid skimming material in a way which leads to a very superficial understanding of issues
that matter in terms of your acquisition of strategic clarity, and do not try dipping into top-
ics out of sequence. Conventional textbooks are often designed to facilitate this approach,
but the likelihood of misunderstanding this book is very high if this approach is taken, even
by experts or directors. The reason is the layered approach to accumulating understanding
because of the interdependencies involved, a key characteristic of this book which you and
all other readers need to appreciate to use it effectively.
Assume you will need to read more and work with experts before this book’s framework
becomes a fully functioning toolset. In an ideal world you would gain experience working
with enlightened experts in an organisation led by enlightened directors, but this is not an
ideal world. You will have to learn to accommodate the implications as best you can.
You are the experts and directors of the future, if you are not already one or both, and
experts and directors are encouraged to see themselves as ‘lifelong students’, if this is not
already the case. ‘Reflective managers’ who are successful directors and experts are usually
lifelong students.
Other target readers as a group includes all those who are not comfortable within at
least one of the director, expert and student characterisations just considered, plus anyone
else who may be interested in how planning might be enlightened or unenlightened. For
example, you might be particularly interested in some of the implications for regulation and
associated government policy developed in later chapters, or ‘third sector’ organisations like
charitable trusts, or you might be interested in planning as an aspect of management from
an academic perspective which is much less practice- driven and more focused on specific
aspects of theory than mine.
You will need to interpret as you see fit when a limited number of suggestions for particu-
lar categories of reader are made. No doubt you would have done so without me suggesting
this, but I did not want you to feel neglected by a simplistic three group characterisation of
target readers.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
In this book ‘telling tales’ means using examples which are case- based stories which have
been retold in a way which makes understanding key messages easier and more memorable.
Tales facilitate understanding by providing practical illustrations of a wide range of
broadly applicable complex concepts, like ‘risk efficiency’. Broadly applicable concepts like
risk efficiency and specific tools like the ‘decision diagrams’ used to achieve risk efficiency
are ‘portable’ in the sense that they can be used in many very different contexts. However,
context always shapes the way conceptual and operational tools are used and the nuances of
their interpretation. A practical approach to planning has to accommodate ‘context’ issues.
This book uses ‘tales’ to deal with a number of context dependency issues.
Part 1 uses a series of short tales. Part 2 use one longer tale per chapter. The rationale is
context always needs to be understood when making decisions, and once we get beyond the
introductory overviews in Part 1, tales which are chapter- sized stories are the most effective
way to convey context related concerns.
Each tale is based on personal practical experience, shaped into a tale to develop key con-
ceptual and operational planning tools in an accessible order. Discussing the credibility of
the tales in Part 2 is an integral feature of the approach adopted.
Drawing on practical experience is crucial in this book, but gives rise to potential prob-
lems of several kinds. This book’s strategy for avoiding or mitigating these problems has
several strands.
In Part 1 some of the relatively short tales directly linked to real organisations are just
slightly fictionalised to keep the examples simple and bring the intended messages up
to date.
Each of the chapter- length tales in Part 2 is a story which mixes experience from more
than one organisation. Four of the five tales were directly shaped by consulting relationships
with one or more named organisations, but none of these tales is just a simple ‘disguised
variant’ of work within named organisations at some point in the past. These Part 2 tales do
not describe any real organisation in a direct way, and they are not dated by the time period
of the tale. They employ a synthesis of organisational features and broader contextual issues
appropriate to communicating messages relevant to a post- 2020 world which is changing
at an accelerating rate.
None of the organisations mentioned have any reason not already in the public domain
to be embarrassed by association with the tales, and in most cases they can be very proud
of their contribution to what has become widely acknowledged as best practice. For the
most part the work with clients this book draws on was a very positive experience, and
when approaches I recommended were not accepted, or some aspects of their implementa-
tion did not work as effectively as intended, I am happy to take my share of any blame and
Overview
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
xxviii Overview
concentrate on positive lessons learned. I have avoided being coy if there are useful lessons
to convey, but used problems positively without dwelling on them.
A chapter by chapter overview ‘road map’ follows – provided as a preliminary to begin-
ning Part 1 which I hope all target readers will find useful.
Chapter 1 explores why formal planning is usually vital but often difficult and frequently
inept. As part of this exploration the meaning of key words like ‘opportunity’, ‘risk’, and
‘uncertainty’ are treated as examples of framing assumptions. This approach means that the
definitions used must be as general (unrestrictive) as possible, supported by working assump-
tions which simplify in the ‘right way’, to clarify what best practice ought to involve in any
given context. Chapter 1 also provides an overview of the other basic framing assumptions
and some related working assumptions the rest of the book builds on.
Chapter 2 introduces a ‘universal planning and complexity management process’ concept, contracted to ‘universal process’ or ‘UP’. This concept is the basis for all process concerns
addressed throughout this book. It is grounded on traditional and ‘soft’ OR (Operational
Research or Operations Research) and broader Management Sciences ideas with the explicit
use of other additional ideas I have found useful plus a provision for any further process
ideas its users may be aware of. This comprehensive and explicitly open nature provides a
systematic basis for thinking about all process choices. Its discussion assumes readers may
have never previously encountered OR, Management Science, or the basis of other key
contributing components, providing an overview of all any reader needs to know about
the historical basis of a conceptual and operational tool which draws on some very different
perspectives and disciplines.
The direct use of this UP concept in practice as a default process is explored in Chapter 5.
It is used to design and develop a ‘specific process’ (SP) in Chapter 6. It is used to adapt
what is commonly referred to as a ‘generic process’ in Chapter 7. It also has an identifiable
role in Chapters 8 and 9. It provides a toolset for ‘keeping it simple systematically’ in terms
of processes.
A core feature of the UP is a ‘capability- culture’ concept. The capability- culture concept
explicitly links the capability and culture assets an organisation has and wants to use during
the development and implementation of plans to the nature of its formal and informal plan-
ning processes. It also addresses associated liabilities – missing or defective assets. Example
assets are requisite skills and information. Example liabilities are untrained or badly moti-
vated staff and incorrect information.
Chapter 2 begins with an overview of the seven phase structure of the advocated form of
the UP concept and the role of the associated capability- culture concept. The origins and
evolution of this UP concept are then explored so that you understand its provenance and
the scope for modifying its basis while adjusting its nature to meet your personal concerns
and the needs of your organisation.
Chapter 3 is a central aspect of Part 1. It is practice- based, but it has been built on a syn-
thesis of theoretical advances by acknowledged leaders in relevant areas, a characteristic of
this book as a whole. It is worth exploring in much more detail than any other chapter in
this Overview because the three efficiency concepts it develops are central to the enlightened
planning approach as a whole and the way enlightened planning departs from a lot of com-
mon practice.
Chapter 3 addresses achieving unbiased estimates of basic attributes like project duration
or cost, using examples which allow exploration of ‘low clarity’ to ‘high clarity’ approaches
while exploring the ‘estimation- efficiency spectrum’. Making use of work by Kahneman
and others on unconscious behavioural estimation biases is part of ‘what has to be done’ to
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Overview xxix
overcome all relevant bias effectively and efficiently, as part of the broader clarity efficiency
concept central to this chapter. ‘Clarity’ in a clarity efficiency context implies understanding
that can be communicated to all relevant people. Clarity efficiency is about achieving any
given level of clarity at the lowest feasible level of cost/effort by systematically using the
right kind of simplicity, avoiding the ‘wrong kind’ of simplicity.
Chapter 3 begins with a ‘minimum clarity’ approach designed to size expected outcomes
and the uncertainty range of actual outcomes at the lowest acceptable level of clarity for a
minimum level of cost/effort. The 1990s context used to provide a simple numerical and
graphical example involves a UK Ministry of Defence (MoD) warship programme manager
asking his team to estimate the duration of a single activity. The tale then explores small
but significant increases in clarity for very modest levels of additional effort. The focus is
on illustrating key issues at the low clarity end of a low to high clarity spectrum. One set
of issues involves understanding why the common practice of using single value estimates
is usually inappropriate and should be scrapped for most purposes, and why range- based
interval estimates are usually essential. Another involves the need to distinguish and clarify
aspirational targets, commitment targets and intermediate balanced targets that may be
expected outcomes (the basic ABCs of targets). A third involves the role of inescapable basic
assumptions which all estimates depend on and everyone involved in both producing and using estimates should understand.
The discussion then moves on to the high clarity end of the range of approaches of
interest. ‘Risk efficiency’ is explored in the context of a tale about the approach to con-
tingency planning which I developed for BP International for their North Sea offshore oil
and gas projects over an eight year period beginning in 1976. This approach was used by
BP on a worldwide basis for about a decade before they moved on to simpler approaches.
These simpler approaches failed them in 2010 in the Gulf of Mexico, with fairly disas-
trous consequences. ‘Risk efficiency’ means a minimum level of risk for any given level of
expected reward. Risk efficiency is a concept Harry Markowitz won a Nobel Prize for in
1990, embraced internationally by economists in a portfolio theory context by the 1960s,
but still not understood or used effectively by many practitioners and theorists in many risk
management contexts. An effective aggressive pursuit of risk efficiency using the conceptual
and operational tools which I developed working with BP provided increased expected
reward via lower expected cost plus less risk (an increase in risk efficiency) while simultane- ously providing duration and cost estimates which were unbiased. This meant that projects could be delivered with both lower expected costs and less risk, usually within budgets and on time, despite some significant surprises. By the early 1980s what needed to be done and
how to do it was understood at board level as well as by everyone engaged in project man-
agement roles. And early confidence in this approach was verified by a decade of experience
and empirical evidence. The BP operational processes and other tools were designed to
cope with what we labelled ‘unknown unknowns’, as well as ‘known unknowns’ which had
been identified but treated qualitatively as ‘conditions’ and ‘known knowns’ which had been
quantified in probabilistic terms.
By the early 1990s a senior MoD colleague responsible for key aspects of MoD procure-
ment processes was arguing that every £1 spent on this kind of approach by the MoD should
save about £100, a massive return on investment in a risk efficiency driven approach which
produced unbiased estimates as a by- product of much more effective and efficient planning.
This kind of risk efficiency driven approach is usefully perceived as an opportunity which
should be seized by all organisations not currently using it, failure to do so because its
benefits and nature are not understood constituting the realisation of a serious corporate
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
xxx Overview
capability risk. My senior MoD colleague had mixed success getting the MoD as a whole to
use it for reasons relevant to central concerns addressed by this book. Many MoD experts
bought into my BP- based risk efficiency driven approach, which was the basis of MoD inter-
nal courses run for several years and widely acknowledged MoD best practice over several
decades. However, from the outset, some MoD ‘experts’ maintained a preference for sim-
plistic, less effective approaches, still a crucial issue in need of resolution within and beyond
the defence context worldwide. Some of the reasons for this, and appropriate responses, are
addressed directly in all Part 2 chapters.
Chapter 3 then employs the risk efficiency framework developed for BP and the BP exam-
ples in a manner used for IBM UK in the 1990s to promote taking more risk in a risk efficient manner at a local decision- making level in a context like bidding for a £20 mil-
lion contract. The goal in this context was increasing overall profitability while decreas-
ing overall corporate risk, understanding the difference between good management and
good luck, bad management and bad luck. This discussion explores what IBM and subse-
quent clients liked to call ‘enlightened caution’ as well as ‘enlightened gambles’. This use
of the term ‘enlightened’ was the origin of the enlightened planning label and other related
uses of ‘enlightened’ in this book. Identifying enlightened caution involves using a high
clarity ‘decision diagram’ tool initially developed for BP to identify risk efficient choices. It
uses what is usually referred to as a ‘stochastic dominance’ generalisation of a Markowitz
approach. Enlightened caution is an issue when one option is a potentially deceptive risk
inefficient ‘unenlightened gamble’ – it involves a higher most likely reward but a lower
expected reward as well as a higher level of risk because a highly asymmetric distribution
with one very long tail is involved. ‘Enlightened caution’ facilitates avoiding ‘unenlight-
ened gambles’ which involve more risk and less expected reward when it is easy to make
errors of judgement if unenlightened intuition is relied on. Without high clarity decision
diagrams, unenlightened gambles may seem to be the obvious preferred choices, but high
clarity decision diagrams make it clear that being capable of identifying the opportunity to
avoid unenlightened gambles is crucial. ‘Enlightened gambles’ involve risk efficient gambles
worth taking for the additional reward. Once these issues are understood, a linked corpo-
rate culture change can then be engineered. Once people get used to working with the new
toolset, simpler diagrams or simple verbal comments with no diagrams are often all that is
needed for most management decision- making practice.
These ideas and their behavioural implications are explained in Chapter 3 using graphs
in a way most people find clear and convincing whatever their background and interests, as
part of explaining, at an overview level, the concept of ‘opportunity efficiency’.
‘Opportunity efficiency’ is a composite concept which is explored in ‘what needs to
be done’ overview terms towards the end of Chapter 3. Opportunity efficiency involves
selecting the most appropriate level of risk– reward trade- off for all relevant attributes. This
requires the exercise of enlightened prudence when relevant, reducing expected reward to
avoid inappropriate risk. Opportunity efficiency also involves making appropriate trade- offs
among all relevant attributes, plus choosing suitable trade- offs between clarity and the cost
of clarity. Opportunity efficiency is an operational definition of ‘best practice’.
My work with IBM UK was central to a culture change programme for all senior and mid-
dle managers based on IBM’s two day Forum 2 programme, run about 40 times in the early
1990s. The IBM chief executive officer opened the morning of the first day by outlining the
strategic changes IBM had to make and explaining why the concepts to be covered by my
contribution were central to those changes. I then outlined the key concepts they needed
to understand until lunchtime. After lunch, I used a case study exercise designed with an
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Overview xxxi
IBM senior executive to put the general conceptual ideas into an IBM context they could
associate with directly and give them feedback on the nature of some of the changes they
had to make in operational terms with cultural implications. The second day they discussed
how they would implement related changes in their areas of operation. This programme was
a significant success, leading to IBM corporate process and culture changes. Several follow-
on consulting assignments were triggered. In addition to its role in Chapter 3, this work for
IBM is central to Chapter 6.
The low to high clarity estimation- efficiency spectrum explored by Chapter 3 provides a
conceptual and operational framework which is central to an enlightened planning approach.
Chapter 3 synthesises and clarifies a core portion of all my earlier work, plus more recent
work with a wide range of colleagues, Stephen Ward’s contributions over the last 20 years
being of particular importance. It builds on earlier work by others like Kahneman, Markow-
itz, and those behind stochastic dominance. It treats unbiased estimates and the pursuit of
opportunity efficiency defined by component risk efficiency and clarity efficiency as non-
separable concepts at a framing assumption level.
Chapter 3 builds on the understanding of a very general approach set out in Chapters 1
and 2, but it uses important explicit simplifying assumptions to avoid difficult complexi-
ties, and it skips over several challenging complexities which need addressing at a founda-
tion level.
Chapter 4 addresses the challenging complexities which need confronting within the
holistic framework this book’s approach has provided thus far. Its starting point is directly
useful operational tools skipped over in Chapter 3, like a high clarity form of ‘sensitiv-
ity diagram’. Sensitivity diagrams were initially developed for BP to understand the rela-
tive importance of different sources of uncertainty in a complex model from the outset of
analysis, helping analysts to build their understanding from the bottom up. Selective use
at board level in a top- down mode for somewhat different purposes is amongst the ideas
explored. The basic ideas are relevant in a wide range of contexts. Effective use of sensi-
tivity analysis is a core best practice analysis capability for most of the people involved in
almost any context – including planners and supporting analysts, their managers at all levels,
and the organisation’s suppliers or customers, as illustrated in Part 2 chapters. A midpoint
Chapter 4 topic is low clarity decision diagrams, used as simple operational tools to address
risk efficiency in terms of a basic primary attribute like cost when trade- offs involving non-
measurable attributes or attributes not worth measuring need consideration and may be
crucial. Again the core capabilities demonstrated are widely applicable. Towards the end of
this chapter the focus shifts to high clarity operational tools and conceptual frameworks for
addressing multiple attributes which may involve complex ethical choices. Chapter 4 closes
by exploring the limitations of formal analysis.
Part 1 as a whole explores a wide range of foundation level framing assumptions and
associated working assumptions as part of beginning to explain why some of the difficulty
encountered in practical decision- making contexts has been generated by both academic
and practitioner ‘experts’ who are unable to distinguish between the right kind of simplicity
and the wrong kind of simplicity. This is a difficulty which needs to be addressed by every- body affected – it cannot just be ‘left to the experts’.
Part 2 illustrates the use of the conceptual and operational toolset outlined in Part 1,
including key operational links between corporate planning, operations planning and project
planning. These links include the roles of operations and project planning in the generation
of bottom- up corporate strategic planning considerations which are driven by new opera-
tional requirements and their strategic relevance for the organisation as a whole. ‘Project
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
xxxii Overview
planning’ is viewed as the management of change in a very general sense whenever this is
useful. So far as I am aware these interdependences have never been addressed in terms
of a holistic and collaborative framework for a wide range of target readers in the way this
book does. This book explicitly acknowledges that planning in different areas can be use-
fully assumed to be ‘separable’ for some purposes, but interdependencies that always matter
need constant direct attention, and interdependencies that may matter need routine testing
for relevance and responding to when necessary. What is meant by ‘separable’ here and in
many other contexts has important practical operational implications, explored throughout
the book.
Chapter 5 uses a very simple supply side operations management context to illustrate
the use of the universal process (UP) developed in outline in Chapter 2. In the context of
Chapter 5 the UP concept is used in a default process mode – in any context if it is not clear
what process should be used, a UP is the default choice. Chapter 5 assumes readers may
have no supply side operations management knowledge and no previous OR, Management
Science, or cognate discipline knowledge, along with perhaps a limited interest in acquiring
new knowledge of any of these areas. This UP concept is a new synthesis of what I have
always understood as a practical view of the classical textbook ‘OR process or method’ and
the more recent ‘soft OR’ ideas, plus further key ideas involving a capability- culture concept
suggested by authors in best- selling recent books whose perspectives are very different –
Gawande (2011) and Kennedy (2014) in particular. It also integrates this basis with process
phase ideas in my ‘uncertainty management’ work with Stephen Ward in a project planning
context. Chapter 5 sets the scene for Part 2 in an initially very simple practical context, keep-
ing the details of the examples as simple as possible, but explaining inherently complex issues
like the role of key assumptions, the crucial role of collaboration between organisations for
mutual benefit, and the difference between simple models used to understand key trade- offs
and plans which go well beyond simple model capabilities.
This focus of Chapter 5 is using a UP concept in bottom- up mode with ‘the top’ at a
very low level initially but successively moving the starting position upwards and gradually
becoming more strategic. It demonstrates the importance of teamwork which is not con-
strained by organisational silos or professional perspectives and the way an effective universal
process can generate ‘propositions’ of strategic importance to the organisation, potentially
leading to transformational changes when integrated with top- down corporate strategic
planning.
As with all other chapters, the focus is what needs to be done, not how to do it. The sim-
plicity of the context helps to keep the requisite how to do it discussion simple. If you are a
director or an expert with extensive experience, your initial impression may be the issues and
tools addressed by this tale are seriously lacking in sophistication and strategic implications.
However, there are some important nuances you are likely to find useful in the context of
the overall approach, and these nuances are further developed when addressing the inher-
ently complex concerns confronted by the rest of the book.
Chapter 6 explores how a sales manager for the regional office of an international organi-
sation might develop a bidding process, building on ideas developed during the IBM UK
culture change programme discussed in Chapter 3. It demonstrates how a capable sales
manager who understood some of the concepts developed in earlier chapters, plus a team
with complementary skills, might address bidding and broader organisational issues when
the context focus is demand- side driven. One of the key roles of this chapter is illustrating
the use of the universal process (UP) concept to design a specific process (SP) for bidding. This approach generalises the use of generic models usually employed by OR processes as
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Overview xxxiii
illustrated in Chapter 5, moving part of the way towards working with ‘generic processes’
as addressed in Chapter 7.
Subtle concerns of importance to an underlying corporate culture change management
programme and a broad capability transformation programme are addressed along the way.
They include identifying when some people are biasing cost estimates to accommodate stra-
tegic concerns not previously considered explicitly, which actually need simple but explicit
qualitative consideration, as well as numerical adjustments to estimates used for decision
making in some cases. They also include using marginal costs rather than standard internal
transfer prices when the difference matters and maintaining a level playing field for internal
staff groups that are competing with external organisations which are crucial strategic part-
ners. Furthermore, they include effective quantitative analysis of likely bid pricing behaviour
by competitors coupled to non- quantifiable ‘value added’ features which potential custom-
ers might be provided with in order to win bids.
This discussion emphasises the importance of non- price concerns for both customers and
vendors and the role of trust when vendors look after their customers’ interests effectively
without overlooking their own best interests. It also emphasises when taking more risk at
the level addressed is essential to reduce the risk of corporate failure, and the operational
implications of this kind of mindset and skill set change. No marketing or other demand-
side management knowledge as assumed.
By the end of this chapter you should see why some of the key ideas explored will need
further reflection in terms of the overall enlightened planning (EP) approach even if the
technical bidding process issues are not a direct interest. The chapter ends with a section
indicating some areas for further reflection as the book progresses to get you started.
Chapter 7 addresses project planning in the context of a water and sewage utility. It
considers planning projects like the construction of new water supply pipelines and sewage
treatment plants. It provides a what needs to be done level of understanding of the how to
do it ideas in the book How to Manage Project Opportunity and Risk (Chapman and Ward, 2011). I believe all members of any board ought to understand these ideas at a what needs
to be done level for a wide range of reasons. One is being capable of exercising their overall
board level governance role when judging the competence of project directors and all pro- ject managers reporting to themselves and to other directors. I believe everybody else in all
organisations who might be tempted to read this book will also find it valuable to acquire
the basic what needs to be done level of strategic clarity about change management provided
by Chapter 7, some also requiring significant further how to do it tactical clarity. Chapter 7
continues to build on all earlier chapters, following the layered approach of the book as a
whole. It does not assume any prior project management knowledge.
Chapter 8 provides a what needs to be done level of understanding of corporate plan-
ning driven by top- down corporate strategic planning integrated with all other planning.
It uses an electricity utility context, initially considering the corporate strategic planning
concerns that I was asked to address as an expert witness for a Province of Ontario govern-
ment enquiry into Ontario Hydro’s strategic plans at the beginning of the 1990s. Ontario
Hydro wanted to build 10 new nuclear power stations over a 25 year period. I was hired by
an ‘official intervenor’ to object on their behalf, funded by Ontario Hydro under Ontario
government rules. My published report explored in outline what Ontario Hydro should
have done in the process of explaining why their corporate strategic planning approach was
inappropriate and their conclusions were unsound. Ontario Hydro withdrew their plans a
week or so before I was due to present oral evidence, for reasons anticipated and addressed
by my report.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
xxxiv Overview
Initially, a radical corporate strategic change, including culture changes, is not discussed.
The focus is given current knowledge about feasible futures in technical terms at a long- term
planning horizon 20 years hence and current understanding of the significant uncertainty
about future costs and prices, what should the utility be aiming for in terms of a portfolio of
power station types, and what commitments should be made for new projects which need
to be started as soon as possible.
After a framework for this kind of corporate strategic planning has been developed, more
radical change is explored. Corporate goals and feasible futures are reassessed in terms of
corporate capability- culture issues as well as potential new technology physical facility issues.
What has been termed ‘designing desirable futures’ is taken as the basis for a much broader
perspective – where organisations ought to start in practice if they want to address what
really matters in an iterative manner driven by goals which may not be feasible, dealing
effectively with priority and precedence relationships.
This chapter also addresses capability- culture concerns which have since driven Ontario
Hydro towards the kind of privatisation visited upon the UK Central Electricity Generating
Board in the 1980s. By the end of this chapter it should be clear why all organisations need a
suitable variant of the kind of corporate planning discussed in this chapter, fully integrated with
the kinds of planning discussed earlier, adapted to the context the organisation has to deal with.
No background knowledge of corporate strategic planning or ERM (Enterprise Risk
Management) is assumed.
Chapter 9 extends the corporate planning framework and integrated operations and pro-
ject planning frameworks provided earlier to very high impact but very low probability
incidents plus a full range of less serious incident variants. It builds on a review of Rail-
track’s approach to a strategic planning framework for UK railway safety management which
I undertook for their head of safety in the 1990s. It also builds on a new framework derived
from a review for the UK MoD over a three year period from 2010 until 2013. My under-
standing is this MoD work has subsequently received attention at a NATO level. The MoD
consultancy addressed justifying high levels of expenditure on a portfolio of preventative and
responsive contingency plans for protecting troops from non- conventional weapon attacks.
Railtrack did not take my advice, but they went bankrupt a couple of years later. My 1990s
work for them would have been significantly enhanced by my 2010– 13 MoD framework’s
approach as generalised in this book, perhaps leading to Railtrack accepting an enhanced
version of my 1990s recommendations. The reasons for Railtrack’s corporate failure were
multiple and debatable, but two very serious accidents within a short period were certainly
crucial. Anticipation of this kind of possible scenario, or an even more serious scenario,
coupled with effective planning to prevent or deal with such circumstances, was central to
my recommendations.
The tale of this chapter integrates railway accident and malicious incident concerns,
addressing potential terrorist attacks as well as accidents for a European railway system.
It explains important practical reasons for this integrated approach, including crucial sim- plifications of the right kind. The key new issues relative to earlier chapter discussions are driven by the role of trade- offs between money and attributes like ‘lives’ (people killed who
might not have been killed if more money had been spent or spent more effectively). It also
addresses associated ‘injuries’ (over a range of levels of seriousness).
Keeping it simple but dealing with very serious levels of complexity that really matters in
a formal planning framework is the central concern of this chapter.
A key aspect of the strategy is a focus on a small set of scenarios which deal with correlated
metrics for attributes associated with different objectives, like levels of fatalities, levels of
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Overview xxxv
injuries, levels of operational disruption to the railway system, and all associated cost and lost
revenue concerns, including reputational risk implications. If all the interested parties can
agree what level of expenditure is appropriate to reduce the severity or probability of each
scenario in this limited set of scenarios, with an initial focus on a ‘worst case’ outcome, the
different parties do not need to agree in detail about their relative priorities for each com-
ponent metric. This is a crucial simplification in a very complex situation, and it provides a
significant opportunity which common practice overlooks.
Furthermore, there is no need to assume that common practice metrics like ‘the value
of an avoided fatality’ have to be constant for any number of fatalities, or that such metrics
need to be based on cost– benefit analysis concepts which are at best debatable, and at their
worst lacking a moral compass and failing any ethically driven political acceptability test.
These are crucial complications which really matter, and addressing them effectively is facili-
tated by the linked simplifications.
Concluding discussions in this chapter explore other even more difficult trade- offs in the
sense of further analytical complexities and complications, generalising the railway context
approach to deal with further relevant metrics like environmental damage. One example
briefly considered is the 2010 Gulf of Mexico Macondo (Deepwater Horizon) incident
which has cost BP about $60 billion to date, largely driven by environmental damage con-
sequences, although everybody on the rig involved died as well. No background knowledge
of safety or security analysis and management is assumed.
Part 2 as a whole might be summed up as follows: Chapter 7 is about delivering change
which meets or surpasses the expectations of all the key stakeholders in that change when-
ever possible. Chapter 8 is about deciding what to change, and the resources needed, includ-
ing addressing capability- culture asset and liability concerns. Chapters 5 and 6 deal with
ongoing operations management issues which interface with the concerns addressed in later
chapters, but they are considered first because they provide basic tools needed to understand
the approaches developed in later chapters. Chapters 5 to 8 are all relevant to all organisa-
tions. Chapter 9 deals with special case issues which are crucially important whenever they
are relevant, and if we generalise very serious accidents or terrorist attacks to include mod-
erately serious environmental, cybersecurity, fraud, or loss of corporate reputation issues of
any other kind deserving moderately serious attention, most organisations need to think
very carefully about addressing potential concerns of this kind.
Part 3 provides further synthesis and reflection in two relatively short chapters.
Chapter 10 provides further synthesis and reflection at an overview level to clarify what
needs to be done concerns in terms of addressing immediate and later priorities. Chapter 10
makes it clear that some of the ideas developed in Chapters 8 and 9 need attention as a first
priority in practice, moving on to less urgent concerns after dealing with starting point con- cerns like overall corporate goals. By the end of this chapter you should have a consolidated
understanding of the strategic clarity all organisations need.
Chapter 10 also reflects in more detail on public/private ownership issues which surface
earlier. Some private sector organisations need effective planning which addresses regula- tory issues and pressures. This may include the need to address the way their competitive
advantage depends on regulator or government interventions which they may be able to
influence. As a special case, ‘nationalisation risk’ may be relevant. Some public sector organi- sations need to understand the pressures driving them towards privatisation and respond
effectively to ‘privatisation risk’ considering all relevant interests.
There is an upside and a downside to both private and public sector status from a vari-
ety of perspectives. How both are implemented determines the best choices for all the
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
xxxvi Overview
interested parties, whose interests may not coincide, with the context being crucial in a
number of ways. You should understand some of the basic issues in a new framework by the
end of this chapter, along with other ideas relevant to all organisations in both sectors. This
framework might be used by readers with different agendas in very different ways.
You may have heard the mantra ‘what cannot be measured cannot be managed’, often
cited by those in both private and public sector organisations. After reading Chapter 10 it
should be clear why those advocating this mantra do not understand what decision making
in most practical planning contexts ought to involve, or they are being deliberately mislead-
ing to serve their own agendas, or both.
By the time you have finished Chapter 10 you should also understand some of the advan-
tages and limitations of:
1 metrics for relevant objectives,
2 quantification in appropriate probabilistic terms,
3 the role of qualitative models which structure and support formal planning,
4 the crucial way informal planning has to support all formal planning approaches,
5 the way corporate capability- culture concerns have to serve as part of the foundation for
the whole edifice, and
6 key limitations of approaches to decision making which do not address these concerns
in an enlightened manner.
Chapter 11 begins with a focus on teamwork and wider collaboration, unbiased decision
making, and contingency planning aspects of ongoing enhancement of strategic clarity and
tactical clarity. Broad planning areas, approaches, perspectives, concerns and issues requir-
ing ongoing strategic and tactical clarity enhancement are then addressed, finishing with a
brief discussion of a diagram which has evolved over many years which portrays the corpo-
rate benefits of an enlightened planning approach in terms of more ‘smiles’ linked to more
‘pleasure’ and less ‘pain’. Reward, usually centred on profit metrics, is a primary concern,
but capability- culture concerns which are crucial need focused attention, and this diagram
is a useful final summary of what is involved in delivering the goals of enlightened planning.
Parts 1 through 3 as a whole are about why some aspects of planning usually need to be
sophisticated, subtle and creative, and why simplistic, inflexible approaches do not work as a
general rule, but appropriate simplifications are essential and can be very powerful. Generally
we all need approximately correct answers to the right questions within the right timeframe which everyone involved can trust. We do not want precise answers to the wrong questions or answers which are too late or answers which cannot be received with well- founded trust. The difference matters greatly, and understanding what the difference involves is what this book is about.
Regulation, politics and ethics sometimes have to be addressed, with examples in four of
the five Part 2 tales. Using a broad interpretation of what regulation might involve can be
very important, including ensuring that all valuable players in a marketplace enjoy ‘a level
playing field’. The complexities involved in planning undertaken by regulators and their
political masters are not addressed in detail, but some key concerns are clarified, and some
of the concepts and tools developed are directly relevant to regulatory concerns. One UK
regulator already moving in the suggested direction is discussed briefly in Chapter 11, and
there is increasing evidence of significant international movement of this kind involving a
very wide range of regulation which needs to be encouraged and facilitated. But it would be
naïve to think that some very serious concerns needing attention are going to be straight-
forward to resolve.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Overview xxxvii
Politics in the most general sense can only rarely be ignored, and ethics are almost always
relevant. Although political decision making is well beyond the scope of this book, as are
appropriate ethics, both are relevant to any enlightened planning perspective. I specifically avoid political dogma and moral imperatives. For example, when ‘good’ and ‘bad’ are dis-
cussed in ethical, political, economic, or any other sense, the boundaries are always assumed
to be debatable. However, this does not mean the issues involved do not matter. It means
we need to be particularly careful about our assumptions and deal with important uncer-
tainty associated with ambiguity. It also means we need to cope in a practical manner with
interdependencies among political, legal, regulatory, and ethical issues which may be inher-
ently extremely complex and controversial.
The need to clarify a generalisation of the ‘uncertainty management’ ideas underlying the
book How to Manage Project Opportunity and Risk : Why Uncertainty Management Can Be a Much Better Approach than Risk Management, (Chapman and Ward, 2011) has been the focus of my efforts on successive drafts of this book for several years. The need to do so in
a carefully considered manner which would be accessible to a very broadly defined target
audience has become increasingly clear each year for several decades. During this period,
Stephen Ward and I have been aware that our evolving uncertainty management approach
was critically acclaimed as helping to define the leading edge of international best practice
by a significant number of people. But we were also aware that many people preferred to
support common practice that they perhaps saw as good practice which was simpler with-
out understanding the implications of the nature of the simplicity employed. And we were
concerned that many others had never come across the basic ideas we believe to be crucial.
What is meant by ‘bad’ practice, ‘best’ practice, and a range of intermediate ‘good’ practices
remains ambiguous and controversial – issues this book confronts directly, to allow you to
judge for yourself, and shape your own future version of best practice.
Much of the bad, good, and best practice controversy is deeply rooted in the difficulties
that we all have when adjusting the framing assumptions and habitual working assumptions
which define our view of a best practice approach to management decision making plus
underlying planning and analysis in any context. Testing and adjusting assumptions that we
have trusted for some time but may need to change is never easy. But it is a core compe-
tency for reflective managers, and it is a central concern and explicit focus throughout this
book. The emphasis on this explicit focus is new relative to Chapman and Ward (2011) and
my earlier publications. Chapter 1 makes a point of explaining some example shifts in my
own framing assumptions, and comparable working assumptions, during the writing of later
drafts of this book, in part to encourage you to think about shifting some of your routine
assumptions throughout. A key message for you to take away from this book is the impor-
tance of the issues you can uncover by an explicit focus on the set of framing assumptions
and comparable working assumptions which define your perspective. They really matter.
My goals when writing this book included providing you with an operational best practice
toolset for all management decision making which you can put to immediate use, plus an
underlying set of conceptual frameworks you can use immediately and then build on. I hope
you will find this book of immediate value and useful as a stimulus for generalising your
toolset, skill set and mindset along some of the lines explored.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
The approach to planning advocated by this book was synthesised from a career based on
pursuing a practice– research– teaching– practice cycle for about 50 years. Improving practice
was always a central goal in this iterative process. Understanding the way my pursuit of this
goal evolved in relation to the content of this book with a kind of detail most authors do
not provide may help you to better understand what this book is about and why it takes the
form employed – the rationale for the approach taken to the next few pages. It should also
help to underpin the candid conversational style used throughout this book, a style which
seemed essential given the nature of some of the intended messages.
I was born and brought up in Toronto. A University of Toronto BASc in Industrial
Engineering (1962) and a University of Birmingham MSc in Operational Research (1964)
provided my initial academic grounding. An Athlone Fellowship funded by the UK Board
of Trade gave me the opportunity to spend 1962 through 1964 in the UK.
I spent 1964– 5 as an IBM computer sales trainee in Toronto. This built on my computer
systems experience working for IBM in Toronto for three summers during my undergradu-
ate degree, plus a year working in a project planning systems development role with Ferranti
in London as the first part of the Athlone Fellowship. Working for IBM Canada initially,
and then seeing where that led, was the ‘career plan A’ adopted while still an undergraduate.
In 1965 I accepted an offer of a lectureship (assistant professorship) in econometrics from
Gordon Fisher – the opportunity was unanticipated but too good to miss. Gordon had
taught one of my University of Birmingham MSc programme econometrics options, and the
following year he founded the Econometrics Department at the University of Southampton
as its first Professor of Econometrics. For nine years, my career focus was managing a new
MSc programme in Economics, Econometrics and Operational Research (OR), designing
and teaching the OR content, and completing a PhD in consumer behaviour theory as a
staff candidate supervised by the economist Professor Ivor Pearce. The PhD shaped my
view of ‘separability’ and several closely coupled concepts which underlie the foundations of
this book as a whole, building on Pearce (1964). It also provided a deep understanding of
the foundations of risk management central to this book, building on Markowitz (1959).
During this period, I developed a passion for research into the issues exposed by practice,
initially centred on the development of consumer behaviour theory to support marketing
decision making for the UK Milk Marketing Board. I also started to acquire a passion for
consulting. One key client relevant to this book was Buckinghamshire County Council.
On Gordon Fisher’s recommendation they hired me as an expert witness to help stop a
‘Third London Airport’ being built at Cublington. The economists David Pearce and John
Wise were recruited to help, working as a fully integrated partnership. The Cublington
recommendation was scrapped. We obviously cannot take full credit, but I believe we were
About the author
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
About the author xxxix
on the right side of a complex cost– benefit analysis argument, and some of the issues and
approaches are directly related to those addressed in Chapter 9. During this period I became
a senior lecturer (associate professor) and served as the assistant dean of my faculty (Social
Sciences). Taken as a whole, this period was a central part of an extensive apprenticeship
which significantly shaped my career and perspective. I had not seriously considered an aca-
demic career or living in England until Gordon’s offer but never regretted my 1965 largely
intuitive change of mind, despite ongoing uncertainty about the permanence of these deci-
sions for 20 years.
For 15 months during 1974– 5 I worked full- time with Acres Consulting Services in
Canada. This opportunity was initiated by a three month consultancy assignment invitation
from Oskar Sigvaldason. Oskar was then head of Acres Special Services Department, later
president of Acres. I learned a lot about consulting, including key team- working and client
management concerns. One key study relevant to this book was leading the risk and uncer-
tainty analysis of a proposal to reduce by one year the construction duration of a pipeline
to bring high Arctic gas to US markets. Another was a comparison of Canadian and US
design regulations for nuclear power stations in relation to seismic (earthquake) risk. I built
a lasting relationship with Acres. However, I declined the offer of a permanent full- time
role with Acres in Canada. I returned to the Department of Economics at the University of
Southampton, which had absorbed Econometrics while I was away. Working for Acres was
immensely stimulating, and my wife, Jean, and I and our two young sons greatly enjoyed
living in Niagara- on the Lake. However, with a young family I was not prepared to accom-
modate commuting to clients in Calgary, Edmonton, Ottawa, and similar locations for a
week at a time for a significant proportion of the year, and for a complex set of reasons on
balance, an academic career based in the UK seemed the best feasible choice.
For the next decade one focal area of my career was consulting to help clients build
processes and embedded model sets which addressed problem areas with no available ‘off-
the-shelf ’ approaches. Through Robin Charlwood, an Acres colleague, I established an
eight year relationship with BP International in London. I helped BP to develop planning
and costing approaches for their North Sea operations, adopted by BP for worldwide use
on all large or sensitive projects for more than a decade. In this period BP projects using
the approaches I helped to develop were generally within time and cost commitments, with
no surprises which could not be accommodated. Through Oskar Sigvaldason, Robin Charl-
wood, Gavin Warnock, and several other Acres colleagues I also worked with Acres teams
for other clients in Canada and the US, building on the BP work, with any lengthy assign-
ments scheduled so that my family could accompany me. Illustrative key clients included
Gulf Canada (Beaufort Sea and Grand Banks oil and gas project design studies, including
the Hibernia oil production platform project off the east coast of Canada, where icebergs
were a key concern, and platform cost uncertainty coupled to oil reserve volume uncertainty
plus oil price uncertainty proved critical), Petro- Canada (a design strategy study for a pilot
liquefied natural gas project on Melville Island in the high Arctic), Fluor Engineers and
Contractors Inc (a design study addressing how best to get a proposed 48 inch gas pipeline
across the Yukon River in close proximity to an existing 48 inch oil line, with a wide range
of relevant threats and interested parties), and Potomac Electric Power Company and the
US Department of Energy (comparison of energy storage via pumped hydro or compressed
air in deep mines). Research driven by my consulting interests was published, and I became
a reader in Management Science. The other focal area of my career during this period was
helping Professor Ken Hilton develop the Department of Accounting and Management
Science, which he extracted from the Department of Economics with my support. As well
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
xl About the author
as teaching, I managed new MSc programmes and undertook various other academic roles.
This period put into practice the maturing practice– research– teaching– practice basis of this
book, initiating and shaping some of the basic ideas.
For nearly a decade the focus of my career then shifted significantly. I had become a
Professor of Management Science with a personal chair, and Head of the Department
of Accounting and Management Science. I now made a full commitment to maintaining
an academic career base. Ken Hilton had increased the size of our department by 50%.
I increased it by a further 100%, adding two new groups with professorial leadership:
Finance & Banking and Information Systems. My consultancy became more UK- focused.
I started accepting invitations to work through UK based consultants, including work with
Sir William Halcrow and Partners of direct relevance to Chapter 9. Some consultancy was
undertaken through the university. Examples central to this book include several studies
helping National Power to develop BP type approaches to building electricity generation
stations, and a series of studies over the period from 1993 to 1995 helping UK Nirex to
plan a repository for nuclear waste disposal and deal with Department of the Environment
(DoE) arguments about deferring the project. The DoE adopted a Her Majesty’s (HM’s)
Treasury mandated real discount rate of 6% when 3% would have been more appropriate
in terms of my arguments at the time. HM Treasury’s own post- 2003 advice is consist-
ent with my 3%, for directly related but different reasons. These issues are important in
private as well as public sector contexts, and the reasons are explored in Chapter 7. I pur-
sued research conventionally funded by research councils and professional bodies as well as
the MoD and other organisations, some directly relevant to this book. Involvement with
professional bodies began, including accepting an invitation to act as founding chair for
the Association for Project Management (APM) Specific Interest Group on Project Risk
Management. I started to spend more time teaching experienced managers, including a
significant culture change programme for IBM UK, its ‘Forum 2’ programme. This was a
two day in- house event run about 40 times, introduced by their CEO on each occasion,
built around my input, central to Chapters 3 and 6. This wider set of activities and concerns
reduced the time available for consultancy, but it did not weaken the practice– research–
teaching– practice basis of my career, and prototype variants of many of the key ideas in this
book matured during this period.
A five year break from university management roles then involved a different slight shift in
focus, centred around two years as president of the Operational Research Society.
My three years as director of the Southampton University Management School (SUMS)
then involved a new university management role. I was appointed director with a transfor-
mation mandate by a new vice- chancellor. While I had been head of the Department of
Accounting and Management Science the University had established SUMS as a separate
Management School to provide MBAs and other post- experience courses. My advice to
avoid making these activities separate was rejected, but SUMS had my support once that
decision had been made. The new vice- chancellor wanted SUMS fully integrated with the
university, located on campus, made profitable, and made reputable in research terms. These
objectives were achieved in the planning horizon which I eventually set myself. My success-
ful exit strategy from my role as director involved recommending that the current the head
of the Department of Accounting and Management Science take over as director of a new
Management School created by a full merger of SUMS plus the department, with me in sup-
porting roles to help complete the transition. Outcomes included doubling MSc and MBA
student numbers, an RAE (Research Assessment Exercise) rating of five (the top rating on
the scale used for UK research assessment at that time) for the new Management School,
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
About the author xli
and strengthened relationships with the faculties of Engineering and Mathematics. The lat-
ter was facilitated by founding the Centre for Operational Research, Management Science
and Information Systems (CORMSIS), with the director’s post alternating every two years
between Management and Mathematics, initially held by the professor of OR in Mathemat-
ics, Paul Williams. The new School of Management has continued to evolve, becoming the
Southampton Business School (SBS) in 2014. CORMSIS has thrived with a series of direc-
tors focused on collaboration within and beyond the university, the sustained joint efforts
of Sally Brailsford (Management/SBS) and Chris Potts (Mathematics) being particularly
important. A separate but overlapping Centre for Risk Research (CRR), founded in 1990
by Johnnie Johnson (Management/SBS), which he led with great success until retiring at
the end of 2018, also thrives and continues to evolve. Both centres embrace an emphasis on
practice and a broadly defined perspective.
From 1991 until 1993 I served as an expert witness providing a critical review of Ontario
Hydro’s strategic plans for the next 25 years, central to Chapter 8.
In 1992 and 1996 I served as a Business and Management Studies panel member for the
RAE. The judgements of these panels determined the distribution of the research fund-
ing component of the UK government’s university funding for business and management
for two four year periods. At the invitation of the panel’s chair as his ‘quantitative analysis
expert’ I unobtrusively but explicitly confronted the management and governance implica-
tions of different people arguing for different weightings when using quantitative measures
of attributes which do not lend themselves to simple metrics plus important non- measurable
concerns when important decisions have to be made by a group of people with very dif-
ferent perspectives and agendas, and the need to use available measures coherently as far as
possible, issues which are central to this book as a whole.
In 1999 I was elected an Honorary Fellow of the Institute of Actuaries. My work on its
joint working parties with the Institution of Civil Engineers on risk management guides
addressing projects, then whole enterprises, then operations, shaped the three component
separability structure for all management decision making adopted by this book.
From 1997 to 2003 I served as a non- executive director of Southern Water, with three
different chairmen of the board and three different ownership structures, useful in terms of
background for Chapter 7 and direct governance experience relevant to Chapters 7 to 9.
There are a number of relevant differences between advising other organisations and
taking your own advice when directly engaged in management and governance functions.
My operations, project and corporate management experience as a head of an academic
department, my change management experience as a management school director, a vari-
ety of other academic and professional roles, and my board level governance experience
as a Southern Water non- executive director, all reinforced my consultancy experience in a
manner relevant to the overall ‘practice basis’ for this book. They were all modest roles in
corporate terms, and you may not see universities as ‘commercial organisations’, but each
helped shape and let me directly test some of the concepts and other tools discussed in this
book, and they all influenced my views on requisite skill sets and mindsets in ways directly
relevant to this book. They were an integral part of the education and practical experience
basis that underlies my current perspective. That is the primary reason for mentioning them
here, as part of explaining ‘where I am coming from’ before you start Chapter 1, to help
you see where this book might take you. One secondary reason is encouraging you to see
‘commercial organisations’ through a broader lens than you might be used to, with a wide
range of objectives. A defining characteristic of this book is encouraging a broader view than
you may be used to of most concepts involved.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
xlii About the author
In 2004 I retired from my full- time academic post, became an emeritus professor on a
part- time contract and accepted an invitation from Mike Nichols to become a senior associ-
ate of the Nichols Group. Three subsequent consultancy studies are directly relevant to this
book.
In 2005 I worked in Venice with Gavin Warnock and Robin Charlwood, renewing our
Acres connections begun in 1974. Both had been Acres vice- presidents but now had their
own consultancy companies, Gavin based in Edinburgh and Robin in Seattle. We worked
through Gavin’s Monitor (International) Ltd. Our client was Consorzio Venezia Nuova,
the contractor for the MOSE flood protection scheme for Venice. We were successful in
persuading the government that the cost estimates had to be significantly increased because
earlier risk provisions based on conventional ‘received wisdom’ estimation methodologies
were biased on the optimistic side. The MOSE project proceeded, with construction due to
complete in 2018. This study helped to shape Chapters 3 and 7.
In 2006 and 2007 I worked with Mike Nichols and a small team to help him write a
report for the secretary of state for transport which explained why UK Highways Agency
cost estimates were consistently optimistically biased, despite following HM Treasury guide-
lines on these issues and what to do about it. I then supported a team of Nichols consultants
help the Highways Agency start to implement our recommendations, initially revising all
current cost estimates in a manner approved and supported by HM Treasury. At that time
Mike Nichols was chairman of the Association for Project Management. We met when he
chaired the joint working party of the Institution of Civil Engineers and the Institute and
Faculty of Actuaries that I served on which produced the RAMP Risk Analysis and Manage- ment of Projects (1998 and 2005) guides. These Highways Agency studies and the RAMP guides also helped to shape Chapters 3 and 7.
From 2010 until 2013 I provided advice to the UK Ministry of Defence (MoD) on
appropriate frameworks for justifying high levels of expenditure on preventative and miti-
gating measures for low probability non- conventional weapon attacks on troops, a form
of analysis also relevant to terrorist activities. A generalisation of the framework developed
underpins Chapter 9, along with earlier work on strategic approaches to safety for Railtrack.
The book How to Manage Project Opportunity and Risk: Why Uncertainty Management Can Be a Much Better Approach than Risk Management (Chapman and Ward, 2011) was the extensively rewritten and retitled third edition of Project Risk Management: Processes, Techniques and Insights (Chapman and Ward, 1997, 2003). The 1997 first edition was a critically acclaimed modest best seller, with roughly a third of its sales in Europe, a third
in North America, and a third in the rest of the world, with Chinese and Greek language
versions. A significant evolution in perspective was involved in the 2003 and 2011 editions.
This book’s Chapter 7 continues the evolution outlined by Chapman and Ward (2011),
developing it from a broader perspective. The rest of this book continues the same evolu-
tion and clarification in terms of a scope which has been further broadened to include all
aspects of planning needing effective integrated treatment within private and public sector
commercial organisations. Its focus is on practice in a rapidly changing post- 2020 world. Its
evolution has been the focus of my professional activities over the last few years.
Promoting the use of enlightened planning ideas in any systematic simplicity form its
users find appropriate in whatever ways I can is central to my future professional goals.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Dale Cooper reviewed the first draft of this book, and many useful detailed changes
throughout this current draft, plus several significant changes in the approach to fram-
ing and presenting key concepts flowed directly from his comments. Particularly valuable
was clarification of areas where there were important differences of opinion, forcing me
to test and develop my understanding, resulting in important new insights in addition to
significantly better explanations. Dale and I worked together as consultants as well as joint
authors in the same academic department for many years in the 1980s and 1990s, before
Dale returned to Australia and founded Broadleaf, his consultancy company. Working with
Dale was always a very enlightening and enjoyable experience, and we have kept in touch
as good friends.
Stephen Ward reviewed the second draft, which incorporated revisions to respond to
Dale’s suggestions. Because Stephen and I have continued to work together on joint and
individual publications since the 1980s, and remained close academic colleagues as well
as good friends, our ongoing differences of opinion on key issues are minimal. However,
Stephen has always brought a highly constructive and candid critical perspective to bear on
anything I write which he has co- authored or reviewed, and many of the ideas used or devel-
oped in this book were initiated by Stephen. One simple illustrative example is the redefined
KISS mantra (keep it simple systematically). But many of the sophisticated underlying con-
cepts which are central to the whole of this book are also Stephen’s creations. The third draft
was completely different from the second. It involved a much more coherent and productive
structure, plus other significant changes in approach, as a direct result of Stephen’s feedback
on the second draft.
Stephen then provided further crucial feedback on the third draft. The key changes this
time were focused on making the ideas more accessible for the diverse set of target readers.
However, the changes were significant, including further restructuring, and further insights
were also important, a spinoff from the focus on communicating ideas.
Paul Thornton also provided very helpful overview comments on the third draft, followed
by insightful detailed comments which further shaped the current draft to a significant
extent. Paul was a founding partner of French Thornton, and his broad experience as a
management consultant operating at the board level was of immense value in terms of sharp-
ening the focus on what matters most for both ‘directors’ and ‘experts’ in a broad range
of organisations and areas of analysis. Our friendship was initially triggered by Paul being
my mentor in the early 1990s, when I became president- elect of the Operational Research
Society (ORS). The ORS had a tradition of ‘practitioners’ alternated with ‘academics’ as
presidents, with two years as a president preceded by a year as president- elect, followed by
a year as immediate past president, and Paul was my practitioner mentor within this cycle.
He revisited a different version of the practitioner mentor role to great effect for this book.
Acknowledgements
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
xliv Acknowledgements
A number of the more than a thousand participants on the IBM Forum 2 programme
provided insights used directly in Chapter 6. For more than 20 years I have been teaching a
course at the University of Southampton on Project Risk Management which now attracts
more than a hundred students per year from more than ten MSc programmes. In the last
few years several students provided ideas in their assessment essays which were incorporated
directly in this book by the completion of the fourth draft. Post- experience in- house and
public course participants, as well as university students, have helped to shape my thinking
since I started teaching, acquiring new insights from ‘students’ with practical experience
being one of the many positive features of teaching which I particularly enjoy. International
Project Management Association (IPMA) advanced training ‘master classes’ in Milan and
Copenhagen are part of the ‘plot’ of two Part 2 tales, and my IPMA course at the end of
2017 in Copenhagen was one of the most stimulating to date.
Colleagues, friends and family have commented on the fourth draft and contributed to
this final fifth draft in a number of ways, including the thoughtful gift or loan of books
which proved useful. My younger son Andy and our joint friend Stephen Cresswell need
specific acknowledgement for early support of various kinds. Stephen Ward was yet again
enormously helpful, with insightful ‘less is more’ pruning, material re- sequencing and other
editorial suggestions.
All colleagues providing contributions to the ‘Comments by other colleagues’ section
need a special thank you for their feedback and encouragement as well as helping to per-
suade you to rise to the challenge of reading this book.
Many of the colleagues I have worked with employed by client organisations as well as
consulting organisations have played a crucial role in the ongoing evolution of the toolset
for planning discussed in this book. Some are mentioned, but all sources of assistance not
identified directly are gratefully acknowledged. Only the misconceptions and mistakes are
entirely mine.
Amy Laurens, my senior editor at Routledge, provided important feedback which shaped
this book’s final form, and both Amy and her senior editorial assistant Alex Atkinson
made the process of completing the book as painless as possible. Their support was greatly
appreciated.
I am very grateful for permission from John Wiley and Sons to use figures and tables from
How to Manage Project Opportunity and Risk (Chapman and Ward, 2011), some directly and some with various degrees of adaptation and development.
This book is dedicated to my wife, Jean. Jean died in 2015, after living more than two
years as fully as she could whilst coping with palliative treatment for cancer. For 50 years Jean
was the centre of my universe. She is still the star that I steer by. Jean set her own agendas
with clarity and pursued them with determination, but she always instinctively understood
what really mattered to other people and responded with empathy and kindness. Jean’s
serious passions included our sons, daughters- in- law, and grandchildren; her many friends;
medical practice and medical education; music; art; and archaeology, but management in
a management sciences sense was never even a remote interest. Asking Jean to read a draft
of this book or discuss technical content issues was never considered. But she had a deep
understanding of what motivated this book, and she inspired and supported its evolution in
crucial ways, as well as playing a central role in shaping the perspective which underlies it.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
1 Why planning is usually vital but often difficult and frequently inept
Formal planning and related informal planning
Planning application areas
What enlightened planning (EP) means at an overview level
Dealing with process design choices
The simplicity dilemma and the enlightened simplicity response
The framing assumptions mantra
Stealth assumptions which matter
Confronting the terminological quagmire
Risk efficiency, clarity efficiency and opportunity efficiency
Framing multiple objective trade- off approaches
Stochastic modelling framework choices
Risk management failures as part of a very complex mess
The good or bad behaviours and practices mantra
Two further enlightened planning mantras
Strategic clarity and tactical clarity
Systematic simplicity as a composite concept
A summary with linked inferences
2 A ‘universal planning uncertainty and complexity management process’ (UP)
A UP overview
Process initiation and termination
Process phase structure
Capability- culture concepts
The ‘closure with completeness’ concept
UP generality in principle and in practice
The provenance of the UP concept
Further clarifying what a UP is and is not from an EP perspective
3 Low to high clarity approaches and the ‘estimation- efficiency spectrum’
A minimum clarity approach to estimation
A two scenario HAT example
More than two scenarios to ‘refine’ an estimate One or more ‘condition’ scenarios
Parametrically defined probability distribution special cases
Part 1
Foundations
Chapters 1 to 4 contents indicating sections within chapters
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
2 Foundations
The synthesis underlying a HAT approach
Systematic decomposition of multiple sources of uncertainty
Higher or lower levels of clarity when appropriate
Clarity efficient boundary concepts
Risk efficiency – with avoiding risk inefficiency as the focus
Risk efficiency – with enlightened caution as the focus
Opportunity efficiency – with enlightened gambles as the focus
Opportunity efficiency – with enlightened prudence as the focus
Interim consolidation of the opportunity efficiency concept
The importance of seeking both specific and general responses
More interim consolidation of the opportunity efficiency concept
4 Confronting challenging complexities usually needing more clarity
Sensitivity diagrams as portals to further opportunities
Source- response diagrams as portals for further clarity
Consolidating unbiased parameter estimation issues
A general conceptual framework for multiple objectives Seeking effective low clarity views of trade- offs
Simple but effective strategies for dealing with multiple objectives
A very dangerous myth and a closely coupled bias
A multiple objectives approach to planning the planning
A key strategy for dealing with multiple objectives
Informal treatment of all objectives not addressed formally
A ‘menu view’ of decision- making process goals
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
1 Why planning is usually vital but often difficult and frequently inept
This chapter begins the process of clarifying why planning is usually vital but often difficult
and frequently inept. Several further goals are also pursued throughout this chapter. The
central further goal is providing an overview of key foundation level concepts, using a lay-
ered approach, with each layer building on earlier discussion. Another important related goal
is beginning to clarify why some significant departures from common practice planning are
needed, including aspects of common practice which are widely believed to be good practice.
The next two sections outline what ‘planning’ means in this book. A third section explores
key implications of ‘enlightened planning’ (EP) at an overview level. This is followed by a
series of sections introducing specific defining features of enlightened planning. The final
section provides a summary with linked inferences. By the close of this chapter you should
have an overview understanding of why planning is usually vital but often difficult and fre-
quently inept, what is implied by enlightened planning terms like ‘strategic clarity’, and why
it may be worthwhile investing the time to read the rest of this book if you already are, or
you would like to become, an effective ‘reflective practitioner’ with an interest in approaches
to planning and associated management decision making which keep it simple systematically.
Formal planning and related informal planning
‘Plans’ sometimes begin as very simple descriptions of a concept or intent. Plans may evolve
into explicit base plans (like an aspirational target ‘plan A’ – what an organisation would like
to do assuming there are no new significant opportunities or big problems) plus contingency
plans (‘plan B’, ‘plan C’, and so on – a set of alternative plans to be used if circumstances arise
which make ‘plan A’ inappropriate). Base plans plus contingency plans usually need to include
ways to capitalise on possible good luck as well as possible bad luck. Contingency plans fre-
quently have to address ambiguous scenarios involving ‘unknown unknowns’ of the ‘any other
unspecified major opportunity or problem’ variety. ‘Plan A’ may need refinement, relabelling
it as A1, and then replacing A1 with A2 followed by A3 to incorporate proactive adjustment of
the base plan to better cope with uncertainty prior to plan implementation. One or more con-
tingent ‘exit plans’ may be prudent, as well as prior plan B and plan C contingency planning.
Sometimes plans are usefully described as ‘policy’, in the sense that plans are broadly
defined strategies for making specific kinds of decisions rather than plans for use on a single
occasion. Sometimes plans are usefully described as ‘process’, in the sense that plans are a
flexible process for implementing policy. Sometimes plans are usefully defined as ‘proce-
dures’ or ‘protocols’, in the sense that plans are sets of rules for implementing policy with
minimal room for creative interpretation normally assumed. Plans may address activities,
resources, designs, operating policies, strategies, operational rules, formal contracts and
informal relationships.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
4 Foundations
The basic role of all the planning addressed by this book is effective and efficient approaches
to management decision making in ‘commercial organisations’. Commercial organisations
are broadly defined to include all product and service providers who charge a price for what
they provide. The example organisations considered in Part 2 span a wide range of examples,
but a much wider range of application areas and planning concerns is within the scope of the
approaches explored in this book.
Planning may involve significant amounts of underlying analysis. In some contexts, you
and many others may regard ‘analysts’ as a more appropriate term than ‘planners’ for those
doing the work referred to as planning in this book. Furthermore, many of the people
involved in planning whom this book addresses as key target readers may prefer to see
themselves as ‘managers’, not ‘planners’ or ‘analysts’, with planners and analysts reporting
to them.
Target readers may serve their organisations in a wide variety of roles, as project manag-
ers, marketing managers, corporate strategy managers, or managing directors, for example.
A key premise this book is based on is all target readers need to be able to communicate
effectively and efficiently about planning, using a shared common language and shared key
assumptions about other important concerns.
You can interpret the word ‘planning’ in plain English terms, along with most other
words used in this book. However, to ‘keep it simple’ this book’s default interpretation of
planning is formal planning, including all associated formal analysis. The roles of formality when planning include ensuring that all the people involved in creat-
ing and shaping the plans make explicit what they have in mind. Associated purposes include
testing assumptions and conveying information to other people who need to understand
the implications of the plans in various ways. These other people may need to understand
the plans because of the need for mutual support when ‘component teams’ of a ‘big team’
adjust their component plans and actions to the plans of other component teams. This is a
key contingency planning issue within a broader concern for coordination and collaboration
as part of effective and efficient teamwork. Other specific reasons for shared understanding
relevant to some players may include their role in helping to execute the plans, their role in
helping to test the plans, and their role in approving the plans. Controlling the relationship
between what planners had in mind initially and evolving development and realisation of the
plans may also be a crucial role for formality. Ongoing learning from experience and more
general corporate knowledge capture may be further reasons of considerable importance for
employing formal planning rather than relying on informal planning.
Assumed ‘non- players’ may become relevant players if they are not prepared to tolerate
the plans – a potentially serious risk. But those initially assumed to be in opposition to or
indifferent to plans may become important allies if mutual opportunities are identified – a
potentially important opportunity.
Making specific aspects of planning formal involves crucially important implications for
associated informal planning. Sometimes these implications are just mentioned in this book,
sometimes they are developed immediately, and sometimes they are developed later at a
more convenient point in the overall dialogue.
All relevant planning concerns must be dealt with via one of three options:
1 formal planning,
2 informal planning, and
3 being ignored.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Why planning is usually vital 5
Clarity about which of these three options is being used to deal with each category and area
of planning, as well as ensuring appropriate option choices, is a central aspect of the ‘strate-
gic clarity’ concerns which formal planning has to address.
Sometimes informal planning of some aspects is more appropriate than formal planning,
and where the line is drawn between formal and informal treatment is often crucial. Usually
effective understanding of informal planning using more sophisticated structures and more
detail is a crucial basis for effective formal planning. All formal planning certainly needs
underpinning by background informal planning. Almost invariably effectively implementing
formal plans involves further detailed informal planning plus creative leaps and real- time
responses to unfolding events. ‘Constructive insubordination’ can be crucial – departing
from prescribed formal plans which are inappropriate, even in military or emergency service
contexts, where disobeying orders or protocols is generally discouraged in fairly harsh terms.
Any holistic formal planning process has to integrate formal and informal components
and anticipate how best to encourage creative implementation in addition to creativity
at the outset and creativity throughout the planning process. Deciding whether for-
mal or informal treatment is the most appropriate option raises complex concerns, and
simply ignoring planning issues raises quite different concerns which also have complex
implications. The core issues include ‘capability- culture’ considerations involving the
capabilities provided by the relevant players and systems and the culture these capabilities
are embedded in.
Two ‘modes’ of planning are distinguished: ‘top down’ and ‘bottom up’. In the top- down
mode ‘goals’ at the current decision- making level are the starting point, asking questions
like ‘What are we trying to achieve, and what matters most?’ In the bottom- up model the
starting point is decomposing what we think needs to be done to get to where we think we want to go, asking questions like ‘How are we going to do it, how long is it going to take,
and how much is it going to cost?’ The use of ‘we think’ implies we may change our mind. Sometimes ‘the top’ needs moving up or ‘the bottom’ needs moving down in terms of the
organisational hierarchy or scope of analysis, as part of considered revisions to the planned
approach through an iterative process.
Multiple iterations are a central process feature of many of the most effective and efficient
formal approaches to planning. Planned iterations can be a cost- effective feature of immense
importance, built into processes with this in mind. But unplanned iterations may prove
essential because of earlier errors of judgement. Unplanned iterations may be very costly.
This kind of planned plus unplanned iterative process is often inherently complex. It is
always important to keep planning processes as simple as possible for all those involved.
However, oversimplification in the ‘wrong way’ can have seriously dysfunctional implica-
tions, and how to simplify in the ‘right way’ may not be obvious unless the approach to
simplicity employed is well informed and systematic.
Planning application areas
To keep this book’s discussion as simple as possible all relevant planning is decomposed into
three very broad application areas:
1 operations management,
2 project management,
3 corporate management.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
6 Foundations
These three application areas need separable treatment because significant differences in
approach need attention, but they are interdependent in important ways.
Each of these three application areas can be further decomposed, and often this is impor-
tant in practice. However, the greater the level of decomposition, the more difficult and
crucial treating interdependency issues becomes, and the importance of commonalities may
be lost.
‘Operations management’ is sometimes interpreted as the day- to- day management of
ongoing operations with a focus on short- term planning for the physical systems used by
an organisation plus input supply chains. However, in this book operations management
and all embedded ‘operations planning’ also addresses the ongoing operation of all other
systems and resources, including output marketing chains. Furthermore, generating cor-
porate strategy propositions to address future operations concerns is a crucial aspect of the
interdependence between operations and corporate planning. Using this broad interpreta-
tion of operations planning, all Part 2 chapters involve a central concern for different kinds
and levels of operations planning, with key roles for a traditional ‘Operations Director’ in
Chapters 7, 8 and 9.
‘Corporate management’ effort and associated ‘corporate planning’ may be focused on
formulating corporate strategy in a top- down planning mode in some organisations. For
some people, this kind of strategic planning is the focus of all corporate planning. But in
this book corporate management and all embedded corporate planning includes five aspects:
1 Top- down corporate planning, which starts by clarifying the corporate mission or vision
in terms of goals and then moves on to strategic planning top- down to achieve those
goals effectively and efficiently.
2 Corporate planning, which involves the strategic scope of operations and project plan-
ning intersecting with the rest of corporate planning.
3 Planning for all corporate resources and capabilities needing attention which may not
be covered fully, if at all, by operations planning or project planning, including corpo-
rate information systems, corporate finance and human resources functions.
4 Dealing with all relevant interdependency concerns not picked up elsewhere.
5 Effective corporate governance addressing everything that matters, including all rel-
evant capability- culture concerns.
Consideration of corporate planning in a top- down sense and its integration with other
corporate planning is deferred until Chapter 8 so that we can build on the earlier discussion
of component concerns. But corporate governance, as well as lower levels of governance, is
directly relevant to all Part 2 chapters, as are interdependencies.
‘Project management’ in this book can be interpreted as the management of change in a
very general sense. ‘Project planning’ led by a ‘Projects Director’ and all comparable project
planning requiring effective integration with both operations and corporate management is
addressed explicitly in detail for the first time in Chapter 7. But the approach developed in
Chapter 7 underlies all the other chapters. Chapters 3 and 4 use examples primarily drawn
from project planning contexts, and the core ideas they illustrate are central to this book as
a whole. An immediate consequence is project planning is usefully explored in a little more
detail now.
Project planning is often given a relatively narrow interpretation which is focused on
delivering ‘assets’ which are physical systems. This is a convenient focus for the con-
texts addressed directly in Chapter 7 and most Chapter 3 and 4 examples. However, this
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Why planning is usually vital 7
discussion requires an interpretation in terms of a much broader scope for project planning,
a scope which embraces delivering any kind of change, with a linked very broad view of the
assets a project may deliver. For example, the asset delivered may be a change in corporate
understanding resulting from a new planning process which helps to drive associated corpo-
rate culture changes, illustrated in Chapter 6.
Sometimes it is very useful to distinguish ‘programmes’ and ‘portfolios’ (of projects or
programmes) from simpler ‘project’ concepts. This book assumes this will be done when
relevant. But it sticks to a simple ‘project planning’ label most of the time, with a broad
interpretation of projects which includes all programme and project portfolio concepts.
Project planning as discussed in Chapter 7 involves four key frameworks which this book
refers to as ‘the four Fs’:
1 a project (asset) lifecycle framework,
2 a seven Ws framework,
3 a goals– plans relationships framework, and
4 a process framework.
The four Fs terminology provides a useful label or ‘handle’ for a key conceptual structure, a
handy contraction of ‘four key conceptual and operational frameworks’. Each of these four frameworks is a ‘gateway concept’, and the way the complete set of
four frameworks which make up the four Fs work together in an integrated manner is an
example of a higher level form of composite gateway concept. In the language of ‘learning
and teaching’ a gateway concept is a portal facilitating entry into a way of thinking about
an important set of lower level basic concepts and operational tools which helps to struc-
ture our overall understanding and make that understanding operational (Biggs and Tang,
2011). In this particular case the four Fs concept is a portal for understanding key aspects
of project planning as an integrated interdependent coherent whole in the sense that it is
central to understanding how to approach all aspects of project planning. Each of the four
component frameworks is itself a portal for understanding a range of concepts and issues
with key interdependencies.
Project lifecycle frameworks are usefully defined in terms of the asset provided by the
change involved, including use of the asset until disposal by the project owners. However, as
an illustration of the practical complications needing effective resolution, everyone involved
has to accommodate the implications of contractors often favouring a different view of pro-
ject lifecycles than their clients. A ‘conceptualisation’ stage followed by three more stages is
a simple common practice portrayal for both clients and contractors. The other three stages
viewed from a client’s perspective are usually ‘planning’, ‘execution and delivery’, and ‘uti-
lisation’. But contractors often use planning, execution, and delivery, dropping the utilisa-
tion stage if this does not involve them and decomposing ‘execution and delivery’ because
interfacing with different groups of people with different objectives is involved.
These common practice four stage project lifecycle portrayals are significantly elaborated
for two basic reasons. One reason is recognising the crucial role of governance and captur-
ing corporate learning at ‘gateway stages’, a further lower level ‘gateway concept’ in several
senses. The other reason is recognising the crucial significance of different people with dif-
ferent concerns being involved as the initial concept evolves into a planned and delivered
functioning asset and that asset is then employed.
The seven Ws framework is a generalisation of Rudyard Kipling’s ‘six honest serving
men’. It begins with the ‘who’ (the parties involved), the ‘why’ (the objectives of all relevant
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
8 Foundations
parties) and the ‘what’ (the design of the asset being created). All the ‘seven Ws’ are formally
developed and linked via an influence diagram to different kinds of plans, building on the
lifecycle framework.
A goals– plans relationships framework has to be built on the project lifecycle and the
seven Ws frameworks. Its purpose is to develop clarity about the relationships between the
very broad top- level corporate goals being pursued by an organisation and the project plans
being used to achieve selected aspects of those goals. For example, the national railway
company central to Chapter 9 is associated with a corporate mission statement like ‘The
best railway travelling experience in Europe, delivered to our passengers by railway staff
who understand and care about what good service means, including comfort, safety, punc-
tuality and convenience at a cost which is fair and good value’. ‘The best railway travelling
experience in Europe’ might be interpreted as its central composite corporate goal, with
everything that follows viewed as interdependent component central goals which need a bal-
anced approach bearing in mind the different preferences of different parties. These central
goals might be associated with a list of attributes like travelling comfort, punctuality, safety
and reasonable fares. A central set of criteria associated with overall costs might aggregate
overall expected annual costs and associated cost risk involving high- side variations from
expected values, assuming that realised costs drive fares and corporate profits. All the other
goals might have various metrics. Assumptions about the central ‘goals’, their component
‘attributes’, and the bottom level ‘criteria’ relationships with plans collectively define the
overall goals– plans relationships of interest for project plans.
Goals, attributes and criteria can all be referred to as ‘objectives’ in a hierarchy involv-
ing crucial and complex interdependencies. You may prefer alternative terminology, but
all those involved in any given organisation need clarity about well- defined relationships
between the plans of interest, relevant top level central goals, all associated attributes, and
component criteria. This book assumes that a framework for ‘goals– plans relationships’ cov-
ers the whole range and all relevant interdependencies, with the goals, attributes and criteria
hierarchy distinctions just made used as convenient working assumptions.
To deliver formal project planning as a fully integrated approach, a project planning pro-
cess framework is the final component of the four Fs. This process framework has to use an
iterative phase structure which builds on the frameworks provided by the project lifecycle
stages, the seven Ws, and the goals– plans relationships. This fourth framework and its inter-
actions with the first three are the central concern of Chapter 7, with implications for all
Part 2 chapters.
Currently you may not think about project planning as the management of change in
terms of explicitly using this four Fs concept. But by the end of this book, the case for
explicit corporate recognition of some variant of this perspective and the four Fs framework
should be clear. Any organisation not using one common explicit four Fs concept is using
one or more implicit variants, and incompatibilities associated with different versions and
missing key features may be the cause of important planning failures.
Corporate and operations planning contexts need variants of the project planning
four Fs concept which is both broadly comparable and compatible. Compatibility is
required to address interdependencies among corporate, operations and project plan-
ning. Variants of the planning horizon aspects of the project lifecycle, the seven Ws and
the goals– plans relationships framework concepts are used throughout Part 2, within a
common overall approach to decision- making processes. By the end of this book the case
for explicit corporate recognition of this role for the four Fs beyond project planning
should be clear.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Why planning is usually vital 9
If you now perceive planning for the purposes of this book from the perspective outlined
so far, it should be reasonably clear already why some aspects of planning in a formal, struc-
tured manner are usually vital, and why a comprehensive understanding of ‘what needs to
be done’ is not always obvious to all those responsible for planning approach choices or per-
ceived in the same way by everybody involved. You may also see why a very broadly defined
approach to planning is a useful perspective on all associated management decision making.
The rest of this book builds on this overview of planning to clarify inherent planning
difficulties and to lay the foundations for exploring effective practical ways to resolve these
difficulties.
What enlightened planning (EP) means at an overview level
The approach to planning advocated in this book is referred to as ‘enlightened planning’,
contracted to ‘EP’.
The scope of EP includes creating, enhancing, shaping, testing, interpreting and imple-
menting all relevant plans from operations, project and corporate management perspectives
using management decision- making frameworks which include process management com-
ponents building upon variants of the four Fs.
‘Enlightened’ is used in a plain English sense. But for the purposes of this book the term
‘enlightened planning’ involves an effective, efficient, robust, flexible and creative approach
to a long list of interrelated planning concerns. These concerns include:
1 obtaining a shared understanding of all relevant uncertainty to the extent this is both possible and useful, with no significant blind spots;
2 obtaining a shared understanding of underlying complexity to the extent this is useful,
with a well- founded overview appreciation of what is not understood;
3 creating a rich set of opportunities and seizing all appropriate opportunities;
4 taking appropriate risk but avoiding imprudent or unnecessary risk;
5 making appropriate trade- offs between all relevant competing objectives;
6 integrating an organisation’s operations, project, and corporate planning roles;
7 collaborating with other organisations whenever this is valuable;
8 making all assumptions as explicit, internally coherent and consistent as possible;
9 ensuring an appropriately shared common understanding of all assumptions that really
matter; and
10 facilitating clarity about which parties are responsible if key assumptions do not hold.
This list implies EP has very ambitious goals, but if any goals of importance to you or
organisations of interest to you have been missed, they can be added to help you shape any
organisation’s own variant of EP. The EP concept has a flexible design to facilitate variant
development in different ways for organisations with different concerns.
The term ‘enlightened planning’ and its contraction to EP are useful for a number of rea-
sons. Your immediate and very reasonable response to this assertion might be ‘this particular
choice of label is too pretentious or too ambitious’. To dampen this concern, enlightened planning was chosen to characterise a plausible but deliberately ambitious ‘aspirational tar-
get’ which is worth seeking. Plausible aspirational targets can be very useful if an opportu-
nity management emphasis is important, and this feature of the EP approach seems worth
emphasis to me. By the end of this book the value of planning to meet ambitious aspira-
tional targets as well as making decisions based on realistic expectations with appropriate
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
10 Foundations
contingencies defining commitments and suitable contingency plans ready to meet crisis
and disaster should be clear. However, if you are not convinced by the choice of enlightened
planning as a label for any reason, change it to something you prefer once you are confident
you understand what EP as I currently see it is about, and how your version might use-
fully emphasise different concerns. There are certainly very good reasons for using different
labels in some contexts. The ‘enhanced planning’ alternative label preserving the EP con-
traction used in Chapter 7 is just one example of a number of suitable options.
The ‘tools’ in the ‘toolset’ for enlightened planning provided by this book range from
broad principles and general processes to specific graphical formats, from conceptual
devices to operational techniques, covering most aspects of formal planning. The associ-
ated requisite ‘skill set’ includes supporting craft skills, which can be crucial. The ‘mindset’
issues which EP has to address have a very broad nature. The toolset issues are the focus of
this book, but skill set and mindset concerns will be touched on and sometimes developed
when relevant.
EP uses a planning perspective and the terminology of planning to address concerns often
considered using other terminology and different conceptual frameworks, such as decision
making, decision analysis, decision theory, decision support, portfolio theory, risk manage-
ment, uncertainty management, problem solving (or resolving or dissolving), and the use of
various approaches to ‘messes’ (used in both the plain English sense and the technical sense
of ‘systems of interconnected problems’) via various forms of systems analysis.
A basic feature of any formal planning is the nature of the assumptions employed in the
planning process. An enlightened approach to planning formally distinguishes two basic
types of assumptions: framing assumptions and working assumptions.
This distinction helps to clarify the process of testing assumptions, which is important
because the ‘right kind’ of simplifying assumption is an opportunity and the ‘wrong kind’ of
simplifying assumption is a risk.
From an EP perspective,
‘Framing assumptions’ define our perspective. ‘Working assumptions’ are assumptions of convenience, selected to make a complex reality tractable in an effective and efficient manner.
Framing assumptions for EP purposes include habitual working assumptions which we do not test because, for all practical purposes, any working assumptions which have an effect
comparable to framing assumptions which are defined in more fundamental terms are indis-
tinguishable in their effect.
One example of a framing assumption from an EP perspective is ‘all organisations need
explicit specific forms for each component of the four Fs concept, defined by suitable work-
ing assumptions’. This is not likely to prove contentious.
Another example of an EP framing assumption is ‘uncertainty’ is appropriately defined
simply as a ‘lack of certainty’ in ‘nominal definition’ terms. This example is contentious
for some people. It will be explored shortly, to couple this simple nominal definition to
an underlying elaborated definition which provides greater clarity at a framing assumption
level. Associated working assumptions will also be explored to further enhance overall clarity
about the meaning of the word ‘uncertainty’.
One example of a working assumption is any particular operational form developed by any
specific organisation for the general project lifecycle concept which was explored briefly in
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Why planning is usually vital 11
the last section as part of the four Fs concept for project planning. Many other examples will
be developed in every chapter.
The EP framing assumption concept and its relationship with working assumptions are
explored throughout this chapter and further developed later. The core role of framing
assumptions is the facilitation of an enlightened approach to choosing working assumptions,
including a robust basis for effectively testing both virtually permanent ongoing working
assumptions and very provisional initial working assumptions at the start of an analysis.
At a conceptual level there may sometimes be a case for distinguishing framing assump-
tions which define fundamental beliefs from working assumptions we habitually use without
testing them. We may habitually use assumptions we know are not strictly true without
testing them because we trust them, because we do not know how to test them, or because
it has never occurred to us there was any need to test them. To ‘keep it simple systemati-
cally’ the EP framing assumption concept explicitly avoids exploring these distinctions, but
if they are of interest to you whatever complexity is deemed useful can be added, as for all
EP framing assumptions.
In Part 1 the focus is framing assumptions which shape basic conceptual, process and
other operational toolset choices. Part 2 tales also explore the implications of organisa-
tions using inappropriate framing assumptions in a much broader sense, involving mindset
assumptions as well as toolset assumptions. For example, the general manager of the organi-
sation which Chapter 5 is centred on begins with the simple long- standing personal mindset
assumption that he is running a family- owned manufacturing company as well as very simple
toolset assumptions, but when he is stimulated into beginning to adopt a less restrictive
mindset and toolset framing assumption perspective, he and the rest of his management
team start to explore significant organisational transformation possibilities. Chapters 8 and
9 address radically rethinking the nature of both mindset and toolset framing assumptions
which are comparatively complex. Other Part 2 chapters also address what might be viewed
as habitual untested working assumptions which include mindset and toolset assumptions
with inappropriate framing assumption implications.
EP involves a new way of looking at a rich variety of mainstream approaches, but the
basis of EP is not new – its considerable ambitions would not be credible if it were. From
the outset, this book acknowledges key contributions to the basis of EP when this is useful
for most readers or particularly helpful for some readers with directly relevant background.
Features of EP explored in this book which are new include significantly enhanced vis-
ibility of:
1 controversial framing assumptions about concepts, processes, and other toolset compo-
nents which need clear understanding for any variant of EP or any coherent alternative
to EP which you or any organisations of interest to you may subscribe to,
2 interconnected fundamental problems with some features of ‘received wisdom’ (con-
ventional thinking and common practice) which need rejecting by any EP variant,
3 useful relationships between decision support processes for managing all relevant uncer-
tainty (framed in terms of associated opportunity and risk plus underlying complexity)
which are ‘universal’ (very general) and other decision support processes which are
‘specific’ (to an organisation or approach or other context features to various degrees),
4 the role of teamwork and broader collaboration considerations as an integral part of the
overall framework for approaches to planning, and
5 the role of capability and culture considerations as part of the relevant context concerns
plus further context issues.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
12 Foundations
Seeing your current toolset for planning from an EP perspective provides a basis for reflect-
ing on your current approaches as well as enhancing your mindset and skill set. Seeing other
people’s approaches from an EP perspective will inform well- founded critiques of received
wisdom and effective synthesis across different ‘schools of thought’. Serving both these
purposes are central missions for this book, and understanding some key areas of contention
may be very important for you. The nature of the cases against particular approaches which
are currently entrenched in your world view may need particular clarification. Assistance
with your critique of received wisdom where its failures are a serious concern for you may
also be important. For all target readers these concerns are addressed as directly as possible.
In some areas this means confronting concepts you may initially see as beyond your usual
comfort zone in terms of ‘how to do it’ detail. However, you should see the point reason-
ably quickly, and as far as possible what any target readers could reasonably see as unneces-
sary detail has been minimised.
Dealing with process design choices
A general planning difficulty is ‘How do we choose or design and develop the most appro-
priate process for planning in any given context?’ At all levels of decision making in all
organisations these process choice concerns require attention, and a systematic approach
which clarifies the issues associated with keeping it simple in the right way can make the
process choices both more effective and easier. These choices are crucial to an organisation’s
success, they are not easy, and there is a lot of scope for errors of judgement if a clear and
coherent framework for making them is not available to everyone involved.
These choices are also highly controversial, with many different competing claims about
both what needs to be done and how to do it. Dealing with the considerable difficulties
involved in these important and controversial process design choices is a central EP concern.
Three closely coupled EP framing assumptions are the following:
All but the simplest of planning should be based on systematic exploratory processes. All planning processes should use a systematic toolset to clarify uncertainty and complexity, building suitable variants of this toolset into the resulting plans as and when this is useful. Opportunity is always a central concern, but risk is often important and sometimes crucial.
These framing assumptions underlie the most general approach to choosing or designing
and developing a suitable planning process which I have found useful in practice, includ-
ing testing the validity of alternative process options as part of the process. It is outlined
and its provenance briefly explored in Chapter 2. It is referred to as a ‘universal planning uncertainty and complexity management process’, usually contracted to ‘universal process’ or ‘UP’. All following chapters build on this UP concept, a key EP tool with both conceptual
and operational roles.
One key role for the UP concept is providing a default planning process if it is not clear
what other approach might be appropriate, illustrated by the tale of Chapter 5. A second
key role is providing a process for building a planning process for a specific context, referred
to as a ‘specific planning uncertainty and complexity management process’, contracted to ‘specific process’ or ‘SP’. This second role is initially illustrated by the tale of Chapter 6.
A third key role is providing a process for adapting what is now commonly called a ‘generic
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Why planning is usually vital 13
process’ (a process developed for general use). This third role is illustrated by the tale of
Chapter 7. Building on all these roles, a UP can be used as a basis for making comparisons
between potentially suitable alternative planning processes and component models, testing
relative effectiveness and efficiency, as illustrated throughout Part 2. In a related but further
additional role, Chapter 8 uses a UP indirectly to define an appropriate overall framework
for corporate planning. Chapter 9 then extends this Chapter 8 role to consider dealing with
low- probability high- impact incident scenarios plus a related set of less serious incident sce-
narios within an appropriate corporate planning framework.
As an aside or footnote on a minor point of detail just in case you wondered, using ‘a UP’
implies saying ‘UP’ using two words (‘u’ and ‘p’), for consistency with SP and EP. Should
you prefer one word and ‘an UP’, that alternative is obviously an option. You may prefer
alternatives for any of my EP terminology choices for good reasons, and what concepts are
called is not an important issue unless the label significantly colours the meaning or com-
mon uses of the concept in ways that really matter.
The simplicity dilemma and the enlightened simplicity response
‘Enlightened simplicity’ is at the heart of a UP and enlightened planning in general – my
current view of what Stephen Ward labelled ‘constructive simplicity’ (Ward, 1989) plus sys-
tematic approaches to using constructive simplicity which we have jointly pursued since the
1990s and others have pursued for many years using various labels. Whatever you chose to
call it, the enlightened simplicity concept has been developed to address a widely understood
and crucial ‘simplicity dilemma’. Stated simply, the simplicity dilemma is the following:
Everybody obviously wants simplicity when making decisions in planning processes, but the issues may be very uncertain and inherently extremely complex, so how do we choose the right kind of simplicity and avoid the wrong kind of simplicity?
The essence of an enlightened response to this simplicity dilemma involves seven key aspects:
1 It depends on a common joint understanding of all the relevant issues in points 2
through 7 at an appropriate level by all the parties involved, although some people may
need to understand some issues in much greater depth than others.
2 It is crucially dependent on the context.
3 It requires an explicit understanding of the working assumptions employed to reduce
complexity.
4 It involves explicitly testing all working assumptions and revising them when appropri-
ate to improve robustness whenever working assumptions may be crucial to important
choices, including decision- making process choices.
5 It requires an explicit understanding of framing assumptions which define an overall per-
spective, including the approach taken to testing working assumptions for robustness.
6 It requires effective and efficient integration of operations, project, and corporate plan-
ning roles.
7 It requires knowledge and craft skills from all relevant disciplines which have been inte-
grated into a coherent and unified whole.
You may decide to add to or rearrange these seven key aspects by the time you finish this
book, but they are a useful starting position to work with in the meantime.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
14 Foundations
Three key concerns follow from this seven- aspect view of enlightened simplicity – again
you may wish to extend or modify this list later:
1 If any organisation does not have a common understanding of the limitations of its
framing assumptions, it will fail to understand the limitations of its collective knowledge
and skills.
2 If those leading decision making use knowledge and skills from relevant disciplines
which employ incompatible framing assumptions without dealing with the implications,
an organisation’s collective perspective will lack coherence.
3 If those involved do not understand the dissonance involved in an incoherent perspec-
tive, or its implications in terms of lack of robustness, or the limitations of the collective
knowledge and skills employed, they will be surprised by the mistakes that are made,
and they will keep making the same or similar mistakes.
This book is about using enlightened simplicity for decision making in formal planning
frameworks and related informal planning frameworks so we can integrate relevant knowl-
edge and craft skills from all relevant disciplines and perspectives and learn from our experi-
ence. It is about hanging our learning and experience on a framework which is as general,
flexible, and robust as possible.
The framing assumptions mantra
A popular quote known as Occam’s razor, attributed to William of Occam (1285– 1349), is
often cited in a form like this:
No more things should be presumed to exist than are necessary.
The basis of both EP and enlightened simplicity is a version of Occam’s razor in the form of
‘the framing assumptions mantra’:
Keep your framing assumptions as general as possible, because you cannot routinely test framing assumptions for robustness, as you can and should for all working assumptions.
A mantra used in this way is a framing assumption, and the framing assumption mantra is a
defining feature of enlightened planning.
As pointed out earlier, working assumptions are assumptions of convenience. Working
assumptions are simplifications to make a complex reality tractable and easier to deal with
in an effective and efficient way. Working assumptions should be explicit, and they should
be routinely tested during EP processes to ensure that they are not misleading, with a clear
understanding that this is much easier said than done. As also pointed out earlier, framing assumptions define our overall perspective. We can
test framing assumptions by hypothesising more general framing assumptions as part of a
process of reflective review of past experience, testing these hypotheses in the future. But
in most operational decision- making contexts it is not practical to propose and test suitable
new framing assumptions in the middle of any given planning exercise.
Deliberately pre- planned development exercises for planning processes may allow exten-
sive testing of framing assumptions, and sometimes this can be a very useful approach, with
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Why planning is usually vital 15
several examples provided later. But in most circumstances, we need to do our operational
planning using a framing assumption perspective we are comfortable with, and it is often
convenient to also use explicit working assumption option sets that we are familiar with.
Because testing framing assumptions is inherently difficult, and routine testing of framing
assumptions in the middle of decision- making processes is not practical, it follows that all EP
framing assumptions should be as general and unrestrictive as we can make them by design.
Furthermore, the greater the generality achieved, the broader the scope for appropriate
working assumption choices, enhancing the scope for choosing the right kind of simplicity
while avoiding the wrong kind of simplicity.
While drafting this book I have hypothesised more general framing assumptions than
those used earlier in a number of areas to see if this might be useful, and continued this
process until further tests did not seem likely to prove productive. The extent to which
framing assumptions and related key working assumptions have or have not been tested
empirically has been clearly identified, to help you judge the validity of any advocated
framing assumptions which differ from those you currently subscribe to. You, and organi-
sations of interest to you, might wish to continue this process in some contexts. But test-
ing framing assumptions is not something anyone will have the time or inclination to do
routinely when trying to make operational decisions in practice – the time pressures will
normally prevent it.
‘Doing better next time’ based on past failures is a common and essential basis for testing
alternative framing assumptions plus key associated working assumptions. Both individuals
and organisation taking a reflective approach to learning from experience routinely test and
evolve their framing and working assumptions based on earlier misjudgements. Part of the
cost of learning by experience in this way is the pain caused by the earlier mistakes, usefully
viewed as lost opportunities to have done better. Reading books and papers or attending
courses which challenge beliefs and discussing concerns with mentors and other colleagues
can play a similar but much less painful role.
One of this book’s goals is to help you learn what you need to know to take an enlight-
ened approach to planning with as little pain as possible. However, to make effective use of
reading this book you will need to test the EP framing assumptions and associated working
assumptions used by this book against your own current framing and working assumptions
as your reading progresses, not always a straightforward or painless task. You will then need
to further test your resulting framing and working assumptions in practice, confirming the
replacement of your own earlier framing assumptions with better (more general) ones and
related working assumptions with better (more effective and efficient ones) whenever you
establish a case for doing so.
To put this in more enticing light, the ultimate goal is ‘more reward’ plus ‘more smiles’
in the ‘more pleasure’ and ‘less pain’ sense discussed using Figure 11.1 in Chapter 11, a
portrayal of systematically searching for opportunities while avoiding risks which reflects
capability and culture concerns in a holistic manner using an EP approach.
Stealth assumptions which matter
If colleagues or any other relevant players in a joint decision- making process are using
framing assumptions or dependent working assumptions which are seriously dysfunctional
because they are restrictive in inappropriate ways, something needs to be done about it as
soon as possible. The same issue arises if we discover that we are using inappropriate framing
assumptions or dependent working assumptions ourselves.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
16 Foundations
‘Stealth assumptions’ which matter are defined as framing or associated working assump-
tions which we or other people seem to be using with seriously dysfunctional implications
when adopting any approach which does not make sense from the broadest feasible EP
perspective. These stealth assumptions may be explicit, consciously adopted because of a
failure to recognise the implications of bad advice, which may be based on conventional
wisdom which is unwise. But stealth assumptions may be tacit, the result of following other
peoples’ bad practice without questioning its basis. Stealth assumptions which matter are viewed as important sources of concern in this book, serious sources of risk which need to
be avoided.
It is generally very difficult to understand some of our own deeply seated stealth assump-
tions, but we can sometimes identify them when we generalise our perspective via new
insights. This happened to me a lot when writing this book, because I was particularly
focused on generalising my framing assumptions in order to develop a coherent and holistic
framework which would help you to generalise yours. However, seeking the most general
perspective feasible has been a conscious goal for many years, and the insights sometimes
revealed are one of the ‘pleasant surprises’ life has to offer.
We cannot know what other people are actually thinking, but we can usually surmise or
hypothesise an explanation for other peoples’ behaviour in the form of stealth assumptions.
Sometimes this is very useful.
Confronting the terminological quagmire
All key words used as technical terms, and some key multiple- word technical terms, are
conceptually important in any planning framework. From an EP perspective, all words given
an important technical terminology role are usefully seen as very basic and important tools,
with interconnected operational as well as conceptual implications.
Some key words are as fundamental to EP as the UP concept and the four Fs. This is
because specific meanings for crucial words which are restrictive can limit perspectives as
much as processes or models or any other conceptual or operational tools which lack gener-
ality. Key words which are conceptually limiting are a form of seriously dysfunctional fram-
ing assumption, stealth assumptions which really matter.
Over several decades the importance of avoiding restrictive technical definitions for key
words has become more and more obvious to me and many colleagues. However, many
other people, including some close colleagues and friends, have strongly defended restrictive
views. In a few of these cases, the definitions involved are not restrictive in obvious ways,
and assuming the differences do not matter may seem reasonable. But in some cases, the
definitions adopted are so highly restrictive their numerous critics have argued vigorously
for decades that they should be abandoned.
Some people defending demonstrably inappropriate definitions insist that their restric-
tive views must be appropriate because they are endorsed (or seem to be endorsed) by
published guides and standards, some of these guides and standards being seen as highly
reputable and beyond reproach. Some of the most vigorous arguments I have been involved
in have taken place on working parties for reputable guides and standards, but none of
these guides or standards reveal this controversy, and some attempt to embrace a range
of views with a serious and regrettable level of ambiguity and inconsistency. This is also
true of other guides and standards I am aware of, although comparable controversy has
been involved in all the many cases I have discussed with contributing authors. This makes
the nature and purpose of all guides and standards controversial and contentious in ways
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Why planning is usually vital 17
which are not generally understood. This section explores some of the immediately relevant
implications for planning processes. You will find it useful to understand these implications
now, whether or not guides and standards and the technical issues they address are of any
direct ongoing interest.
The terms ‘uncertainty’, ‘opportunity’, ‘risk’ and ‘complexity’ are a core quartet, par-
ticularly in need of clarity. From an EP perspective, what is meant by these four key words
is a very good example of why being clear about framing assumptions matters, why the way
framing assumptions relate to associated working assumptions matters, and why under-
standing these issues helps to clarify the nature of the stealth assumptions which matter
associated with common practice you need to avoid.
This section explains how we can escape the terminological quagmire which common
practice is bogged down in, and why doing so is essential. Relying on guides and standards
and textbooks or papers by authors whose opinions are based on guides and standards is not
a viable route to safe ground.
The EP approach developed in this section may initially seem unnecessarily pedantic – but
the concern is keeping it as simple as possible with the flexibility and nuanced clarity which
all planners and decision makers need in practice. It is an example of the more general con-
cern for keeping it simple in the ‘right way’, avoiding keeping it simple in the ‘wrong way’.
It illustrates the value of an initial investment in capability to achieve massive ongoing pay-
offs in terms of more clarity for less effort.
Uncertainty
Uncertainty is inherent in all aspects of planning and associated management. You need
a clear understanding of how to perceive and deal with uncertainty, and organisations
need a clear shared understanding. Consider the closely coupled pair of simple nominal
definitions:
‘Uncertainty’ means ‘lack of certainty’, and ‘uncertain’ means ‘not certain’.
When discussing any aspects of ‘uncertainty’, the ‘lack of certainty’ definition has been the
simple plain English interpretation of preference for me and a number of colleagues for
many years. We have not explicitly used ‘uncertain’ means ‘not certain’ quite so much, but
its usage was always implicit because they are just two ways to define the same concept.
Over the past decade, several colleagues have explicitly rejected these ‘one- part’ lack
of certainty or underlying not certain definitions for the purpose of defining professional
guidelines on working parties I have served on. Their grounds for doing so have been
that this kind of simple definition is circular. My strenuous objections have been over-
ruled. There were no hard feelings about it, but I remained unconvinced and unrepentant,
without really understanding the nature of the underlying issues driving our differences in
opinion.
From an EP perspective the underlying problems are now clear to me, confirming that
my intuitive concerns were valid. This EP perspective did not emerge until the fourth draft
of this book, and understanding the nature of the clarification process should help you to
understand the rationale of the EP perspective now advocated.
The starting point for this clarification was linking the one- part, very simple nominal defi-
nitions provided earlier to an elaboration definition developed using dictionary definitions
as well as technical definitions, testing to see what this elaboration definition might reveal.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
18 Foundations
Standard dictionary definitions (using several recent editions of the Concise Oxford Diction- ary to get started) suggested three quite different meanings for uncertain:
1 not certainly known or knowing;
2 unreliable;
3 changeable or erratic.
Circularity is clearly involved in the one- part nominal definitions provided earlier – there is
no disagreement on this issue. However, that is also the case for the ‘not certainly known
or knowing’ dictionary meaning of ‘uncertain’. Amongst the colleagues rejecting ‘lack of
certainty’ or ‘not certain’, the common preference for ‘unknown’ or ‘not known’ misses the
extremely important nuance ‘may be partially known’ interjected by ‘not certainly’ in the
first dictionary meaning, as well as completely missing the second two dictionary meanings.
These omissions matter.
The EP position now advocated is using the one- part nominal definitions ‘lack of cer- tainty’ or ‘not certain’ as usefully simple summary components of the ‘overall definition’, but explicitly interpreting these one- part nominal definitions using a ‘three- part’ ‘elabo- ration definition’ which involves all three dictionary definitions for ‘uncertain’ so that no important dictionary meanings or nuances are lost. This makes the EP ‘overall definitions’ approach (nominal plus elaboration definitions) as general as possible.
Having established the generality of this overall definition approach for EP framing
assumption purposes, briefly consider what lies behind some alternative common restric-
tive definitions which everyone needs to avoid. There are technical definitions of ‘uncer-
tainty’ and ‘uncertain’ which insist that probabilities cannot be associated with uncertainty.
For about 50 years this idea has been explicitly rejected as inappropriately restrictive by
most well- informed experts, and an EP approach follows suit. However, a dysfunctional link
between uncertainty and a lack of ‘probabilities’ lingers in the sense that many people still
regard a lack of objective probabilities and uncertainty as closely coupled if not synonymous,
and a lot of inept planning is the direct result.
The reasons most well- informed experts now reject this connection may be clear to you
already. If not they should be clear by the end of Chapter 3. But it may be helpful for you to
understand now that arguing probabilities cannot be associated with uncertainty is usually
grounded on a classical decision analysis perspective rejected by modern decision analysis
in the 1960s, when subjective probabilities became accepted as proper probabilities which
should embed objective probabilities as a component part if relevant data was available and
worth using, but a lack of data should not preclude using probabilities defined as statements
of belief, following acknowledged experts like Howard Raiffa (1968). The economist Frank
Knight (1921) is often blamed for the initial confusion, but the classical decision analysis
framing assumptions are probably a fairer target, if blame is an issue.
From an EP perspective, decision makers can always use subjective probabilities to describe
uncertainty whenever that is helpful, but sometimes not doing so is preferable, for reasons
outlined in Chapter 3.
It is often extremely useful – arguably crucial – to use working assumptions which build
on this comprehensive overall definition of ‘uncertainty’ and ‘uncertain’ to clarify what
is involved. These working assumptions provide valuable conceptual tools with important
operational implications.
During the 1970s and 1980s I used working assumptions involving what I now see as
three explicitly envisaged ‘portrayals’ of uncertainty, and for about two decades prior to the
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Why planning is usually vital 19
fourth draft of this book, four portrayals of uncertainty seemed useful. I am now convinced
that adding a fifth portrayal is very important.
The EP perspective advocated in this book uses the framing assumption that multiple por- trayals of uncertainty are crucial. For present working assumption purposes, five portrayals of uncertainty can be operationally useful:
1 ‘event uncertainty’,
2 ‘variability uncertainty’,
3 ‘ambiguity uncertainty’,
4 ‘capability- culture uncertainty’, and
5 ‘systemic uncertainty’.
The most recent addition is number four, a need for number three emerging in the 1980s.
The ordering of the list was designed to simplify the explanations to follow.
Event uncertainty is inherent in some contexts, in the sense that ‘a machine failing or
not’ may be seen as a form of uncertainty associated with a simple event happening or not.
However, event uncertainty may also be associated with a scenario representing a complex
range of outcomes involving many attributes which is treated as a discrete event on an
outcome branch of a decision tree. For example, Chapter 9 considers a ‘catastrophic inci-
dent scenario’ associated with a railway involving a conditional expectation of 350 fatalities
(associated with a range from 200 to 1,200+), plus an associated conditional expectation
of injuries with degrees of seriousness on many scales, plus a further associated conditional
expectation of physical damage to the railway system with ongoing operational cost and lost
revenue implications, plus reputational damage and the implications of lawsuits and fines
which may lead to bankruptcy of the railway company. This incident will be treated for some purposes as a discrete event which may result from earlier decisions, but that is a working assumption, a convenient portrayal of extremely complex circumstances, not an inherent
event concept. Chapter 3 develops a range of much simpler ‘scenario’ examples of event
uncertainty using a ‘histogram and tree’ (HAT) framework to provide coherent framing and
working assumptions for an approach based on an ‘event uncertainty’ portrayal which links
it to all the other portrayals of uncertainty.
Variability uncertainty may be inherent in the sense that uncertainty about ‘good or
bad weather’ in terms of a lay- barge being able to lay an oil or gas pipeline in the North Sea
during a one month period may be viewed on a continuous variable scale over the range 0
to 31 ‘lay’ (working) days. However, variability uncertainty may also be a very useful way
of portraying uncertainty about the total cost of an offshore project – which is a mixture of
inherent variability uncertainty, event uncertainty, ambiguity uncertainty, capability- culture
uncertainty and systemic uncertainty.
Ambiguity uncertainty can be a particularly useful portrayal of uncertainty associated
with knowledge we would like to have which could be acquired at a low cost relative to the
expected cost of carrying on without this knowledge. However, it can also be an important
way to look at some uncertainty which will be partially reduced as a result of ongoing plan-
ning processes, the implementation of plans and the passage of time. Ambiguity uncertainty
is a generalisation of what some decision analysis literature approaches address via ‘the value
of information’ in ‘perfect information’ and ‘imperfect information’ forms.
Capability- culture uncertainty encompasses the uncertainty underlying all relevant
‘human error’ issues, interpreted to include the underlying basis of all associated lack of
capability and behavioural risk concerns plus relevant supporting systems concerns. Gawande
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
20 Foundations
(2011) provides an example discussed in more detail in Chapter 2 which links human capa-
bility issues to associated systems – Boeing’s approach to supporting aircrew faced with
very serious low probability incidents which uses an on- board computer- based information
system. A well- trained co- pilot can use the system to access the best advice available in a very
short period to prevent an accident. A system of this nature can cover a huge range of issues
most pilots will never encounter, and it can be kept updated from a wide range of sources.
All four of the alternative portrayals of uncertainty may be viewed within or overlap this
portrayal. This way of looking at uncertainty draws on all three of the dictionary meanings,
incorporating associated nuances as well.
Systemic uncertainty may be reasonably straightforward or exceedingly difficult to charac-
terise. At the relatively straightforward end of the spectrum, systemic uncertainty may underlie
the sum of the uncertainty associated with a sequence of activities in a project in the sense that
if one activity is delayed, a second following activity may be more likely to be delayed, and
knock- on effects including cascade effects may make the delay worse. Two somewhat differ-
ent types of dependence may be involved, one associated with the chance of initiating condi-
tions, the other associated with consequences and responses to consequences. Both might be
addressed in statistical dependence terms or in causal dependence terms. While early responses
to specific problems and general responses to sets of accumulating problems may dampen
accumulating delays, consequential positive and negative feedback loops can be complex and
difficult to disentangle. The relationships involved can become very complex very quickly if
we keep digging, even in a relatively straightforward project planning context limited to how
long a sequence of activities might take. Much more complex situations can be associated with
contexts involving concerns like macroeconomics and geopolitics.
All five portrayals of uncertainty may be involved, with important interdependencies which
systemic uncertainty might portray. Concepts involving uncertainty which may be beyond the
bounds of conventional analysis, like ‘black swans’ (Taleb, 2007), ‘exceptional uncertainty’
(Marshall, 2015) and ‘radical uncertainty’ (King, 2016), need to be associated with all five
when relevant, sometimes as components of what this book calls ‘unknown unknowns’.
Unknown unknowns is a term which I initially used in 1970s work with BP, in a way
discussed in Chapter 3. I now see ‘unknown unknowns’ as a useful EP term for uncertainty
which may be unknowable but might be knowable if we were more capable or had more time
and often is simply not worth understanding any better in the current context at this particu- lar time. This unknown unknowns kind of uncertainty might be associated with any of the
five portrayals just outlined and linked to uncertainty beyond the bounds of conventional
analysis. For example, ‘black swans’ is now a well- known term for uncertainty missed by con-
ventional thinking in a financial analysis area (Taleb, 2007), with wider implications. ‘Radical
uncertainty’ is a term formally defined to go beyond conventional economic thinking in a
book on global economics including underlying geopolitical issues by Mervyn King (King,
2016), a distinguished economist who was governor of the Bank of England from 2003 to
2013. There are many ways of looking at unknown unknown concepts, and you may have
your own preferred terminology, but it is convenient to have one unifying concept which
serves all the purposes of the unknown unknows EP concept. Forms of systemic uncertainty
which may be extremely difficult to clarify include links between capability- culture uncer-
tainty and ambiguity uncertainty – like an unknown ability to deal with non- predictable
initiating events which can turn a potential crisis into a catastrophe very quickly, the role
of feedback loops between parties to a project whose objectives are not aligned which can
lead to a complete breakdown in relationships which means everybody is in serious trouble,
and economic or geopolitical issues. Forms of uncertainty which simply may not be worth
resolving at present include the events which might disrupt the detailed plans for executing
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Why planning is usually vital 21
a project which we have not yet decided to undertake and may reject if better opportuni-
ties are identified. Dealing with some of the relatively straightforward kinds of unknown
unknown is not that difficult if we understand what needs to be done and how to do it, but
the capability required to do so effectively is not currently as widely available as it might be.
Risk/uncertainty distinctions are important in this book, with uncertainty underlying
risk. For example, ‘systemic uncertainty’ is an uncertainty management term for the uncer-
tainty underlying what some people call ‘systemic risk’. From an EP perspective, risk and
uncertainty should not be confused, for reasons we will start to explore shortly. Systemic
risk, risk associated with variability and risk associated with events are all well- known risk
management concerns, although they are not always considered in the same framework by
all risk management experts. Risk associated with human factors and related systems fail-
ures are also well- developed areas by those who focus on these areas. Risk associated with
ambiguity is often implicit, but addressing ambiguity explicitly is not a new idea. Explicitly
addressing all five of the portrayals of sources of uncertainty which this book assumes may
underlie risk, and linking this to a broad EP interpretation of unknown unknowns, may be
novel – I have not encountered it before. But others have used variants of the same basic
ideas, and what really matters is that everyone who needs to explicitly employ these five
portrayals of uncertainty, plus an effective and efficient overlapping unknown unknowns
concept, can whenever this is useful, avoiding any relevant blind spots. Ways of doing so are
demonstrated later.
Four general observations about key words are worth brief reflection before moving on.
First, we all want simple definitions, but useful simple definitions with an element of cir-
cularity may be in need of elaboration. One kind of elaboration is illustrated by the explicit
referral to all three dictionary definitions for ‘uncertain’ discussed earlier. More generally,
an elaboration definition approach to the ‘overall definition’ for EP framing assumption
purposes may be a useful approach. Second, this kind of simple nominal definition is nominal in two senses: it needs a closely
coupled elaboration definition, and even the elaboration definition is nominal in the sense
that if you prefer a slightly different dictionary definition this is not a problem.
Third, as we will see shortly, some nominal definitions may need to deal directly with mul-
tiple meanings because there is no useful simple one- part definition, circular or otherwise.
Fourth, having obtained the framing assumption generality that we need in terms of the
simplest feasible overall definition, we may also need elaboration in terms of specific identi-
fied working assumptions – like five portrayals of uncertainty – to provide operational ways
of looking at the general concepts delivered by unrestrictive framing assumptions.
A general concern addressed by this EP approach to definitions is the need to avoid
a ‘Red Queen syndrome’. In Alice in Wonderland (Carroll, 1865), the Red Queen declares that ‘words mean exactly what I want them to mean’ (because I am the Queen, and anyone
who disagrees will lose their head). None of us are Red Queens, even when serving as the
authors of standards and guides, and we do not live in Wonderland.
Opportunity
Recognising and exploiting all feasible opportunities to improve plans is a key aspect of all
enlightened planning. Consider direct use of the three- part nominal definition:
1 ‘Opportunity’ means ‘a favourable situation’; or 2 ‘a situation with a good chance of a favourable outcome’; or 3 ‘a potential favourable outcome’.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
22 Foundations
The first two meanings are key dictionary meanings which are clearly different, while the
third is arguably a very special case of the first, and there is no simple one- part definition in
a circular form or any other form.
Not until the fourth draft of this book was underway did it become clear to me that none
of these meanings on its own is a suitable overall definition in the sense an EP approach
requires; there is no simple definition that captures all three of these meanings with requisite
clarity, circular or otherwise, and all three meanings are actually crucially important aspects of what we all mean by ‘opportunity’ in different contexts. All three meanings are worth
including separately in the direct three- part approach to the nominal definition used earlier, further enhanced by any alternative variants you think are important.
Along with some colleagues I have subscribed to the third definition as a technical defi- nition of ‘opportunity’ for some purposes for many years. But it is now clear that we were implicitly using a more general interpretation which included the first two parts of this three- part definition for other purposes, and this inconsistency was unhelpful. We did not
have a simple one- part circular equivalent to the uncertain and uncertainty definitions, and
we do not need one. What we do need is a three- part (or more) approach to a nominal
definition concept for opportunity.
My colleague Matthew Leitch has in effect been telling me this for some time. For exam-
ple, when providing feedback on a draft of Chapman and Ward (2011) he argued that he
saw opportunities as ‘a set of circumstances which made it relatively easy to do what you
wanted to do’ – a very useful illustration of a variant of the first dictionary meaning in an
‘opportunity efficiency’ context which is ignored by the third meaning. It took Stephen
Ward’s suggestion to carefully study dictionary definitions linked to also reappraising use-
ful technical definitions for ‘the penny to drop’ – I have been a slow learner on this issue.
However, I have not been as slow as those in the risk management ‘experts’ community
who have continued to insist that ‘opportunities’ are ‘favourable events’ in an ‘upside risk
event’ sense or continued to promote tools which imply this perspective without careful and
explicit health warnings.
Limiting ‘opportunities’ to ‘favourable events’ is obviously too restrictive. It results in
a focus on event uncertainty which ignores variability uncertainty, ambiguity uncertainty,
capability- culture uncertainty, and all forms of systemic uncertainty. Indeed, what actually
prompted my formal ‘favourable potential outcome’ interpretation was an attempt to explic-
itly avoid a ‘favourable event’ interpretation when its promoters became vocal and strident
in the 1990s, keeping it simple in a comparable manner. But a one- part definition restriction
involved a serious mistake. I had fallen into the common trap of being simplistic because of
a well- intentioned wish to keep it simple, avoiding simplifying in one wrong way, but still
simplifying in another ‘wrong way’. Explicitly embracing the first two definitions as well as
the third helps to clarify some of the very important more complex roles and nuances of
‘opportunity efficiency’, addressed in outline shortly, in more detail in Chapter 3 and all fol-
lowing chapters. This clarification is very useful, and a three- part nominal definition which
needs no elaboration definition is the simplest feasible way to achieve the needed generality
of ‘opportunity’ from an EP perspective.
Risk
Recognising and deciding what to do about risk is an important part of all management
activity, and risk should be a key concern in all planning processes. Risk is not just the prov-
ince of specialist risk management experts.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Why planning is usually vital 23
Consider the two part nominal definition:
1 ‘Risk’ is ‘the possibility of unfavourable outcomes’; or 2 ‘a person or thing that could cause unfavourable outcomes’.
Until the need for the definitions of uncertainty and opportunity just discussed became
clear, my focus was on the first part of this two part nominal definition of ‘risk’. But both
these two quite different meanings for ‘risk’ are well worth making explicit.
The second part of this definition is particularly rich in implications when coupled to
recognising the crucial role of ‘capability- culture uncertainty’. For example, a ‘risk’ may be
a thing or a person which is an inherent source of risk, ‘an accident waiting to happen’ in
common parlance, a ‘hazard’ in some peoples’ terms. One key associated uncertainty may
be how long it will take to happen, but all five portrayals of uncertainty may be relevant, not
just events. For example, if a person is unreliable, badly trained, poorly motivated, or not
suitably supported by the systems they work with, in plain English we can say they are a risk
or a source of risk, even if nothing has happened yet, and we have no real idea what sort of
misfortunes might be forthcoming.
From an EP perspective there is no need for an underlying elaboration definition, but
working assumptions which build on this two part nominal definition of ‘risk’ to provide
additional clarity are very important. One aspect of desirable clarity is developing an effec-
tive understanding of the issues involved in ‘risk appetite’ concerns. These concerns can be
fairly simple or very complex. A suitably general framework for thinking about risk appetite
in terms of an attribute like profit or loss with a monetary metric requires recognition that
as the probability or size of a potential loss increases, the appetite for more risk may change.
Furthermore, different relevant parties may have very different risk– reward preferences.
Explicitly accommodating these issues effectively may be important. Relatively simple prac-
tical ways of implementing this general framework are available in some contexts, but others
require more sophistication.
From an EP perspective, a lot of common practice is based on definitions of ‘risk’ which
are far too narrow. For example, briefly consider the definition ‘risk = probability × impact’.
This is the most unacceptably narrow technical definition of ‘risk’ available. It is what many
‘risk experts’ meant by risk 50 years ago, and some still do. The directly linked definition of
‘opportunity’ is an ‘upside risk event’ variant of the ‘downside risk event’ definition. Joint
use of these definitions implies risk and opportunity necessarily involve events (or condi-
tions) which either happen or do not happen. That is, risk and opportunity are limited to
‘event uncertainty’ from an EP perspective, which involves unacceptable stealth assump-
tions which matter greatly. Variability uncertainty, ambiguity uncertainty, capability- culture
uncertainty, and systemic uncertainty concepts may overlap, but they collectively cover much more than just ‘event uncertainty’. Sometimes fundamental changes in the nature of the issues which need addressing become very clear if we replace an event uncertainty basis for
risk with an EP nominal definition, in addition to the order of magnitude increases in the
scale and the importance of the potential uncertainty involved.
If you understand the operational implications of the five portrayals of uncertainty, it will
be obvious that a focus on risk associated with event uncertainty is seriously myopic because
it excludes the other four kinds of uncertainty. Basic problems remain even if advice linked
to using an event- based risk concept urges concern for causes. Skilled users of event- based
approaches may address causal uncertainty structures in great detail, but they may be blind
to other crucial aspects of uncertainty. All risk management approaches which use framing
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
24 Foundations
assumptions limiting ‘risk’ to event uncertainty need to be reframed, and some associated
tools need to be scrapped.
Furthermore, ‘risk = probability × impact’ implies that ‘risk’ can be measured by the
expected outcomes of possible events as defined by this equation, completely ignoring the
variability uncertainty associated with events, never mind the variability uncertainty incorpo-
rating correlation (systemic uncertainty) and embedded ambiguity which the mean- variance
formulation of relevant risk developed by Harry Markowitz (1959) captures via variance
and underlying covariance, and alternative working assumptions can address in other frame-
works. Completely ignoring variability in the sense of possible departures from expected
outcomes when defining ‘risk’ should never be tolerated – it is almost beyond belief that anyone currently claiming to be a ‘risk expert’ could seriously support such an approach.
Surprisingly, many risk experts still seem to believe this is an appropriate starting position
for considering risk, even in contexts where the variability aspects of risk matter enormously.
Chapter 9 provides one key example.
Some risk management experts think that an event- based definition of risk requires a
linked restrictive definition of ‘issues’, whenever a probability of one is involved. Some also
think that an event- based definition of risk implies that objective probabilities of the events occurring are required – otherwise, we are talking about uncertainties with unknown or
unreliable subjective probabilities. Those who subscribe to this latter view see risks and
uncertainties as mutually exclusive, although some others see the terms ‘risk’ and ‘uncer-
tainty’ as interchangeable equivalents.
This kind of limiting and confusing thinking about risk, opportunity, and uncertainty is
clearly interconnected in complex ways, linked to underlying assumptions about what is
meant by other key concepts – like probability.
The confusion which widespread support for these restrictive framing assumption posi-
tions adds to an already difficult set of planning concerns is demonstrably unhelpful. The
inevitable result is unnecessarily inept treatment of risk, uncertainty, and opportunities, usu-
ally adding to, compounding, and confusing the nature of further reasons for widespread
planning difficulties. Illustrative examples will be explored to clarify what is involved later.
Complexity
Effective and efficient planning requires a ‘fit- for the purpose’ level and form of understand-
ing of complexity associated with both the planning context and the planning process being
used. Consider the very simple one- part nominal definition:
‘Complexity’ means ‘lack of simplicity’.
This definition is clearly circular in the same sense as ‘uncertainty’ defined as ‘lack of certainty’
is circular. Dictionary definitions of complexity mention ‘component parts’ and ‘compli-
cated’. Experts with a focus on complexity often make distinctions between situations which
are ‘complex’ and situations which are just ‘complicated’. For example, Gawande (2011)
illustrates the difference using ‘putting someone on the moon is complicated’, but ‘bringing
up children is complex’. These are just different kinds of complexity for present purposes,
related to the number of relevant factors and the extent of their interactions. There is no
reason why you should not make these distinctions using multiple part elaboration defini-
tions or working assumptions if and when doing so is useful. But for the purposes of this
book, there are no obvious disadvantages to limiting the ‘overall definition’ of ‘complexity’
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Why planning is usually vital 25
to a simple one- part nominal definition, with a provision for further working assumptions
where this is useful.
Complexity is assumed to underlie uncertainty when we do not fully understand com-
plexity, and complexity is viewed as pervasive and not easily characterised or separated from
uncertainty. Uncertainty is assumed to underlie opportunity and risk.
Some further nominal definitions
Having explored the meanings for three of the four crucial key words in considerable detail,
followed by a relatively light- touch approach to the fourth, we can now use a relatively light-
touch approach to generalise this nominal definition concept for six further key terms which
need nominal definitions to clarify the discussion to follow.
Each of these further nominal definitions began as a simple plain English definition, used
to capture associated technical definition concerns for many years. But each has now been
carefully tested against dictionary definitions and common technical interpretations, to
maintain as much simplicity as possible without being restrictive or excluding any useful
nuances. The six further nominal definitions are
1 ‘probability’ means ‘a subjective measure of likelihood which may embed data- based
objective measures’,
2 ‘robustness’ incorporates ‘resilience’ and comparable concepts when they do not need
separate treatment,
3 ‘scenarios’ are qualitative or quantitative analysis constructs,
4 ‘separable’ means ‘able to be separated’,
5 ‘synthesis’ incorporates ‘synergistic synthesis’ in the usual sense of going beyond a sim-
ple sum of the parts plus going beyond the current scope of analysis in an imaginative
and creative manner whenever this is appropriate, and
6 ‘analysis’ incorporates ‘synthesis’ in the broad sense just defined and includes under-
standing all key aspects of interconnectedness and context.
Restrictive technical definitions for any of these words should be avoided. You may prefer to
view some of these further nominal definitions as explicit working assumptions, not defini-
tions in the framing assumption sense associated with the four nominal definitions consid-
ered first. For example, I see the ‘probability’ definition given earlier as a crucially important
framing assumption but the ‘robustness’ definition as a convenient working assumption.
You or your organisation may want to add to these six.
It was convenient to provide these nominal definitions in the order used, but it is conveni-
ent to provide brief clarification of each working from the bottom up.
Analysis explicitly avoids precluding any specific approaches or separating analysis from
linked synthesis in terms of the broad interpretation of synthesis just defined. Sometimes
it will be useful to use the word ‘synthesis’ separately or in conjunction with analysis for
emphasis, usually implying a broad interpretation of synthesis without being tedious by
repeatedly saying so.
Synthesis in the broad sense defined earlier is usefully illustrated with a simple example.
You might arrange an afternoon business meeting of importance several weeks in advance
in a city some distance away which you do not know. When you come to plan the trip to
get there, you might discover your initial base plan for travelling during the morning of the
meeting day involves unanticipated complications which could lead to you being late for
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
26 Foundations
the meeting. Train or airplane leg ‘A’ followed by a connection to ‘B’ and then ‘C’ might
involve a significant chance of a failed connection, for example. Standard analysis approaches
to contingency planning might involve making plans to tackle leg C differently depend-
ing on the outcome of legs A and B. But contingency planning approaches using analysis
incorporating a broad view of synthesis might suggest going a day early, perhaps spending
time doing some work that could be particularly productive while you have no other distrac-
tions, or perhaps going several days early and making a mini- holiday of the extra time in an
unfamiliar city or nearby countryside, possibly taking someone with you so you can both
enjoy the opportunity. An imaginative creative leap well beyond the initial analysis is trig-
gered by a problem initially addressed via conventional analysis and synthesis. This approach
to synthesis is a core characteristic of EP in all contexts. Chapter 5 provides more extensive
examples, as do Chapters 6 to 10.
Separable adopts the basic dictionary plain English interpretation ‘able to be separated’,
but it also facilitates explicit use of a much broader technical definition than usual, of the form
‘a working assumption implying a sequential treatment selected to make interdependence
assumptions which are as appropriate as possible’. For example, if an iterative process has
sequential component phases 1, . . . , n, phase separability implies that the phase partitions are
useful because different kinds of issues are conveniently separated in this decomposition struc-
ture, and it also implies that the ordering reflects useful precedence relationship assumptions.
The first pass through the process assumes that phase 1 can be addressed without knowing
the outcomes from phases 2, . . . , n, the first pass through phase 2 can be addressed knowing
the outcome of phase 1 but not the outcome of phases 3, . . . , n, the first pass through phase
3 can be addressed knowing the outcome of phases 1 and 2 but not the outcome of phases
4, . . . , n, and so on. A second complete pass can readdress phase 1 knowing the first pass
outcomes for all n phases, in some circumstances in effect just addressing marginal changes.
Provided an iterative process converges to a stable overall assessment, full interdependence
between all the issues addressed in all the phases is a viable framing assumption. Separability is
just a working assumption, which employs a structure chosen for effectiveness and efficiency.
The underlying basis of this interpretation is a form of pairwise separability used throughout
this book. A what needs to be done intuitive understanding developed via examples will be
provided for all readers as needed as this book progresses. For those interested in more how
to do it detail, the initial exploration in Chapman (1974, 1975), based on Pearce (1964), is
further developed and updated in Chapman and Ward (2002).
Scenarios are usefully visualised using a wide range of approaches, some involving ‘quan-
titative’ interpretations (probability- based) and some involving purely ‘qualitative’ interpre-
tations (no probabilities), as advocated by authors like MacNulty (1977), Becker (1983),
Schoemaker (1992, 1995), and van der Heijden (1996).
Robustness is defined in general terms to include ‘resilience’, and comparable concepts
like ‘redundancy’, whenever it is useful to do so for simplicity and comprehensiveness.
A concern for robustness in a very broad sense is central to an EP approach, as a key aspect
of a ‘keep it simple’ perspective. Distinguishing concepts like resilience and robustness or
using several words when one will do is not helpful much of the time, but sometimes helpful
distinctions can be made. For example, there are times when it is useful to associate ‘robust-
ness’ with effective testing of prior planning analysis in a general sense, to not only tease out
any working assumptions which are not appropriate, including assumptions driving a lack
of resilience, but also associate ‘resilience’ with plans which embody a post- implementation
ability to respond to circumstances which could not have been anticipated. This may help to
clarify how resilience will be used as part of an overall strategy to achieve robustness.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Why planning is usually vital 27
Probability is a tool for quantifying some uncertainty when this is useful, using a modern decision analysis view of subjective probabilities for framing assumption purposes which
embeds an objective probability basis when this is appropriate. It is important to recognise
there is no such thing as an objective probability for making decisions if anything other than
the data is relevant and any of the statistical analysis assumptions may not be strictly true.
For example, the future may not replicate the past associated with the available data. Using
an objective data- based probability involves a statement of belief, whether or not decision
makers and their analysts care to recognise this is the case. It is also important to recognise
that no data may be available, if data are available its quality may be debatable, and it may
not be cost- effective or feasible to use data in the time available. While some people like a
0– 1scale for probabilities all of the time, others find a percentage scale or a ‘1 in x chance’ or
a ‘return period of y years’ useful some of the time, and whatever approach is judged most
appropriate for the context is the working assumption adopted by this book.
Nominal definitions as a whole
The high level of clarity sought when addressing these ten general (unrestrictive) nominal
definitions usefully clarifies the use of these words in the rest of this book. But this does
not mean that the use of these words cannot be given alternative interpretations using
clearly articulated assumptions whenever this is helpful. Provided the differences do not
lead to different decisions they do not matter. However, if they do lead to different deci-
sions you need to test the implications, looking for stealth assumptions that really matter.
What always matters is framing assumptions which are as general as possible, plus working assumptions which are fit for purpose, with clarity about the role and implications of all
relevant assumptions.
Risk efficiency, clarity efficiency and opportunity efficiency
Three interdependent ‘efficiency’ concepts frame an EP approach to overall planning effi-
ciency and effectiveness. Because routine repetition of ‘efficiency and effectiveness’ would
be tedious, this book follows the convention established by the Markowitz ‘risk efficiency’
term, explicitly embedding ‘effectiveness’ as well as efficiency concerns in each of the three
EP terms:
1 ‘risk efficiency’,
2 ‘clarity efficiency’, and
3 ‘opportunity efficiency’.
Consider ‘risk efficiency’ first, because in the context of this section risk efficiency is the
simplest place to start to understand all three EP efficiency concepts.
Risk efficiency
As explained in the Overview when discussing Chapter 3, the use by BP of an approach to
planning North Sea projects based on risk efficiency delivered lower expected costs, plus less risk, plus reliable estimates of project cost and duration, simultaneously. Risk efficiency is also the basis of many other planning benefits explored in Chapter 3 in relation to IBM UK and
the UK MoD, elaborated in Chapter 4 and all Part 2 chapters.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
28 Foundations
From an EP perspective,
‘Risk efficiency’ means a minimum level of risk for any given level of expected reward.
The term ‘risk efficiency’ is usually associated with Markowitz and a mean– variance interpre-
tation of risk, but EP requires a much more general interpretation.
To begin to appreciate what risk efficiency means in generalised EP terms, it can be useful
to start with the common simplifying working assumptions in a Markowitz portfolio analysis
context:
1 only one attribute is of interest, referred to as ‘reward’;
2 reward is measured by total profit;
3 we can estimate probability distributions for all the components of expected profit and
associated risk; and
4 all these probability distributions are Normal (Gaussian) and unconditional.
Given these working assumptions, ‘expected reward’ is a ‘best estimate’ of what should
happen on average, formally defined as the first moment about the origin of the associated
probability distribution. Overall portfolio expected reward can be expressed as a linear func-
tion of component investment values multiplied by their expected rewards per monetary
unit invested. ‘Variance’ associated with expected reward is a suitable surrogate measure
for risk associated with any given expected reward, because Normal distributions are fully
defined by their mean (expected value) plus their variance (a measure of spread relative to
the mean, formally defined as the second moment about the mean). That is, as the vari-
ance gets bigger, the probability of any given downside departure from the expected value
increases, so risk in terms of the EP nominal definition unambiguously gets bigger. Overall
portfolio variance involves a quadratic functional form for combining all component vari-
ance and covariance terms.
Chapter 3 explores approaches which are simpler and less restrictive assuming that one measurable reward attribute is relevant, maintaining only the first of these four working
assumptions. Chapter 4 starts to explore the more difficult situations when multiple attrib-
utes are involved. All further chapters build on this basis, with some particularly complex
and contentious multiple attribute concerns addressed in Chapters 8 and 9.
Clarity efficiency
Now consider the second aspect of EP efficiency – ‘clarity efficiency’, a focal point for an EP
implementation of enlightened simplicity:
‘Clarity efficiency’ means a minimum level of planning cost/effort for any given level of relevant clarity.
The word ‘clarity’ in this book can be given a plain English meaning if the context is a gen-
eral discussion, but if the context is the role of clarity efficiency or the ‘strategic clarity’ and
‘tactical clarity’ concepts discussed shortly, an associated working assumption is that ‘clarity’
means ‘relevant insight which can be shared’.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Why planning is usually vital 29
If relevant understanding of key issues of concern which require a degree of common
understanding cannot be communicated effectively to all those who need to understand, we
do not have ‘clarity’ in a clarity efficiency sense.
Achieving a higher level of clarity while simultaneously expending less planning cost/
effort and then making sure that the trade- offs between clarity achieved and the marginal
planning cost/effort of more clarity is appropriate is what enlightened simplicity is about, a
central feature of an EP approach.
The practical issues requiring confrontation include the following:
1 clarity is not a single attribute,
2 most of the relevant multiple attributes are not measurable, and
3 what ‘relevant clarity’ implies is complex in a dynamic sense.
Chapters 3 and 4 deal with these difficulties by carefully ordering the issues addressed to
keep the discussion as simple as possible. Part 2 confronts the practical implications of these
difficulties in a range of contexts, systematically exploring how non- measurable multiple
attributes and other aspects of the complexity involved can be accommodated with a mini-
mal level of planning cost/effort, tailoring the approach to the contexts involved.
The dynamic nature of the complexity associated with relevant clarity arises because the
clarity which is relevant depends upon what our planning process is trying to achieve, and
our understanding of our planning goals may evolve as the planning process proceeds in a
manner which is partially but not wholly predictable. For example, in the concept stage of a
project we need initial clarity about the robustness of the business case for proceeding with
a project concept, and we can predict approximately how that robustness will grow if there
are no really big surprises as project definition proceeds. However, big surprises are both
frequent and inherently unpredictable. Chapters 3 and 4 explore some predictable aspects in
a preliminary way, further developed in all later chapters. Chapter 5 will initiate exploration
of some unpredictable aspects which the rest of Part 2 builds on.
Opportunity efficiency
The third aspect of EP efficiency, ‘opportunity efficiency’, is formally defined in terms of
three key component assumptions.
‘Opportunity efficiency’ requires risk efficiency with respect to all relevant attributes, plus appropriate trade- offs between risk and reward for each attribute, plus appropriate trade- offs between all attributes.
Opportunity efficiency is clearly a very complex composite concept. But it would be unac-
ceptably naïve not to face this complexity and deal with all aspects of it to the best of our
ability.
One aspect of this complexity is clarified and amplified if we recognise that clarity effi-
ciency attributes are directly relevant, and we also insist on the explicit working assumption:
Both ‘relevant attributes’ and ‘appropriate trade- offs’ should be judged by all parties with legitimate concerns.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
30 Foundations
Opportunity efficiency builds on both the risk efficiency and the clarity efficiency concepts,
neither of which is simple and straightforward. Opportunity efficiency can be seen as the
basis of effectiveness as well as efficiency, in the sense that achieving both effectiveness and
efficiency in an appropriately balanced manner with a balance that may vary over time is the
same thing as achieving opportunity efficiency. Opportunity efficiency can also be seen as a
crucial way of viewing ‘best practice’, in the sense that opportunity efficiency is an opera-
tional definition of what best practice ought to mean from an EP perspective. It would be surprising if you did not have an initial intuitive feel for most aspects of what
opportunity efficiency involves. If you are mathematically inclined, intuitively it is a kind
of ‘universal optimality’ concept, going beyond conventional ‘global optimality’ concepts
used by mathematicians in the sense that it considers trade- offs between the cost of seek-
ing optimality and the virtues of ‘satisficing’ behaviour and the use of ‘coping strategies’
in a stochastic framework which goes beyond the limitations of probability distribution
representation of uncertainty in a quantified form. If you are not mathematically inclined,
you may find the rejection of conventional ‘global optimisation’ in favour of a practical
emphasise driven by clarity efficiency usefully encouraging, and the decomposition structure
intuitively attractive.
However, it would also be surprising if you were not concerned about the multi- faceted
nature of opportunity efficiency, and the operational implications of the implied approaches
to seeking opportunity efficiency in practice. These concerns will be addressed – that is what
this book is about. But there is no easy direct route to resolving your concerns – they will
have to be addressed gradually as this book progresses.
The essence of what this book is about is using opportunity efficiency as an operational
approach to achieving best practice, with a view to avoiding the common ambiguity if not out-
right confusion about what best practice and good practice actually mean. We all know very
bad practice when we see it, if it is bad enough. Most of us know very good practice when we
see it, if it is good enough. What is much more difficult is deciding what best practice ought to mean in advance in the context of a particular organisation, and then getting a group of people
to agree to pursue it with a common vision and shared conceptual and operational tools, with
effective leadership from the top, the middle, and the bottom as appropriate.
To reinforce your initial intuitions, to outline in broad terms where we are going, and to
indicate why getting there matters, consider a one- paragraph overview of a few clues.
Clarity about what is meant by ‘risk appetite’ in terms of all relevant attributes can be an
important part of opportunity efficiency. Relevant concerns include everyone in an organisa-
tion understanding why an aggressive approach to financial risk up to a limit determined by
each individual’s role may be both sound and essential practice, but an aggressive approach
to reputation risk will not be tolerated. In terms of key trade- offs between attributes, impor-
tant concerns may include issues like appropriate trade- offs between expected levels of envi-
ronmental security, expected levels of safety in terms of fatalities and injuries, expected levels
of cost or profit, and all associated potential departures from expectations. At simpler levels,
opportunity efficiency addresses trade- off questions like ‘How much is it worth to finish a
project earlier than currently expected, perhaps by increasing the cost, or perhaps by reduc-
ing the functionality of what is being delivered?’ At the simplest level, the concern may be
‘if all we want is an unbiased estimate of an activity’s duration to plan the rest of a project,
what does a minimum-clarity approach to estimation involve, and what kinds of additional
clarity might be worthwhile if they were available for a modest level of additional effort?’
This very simple overview of what opportunity efficiency addresses is not going to resolve
all your concerns, and there is no simple way to do so because opportunity efficiency is an
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Why planning is usually vital 31
inherently complex concept. However, Chapters 3 and 4 explore some basic aspects success-
fully addressed by organisations like BP International and IBM UK in the 1970s, 1980s,
and 1990s, successfully employed by many other organisations during that period and since,
along with some key complexities, all further developed in later chapters. Achieving oppor- tunity efficiency and achieving enlightened planning are synonymous, involving the same goals viewed through a different lens. Both need understanding developed in a layered man-
ner. The rest of this book will gradually build your understanding of opportunity efficiency,
layer by layer, as an integral part of building your understanding of EP.
You may find it helpful to see opportunity efficiency as another composite gateway con-
cept, at a higher level of composition than the four Fs discussed at the outset of this chapter.
Risk efficiency and clarity efficiency are component gateway concepts.
Framing multiple objective trade- off approaches
One section in Chapter 4 briefly outlines the nature of ‘goal programming’ as a general con- ceptual basis for framing all multiple objective planning approaches. Goal programming uses a mathematical programming approach to providing a consistent basis for assessing alterna-
tive operational procedures for achieving opportunity efficiency trade- offs using ‘shadow
price’ and ‘shadow cost’ concepts. Mathematical programming tools are never used directly
in this book, and those who are not mathematically inclined should be comfortable with the
Chapter 4 treatment of this issue and the way it is used throughout this book. However,
shadow price and cost concepts are crucial components of an EP understanding of both
simple and complex trade- offs in opportunity efficiency terms, another important gateway
concept or portal which provides a very general conceptual basis for simple operational tools
which all target readers need to understand.
Risk efficiency in Markowitz mean– variance terms can be interpreted in terms of a goal
programming perspective to address the trade- offs between the criteria expected reward and
reward risk when reward involves a single attribute like profit – two criteria associated with a
single attribute. But more general multiple attribute contexts can also be addressed.
One important example of the role of this goal- programming perspective is the EP
approach to discounted cash flow assessments outlined in Chapter 7. Common practice
approaches to discounted cash flow analysis are usually based on framing assumptions that imply a single ‘hurdle rate’ test associated with the discount rate is appropriate. In Chapter 7
a case is made for a multiple ‘hurdle rate’ test approach within a simple iterative procedure
which uses a discount rate based on the expected cost of capital and avoids embedding
consideration of any other criteria in the discount rate. It argues that it is a serious error to
embed the implications of risk premiums or opportunity costs associated with other possible
uses for the capital in the discount rate. Multiple hurdle rate tests in an iterative process
are required for all concerns beyond an expected return which covers the expected cost of
capital. A single hurdle rate discounting test is often seriously misleading, even if return on
capital is the only concern. For example, including an inappropriate ‘risk premium’ in the
discount rate can drive organisations away from key opportunities involving very low risk
long pay-off profile projects towards very high risk ‘quick- buck’ projects, the exact opposite
of what is intended and needed.
HM Treasury (2003a) explicitly recognised the multiple hurdle rate test requirement
in a public sector context, the first publication to do so that I am aware of. Chapman,
Ward and Klein (2006) clarifies the goal programming perspective underlying the 2003
HM Treasury position in a public sector context, using disposal of nuclear waste by UK
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
32 Foundations
Nirex as the example context, and Chapman and Ward (2011) extends the approach to
private sector decisions in the way it is used for a private sector water and sewage utility
in Chapter 7.
Chapter 9 employs the same underlying goal programming conceptual framework
approach, plus a ‘revealed preference’ interpretation of trade- offs, leading to a more com-
plex but still practical operational approach, the most complex level of generalisation of this
framework used in this book.
Most Part 2 chapters use simpler interpretations within the same overall conceptual
framework, beginning with simple conventional opportunity cost approaches in Chapter 5.
Shadow price and cost concepts are a generalisation of ‘opportunity cost’ concepts in com-
mon widespread use long before the mathematical programming basis of goal programming
was developed in the 1950s.
This very general framing of all multiple objective approaches provides EP with a coher-
ent framework for assessing different operational procedures in different contexts, to
achieve clarity efficiency as well as risk efficiency and overall opportunity efficiency. This
avoids a range of common practice pitfalls. The discounting issues in Chapter 7 and the
trade- offs among money, lives and injuries in Chapter 9 are the most obvious examples
of common practice pitfalls discussed in this book, but there are further less obvious
implications.
Stochastic modelling framework choices
Chapter 3 introduces the use of a very simple ‘histogram and tree’ (HAT) framework for
understanding all the individual sources of uncertainty which are modelled in quantitative
(probabilistic) terms for stochastic modelling purposes in this book. The HAT framework
facilitates thinking about uncertainty in continuous variable histogram forms or discrete
variable probability tree forms, using graphs or tables or both in ways most people can fol-
low fairly easily even if they are not numerically inclined. It makes the use of parametrically
defined specific probability distribution choices an option when this is clarity efficient, but it
avoids ever making them a required framing assumption.
Chapter 4 uses a single section to explain how this framework can deal with multiple
sources of uncertainty involving a range of dependence structures using discrete probability
arithmetic approaches. These approaches facilitate understanding stochastic modelling using
numerical examples portrayed by graphs and linked tables which are kept as simple as pos-
sible in Part 2. This section also briefly explains how standard Monte Carlo simulation pro-
cedures can be employed to produce comparable results and the role of other computational
approaches and related conceptual frameworks.
This HAT conceptual and operational framework is another EP tool which contributes
to the opportunity efficiency of the approach as a whole, helping to overcome common
practice shortfalls like ineffective treatment of dependence and weak treatment of sen-
sitivity analysis by both analysts and the managers using their analysis. Without a HAT
framework many people are confused about how to choose between alternative approaches
to stochastic modelling, with competing claims from advocates of components of a HAT
approach which can seem incompatible alternatives, like decision trees, methods based on
moments, and Monte Carlo simulation. All target readers need a basic common under-
standing of clarity efficient and opportunity efficient option choices, even if they are senior
managers who are users of analysis provided by others and they are usually not interested
in analysis details.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Why planning is usually vital 33
Risk management failures as part of a very complex mess
Understanding what opportunity efficiency involves, why it matters, and how to achieve it,
is not easy. However, ignoring the implications is not a good idea, and although failing to
achieve opportunity efficiency is frequent, this need not be the case.
To illustrate why the pursuit of opportunity in an opportunity efficiency sense is worth
the effort, whatever your role and organisational context, consider the nature of the case
against a common-practice risk management view of ‘risk’ which goes well beyond specific
risk management technical problems in areas like the Chapter 9 safety concerns, taking a
much wider and deeper view.
Components of this generalised perspective are developed in every Part 2 chapter. For example, the generalisation of risk efficiency in this book goes well beyond the reasonably
obvious generalisations of a traditional mean– variance approach like the use of ‘stochastic
dominance’. It includes the need for a clear view of all appropriate objectives in multiple attribute terms in all contexts. It also includes addressing objectives which cannot be meas-
ured in any direct sense or are not worth trying to measure. Furthermore, it includes rec-
ognising that sometimes making decisions with a focus on expected cost, revenue, or profit
values, explicitly ignoring all associated variability and risk, may be opportunity efficient, any effort devoted to risk management in these circumstances wasting time and money as well
as distracting people from what really matters.
Chapter 5 reflects this perspective using a simple risk management perspective which does
not even employ the term ‘risk management’ – it is simply not relevant. Chapter 6 uses a
more sophisticated risk management perspective, but the term ‘risk management’ is still not
relevant. Chapter 7 addresses dissolving an existing ‘project risk management’ department
to formally integrate a transformed version of what its members currently do into the tool-
set, skill set, and mindset of a single fully integrated Projects Group team. This integrated
team treats ‘estimating’, ‘risk management’ and ‘planning driven by a search for opportuni-
ties’ as inseparable because the problems needing confrontation involve a mess of intercon-
nections which are too complex to address successfully in any other way. The associated
‘mess’ becomes more complex in Chapter 8, well beyond the scope of most Enterprise Risk
Management (ERM), still more complex in terms of some particular issues in Chapter 9.
Those insisting on using framing assumptions for risk management which are too narrow should be much more vigorously publicly criticised than is currently the case, with all the rel- evant people whose concerns are impacted understanding why risk management is an issue
which cannot be left to the risk management experts in specific areas. But the mess involving
risk management is by no means attributable to risk management experts on their own – it
is also attributable to everyone else who has failed to see that risk management issues are
part of a much bigger mess, the tip of a proverbial iceberg. The volume of ice underwater
represents the extent of further more general concerns which go well beyond a collective
view of all common-practice risk management concerns, as illustrated in all Part 2 chapters.
The good or bad behaviours and practices mantra
An EP approach uses the general notion that ‘bad practices and behaviours can drive out
good practices and behaviours unless good practices and behaviours are promoted and pro-
tected and bad practices and behaviours are contested and constrained’.
As a simple example, an organisation setting up a bidding process to deal with contrac-
tors providing or modifying a corporate asset needs to ensure that the ‘good’ potential
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
34 Foundations
contractors have a level playing field, and the bidding process does not favour ‘bad’ potential
contractors lacking expertise or ethics. This is part of a broader concern for good client–
contractor relationships from a client perspective. If the least competent and biggest liar
wins, that is in part a consequence of bad planning by the client.
The interpretation of this notion as illustrated by this simple example can be generalised,
and its implications extended in several ways. For example, if good contractors keep losing,
good contractors may need to take collective action to create a level playing field or do
much more to help their potential clients become more enlightened or both. In some cases,
good contractors and good clients may need to take collective political action, to ensure
that marketplaces are not permanently damaged by unscrupulous players who a majority
of the population would be better off without. In some contexts, this kind of response by
the suppliers or customers of goods or services may drive important regulatory and legal
changes. In general, markets which involve a level playing field are not a simple matter. An
EP approach can help address the complexity in a clarity efficient and opportunity efficient
manner, explored briefly in later chapters.
Generalising this notion means encouraging good behaviours and practices in any one
of a diverse variety of ways which are central to the EP concept as a whole – including
making effective and efficient use of good teamwork, appropriate collaboration, and cor-
porate understanding of the difference between good management and good luck, bad
management and bad luck. The role of this form of generalisation is important enough
to state the general notion as an explicit EP mantra. For simplicity we can contract ‘good
behaviours and practices’ to ‘the good’, with a comparable interpretation of ‘the bad’.
The ‘good or bad behaviours and practices mantra’ can then be stated simply and con-
cisely as
Promote and protect the good, contest and constrain the bad.
What is good and bad is, of course, debatable, but this does not make the issues unimpor-
tant; it just increases the difficulties involved which makes the need for an effective approach
more important.
A mantra of this kind can be viewed as a framing assumption. It is a framing assumption
which is well worth keeping in mind in a world with an uncomfortably high level of bad
behaviours and practices. It is also worth bearing in mind that a less obvious but sizable
proportion of the players at both individual and corporate levels are very open to being
motivated to deliver good behaviours and practices, and very willing to respond positively to
effective promotion and protection of the good, provided the bad are effectively contested
and constrained.
Two further enlightened planning mantras
‘KISS’ is commonly interpreted as ‘keep it simple, stupid’. A mantra which is central to EP
is ‘the redefined KISS mantra’:
Keep it simple systematically.
This redefinition was Stephen Ward’s idea to capture systematic use of his ‘constructive
simplicity’ concept as it was evolving towards the EP concepts of enlightened simplicity
and opportunity efficiency embracing risk efficiency and clarity efficiency in our early work
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Why planning is usually vital 35
together. Our joint project risk management publications have used it for many years. It is
central to the clarity efficiency concept and the opportunity efficiency concept as a whole.
We all have a natural desire to keep it as simple as possible, which induces a natural bias to
oversimplify even if we avoid a simplistic ‘keep it simple, stupid’ approach and ‘keep it simple systematically’ in a clarity efficiency sense. To emphasise the role of opportunity efficiency,
another very useful mantra is ‘the second- mile mantra’ in the following form:
Always go the second mile in terms of erring on the side of more clarity when making decisions with important implications, making a point of avoiding the wrong level of simplicity as an inherent aspect of avoiding the wrong kind of simplicity when misjudgements may have serious consequences.
Part of the rationale is a practical need to deal with the asymmetric implications of too lit-
tle versus too much clarity when trying to assess an appropriate trade- off between further
clarity and the associated cost/effort. Viewing the wrong level of simplicity as an inherent
aspect of the wrong kind of simplicity is a useful general reminder we need opportunity
efficiency as well as clarity efficiency as part of the ‘right kind’ of simplicity.
Strategic clarity and tactical clarity
Now consider two very high level composite concepts relevant to operations, project and
corporate planning contexts from two closely coupled but quite different perspectives.
The two concepts are strategic clarity and tactical clarity. These two concepts have an
inherent complexity driven by their composite nature. Like opportunity efficiency, they
require gradual clarification as this book progresses. But they need understanding in terms
of overview definitions before the end of this opening chapter.
The two perspectives are your personal perspective and the perspective of your organisa-
tion, using ‘your organisation’ as a convenient contraction of ‘any organisation of interest
to you, now or in the future, with a focus on one you are currently embedded in if this is
relevant’:
Strategic clarity from the perspective of your organisation means ‘everyone in the organi- sation has appropriate clarity (shared understanding) about “what needs to be done” by themselves and by those they interact with in all relevant teams to achieve all relevant cor- porate goals in EP terms’.
Tactical clarity from the perspective of your organisation means ‘everyone in the organi- sation has appropriate clarity (shared understanding) about “how to do it” for all the com- mon ground EP tasks which have to be understood for strategic clarity plus further “how to do it” toolsets, skill sets and mindsets for all the tasks they have to take responsibility for or assist with’.
Strategic clarity from your perspective means ‘as a target reader you have appropriate strategic clarity about “what needs to be done” including key common ground “how to do it” concerns relevant to all your roles in your organisation’.
Tactical clarity from your perspective means ‘as a target reader you have appropriate clarity about “how to do it” for all the common- ground EP tasks which have to be under- stood for strategic clarity plus further “how to do it” toolsets, skill sets, and mindsets for all the tasks you have to take responsibility for or assist with’.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
36 Foundations
These terms are bound to seem even more ambiguous than opportunity efficiency at this
stage, in part because they are even higher order gateway concepts than opportunity effi-
ciency. But by the time you finish Part 1 they should be starting to become useful shorthand
labels or ‘handles’ for what this book aims to provide.
Strategic clarity is explored in this book in terms of key what needs to be done issues
which need a holistic perspective appropriately shared by all members of an organisation.
Each organisation will have to address its own complete set of what needs to be done con-
cerns to achieve overall strategic clarity, drawing on their current strengths and confronting
their most pressing weaknesses. Very different organisations will need perspectives which
may have novel and unique features. One vision of strategic clarity will not fit all organisa-
tions, and your personal vision of strategic clarity will have to depend on your roles within
an organisation as well as your experience and knowledge to date and your career plans for
the future. The focus of this book is providing a strategic clarity concept which facilitates
adaptation to the context as needed.
Tactical clarity, as explored in this book, is about key how to do it issues which need a
holistic perspective in order to understand relevant strategic clarity concerns. Strategic and
tactical clarity concerns are not separable – they involve complex overlaps and interdepend-
ences. If people do not have an overview understanding of what is involved in how to do it
terms in some key areas they will not understand what needs to be done. But each of the key
how to do it areas explored in this book involves many further issues which experts in these
areas will have to address.
Put slightly differently, this book’s approach to strategic clarity is as comprehensive as pos-
sible, while its focus on tactical clarity is limited to the common ground essential to strategic
clarity, avoiding tactical clarity detail which target readers do not need to understand for
strategic clarity.
As an illustrative example of what the rest of Part 1 aims to achieve in these terms, Chap-
ter 3 begins with a very simple ‘low clarity’ approach to unbiased clarity efficient estimates of
how long a project activity might take, gradually adding more clarity for a minimal increase
in cost/effort. It then begins to explore interdependent ‘efficiency’ issues, involving risk
efficiency, clarity efficiency, and opportunity efficiency concerns. This ‘estimation- efficiency
spectrum’ is explored throughout Chapter 3 maintaining working assumptions which keep
the discussion as simple as possible. Chapter 4 addresses relaxing some of these assumptions.
By the conclusion of Chapter 4 you should understand the basis of strategic clarity, includ-
ing requisite tactical clarity. The focus of Chapters 3 and 4 is estimating and closely coupled
efficiency concerns which everyone involved in providing estimates or using estimates ought
to have in any organisation aspiring to strategic and tactical clarity. This Part 1 focus is rel-
evant to all target readers, and each chapter in Part 2 enriches this understanding in different
ways for very different contexts and kinds of organisations.
Your perspective and your organisation’s perspective on strategic and tactical clarity are
gateway concepts at a very high level of composition. If you have strategic clarity you have
an operational understanding of all the gateway concepts like opportunity efficiency and
the four Fs which are needed to make a significant contribution to your organisation in EP
terms, but you may need a lot more tactical clarity in particular areas of responsibility.
Your organisation’s needs are much more complex than your needs because they involve
capability- culture issues for the organisation as a whole.
For some readers who usually associate ‘strategy’ exclusively with corporate planning
there may be a risk of possible confusion because strategic clarity plus tactical clarity con-
cepts are being used in project and operations planning contexts as well as corporate strategy
contexts. However, we have to think both strategically as well as tactically in all planning
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Why planning is usually vital 37
areas, so we cannot avoid confronting this complication. Indeed, organisations need both
strategic and tactical clarity about the key interdependencies between corporate, project,
and operations planning to make an enlightened planning approach fully operational. In
part this is demonstrated by the need to address bottom- up strategic planning in operations
and project planning contexts, but it is also demonstrated in other ways throughout this
book. You may find it useful to regularly remind yourself that strategic clarity is about what
needs to be done, whereas tactical clarity is about how to do it.
Systematic simplicity as a composite concept
Systematic simplicity is the basis of strategic clarity plus tactical clarity as you see it, as
your organisation sees it, and as others see it, assuming that an appropriate variant of an
enlightened approach to all operations, project and corporate planning incorporating a sys-
tematic approach to keeping it simple has been adopted. Enlightened planning including
enlightened simplicity is my particular variant of the systematic simplicity concept. You and
your organisation may need your own versions of some of the conceptual and operational
frameworks, including alternative terminology, adapting my advocated enlightened plan-
ning approach to better suit your context.
A ‘common ground’ term for the composite concept which defines what we are seek-
ing from different but closely coupled perspectives is useful for several purposes. One key
purpose is to emphasise that you and your organisation should feel free to tailor both con-
ceptual and operational tools as well as their terminology to best suit your own context.
Assuming this is done, we need a label for the implied common ground, and ‘systematic
simplicity’ is my suggestion, because ‘keeping it simple systematically’ is at the core of any
best practice systematic approach to planning.
Further purposes are explored later via the website discussed briefly at the end of this
book, accessed via ‘enlightenedplanning.uk’. These further purposes will involve ongoing
adaptation and development of EP concepts using a shared view of the need to ‘keep it sim-
ple systematically’ as the starting point.
A summary with linked inferences
A summary of why planning is usually vital but often difficult and frequently inept concludes
this chapter, along with linked inferences about how EP can help to ease the difficulties and
reduce the level of inept planning to more acceptable levels.
Why planning is usually vital
For reasons initially explored in the opening two sections on what this book means by the
word ‘planning’ and the role of planning in different areas, it should be reasonably clear why
some aspects of planning in formal terms are usually vital, as is related informal planning.
Consolidating these reasons with those touched on in later sections suggests the following:
1 Plans need common understanding by a significant number of people for a range of
different reasons.
2 A lot of planning is done in different ‘silos’ for very good efficiency- driven reasons, but
corporate effectiveness means interdependences need special care in terms of commu-
nication between components of formal planning process structures, as in the need to
coordinate operations, corporate, and project planning.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
38 Foundations
3 Governance and control as plans evolve and then get implemented is crucial, usually at
several levels: project managers, boards, and regulators, for example.
4 Organisational learning and knowledge capture as experience is gained is essential, usu-
ally at several different levels, with crucial common ground.
Why planning is often difficult
You may have begun this book thinking that what needs to be done and why it is often dif-
ficult are reasonably obvious. But some new concerns may have surfaced already, and both
new and familiar concerns involve difficulties that you may be able to address in new ways
based on this book’s conceptual and operational tools.
Consolidating key issues raised in this chapter, reasons why planning is often difficult
include the following:
1 Different people need a common understanding of plans, but their information needs
and their perspectives and conceptual frameworks can vary significantly, depending on
their role. Ensuring that plans address the right questions on behalf of all the relevant
players, and provide approximately correct unbiased answers in a suitable timeframe, is
rarely straightforward.
2 People who work in silos can become acclimatised to working in separate bubbles with
minimal communication between groups. Making sure that they collaborate effectively
is crucial. All relevant people need to understand the full range of outcomes associated
with sources of uncertainty others face which they may be able to help to manage for
better overall corporate performance. This is usually easier said than done. However,
the pay- offs from doing so are huge, and there is no excuse for ignoring this or any of
the other difficulties planning practice has to confront.
3 Testing working assumptions effectively while plans are being evolved by planners as
well as during governance gateway stages prior to the implementation of plans is rarely
straightforward, and effective and efficient contingency planning to deal with working
assumptions which may not hold is almost never straightforward.
4 Correctly inferring exactly what needs learning from experience, and capturing the knowledge that really matters, is inherently difficult.
5 Even if an organisation has a fully developed enlightened planning capability and cul-
ture, the collective implications of inherent difficulties can be very daunting.
6 Confusion caused by the multitude of conflicting and restrictive concepts and tools
promoted by different interest groups involved in decision making and associated plan-
ning makes an inherently difficult situation significantly worse, and the resolution of
this self- inflicted set of difficulties in the near future is unlikely. Confusion about what
risk management ought to mean is just the tip of a very big iceberg, with all sorts of
associated ‘growlers’ (little icebergs), like how can organisations distinguish between
good luck and good management, bad luck and bad management.
Why planning is frequently inept
Meeting all the vital planning needs for organisations in a way which overcomes all the
important difficulties in a clarity efficient manner with the most appropriate trade- off levels
between different clarity attributes and the effort/cost involved is demanding.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Why planning is usually vital 39
If everybody involved is not clear about what needs to be done in broad ‘big picture’ terms, plus their role in doing it and how to execute their role, inept planning is frequently the
almost inevitable result. Only one weak link in a chain will cause a chain to fail, but multiple
people not having clarity in terms of this basic requirement will have a cumulative effect.
Some of the people involved have to ensure that appropriate risk– reward trade- offs are
made using suitable decision- making tools effectively and efficiently, coping with both good
luck and bad luck positively within a culture which understands and pursues opportunity
efficiency in an effective manner. If any key players providing these toolsets, skill sets, and
mindsets do not understand what needs to be done and how to play their role effectively,
inept planning will be the consequence.
Achieving opportunity efficiency may be difficult, but it can be reasonably simple in some
contexts. We have to seek simplicity systematically and confront complexity when it really
matters. This is at the heart of what ‘best practice’ ought to be about in effective operational
terms. If we pretend the issues facing us are simpler than they really are, this does not make them simpler. It just confuses people and makes overcoming the results of this confusion an additional set of problems and concerns which could have been avoided.
Some people are routinely hampered and confused by best or good practice standards
based on assumptions which are inappropriately restrictive, seriously aggravating what are
already inherently difficult planning concerns. In principle this problem is not necessary, but
in practice, until it is resolved you will have to confront the implications as best you can.
Key inferences from this section and earlier aspects of this chapter
If your organisation is already highly effective and efficient, exploring the scope for your
organisation finding further opportunity efficiency improvements via an EP approach may
provide only marginal benefits. But the reassurance that a fresh perspective on your organi-
sation’s approach confirms that best practice is already in place may still prove extremely
useful. In particular, it may help to clarify what your organisation means by best practice
and its ability to maintain its current high levels of performance when circumstances change,
including coping with a run of bad luck. Simply assuming your organisation does not need
to test its capability and culture in this way may be very risky.
If your organisation is well short of opportunity efficiency, the scope for improvement
will mean the effort involved in developing an effective opportunity efficiency driven EP
approach more than pays for itself.
EP is not a ‘silver bullet’. But it offers a fresh perspective, based on synthesising many
tried-and- true perspectives. In particular, it offers a way forward which will resolve some of
the unnecessary difficulties and easily avoidable sources of inept decision making for those
prepared to invest the effort. If key players in an organisation make a collective effort to
understand the overall EP agenda, they can begin implementing the key ideas that matter
the most for their context with a view to what is easiest and most productive first. However,
before they even make preliminary implementation plans, they need an understanding of the
available possibilities. You may be able to help your organisation with this.
One way to look at this overall goal is you achieving strategic clarity in order to assist your
organisation to achieve strategic clarity. You and everyone else will also need to acquire the
further tactical clarity needed to play your roles, but strategic clarity needs to lead tactical
clarity, so starting with a focus on strategic clarity is appropriate.
There are ‘no free lunches’. A significant investment in time and effort may be required to significantly enhance EP competence by individuals and organisations. But the pay- offs
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
40 Foundations
can involve a massive return on the investment, with no significant risk. The risk associated
with a failure to seek feasible competence enhancements may be very high, and it will not be
understood if the issues are not addressed.
This chapter has laid part of the foundations for Parts 2 and 3, by introducing key con-
cepts which frame the recommended approach to planning at an overview level as a first step
in explaining what enlightened planning addresses and involves.
This chapter has also provided brief initial outline answers to the question, ‘Why is plan-
ning usually vital but often difficult and frequently inept?’ You might reasonably see these
initial answers as little more than the source of a set of further questions which need much
more developed answers. But the complex interdependency of the issues to be considered in
a layered manner in the following chapters makes this inevitable.
One aspirational target for this chapter was encouraging you to pursue more devel-
oped answers to the new questions raised by reading the rest of this book. Even if you are
‘daunted easily’, it is important to avoid being too daunted by the scope of the issues which
need to be addressed, the complex interdependencies involved, and the combined length
of the remaining ten chapters. You will find it useful to develop a full understanding of the
what needs to be done aspects of strategic clarity. Difficulties understood and overcome are
transformed into opportunities, and the processes involved can be very satisfying. Difficul-
ties misunderstood or ignored do not go away. They simply become chronic liabilities and
threats, ongoing sources of anxiety, possibly leading to disaster.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Planning processes with a very wide range of forms, purposes and provenances are a defining
feature of ‘common practice’ planning. A closely coupled feature is the use of assumptions
which involve potentially important restrictions without trade- offs between the benefits and
costs of limitations imposed by these restrictions being clear.
One set of resulting challenges is dealing with the questions ‘Could this planning process
and the resulting outcomes be significantly improved by using a different approach based on
different assumptions, and if so, how?’ The basis for answering these questions from an EP
perspective is addressed in this chapter, starting with the more general question ‘what gen-
eral process should we use to plan a way forward if we do not have a specific process which
is clearly appropriate for the issues needing attention?’ An intermediate question is ‘What
process should we use to develop a specific process for repeated use in a particular context?’
Two basic planning process types are key tools for ‘keeping it simple systematically’. One
is as general as possible, by definition and design, the focus of this chapter and Chapter 5.
All other planning processes are specific, by definition and design, the focus of Chapters 6
and 7.
The general form addressed in this chapter is a ‘universal planning uncertainty and com- plexity management process’, usually contracted to ‘universal process’ or ‘UP’. The purpose of the full name is a clear description of its role, but contraction is more convenient for
routine usage.
The next section provides an overview of the UP recommended by this book. Several
following sections outline key components and concepts. A penultimate section briefly
describes the origins and evolution of the recommended UP. The concluding section fur-
ther clarifies key aspects of what a UP is and what a UP is not from an EP perspective.
This book assumes that you may not be familiar with Operational Research or Operations
Research (OR), Management Science (MS) or cognate disciplines. This chapter explains all
that you need to know about the basis of their contributions to the UP concept to follow
the discussion in the rest of this book.
A UP overview
‘Universal process’ signifies a general nature in more than one sense. For example, a UP can be used on its own, inserted in other processes, draw on other processes which might serve
as useful components, and be employed to design or test other processes.
The UP concept advocated in this book has been designed for maximum generality (flex-
ibility due to lack of limiting restrictions) and power (effectiveness and efficiency) in any
relevant context. A key design criterion was clarity about all the key working assumptions
2 A ‘universal planning uncertainty and complexity management process’ (UP)
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
42 Foundations
used by the UP, including clarity about the relationships between working assumptions and
the objectives of each component of the UP portrayed by Figure 2.1.
The flow chart of Figure 2.1 provides an overview of the recommended universal process
depicted as a set of seven sequential phases with an iterative structure providing maximum
flexibility plus supporting capability- culture assets and capability- culture liabilities which
require attention.
The basis of Figure 2.1 is my current synthesis of all the relevant planning and decision
support processes that I am aware of which have proved useful in practice with an explicit
provision for further processes which you and your organisation might find useful. If you
start with the UP of Figure 2.1, the basis of any tentative modifications can be tested within
this framework.
select and focus the process for appropriate clarity
create and enhance the plans for all relevant concerns
shape the plans using models of key issues
test the plans to ensure robustness
interpret the plans to exploit creativity
implement the plans as and when appropriate
a trigger in some
other process
return to an
appropriate process
capability-culture
assets
capability-culture
liabilitiescapture the context with appropriate clarity
Figure 2.1 A universal planning uncertainty and complexity management process, contracted to ‘universal process’ or ‘UP’.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
A ‘universal process’ 43
One key contention underlying this book is that all organisations without an explicit
corporate UP variant have an implicit set of different UP variants, independently defined by
different individuals and groups of people making an organisation’s decisions. One related
contention is these implicit UP variants are probably incompatible in some areas. Another
is some of them are probably not as effective or efficient as the UP concept advocated by
this book. An important follow- on implication is some decision makers are probably making
inappropriate ‘stealth assumptions’, and careful examination of these stealth assumptions
would reveal why a more effective and efficient explicit UP would be preferable.
Recall from Chapter 1 that ‘stealth assumptions’ which matter are framing assumptions
which by definition are unhelpful because they are restrictive in seriously inappropriate
ways, and they are invisible or inappropriately understood. The result of a set of stealth
assumptions which matter in the context of the UP concept used by an organisation is an
inept mix of incompatible approaches to making decisions with defects which may be ‘off
the radar’. They may be affecting all current management decision- making processes in all
planning areas – an operations, project, and corporate management concern for everyone
involved.
By implication, one single agreed- on UP concept which is as general, flexible, effective
and efficient as possible ought to provide an opportunity for improvement. How much that
opportunity matters will depend on the diversity, effectiveness and efficiency of the current
processes consolidated in a single explicit UP.
The ‘universal’ nature of this UP concept does not mean that other processes are not
required. UPs need to work alongside SPs (specific processes) as well as drawing upon
other processes as part of the UP ‘select and focus the process’ phase, and UPs need to be
used to routinely test the effectiveness and efficiency of all SPs as they evolve with changing
circumstances.
Using a UP requires explicit leadership or coordination of all the phases by one or more
of the people involved. There are good reasons why the first three phases might be led by
different people in some contexts, but joint management requires coordination. All the fol-
lowing phases involve the possibility of leadership and management by further new players,
but an ongoing need for coordination. An overall facilitator/coordinator role might be use-
ful, but everyone involved needs to know how to play their role effectively and efficiently,
and teamwork is usually a crucial success factor.
Process initiation and termination
As indicated by the circle at the top of Figure 2.1, a UP can be entered from any other
process, initiated by a ‘trigger’ like recognising an issue which needs to be addressed is
beyond the scope of the current planning process. As indicated by the circle at the bottom,
completing a UP can involve returning to the original process or moving on to some other
appropriate process. For example, a problem associated with implementing a current plan
might be recognised, triggering a decision to address the problem using a UP, returning to
the implementation of the revised plan when the UP intervention is complete.
Process phase structure
As may be self- evident from the diagram, phase separations signal changes in the nature and
purpose of process components worth explicit recognition in terms of a separate phase. This
section considers each phase in sequence.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
44 Foundations
Phase 1: capture the context with appropriate clarity
Sometimes ‘capture the context with appropriate clarity’ is usefully contracted to ‘capture
the context’ or just ‘capture’ if the meaning is still clear. A comparable approach to con-
tractions will be used for all phase descriptors, although sometimes a suitable single word
contraction is not feasible.
‘Context’ is used in a plain English sense, with a scope which is wide enough to address
everything of relevance. Everybody directly involved in a UP needs a common understand-
ing of the circumstances which define the context at a level of detail appropriate to their role
and concerns. The capture the context phase involves making explicit a common under-
standing of working assumptions which appropriately capture the relevant key aspects of the
context. These working assumptions are assumptions of convenience to reduce complexity.
All working assumptions in this and later phases should be made explicit for two reasons.
First, they need to be formally identified and then clarified for everyone involved who needs
to understand them to interpret the plans for a wide range of purposes as and when appro-
priate. Second, they need to be tested for robustness, immediately and later in the process.
It is important to appreciate that identifying, communicating, and testing working assump-
tions is much easier said than done. The reasons are worth understanding, and they are explored in some detail in Chapter 5. All following Part 2 chapters build on this foundation.
Working assumptions identified in the capture the context phase and the associated shared
understanding of the context will need to evolve as the process proceeds, one reason ongo-
ing testing of assumptions is necessary.
If more than one person is involved, one or more of the people involved must coordinate
or lead the process of identifying, clarifying, and sharing information about the working
assumptions which capture the context during this first phase.
Phase 2: select and focus the process for appropriate clarity
The ‘select and focus the process for appropriate clarity’ phase chooses between different
types of available processes and then tailors the way the selected process operates within a UP
based on the context as captured by the preceding phase and further working assumptions.
This select and focus the process phase is a higher order ‘planning the planning’ compo-
nent of the UP concept. It is central to managing a UP as an integrated whole but usefully
separated because it involves higher order planning with a different kind and level of com-
plexity. The UP concept is effective when it is not clear what working assumptions about
process choices would be appropriate at the outset of the capture the context phase because
of its generality, but the cost of this generality is a lack of efficiency. Preliminary working
assumptions about a suitable process selection and focus are needed almost immediately,
using restrictive assumptions to increase efficiency. These assumptions need immediate and
later ongoing routine testing for robustness, with different concerns as an analysis matures.
The process approaches which might be selected and then focused must lend themselves
to clarity efficiency – this is the only constraint. That is, the initial process choice and its
subsequent focus during its ongoing use must facilitate maximising the level of appropriate
insight which can be captured and communicated for any given level of effort/cost associ-
ated with acquiring that clarity.
Using a UP requires explicit leadership of process choices and ongoing tailoring of the
process selected by someone who understands the feasible options. At this point you may
have no idea what options involving standard planning tools of various kinds or more
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
A ‘universal process’ 45
specialist planning tools drawing on disciplines like OR might imply, and limited interest in
what might be involved. However, in ‘what needs to be done’ terms you need to appreci-
ate that someone who understands the feasible option choices needs to be involved dur-
ing the select and focus phase of a UP for this universal planning process concept to work
effectively.
To illustrate one possible option, the context as captured in the first phase may suggest
that a simple ‘project’ is involved, requiring a simple process for planning a sequence of
tasks. There may be a simple ‘off the shelf ’ project planning process that could be used
effectively, associated with a suitable family of widely used project planning models, the
focus aspect further narrowing the model set proposed.
Alternatively, the context may suggest a traditional OR process, and a particular family of
associated OR models may seem to suit ‘the problem’, defining ‘the problem’ as ‘specific
concerns requiring decisions when the most appropriate choices are not obvious but rel-
evant generic tools are available’.
However, a very general ‘soft OR’ situation structuring process for exploring ‘messes’
might be more appropriate, defining ‘messes’ as ‘interconnected systems of problems’. This
soft OR approach might be used as a front- end analysis stage process choice to be followed
later by further stages making use of other processes, but the need for a soft OR stage might
not become evident until a simpler approach is tried first.
A series of evolutionary steps which underlie the UP of Figure 2.1 are worth understand-
ing in outline, provided later. The key thing to grasp at this point is the Figure 2.1 UP
concept is a flexible process framework by design, the power of that flexibility depending
upon the skills of the people making the process choices. A direct implication you need to
bear in mind is there are ‘no free lunches’ in this context as in many others – general and
powerful processes require skilful process choices which require well- informed and skilful
people. Pretending this is not a concern simply means crude and ineffective choices become
the norm – a potential stealth assumption of crucial importance.
Adjusting select and focus phase process choices may be triggered as the overall or partial
process iterations identify unexpected issues – the select and focus phase involves working
assumptions about the most appropriate process which may need adjusting many times as
the overall process unfolds. These changes may affect the approach taken to the preced-
ing capture phase as well as later phases. Part of the rationale for the feedback structure
of Figure 2.1 is meeting this need for possible adjustments to earlier phases as the process
unfolds in a flexible manner, without waiting until the end of the current pass and making
a complete further pass through the process. In practice the approach to early phases may
approximate proceeding in parallel because of very frequent partial iterations.
Adjusting the nature of a UP as the context becomes clearer may involve more or less
emphasis on top- down thinking relative to bottom- up thinking, more or less concern for
problems versus messes, more or less focus on overall strategy versus specific tactics, and
so on.
Phase 3: create and enhance the plans for all relevant concerns
The central concern of initial plan creation may be formulating tentative first thoughts on
goals to be achieved and how to proceed. The plans created initially will be tailored by
current understanding provided by the capture the context phase plus the select and focus
the process phase. Initial plans may be an aspirational target, a robust safe option, or an
appropriate intermediate option.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
46 Foundations
In some situations, initial plans may be focused on learning more about the context, using
an exploratory approach explicitly driven by this concern. In other situations, a very specific
initial plan may be assumed.
Proactively seeking inspiration and the use of imagination can be crucial when identifying
and choosing between initial plan options. What some people see as craft- based understand-
ing of an ends– means synthesis may be crucial – knowing how to get from where we are to
where we want to be. To some extent relevant practical experience is usually important to
synthesise what works and what doesn’t work.
The people best suited to lead plan creation and enhancement may not be the same as
those best suited to lead earlier or later phases, and the issues involved may include compli-
cations like a problem ‘owner’ who identifies a need to resolve concerns directly having an
impact on his or her areas of responsibility may wish to continue to exercise control over
what is done, even when it becomes clear other people’s areas of responsibility are also
affected.
Once plans have been created in an initial form, enhancing plans includes enriching the
plans using a broader scope or providing more detail. For example, the initial planning of
what needs to be done might be followed by considering what resources are needed and
how these resources might be obtained and deployed, this kind of detail enhancing the plan-
ner’s understanding of what the plans imply.
‘Enhancement’ to address wider concerns and new issues at any stage in the process is
crucially dependent on creativity, building on and complementing the creativity of the initial
planning in a supportive manner.
Revisiting the create and enhance the plans phase until all relevant concerns have been
addressed as part of the iterative nature of a UP may lead to adjustment and further enhance-
ment of any of the initial plans – planning on a different basis or dealing with a wider set of
concerns or specific issues not addressed earlier.
The create and enhance the plans phase involves a third set of working assumptions,
building on the working assumptions developed in the first two phases. This time the
working assumptions are about the nature of the plans and how the plans relate to the
context. These plans and associated working assumptions should reflect anticipated issues
as well as current concerns. Sharing appropriate associated information between the parties
involved to the extent that this is useful is part of the create and enhance the plans phase,
as for all other phases. But in this phase using structured documentation linking some of
the key outputs of successive phases can begin to become crucial. A process facilitator may
be involved to help with the sharing and documentation, but plan creators and enhancers
have to play a central role. Training to help everyone work as a team can be invaluable.
Everyone involved becoming fully engaged in ‘learning by doing’ may be the key, as illus-
trated in Chapter 5.
During this phase, addressing resilience may be worth explicit initial treatment. Building
resilience into the plans themselves addresses issues like creating imaginative contingency
plans to facilitate recovering from situations which may be inherently unpredictable.
Simply providing ‘slack’ in the system is a basic approach to providing one form of resil-
ience. Systems which are highly optimised in a deterministic sense with no slack are inher-
ently fragile. ‘Organisational slack’ is a useful more general form of this idea, implying
people can find the time, energy, and other resources to deal with disruptive unpredictable
opportunities, crises and disasters. These kinds of distinctions can become even more rel-
evant in the next phase if specific explicit approaches to resilience are modelled.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
A ‘universal process’ 47
Phase 4: shape the plans using models of key issues
‘Shape the plans using models of key issues’ usually involves developing explicit formal
‘models’ of some selected core issues which have been partially or fully identified dur-
ing plan creation and enhancement. Generally, it is not sensible to even contemplate
trying to model all relevant concerns. Examples of simple common ‘generic’ planning
models include the discounted cash flow models used by financial planners for business
case plans, the bar charts used by project planners, the checklists used by airline pilots
to ensure that no crucial steps are overlooked, and the ‘to do’ lists most of us employ.
These models may have mathematical, graphical or verbal forms. A joint mathematical
and graphical form can be useful – an approach strongly promoted by most of the OR
community. However, some particularly useful qualitative structuring models with no
quantitative meaning also have OR roots, like the seven Ws model explored in Chapter 7.
Furthermore, some very simple purely qualitative models which have no OR roots and
are common practice in many contexts can also be useful, as illustrated by a red to green
to blue traffic light model in Chapter 6. These ‘models’ are all abstractions designed to
simplify some aspects of a complex reality in formal ways which facilitate communication
and may allow useful manipulation. For example, activities in a bar chart can have their
durations or precedence relationships changed to see what effect this has on project dura-
tion, the numbers used in a discounted cash flow model can be adjusted to see how this
affects a business case, and tasks in a to- do list can be reordered to consider the implica-
tions of alternative priorities.
‘Plans’ are invariably more comprehensive and sophisticated than the models used to
formally address selected key aspects of the plans. The associated informal planning content
may be crucial. For example, an experienced project planner using a bar chart showing activ-
ity timing as the only formal model employed will intuitively address key resource availability
assumptions which may underlie some activity timing and precedence relationship assump-
tions, and a simple ‘to do’ list may imply what needs to be done and how to do it plans
which have been developed without using formal models.
Not all key issues lend themselves to effective and efficient modelling, but it can be impor-
tant to understand when important exclusions are involved. For example, if those respon-
sible for business case plans are not clear about all their objectives, including ‘soft’ and
‘qualitative’ concerns, trying to get them to articulate and possibly quantify objectives may
be a useful modelling exercise. But if they refuse or are unable to agree on a coordinated
approach to all the objectives of actual concern, perhaps because some objectives involve
concealed agendas and hidden motives, then modelling is not going to resolve one of the
most crucial and fundamental of all planning issues.
Modellers are often associated with a facilitator role, but all the people involved in a
planning exercise must share a common understanding of what they are doing. A facili-
tator to ensure that everyone involved understands all working assumptions appropri-
ately may be useful, but the individuals involved need to consciously seek a common
understanding.
Shaping the plans using models in this fourth phase involves a fourth set of working
assumptions, building on earlier assumptions in the preceding three phases. This fourth set
of working assumptions must reflect the implications of key aspects of the plans which are
not modelled as well as the directly modelled aspects. What is not modelled does not go
away, and failing to understand the implications can be serious.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
48 Foundations
Phase 5: test the plans to ensure robustness
This phase involves testing all four sets of working assumptions made in the four preceding phases to ensure robustness in holistic terms. It must trigger looping back to modify work-
ing assumptions if tests suggest a lack of robustness which implies the emerging analysis and
synthesis conclusions might be misleading. In some circumstances, a loop back to explore
the implications of alternative working assumptions may be the best test of current assump-
tions which look suspect. In the ‘test’ phase, all the working assumptions must be understood in terms of an underlying set of more general framing assumptions. This is always easier said than done, and someone involved in a UP needs to have the capability and inclination to
ensure that this happens.
As pointed out earlier, framing assumptions shape the way we look at the complexities
of reality. Framing assumptions which lack generality limit and impair the way we look at
the robustness of working assumptions, limiting and impairing the way we look at reality.
A popular saying which reflects this idea is ‘if the only tool in your toolkit is a hammer, every
problem looks like a nail’.
As also pointed out earlier, the recommended UP concept and all the other key EP tools
are as general as possible by definition and design – the EP toolset is as unrestrictive as pos-
sible. An implication is the team involved in using a UP must have the collective skill set and mindset needed to take full advantage of the flexibility and power of the available collective
toolset. Different team members may contribute very different component toolsets, skill
sets, and mindsets, but they need a common basis for mutual understanding.
Phase 6: interpret the plans to exploit creativity
This phase involves drawing on the analysis and synthesis just undertaken and further syn-
thesising it with all other relevant understanding. A broad form of synthesis which exploits
all available creativity and inspiration is essential. The ‘interpret’ phase has to be linked as far
as possible to everything that matters, possibly drawing on new people with fresh perspec-
tives. Looping back to earlier phases to revisit understanding of the context or any interme-
diate aspect of the process from a perspective enhanced by new contributors to the process
can be a crucial option in this phase.
At this point in the process those involved need to keep looping back to earlier phases on
a selective basis or using complete further passes of the whole process until they believe the
process should be terminated. Iteration control involves a final testing of the complete set
of relevant assumptions before terminating iterations and moving on to implementation.
Iteration control is one of the key functions of the interpret phase, a final version of ear- lier iteration control in the test phase. However, the basic purpose of the interpret phase is
making sense of all previous process activity, using what has been learned to make further
creative leaps, and then assessing what else may need to be done.
All relevant players in the process need to contribute in an appropriate manner at an
appropriate time. The scope of this interpret the plans phase means members of decision-
making teams who may not have contributed earlier might make highly significant contri-
butions now. One example is a senior manager approving earlier analysis by his or her team
with modest- looking but important amendments before their analysis goes beyond that
team. Another example is board level decision makers, whose primary role might be seen as
governance, suggesting that board approval must be conditional on key modifications being
incorporated in the plans submitted to the board.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
A ‘universal process’ 49
Phase 7: implement the plans as and when this is appropriate
This final phase involves making use of UP analysis and synthesis when implementation
becomes appropriate assuming no further iterations are called for at present. The option to
loop back from the implement the plans phase in Figure 2.1 allows for adjustments to the
plans when implementation is underway. This is a central component of ongoing control and
respond functions which are needed until the planning process is terminated or replaced.
Capability- culture concepts
Using a UP involves making working assumptions in a direct manner during each of the first
four phases, plus further working assumptions in the following three phases. It also requires
making use of other analysis skill sets and mindsets in all seven phases. The complete set of
all relevant skill set and mindset working assumptions involves organisational capability and
culture issues which may not align with individual phases. This means that considering a
‘capability- culture’ concept separately is useful.
Approaching some of the full set of all working assumptions requiring identification and
clarification using separate ‘capability- culture assets’ and ‘capability- culture liabilities’ com-
ponents of this overall capability- culture concept can help to clarify what is involved while
‘keeping it simple’ in a systematic and robust manner.
The capability- culture ‘assets’ are all the ‘good’ capability- culture ‘elements’ embedded
in the organisation which planning assumptions reflect, like ‘immediately available requisite
knowledge and skills required for immediate use’.
A key working assumption is that those involved in using a UP know about these assets
when constructing plans and will plan to make the best possible use of these assets during
the whole of the universal planning process, including the creation, shaping, testing, and
implementation of the plans.
If those using a UP do not have assets which would be useful, liabilities are generated.
If assets are available but they are defective in important ways, liabilities are generated. If
none of those involved knows about the missing assets which would be useful if they were
available, or the asset defects, they generate unknown liabilities – unknown variants of the
missing and defective asset liabilities.
Somewhat different ‘liabilities’ may be associated with all the ‘bad’ capability- culture ele-
ments embedded in the organisation which may or may not be known – like employees who
are wilfully dishonest or computer- based systems which are seriously full of ‘bugs’ or any
other systems which are dysfunctional for many other reasons.
All known capability- culture liabilities are problems which may need resolution involv-
ing risks which may need managing. All unknown capability- culture liabilities are poten-
tially important capability- culture uncertainty sources which may give rise to risk with an
unknown unknown character which needs attention.
Capability- culture liabilities may require ‘accommodations’ – a clearly defined means of
managing the implications of these liabilities. As an illustration, the Projects Director who is
the focus of Chapter 7 believed that the required number of in- house planning staff to deal
with detailed planning was not a cost- effective proposition, and he fully supported the policy
his predecessor adopted to outsource all detailed planning work to their contractors, a form
of accommodation for the absence of in- house capability. Only liabilities that an organisa-
tion is aware of and understands can be addressed via accommodations, but sometimes
robust accommodations for ill- defined liabilities may be feasible.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
50 Foundations
Four example capability aspects of the capability- culture asset set are:
1 immediately available knowledge and skills required for immediate use,
2 pathways for drawing upon and integrating all requisite knowledge and skills,
3 pathways for feeding back and systematically accumulating requisite learning, and
4 pathways for all other relevant communication.
The matching capability- culture liabilities include knowledge and skills which an organisa- tion knows are not available although they are needed immediately, and so on. Liabilities
also include those generated by ‘knowledge’ which is assumed to be correct but is false
and misleading, skills assumed to be in place which are substandard, pathways which are
substandard, and so on. As an example, if in the create and enhance plans phase a busi-
ness plan model is addressed in the early stages of planning a project, an organisation may
know that some key knowledge is not available at present, but it will be at a later stage in
the planning process, and other key knowledge will not be available until the project has
been completed. These two different types of unavailable knowledge may require different
kinds of accommodations, but both might be addressed as ambiguity uncertainty for some
estimation purposes.
Four example culture aspects of the capability- culture asset set are:
1 routine encouragement and empathetic governance initiated when appropriate from
the top downwards and from the bottom upwards,
2 fully developed teamwork and cooperation at all times and at all levels,
3 a spirit of continuous improvement at all levels, and
4 encouragement of innovative leaps at all levels.
The matching capability- culture liabilities include those generated by lack of routine encouragement and empathetic governance initiated from the top downwards, and so on,
whether or not they are seen as missing assets. It also includes those generated by related
‘bad’ capability- culture characteristics, like self- serving, inflexible, or unsympathetic govern-
ance from the top downwards which goes beyond indifference to become dysfunctional
to an untenable degree. For example, if during the shape the plans phase one member of
the planning team identifies a major misconception in the assumptions underlying the cur-
rent strategy and a way to deal with it effectively which some higher level managers want
to ignore for inappropriate reasons, effective accommodations may not be possible short
of more senior managers dealing with the managers who are being dysfunctional. If this
is not done everyone associated with the implications may suffer: members of the board,
employees, customers, subcontractors, shareholders, the public and the regulators who are
supposed to be looking after the best interests of some of these other parties.
Two middle ground illustrative example aspects of the capability- culture asset set are:
1 corporate mechanisms which encourage positive behaviours for all relevant parties, and
2 corporate mechanisms which discourage negative behaviours for all relevant parties.
The first category of examples are mechanisms which encourage positive behaviours like
being fully engaged in teamwork, supporting colleagues who need help, and delivering on
promises. The second are corporate mechanisms which discourage negative behaviours like
being unhelpful and dishonest.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
A ‘universal process’ 51
Elaborating briefly on the role of these two middle ground elements, the most general
way to frame some of the relevant behavioural, political, and ethical planning concerns which I find useful in practice is adopting the ‘good or bad behaviours and practices’ mantra, dis-
cussed briefly in Chapter 1 and stated as
Promote and protect the good, contest and constrain the bad.
Looking for effective corporate mechanisms to make this mantra operational within the
use of UP based corporate processes (payment systems and rules defining required cor-
porate behaviour for example) as an integral part of all enlightened planning processes is
a central EP concern. However, a key associated framing assumption is leaving what is
meant by ‘good’ and ‘bad’ open for discussion. Different people involved in any decision-
making process will inevitably see the boundaries between good and bad in different
places. But if they are going to work together in an effective manner, they may have to
negotiate a working definition of ‘good’ and ‘bad’ to enable implementation of this man-
tra in proactive terms. Not addressing these concerns is a source of obvious problems,
rarely an effective solution.
‘Contest and constrain’ has to address a complete spectrum from gentle but effective
proactive discouragement to whatever limits are allowed by the politics and ethics of key
players. There may be legal implications. This range has to include all feasible disincentives
and sanctions, going well beyond simple financial remedies.
‘Promote and protect’ has to address a similarly broad and exhaustive spectrum of incen-
tives, encouragements and facilitation, with further potential political, ethical, and legal
implications.
‘Carrots and sticks’ in the usual sense may be part of this. Carrots (positive incentives) are
widely held to be more effective and efficient than sticks (negative incentives), but both may
be needed, and it is the effectiveness and efficiency of the joint effect that matter.
The ten elements identified above might be thought of as my current ‘top ten’ partially
separable elements, worth separate identification for emphasis and focus, referred to as ‘ele-
ments one to ten’. You or your organisation may find it useful to modify these choices.
‘Element 11’ of the asset set is a residual element formally defined as a generalisation of
what some people call the ‘right stuff ’, including a synthesis of whatever ‘top ten’ individual
elements your organisation may prefer plus any missed out. Using my top ten elements the
missing element concepts might include:
1 professionalism in an appropriate form at all levels for all the people executing the
organisation’s responsibilities;
2 loyalty and trust in appropriate forms;
3 character, creativity, and courage as needed;
4 effective accommodations for all capability- culture liabilities; 5 an explicit residual involving all other relevant elements; 6 plus all eleven elements put together in a creative manner.
Element 11 of the liability set is a matching set of any missing element eleven assets. For
example, an organisation which lacks appropriate leadership and professional standards
clearly has a ‘wrong stuff ’ variant of the ‘right stuff ’.
Table 2.1 provides a summary of the eleven elements of the capability- culture concept
described as assets which might serve as a useful checklist.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
52 Foundations
Table 2.1 The 11 elements of capability- culture assets.
1 immediately available knowledge and skills required for immediate use 2 pathways for drawing on and integrating all requisite knowledge and skills 3 pathways for feeding back and systematically accumulating requisite learning 4 pathways for all other relevant communication 5 routine encouragement and empathetic governance initiated top- down and bottom- up 6 fully developed teamwork and cooperation at all times and at all levels 7 a spirit of continuous improvement when appropriate at all levels 8 encouragement of and valuation of innovative leaps at all levels 9 corporate mechanisms which encourage positive behaviours for all relevant parties 10 corporate mechanisms which discourage negative behaviours for all relevant parties 11 the ‘right stuff ’ as a synthesis of all the above plus any other relevant element concepts
This 11 element capability- culture concept can be seen as a way of separating issues which
are largely generic to an organisation’s competence and culture from issues which are spe-
cific to each particular application of a UP, further distinguishing between competence in
‘capability’ and ‘culture’ terms in an individual ‘assets’ and ‘liabilities’ framework. However,
this separability involves useful working assumptions, not framing assumptions. Key element
11 issues usually need attention, including overlaps, omissions and synergy concerns.
This capability- culture concept approach can also be viewed as a variant of the traditional
idea of ‘corporate strengths and weaknesses’. When doing so it may be important to under-
stand that some ‘corporate characteristics’ in a strengths and weaknesses framework may be
both strengths and weaknesses (assets and liabilities), and they might be framed in a way
which is not quite the same as the elements concept used here. Furthermore, the distinction
between ‘capabilities’ and ‘culture’ is not always clear- cut in any framework.
You or your organisation may well find additional capability- culture example elements
useful or alternative groupings preferable. However, the addition of the element 11 con-
cept to some broadly equivalent grouping of your top ten elements is crucial – an element
defined as an explicit comprehensive ‘everything else’ composite using synthesis in its broad-
est sense for effective holistic approaches to dealing with what we know is important but
cannot easily define – the right stuff in common parlance. Furthermore, a failure to ensure
the ‘right stuff ’ is in place may imply the ‘wrong stuff ’.
The generality of the capability- culture concept and the way this feature of the Figure 2.1
portrayal of the UP concept contributes are important. However, the generality of this
capability- culture concept’s nature and the limited degree of definition provided means it
will need specific examples to clarify its use and roles, and tailoring it to your concerns in
particular contexts may be crucial.
The roles this capability- culture concept can play might be decomposed in several ways.
However, the basis is being clear about assumptions made about the capabilities available
and the culture shaping their use when creating plans, when implementing plans, and all the
intermediate aspects of planning, including dealing effectively with missing assets which may
or may not be known about.
Underlying this capability- culture concept are influences and values generated by ‘organi-
sational structure’ characteristics, which include ‘ownership structure’ (e.g. various forms
of private and public sector ownership as explored in later Part 2 chapters and Chapter 10),
‘employee structure’ (e.g. various levels of outsourcing and insourcing using related vari-
ants of ‘strategic partnership’ approaches, as explored in Chapter 6), ‘integration structure’
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
A ‘universal process’ 53
(e.g. various forms of horizontal and vertical integration as explored in Chapter 5), and
‘diversification structure’ (e.g. to facilitate various kinds of opportunities as well as manag-
ing risk via various forms of self- insurance as explored in Chapter 8). Why this is the case
needs clarification via examples in later chapters in conjunction with clarifying the roles of
the capability- culture concept itself.
When looking at a UP in overview terms, some uncertainty can be usefully interpreted
as ‘incomplete knowledge or skills which could be acquired by drawing on what is known
about decision making in general and this kind of decision- making context in particu-
lar’. Relevant knowledge and skills include tacit knowledge and craft skills, and practical
understanding about what works and what does not work based on experience as well as
information.
Often organisations need to feedback and capture what has been learned for future use.
An explicit UP should operate as a learning process, and to some extent a UP may involve
investing in future decision- making capability as well as making immediate decisions. Even if
learning from feedback and structured accumulation of knowledge is not relevant, gathering
relevant knowledge from different people and other sources is usually essential, and doing so
effectively and efficiently using appropriate facilitation capability is invariably crucial.
Effective communication is central to drawing on and using knowledge held by different
people and organisations and to feeding back learning. But effective communication also
has wider roles, which include keeping everyone who needs to be informed up to date, for
governance purposes or public relations purposes for example.
As more than one person is usually involved in the use of a UP, the implications of more
than one point of view usually have to be addressed. Simply assuming that a fully informed
collaborative joint effort will be involved throughout is usually inappropriate and often dan-
gerous. Even if only one person is involved, unconscious sources of bias may be crucial. In
addition, organisations also have to be prepared to deal with ‘gaming behaviour’, including
parties prepared to wilfully manipulate the rules of engagement and behave in a criminal
manner. Organisations can exploit relevant ‘game theory’, ‘market behaviour knowledge’,
‘agency theory’, and more general behavioural knowledge. Furthermore, they can use their
experience and skills to develop appropriate internal and external contracts, as well as other
rules or laws, to constrain behaviour. But organisations may have to make explicit adjust-
ments and provisions for behaviour which could go beyond current knowledge and models,
to cope with the realities of all relevant behaviour. Further still, all other gaps in knowledge
and skills have to be coped with. Ensuring all these concerns are addressed effectively may
require explicit monitoring and review, perhaps separate governance processes. Assuming
otherwise as working assumptions without testing the robustness of these assumptions may
be very unwise, in some cases negligent in a legal sense. Unanticipated market failures,
political crises, or frauds of direct relevance are examples of some of the extremes which may
be involved.
You and your organisation do not want to exclude anything from the uncertainty and
complexity concepts employed without making a conscious working assumption which can
be tested for robustness when appropriate, and achieving this in practice is never easy.
The ‘closure with completeness’ concept
The element 11 aspect of the capability- culture concept illustrates a general approach to
decomposition/composition which is usefully clarified and given a contraction label at this
point in the development of EP foundation concepts.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
54 Foundations
If n separable components 1, . . . , n are used in a decomposition/composition process
for working assumption purposes, whenever doing so is appropriate component n can be
defined to provide ‘closure of the decomposition/composition process using an “everything else” interpretation which includes a holistic approach to synergy to ensure completeness’ – contracted to a ‘closure with completeness’ component or concept. In the capability- culture
context element n = 11 serves as a closure with completeness component.
This closure with completeness concept will be used in a variety of ways in what follows.
The capability- culture element n = 11 context provides a usefully simple initial illustration
in the sense that my working assumption that a top 10 separable set of elements might have
been a top 5 or a top 20, and other choices within the top 10 might have been made, but the
right stuff element n which completes any alternative choices that you or your organisation
may prefer should still provide closure with completeness.
UP generality in principle and in practice
As indicated earlier, one way to use a UP is as a default choice for planning in novel contexts,
illustrated in Chapter 5. Chapter 5 illustrates the details of formally and explicitly using a
UP with a focus on a bottom- up mode of use. This starts to make Figure 2.1 an operational
concept in terms of its phase structure and the capability- culture concept. It does so using
simple, traditional generic models of manufacturing process inventories as a starting point.
Part of the purpose of Chapter 5 is to demonstrate the creative aspects of a UP, the potential
strategic flavour of a UP even when we start with simple tactical concerns, the top- down
implications of a UP even when a bottom- up starting position is involved, the corporate col-
laboration it can help to generate, the need for teamwork in process terms across all levels of
management, and the value of cooperation with other organisations.
Further ways of using a UP and top- down modes of operation are illustrated by all the
other Part 2 chapters. The range of UP uses and the range of organisations which can use a
UP demonstrated by Part 2 is very wide. Part 2 limits itself to ‘commercial organisations’,
but much broader bounds could be explored. In principle there are no real limits that I am
aware of with regard to the core concepts, in the sense that they have a very broad prov-
enance based on ideas which have a track history of portability, including extensive use for
military planning, police service planning, health service planning and other government
planning.
The provenance of the UP concept
This section explains how various prototype variants of the UP of Figure 2.1 and its Table 2.1
capability- culture components have been used in practice in the past, its empirical basis.
The primary aim is an overview of the provenance of the recommended universal process
for those who may have never heard of the disciplines associated with its starting point and
evolution and may have limited interest in learning more. A closely coupled and important
secondary purpose is clarifying the EP perspective on aspects where different people have
different views.
Operational Research (commonly contracted to OR) can be viewed as the starting point
contributor to the UP concept of Figure 2.1, the perspective adopted in this section. The
evolution from this starting point involved contributions from a group of cognate disciplines
centred on ‘Operations Research’ (also contracted to OR) and MS. A common joint con-
traction is MS/OR (or OR/MS). This book associates the MS/OR label with the broadest
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
A ‘universal process’ 55
possible interpretation of an MS/OR perspective, but relevant cognate disciplines which can
overlap but are usually viewed as separate disciples with their own decomposition structures
include engineering, mathematics, statistics, economics, and all the behavioural and social
sciences, to mention just a few key ones.
All target readers need an overview understanding of the provenance of the Figure 2.1
universal process to begin to clarify the nature of the roles a UP can play at this point. Later
all readers will need a richer understanding of the range of roles a UP can play and what
these roles involve in ‘what needs to be done’ terms, addressed in the rest of this book.
Some readers may want to develop still further understanding of selected subsets of these
issues, going into ‘how to do it’ territory beyond the scope of this book.
Those who already have a deep understanding of MS/OR may find this section useful
in terms of reinforcing shared views and testing views where we differ. However, exploring
the implications of alternative views or perspectives is a matter for another time and place.
Eight ‘contributors’ (contributing schools of thought or perspectives) which embrace
some of the sciences, arts and crafts relevant to an EP interpretation of a UP are outlined in this section in eight separate subsections. A ninth ‘contributor’ makes an explicit provi-
sion for anything not covered in the first eight plus a synthesis of all contributions, another
closure with completeness concept. Many of the later contributions have only limited con-
nections with MS/OR, however broadly this term is interpreted.
Contributor 1: 1930s/40s OR
A British creation in the 1930s and 1940s, OR was widely acknowledged as an important
contributor to the victory of the Allied Forces in World War II. As war approached in the
late 1930s, OR groups making use of British scientists from a range of backgrounds started
to address pressing military planning problems. The success of early efforts led to a signifi-
cant expansion of OR activity when war began.
Some of the planning problems addressed were very big and very complex – like given
radar was technically operational, how should radar be deployed effectively as part of a com-
mand and control system designed to get British fighter aircraft into the air at the best time
and in the best place to meet incoming enemy aircraft? What became known as ‘the Battle
of Britain’ depended on this system, and assisting with its development was an important
early OR contribution to British military planning. According to Patrick Rivett (1994), the
first OR scientists were the physicists who had invented and developed radar who were then
given the task of introducing the radar sets to the army units who would use them. In doing
so they were drawn into planning the operational use of radar. This activity was initially
called ‘radar operational research’, later generalised and contracted to ‘operational research
(OR)’. It was estimated at the time that radar doubled the effectiveness of the Royal Air
Force fighter squadrons in the Battle of Britain, and that OR doubled it again.
Other planning problems were relatively simply stated but still big and complex – like
given a limited number of anti- aircraft guns were available as ‘the Battle of the Atlantic’
began, should merchant ships be given significant priority or should priority be given to
shore- based batteries. This relatively simply stated planning problem can be used to briefly
illustrate the spirit of one approach sometimes taken, as described by one of my University
of Toronto professors when I was first introduced to OR at the beginning of the 1960s.
Statistical analysis comparing the proportion of enemy aircraft destroyed when a ship
or shore- based anti- aircraft gun opens fire on incoming enemy aircraft confirmed the
anticipated ‘productivity’ superiority (in terms of enemy aircraft damaged or destroyed) of
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
56 Foundations
shore- based guns and measured its extent. Common sense had suggested that a stable shore
base would make hitting enemy aircraft easier, but by how much was uncertain.
Common sense had also suggested that aircraft attempting to bomb a merchant ship are
more likely to drop their bombs at a greater distance and miss if they are being vigorously
shot at by effective anti- aircraft guns on the ship being attacked, and the morale of those
serving on the ships would be greatly enhanced if they could defend themselves effectively.
Further statistical analysis comparing the proportion of ships sunk by attacking enemy air-
craft when they were or were not equipped with anti- aircraft guns confirmed and measured
the extent to which attacking aircraft were more likely to miss when they were being fired
at. Morale was recognised as important but not measurable.
Because winning the Battle of the Atlantic was essential to avoid losing the war, the
merchant ships were given significant priority. What could be usefully measured in the time
available was measured, but the politics and ethics of different competing objectives were
managed by senior military commanders interpreting the results with a clear understanding
of both the power and the limitations of the quantitative analysis. Senior military com-
manders made the choices and considered non- measurable issues like morale on the ships
and related operational effectiveness as well as the ethical implications of important conse-
quences for those targets not protected by anti- aircraft guns. OR support staff provided sys-
tematic statistical analysis and complementary non- quantitative analysis and synthesis where
this was feasible and useful.
These early British OR groups were so successful they were soon followed by similar
groups supporting military planning in all Allied Forces. In the US and some other coun-
tries, OR was referred to as Operations Research rather than Operational Research because
of differences in military terminology, preserving the common OR contraction.
At the end of the war many of those involved in British OR went back to doing what they
had been doing before the war. For example PMS Blackett, the Admiralty’s chief of OR
who led a group referred to as ‘Blackett’s Circus’, went back to an academic life at the Uni-
versity of Cambridge, although he championed ongoing OR use and had political interests
and ambitions. Blackett is a useful example because he won a Nobel Prize in Physics, which
clearly illustrates the quality of the thinking involved, undoubtedly of significant impor-
tance. Many others involved in British wartime OR groups were instrumental in setting up
OR groups in industry after the war, especially in the newly nationalised industries like coal,
steel and railways. Patrick Rivett (Rivett, 1994) first applied OR to armaments, but became
the head of OR at the National Coal Board and then the manager of OR at Arthur Anderson
before becoming the first UK professor of OR at Lancaster University in 1962. Industrially
based OR also started in other counties.
A key characteristic of OR in this period was a highly successful practice- driven develop-
ment which drew on people with backgrounds often lacking mainstream management or
military expertise but strong on systematic analysis and synthesis working in highly moti-
vated collaborative teams with people who had the military technical knowledge and the
military management experience to jointly create successful plans.
Some civilian scientists and military decision makers did not mix easily. When the zoolo-
gist Solly Zuckerman produced an OR report recommending bombing nodal points on
the continental railway network to slow down German reinforcements after D- Day, his
recommendation was famously ridiculed by Arthur ‘Bomber’ Harris as ‘a panacea by a civil-
ian professor whose forte is the sexual aberrations of apes’ (Kennedy, 2014). But history
confirms that Zuckerman was clearly right, and Harris was demonstrably wrong as well as
being ridiculously rude.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
A ‘universal process’ 57
To make sense of all corporate processes from an EP perspective, you do not need a deep or detailed understanding of OR, but you do need a broad understanding of the issues
which can be addressed by OR broadly defined plus all cognate disciplines in the kind of
synthesis provided by the UP concept.
Contributor 2: 1950s/60s OR
A key feature of 1950s/60s OR was the attempts to capture what the 1930s/40s OR
groups had been doing in terms of process – usually referred to as the ‘OR method’ or the
‘OR process’. A central feature of these 1950s/60s OR process models in terms of their
textbook portrayal was a synthesis with a limited selection of earlier relevant generic models
for making decisions, like the inventory models explored in Chapter 5. Early OR textbooks,
largely a creation of academics and practitioners based in the US, presented OR as a practical
approach to decision making using this new OR process plus previously developed relevant
generic models. This led to significant further developments in OR theory, new areas of OR
practice based on this new OR theory, and the first university courses teaching OR theory
and practice from a range of perspectives.
To help put this book’s approach into context and clarify my perspective, issues illustrated
using some of my background may be helpful.
My BASc in Industrial Engineering at the University of Toronto (1962) was redesigned
just before I started it. It was the first undergraduate programme in North America with a
significant and central OR content. Part of its new core was based on these 1950s/60s OR
developments, linked to current developments in MS from a North American Industrial
Engineering perspective. By the 1980s many people saw the term Management Science as synonymous with OR, but in the context of my BASc, ‘Management Science’ was broadly
defined to include earlier Industrial Engineering and Scientific Management ideas, plus
Industrial Psychology (how individuals behave in an organisational context) and Sociology
(how groups of people behave in an organisational context). At that time the term Manage- ment Science was replacing earlier use of the label ‘Scientific Management’ as a behavioural science critique of Scientific Management was developed. The term Scientific Management had been introduced by innovators like Frederick Taylor and Henry Gantt, in books like
The Principles of Scientific Management (Taylor, 1911). Although the basis of some of their work was rejected, some early contributions by both Taylor and Gantt were still relevant in
the 1960s, and some will remain relevant in the 2020s and beyond. Much of the existing
core of this BASc was basic engineering sciences. A course in economics was taken by all
University of Toronto engineering students, and in my case a very charismatic economics
professor stimulated a transformative interest in economics.
My MSc in OR at the University of Birmingham (1964) was the first postgraduate course
in the UK centred on OR. This course was already well established when I took it, with a
very relevant option in Economics and Econometrics. Taking this MSc enriched my under-
standing of OR in the 1950s/60s OR sense from a British perspective, and it provided an
important set of different perspectives from economics and econometrics.
My first academic post, starting in 1965 as a lecturer (assistant professor) in Economet-
rics, involved managing an MSc in Economics, Econometrics and OR and developing and
teaching the OR content.
The starting position for my understanding of a UP was the OR processes described in
1950s/60s OR textbooks like Operations Research: Methods and Problems (Sasieni, Yaspan and Friedman, 1959). Table 2.2 is a common form of a brief summary of this OR process.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
58 Foundations
Table 2.2 A five phase portrayal of the OR process.
1 describe the problem, 2 formulate a model of the problem, 3 solve the model, 4 test the solution for robustness, and 5 implement the solution.
Comparing the five phase structure of Table 2.2 with Figure 2.1 assuming a comparable
iterative structure, there are fairly obvious similarities and some very important differences.
By the end of the 1960s, I was aware that some practitioners who were pioneers of early
British OR practice, like Patrick Rivett, had serious reservations about the oversimplicity of
this textbook portrayal. Rereading parts of The Craft of Decision Modelling (Rivett, 1994) recently has helped to clarify why. Patrick and I never had a conversation about this topic,
and I can no longer discuss his views at the convivial lunch discussions he used to chair in
London at Blackett Club meetings he instituted, but I would like to think Patrick and earlier
pioneers like Blackett and Zuckerman would prefer the UP of Figure 2.1 to the basic five
phase OR process of Table 2.2. I think they would prefer the formal planning process focus instead of a problem modelling and solving process focus, the UP’s front end which flows from its formal planning process focus, and other changes which flow from this revised front
end. Furthermore, I believe they would appreciate the role of the capability- culture concept.
I also think the senior decision makers and all the other people they worked with would be
much more comfortable with the Figure 2.1 portrayal of what was involved.
Using the anti- aircraft guns example, the Figure 2.1 front- end formal planning process
emphasises the importance of the plans as a whole in a broad view of the context, with the
statistical modelling employed focused on a subset of the relevant issues which are suitable
for this kind of quantitative modelling and in urgent need of immediate attention as part
of a much broader overall planning approach. This avoids overplaying the role of quantita-
tive modelling, and it paints a more realistic picture of how formal analysis using quantita-
tive analysis can assist senior decision makers plus everyone else involved. It was of crucial
importance to see ‘the overall planning concern’ as winning the Battle of the Atlantic in
the context of not losing the Second World War. Any attempt to see this overall concern in
terms of the quantitative OR models used would have seriously oversimplified and inevitably
failed, because not losing the Second World War was in both plain English and technical
language terms a huge ‘mess’ – a very complex set of interconnected issues and associated
problems which were certainly well beyond the direct modelling capabilities of OR at that
time, arguably still well beyond MS/OR capabilities.
However, for many years I had no serious reservations about simple OR process portray-
als like Table 2.2 as the direct equivalent for my implicit UP. With hindsight I may have
been implicitly presuming that other people would see formal quantitative modelling in
the broader perspective of some prototype variant of Figure 2.1 without really exploring or
properly understanding the implications of my assumptions – a personal stealth assumption
I failed to identify and test. Whatever the explanation, I now believe explicitly exploring
and clarifying the implications of Figure 2.1 instead of just using a textbook OR method
perspective is useful. In the process of writing this book I have uncovered other significant
stealth assumptions in my own earlier thinking, now resolved via generalisation of a range of
concepts as a direct consequence of applying a UP perspective shaped by Figure 2.1.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
A ‘universal process’ 59
Another key feature of 1950s/60s OR was that different people saw it very differently – an issue which may be less obvious now, but it persists, with important ongoing implications.
For example, OR practitioners who were managers or consultants with an OR background
and perspective tended to see OR as drawing on many disciplines, including all those in
their own background which they found useful plus all those in the background of their col-
leagues which they found useful. However, academics tended to claim what they saw as the
useful aspects of OR as part of their own academic discipline domain, sometimes underplay-
ing or ignoring the rest, for what might be interpreted as academic group cultural reasons
driven by territorial concerns.
For example, Gordon Fisher, the Foundation Professor of Econometrics and the head
of the Department of Econometrics at Southampton University, who taught one of my
econometrics courses at the University of Birmingham and gave me my first academic job
in his new Department of Econometrics, saw OR as a subset of econometrics. He referred
to this view as ‘the Dutch perspective’ because he shared it with several Dutch econometri-
cian colleagues. My PhD supervisor, Ivor Pearce, professor of Economics and later the head
of the Department of Economics at Southampton, saw OR and econometrics as separate
subsets of economics. The Mathematics Department at Southampton saw OR as a subset of
applied mathematics and created a separate OR group with a professor of OR to lead it and
develop this perspective. My MSc from the University of Birmingham implied economics
and econometrics were a subset of OR which was a subset of engineering. My Industrial
Engineering professors at the University of Toronto probably saw OR as a subset of what is
now commonly referred to as MS/OR plus a broader view of MS and the integration of this
very broadly defined MS/OR concept as a subset of engineering.
A perspective you might find attractive, especially if you are an engineer, is using my
favourite definition of ‘engineering’ – creative synthesis of ends and means – and defining a UP as a universal process for ‘engineering planning’ (EP) in any context requiring ‘oppor-
tunity engineering’. This provides an ‘engineering flavour’ to the EP perspective and the UP
without changing the contraction labels. Opportunity engineering generalises the ‘risk engi-
neering’ term Dale Cooper suggested which we used for joint publications in the 1980s, and
it may appeal to people who do not see themselves as engineers in the usual sense.
More generally, you can develop your own flavour, and you can embed an EP perspective
in whatever disciplines shape your own outlook and approach to the UP concept.
My 1960s prototype UP accepted the limitations of the OR process description. But it
explicitly embraced the notion that all the sciences, arts and crafts which were relevant to
planning in the widest sense used in this book needed to be drawn upon so far as possible.
A key feature of the UP of Figure 2.1 is drawing on all relevant disciplines, without com-
peting for territory, in a practical planning tradition incorporating problem solving and mess
resolution which predates OR, MS and cognate disciplines.
Contributor 3: Ackoff’s creativity and strategic vision critiques of OR
Russell Ackoff was one of the pioneers of 1950s/60s OR based in the US. He co- authored
one of the first international textbooks on the subject, Introduction to Operations Research (Churchman, Ackoff and Arnoff, 1957). His book Fundamentals of Operations Research (Ackoff and Sasieni, 1968) became a standard international text. A Manager’s Guide to Operational Research (Rivett and Ackoff, 1967) was an early transatlantic book.
Ackoff was also one of the most perceptive critics of mainstream 1970s OR. Redesigning the Future (Ackoff, 1974) develops a number of ideas that the UP concept and associated
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
60 Foundations
SPs attempt to capture – in particular, the idea that top- down strategic planning ought to
start by being clear about goals and formally address designing the future we want, as devel-
oped in Chapter 8 of this book.
Ackoff also emphasised the role of creativity, adopted as a fundamental feature of most
phases of a UP or derivative SP. The create and enhance plans phase of Figure 2.1 and the
creative aspects of other phases reflect in part my long- standing view that creativity is crucial.
Creativity is crucial in the models used for quantitative analysis, but creativity is even more
crucial in the much broader plans themselves and in the way the analysis plus the broader
plans are interpreted. Good planners have relevant creative craft skill sets as well as effective
and efficient toolsets. Ackoff stimulated the clarification and development of those views.
Ackoff ’s creative design ideas (he was an architect by initial training) are of central impor-
tance to a UP concept, as are his concerns for promoting a strategic vision which embraces
‘designing a desirable future’. He was an American who had a deep understanding of OR
developments on both sides of the Atlantic and had a profound worldwide impact.
Contributor 4: Soft Systems, soft OR and related process concerns
Qualitative (soft) factors, as well as quantitative (hard) factors, are essential aspects of an EP
perspective.
‘Soft Systems’ was the creation of Peter Checkland, who was concerned about the practi-
cal limitations of traditional ‘hard systems’ approaches. Systems Thinking, Systems Practice (Checkland, 1981) is a seminal text, and Checkland and Scholes (1990) is a useful compan-
ion volume.
Checkland found a particularly appreciative audience amongst those who saw his approach
as a very general form of ‘soft OR’ process, often viewing soft OR as an important alterna-
tive to a ‘traditional’, or ‘hard, OR’ approach.
Rational Analysis for a Problematic World (Rosenhead, 1989) is a seminal text on soft OR, addressing Checkland’s approach plus a range of other soft OR alternatives, and Rosen-
head and Mingers (2001) is a useful update. In part, the critique soft OR applied to main-
stream 1970s/80s OR was the need to address complex messes – interconnected systems
of problems as distinct from problems suitable for independent treatment. I would argue
that 1930s/40s OR often took an integrated hard and soft systems view of messes, but
1970s/80s OR seemed to have forsaken a systems perspective as those advocating soft
approaches saw it. In part, the critique soft approaches applied was the need to avoid a
‘reductionist’ form of analysis – to always see the big picture and the interconnections – to
explore the interconnectivity of messes rather than to solve problems. In part the critique
it applied was the need to maintain an interest in concerns which could not be quantified,
including ethical issues.
My personal use of soft OR tools in practice has been limited, probably because my formal
training did not include them. But I have used some simple soft OR tools myself and with
colleagues successfully (the seven Ws structure in Chapter 7 initially created by Stephen
Ward using a technique developed by Colin Eden is an example – see Eden [1988] for
an overview). And I have suggested clients should bring in experts with requisite soft OR
experience when sophisticated use of soft OR tools looked promising, advice followed very
successfully by one client (Bombardier – who subsequently built an ongoing relationship
with Colin Eden and his colleagues at Strathclyde University). Soft OR processes and tools
should be an optional choice in the select and focus the process phase at any stage in an
analysis if they are relevant to an organisation’s concerns.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
A ‘universal process’ 61
It is useful to interpret all soft approaches as systematic qualitative approaches which are not immediately concerned with quantitative analysis, as part of a broader set of qualitative analysis approaches which may or may not be concerned with eventual quantitative analysis.
It is also useful to be very clear that there is no such thing as a quantitative model which
does not need very careful qualitative structuring and interpretation, and this was a universal truth long before any kind of MS/OR/Soft Systems approaches.
Put slightly differently, quantitative and qualitative analysis are not separable concepts in
the sense that we can have qualitative analysis without quantitative analysis, but we cannot
have quantitative analysis without associated qualitative analysis.
Contributor 5: behavioural, political and ethical concerns relevant to planning
As noted earlier, behavioural science critiques of ‘Scientific Management’ led to ‘Manage-
ment Science’, defined to include the development of Industrial Psychology and Sociol-
ogy. My Industrial Engineering BASc had a significate behavioural science emphasis which
addressed Management Science in this sense. However, the OR process in its 1950s/60s
form and subsequent variants do not usually embed any behavioural or cultural content in
a direct and explicit form. They certainly do not emphasise political and ethical content or
concerns. This is an oversight which any UP concept needs to avoid.
Some authors associated with soft OR and underlying systems ideas, like Rosenhead and
Churchman, have been very explicit about the need to address behavioural, political, and
ethical concerns, and Churchman was the lead author of the 1957 OR text cited earlier
which shaped the foundations of textbook OR. I am not suggesting there is anything funda-
mentally new about addressing these concerns in a UP concept. However, those who want
to keep such issues out of planning are potentially distorting an effective interpretation of
analysis in a way which needs general understanding by all potential users of their analysis.
Part of the role of the capability- culture concept is helping to frame planning in a manner
which may offer some fresh insights about the pertinence of behavioural, political and ethi-
cal matters, and how to approach them in practice.
The mantra ‘promote and protect the good, contest and constrain the bad’ was associated
with the basis of the two ‘middle ground’ example elements of the Table 2.1 capability-
culture concept. This mantra can be applied to planning to accommodate behaviour as
simple and ubiquitous as conscious (deliberate) bias when estimating parameters. As an
example, Megaprojects and Risk – an Anatomy of Ambition (Flyvbjerg, Bruzelius and Roth- engatter, 2003) uses the euphemism ‘strategic misrepresentation’ to avoid directly accusing
people of lying, arguably very useful to encourage discussion but potentially dangerous if
it encourages rather than discourages deliberate bias. What is crucial is encouraging open
discussion of conscious bias and explicitly discouraging tolerance of dishonesty which can
shade into fraud which should not be tolerated.
The mantra ‘promote and protect the good, contest and constrain the bad’ can also be
applied to avoiding the implications of behaviour which is more complex. Elaborating on
the Chapter 1 example, it may be important to help a project owner planning a bidding
process to avoid a situation which results in the winning bidder being the tenderer who least
understands the task being bid for (the most incompetent) or is most prepared to set up
their contracts in a way which lets them add costs as the work progresses (‘claims engineer-
ing’) or is simply the one prepared to tell the biggest and boldest lies (the biggest crook).
In this case, the immediate issue is protecting a given client, but an underlying problem
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
62 Foundations
may be honest and competent bidders will go out of business unless the bad is contested
and constrained. Dishonest and incompetent bidders need to be dealt with effectively by
the combined efforts of clients planning bidding processes and other interested parties –
otherwise, they will ruin the marketplace for everyone else. Honest and competent bidders
may have to take the lead with prospective clients, trade associations, the press, government
regulators, and public prosecutors. The line between honest and dishonest is wilfully made
fuzzy and ambiguous by those who believe sharp practice is just acceptable business practice,
but there is no fundamental reason why everyone else should not prevent them from getting
away with unacceptable practice.
Furthermore, this mantra applies to dealing with the implications of important political
and ethical issues which can be very complex and highly controversial, like should an organi-
sation providing a basic utility service like water supply or electrical power supply be private
sector, public sector, or some combination. What is bad and what is good involves ethical
concerns at personal, organisational, and political levels.
The mantra ‘promote and protect the good, contest and constrain the bad’ may involve
consciously adapting processes to cope effectively with some people pursuing self- interest
in a straightforward way, other people pursuing the interests of others in genuine altruistic
terms, and a range of intermediate positions, including what some people see as ‘enlight-
ened self- interest’. However, and this is crucial, there may be no need to seek agreement
between those involved about some key decision- making issues beyond what is necessary, provided an enlightened structuring of components of the supporting analysis is deployed.
Chapter 9 addresses this aspect of decision making explicitly when it is especially important,
but it is central to EP as a whole. Society has grave difficulties attempting to agree about
what should be legal or illegal, and what is moral or immoral is much more complex. But
this does not mean we can give up addressing related concerns, which can be crucial. When
it matters, we have to try very hard to agree what we mean by ‘good’ and ‘bad’ and how we
are going to promote and protect the good and contest and constrain the bad to the extent that it matters. If we fail to agree, we still have to confront the need for working assump- tions addressing effective and efficient accommodations to cope with our ongoing failure to
agree. The issues do not go away if they are not dealt with because they are ‘too difficult’.
Associating requisite agreement about appropriate decisions with the capability- culture
asset set and any linked failures needing accommodations with the capability- culture liability
concept of Figure 2.1 serves several purposes. One important role is acting as a reminder to
deal effectively with relevant behavioural, political, and ethical concerns by testing crucial
assumptions and taking appropriate proactive action when necessary. Another is ensuring
any failures are recognised and treated via effective accommodations.
Contributor 6: Paul Kennedy’s and Atul Gawande’s approaches to capability- culture
A useful reminder of the importance of culture to making plans work is one of the important
roles for the capability- culture concept. Including the capability- culture asset examples like
‘encouragement and empathetic governance top- down and bottom- up’ and ‘a spirit of con-
tinuous improvement when appropriate at all levels’ was used earlier to illustrate what was
involved. I have been conscious of the need to consider these issues for most of my career,
but the incentive to do so in this particular way was triggered by reading the final chapter
of Engineers of Victory: The Problem Solvers Who Turned the Tide in the Second World War (Kennedy, 2014). The starting point for the structuring of the capability- culture concept of
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
A ‘universal process’ 63
Table 2.1 was the discussion starting on page 365 of what Paul Kennedy terms the ‘culture
of encouragement’ or the ‘culture of innovation’ and related capability and culture issues
associated with why the Allies won the Second World War.
Paul Kennedy is an influential and bestselling military historian based at Yale, with strong
British links. His 2014 book argues from a very broad perspective the importance of the col-
lective role of the ‘forgotten’ scientists, designers, engineers, technicians, test pilots, military
and civilian planners, businessmen, and others who helped to deliver victory via all three
services in all theatres of the Second World War. He mentions OR (and Blackett) but just in
passing. Some of his more developed examples include
1 the PhD students who initially developed a prototype magnetron in Britain at the Uni-
versity of Birmingham and the chain of people who ensured that it was taken to Massa-
chusetts Institute of Technology (MIT) to be further developed and then produced in
the US as the basis of all compact radar systems used on aircraft and ships by the Allies;
2 the British test pilot who first suggested that the American P51 Mustang’s original
engine (US designed and built) made it an indifferent performer but that with a Merlin
engine as used in the Spitfire, the Mustang would be a world- class long- range fighter
(with a much longer range than the Spitfire) and the chain of people who ensured that
Packard- built US Merlin engines were used;
3 the American tank designer whose ideas became the basis of the Russian T34 tank and
the chain of people who ensured it progressed through initial and in- service develop-
ment in Russia;
4 the ‘Seabees’ (US Navy Construction Battalions) who were crucial to all the US cam-
paigns against the Japanese; and
5 the planners who made D- Day work as perhaps the best example of highly successful
planning in the whole war.
Having borrowed from Kennedy’s ideas to recraft earlier prototype versions of the Table 2.1
capability- culture concept during the third draft of this book and tried them out, redrafting
the tales developed in later chapters, it now seems obvious to me that the revised structure
of the capability- culture concept greatly improved the usefulness of the UP concept as a
whole, and the basic idea of a capability- culture concept of some kind embedded in a UP is
crucial. Ongoing further testing and development are clearly needed, but the basic nature
and the role of a capability- culture concept are clear and robust.
The Checklist Manifesto: How to Get Things Right (Gawande, 2011) was another impor- tant contributor to shaping the capability- culture concept underlying the UP concept. It has
been coupled to Kennedy’s contribution because it was encountered at the same time as well
as addressing directly related issues.
Atul Gawande is a US based surgeon, an advisor to the World Health Organization, plus
many other bodies, and a best- selling author. He came to my attention when he was invited
to give the 2014 Reith Lectures on the BBC. He treats checklist concepts borrowed from
aircraft safety, construction industry planning, and other contexts as planning and com-
munication tools in a medical context, addressing capability- culture issues in surgical teams,
plus some other contexts, with very wide implications for planning in all contexts.
Gawande uses the word ‘checklist’ when I would use ‘plan’, with checklists as a special case
of plans in some cases and with checklists as procedures for testing the robustness of plans
in other cases. His checklist concept is arguably overstretched in a way which confuses some
of his messages by making them seem less general in their applicability than they actually
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
64 Foundations
are. However, this is a minor presentational quibble, ‘checklist manifesto’ has a much better
ring about it than ‘plan manifesto’, and his key messages are very important. Many of his
key messages chime with those taken from Kennedy, and they helped to shape the emerg-
ing structure of the capability- culture concept. For example, Gawande explains why it is
important for members of a surgical team to know each other’s names, spend a minute or
two discussing what they are planning to do and key sources of uncertainty before starting,
and for nurses and junior doctors to feel able to point out potential procedural errors by a
senior surgeon which are anticipated because of prior actions before the errors are made – the essence of good communication pathways, teamwork, and enlightened governance.
As another example, Gawande explains in considerable detail how the knowledge needed
by the aircrew of an airplane to deal with a low- probability potentially catastrophic fault
was updated in weeks for all aircrew flying Boeing aircraft after one incident demonstrated the need to change the advice and accessed when it was needed soon after the update in a
way which probably saved many lives. This example illustrates the essence of good pathways
to feedback learning, to accumulate knowledge in a structured manner, and to access key
knowledge about detailed contingency plans when needed, as well as a procedure to test
plans being put in place to deal with an emergency.
As a further example, Gawande indicates the kind and degree of ‘professionalism’ we
expect of aviators includes a concern for ‘discipline’ (following a prudent procedure in func-
tioning with others) which is currently not expected of surgeons and some other professions
society depends upon. He argues this kind of discipline should be not only expected; it
should also be demanded. This is linked to the need to deal with complexity which requires
a team effort to an extent which challenges conventional wisdom about the autonomy
required or simply desired by some players in a team, discussed explicitly in terms of a
change in what society should mean by the right stuff, a term made popular by Tom Wolfe’s
1979 best- selling book The Right Stuff about the history of the Mercury space programme and the early development of NASA.
The focus of this book is planning in private and public sector commercial organisations,
but the starting place for the UP phase structure and the current form of the capability-
culture concept are both attributed to ideas drawn from warfare. This may raise concerns
for you about the transfer of warfare- based ideas into the contexts of direct interest. Fur-
thermore, the surgical team emphasis of Gawande’s work may also seem remote from a
common interpretation of the focus of this book. This may raise more general concerns
about the relevance of some sources of UP concepts. However, this kind of transferability or
portability of concepts is a characteristic of MS/OR grounded ideas. The generality of the
UP concept is one of its defining characteristics, with significant earlier precedents in terms
of all ‘contributors’ and all associated application areas. Portability is simply an aspect of the
generality of the UP concept in the MS/OR tradition.
Contributor 7: the SP for projects perspective of the first four UP phases
The structure, function and labels of the first four phases of the UP were derived by gener-
alising the front- end phases of the SP for projects discussed in Chapter 7, following a sug-
gestion by Stephen Ward. This suggestion was important for several reasons, and the project
management provenance involved has useful implications.
First, it helps to clarify the interpretation of the front end of the UP at an overall what
needs to be done level simply because it facilitates effective integration of all the concerns
outlined earlier in this section.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
A ‘universal process’ 65
Second, when we come to considering the specific process for projects which is developed
in Chapter 7 at a what needs to be done level, the common structure is an important basis
for understanding the relationship between the universal process concept and specific pro-
cesses in a general sense as well as understanding the specific process addressed in Chapter 7.
Third, when we come to use the UP in practice in any context it has significant how to do
it implications with directly related and very important what needs to be done implications.
This is because there is a rich experience base associated with the front end of the specific
process for projects of Chapter 7 which is directly relevant to making operational use of the
UP concept.
Fourth, by decomposing the process of making working assumptions into four phases
the UP concept clarifies four significantly different kinds of working assumptions which are
required as the process progresses, orders their consideration in a useful manner, and reminds
us to address all four sets in relation to a very unrestrictive set of framing assumptions.
Finally, separating the four sets of working assumptions associated with the first four
phases helped to trigger the recognition that we need to explicitly consider a further sepa-
rate set of working assumptions associated with all the capability- culture concerns plus their
overall role in the UP portrayed by Figure 2.1. In retrospect, this explicit recognition is
clearly needed. This further set of working assumptions is usefully seen as applying to the
way all seven phases work with the capability- culture concept to give the recommended UP
concept an operational form which might be adapted by you or your organisation.
The UP concept is explicitly associated with the key framing assumption:
Some explicit operational UP concept is essential to avoid implicitly accepting a situation where more than one UP variant is probably in use and they are likely to be incompatible.
Consider what this implies in terms of each of the first four phases of the UP, the UP’s
capability- culture concept, and the UP concept as a whole.
‘Capture the context with appropriate clarity’ is a generalisation of an extensively tested
‘define the project’ phase. This generalised capture the context phase is about forming a
strategic vision of the context in ‘big picture’ terms. This vision must include key concerns
and their interconnectedness while the process is as general as possible. There must be a
clear focus on what seems to really matter. When any important limiting assumptions about
the context are detected via testing later in the process, the key defining features of this
strategic view of the context need to be adjusted and elaborated as necessary as the process
evolves. Even in a relatively limited project planning context the equivalent phase involves
consideration of all seven Ws, starting with the who, the key players. It addresses all objec- tives of all the key players explicitly, as well as all the other important seven Ws concerns. The
reason the ‘who’ and the ‘why’ are the starting points is that they jointly drive what it is that
has to be done and how it needs to be done. Misaligned or misunderstood objectives are
crucial. These and other related ideas are developed in more detail with a focus on projects
in Chapter 7, with extensive backup in Chapman and Ward (2011). Of central importance
in this subsection, these ideas can be built on in any context, as demonstrated throughout
Part 2.
‘Select and focus the process for appropriate clarity’ is about making conscious and explicit
planning the planning decisions to start gaining efficiency with a minimal loss of effective-
ness. This must begin as soon as the key context issues have been captured in a way which
allows the key planning issues to begin emerging. The select and focus the process phase
addresses a big picture view of all potentially useful processes, a very different set of concerns
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
66 Foundations
and working assumptions to those considered in the capture the context phase. When cru-
cial working assumptions about process choices look debatable during robustness testing
later in the process, revisiting the process assumptions used to gain efficiency is crucial, as
is adapting the process as a result whenever this is appropriate. There is no one best way to
proceed, and initial working assumptions may need revising. As a generalisation of an exten-
sively tested focus the process phase in a project planning context, the select and focus the
process phase in the UP opens a portal to all useful processes. This includes the whole range of soft and hard OR methodologies. But of crucial importance, it also includes any other
methodologies which may be useful. A rich range of methodologies associated with plan-
ning in particular areas offers a whole universe of particularly attractive and relevant options.
Some of these planning process option choices may have some MS/OR in their background
DNA, the CPM (Critical Path Method) process for example, but others may be entirely dif-
ferent in their origins. Drawing on process ideas entirely unrelated to mainstream MS/OR
is potentially very important, and an unrestrictive portal into process choices from anywhere
is an important feature of the select and focus the process phase which can be attributed to
contributor 7. Using very different methodologies on different passes may prove crucial,
and the ideas developed in Chapter 7 and Chapman and Ward (2011) can be generalised
and built on in any UP application context, as demonstrated in Part 2 chapters.
‘Create and enhance plans for all relevant concerns’ then uses the context and process
assumptions made in the first two phases to make further dependent plan creation and
enhancement assumptions. Separating this phase opens an additional unrestricted portal,
this time to all plan creation and portrayal tools relevant to the aspects and areas of planning being addressed. This create and enhance the plans phase is a central concern of the UP, it is
well beyond the bounds of mainstream MS/OR processes by design, and it takes us beyond
Chapman and Ward (2011) in some important respects which are explained in Chapter 7
and illustrated in all Part 2 chapters.
‘Shape the plans using models of some key issues’ leaves until the fourth phase the
introduction of modelling which the basic 1950s/60s OR process addressed in the second
phase of the Table 2.2 portrayal. This delayed introduction of modelling facilitates a less
restrictive and more powerful perspective overall than the Table 2.2 approach because of
the enhanced clarity provided by the earlier Figure 2.1 phases. Again, the portal involved is
by design not restricted to MS/OR models. By this point we may be using a process which
is very different from a traditional OR process, drawing on a different vision of ‘plans’ and
a different vision of associated model sets. For example, in Chapters 8 and 9 you may find
it useful to visualise the planning involved using a range of conventional strategic planning
or safety and security planning tools within a partially structured mess, the partial structur-
ing of the mess being the result of earlier use of the UP concept in comparable situations
with very limited obvious direct links to traditional MS/OR visions of what an OR process
is about.
The further set of working assumptions associated with capability- culture concerns will
influence all seven phases of the process, with implications which can obviously have an
impact on and interact in specific ways with the four sets of working assumptions just con-
sidered plus those associated with phases 5, 6, and 7. For example, if someone leading the
application of a UP approach does not understand soft OR processes, they are not going to
be part of the toolset which might be applied, with similar observations applying to more
basic planning tools like basic project planning techniques or cash flow models.
A ‘closure with completeness’ concept which includes the proposed specific operational
form of the UP can be associated with the capability- culture set of working assumptions.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
A ‘universal process’ 67
Contributor 8: a ‘learning organisation’ as an explicit corporate goal
The Fifth Discipline: The Art and Practice of the Learning Organisation (Senge, 1990) is a popular book in the field of organisations and their management which is focused on
achieving ‘a learning organisation’ using a systems dynamics perspective. Rethinking the Fifth Discipline (Flood, 1999) takes a systems overview of ‘the learning organisation’ which generalises the systems dynamics basis of Senge’s approach to embrace further systems ideas
originated by Ackoff, Beer, Bertalanaffy, Checkland, and Churchman. Flood is using sys-
tems ideas from authors I have drawn on earlier in this section in a different way to do
something rather different, but there are overlaps, and the idea that a learning organisation
ought to be an explicit corporate goal is widely accepted as a sensible proposition for many
people working in a wide range of organisations. An organisation which explicitly promotes
‘reflective practice about systematic approaches to planning’ as part of the best practice
being sought is the EP aspect of direct interest.
A ‘learning organisation’ in the particular sense developed by Senge and Flood, pur-
sued using variants of their approach, is not a central part of the EP agenda. However,
pursuing a learning organisation corporate goal in a manner suited to the organisation
being considered using the capability- culture concept as an explicit tool for promoting
‘reflective practice about systematic approaches to planning’ is part of the UP agenda.
This idea is developed in all Part 2 tales. It is not given a lot of attention, but it is a good
example of a UP feature which might be much further developed if it is relevant to your
organisation.
Teaching for Quality Learning at University: What the Student Does (Biggs and Tang, 2011) provides a conceptual framework for teaching referred to as ‘constructive alignment’.
This involves starting with the intended learning outcomes for different kinds of students
with different levels of motivation to achieve different goals, and given these intended learn-
ing outcomes for the parties directly affected, seeks to align course design, delivery, and
assessment. This approach is consistent with an EP approach to all planning, and generalis-
ing some of their ideas to address corporate learning processes might prove useful in some
organisations. Their notion of gateway concepts has already been drawn on as a characterisa-
tion of concepts like strategic and tactical clarity, opportunity efficiency, and the four Fs for
project planning.
Contributor 9: all other relevant concerns, plus a broad synthesis
A formal provision for further contributors which you or your organisation may find useful
plus a broad synthesis of all contribution sources is the final ‘contributor’ category in my
list, a further ‘closure with completeness’ concept in the same spirit as element 11 of the
capability- culture concept. The role of contributors unrelated to MS/OR may build signifi-
cantly on portals of the kind opened by contributors 6 and 7.
You may wish to alter the UP in terms of structure or detail as a consequence of past
experience or as new experience is gained and lessons are learned from other sources.
The UP concept of Figure 2.1 is flexible by design. However, don’t rush to change
the Figure 2.1 and Table 2.1 portrayals and interpretations too quickly, and develop
some feel for how the UP concept of Figure 2.1 works before you commit to any
changes which are not tested routinely. And think about proposed changes as alterna-
tive working assumptions in relation to all the working assumption sets discussed in this
chapter.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
68 Foundations
Further clarifying what a UP is and is not from an EP perspective
At an OR Society Annual Conference in the 1990s an intriguing demonstration of soft OR
was a central attraction. It involved extended parallel sessions in a number of rooms which
delegates could wander in and out of. The session in each room employed one of about four
different soft OR approaches to address the questions, ‘Should OR be rebranded, and if so,
what should it be called?’
It was an extremely interesting live demonstration of different soft OR approaches at
work. It did not resolve the OR branding controversy, but it is probable that no approach
could, then or now. Most members of the OR community agreed a long time ago that the
OR label is seriously unhelpful from a marketing perspective, but they still cannot agree
on what to change it to, because of the wide range of perspectives on what it is about.
‘Decision Analytics’, ‘Business Analytics’, and just ‘Analytics’ are one illustrative set of
current propositions, but none of these options enjoys unanimous support. Management
Science and the composite MS/OR term are alternatives with different but related brand- label problems.
The spirit of 1930s/40s’ OR is deeply embedded in the UP of Figure 2.1, and this spe-
cific operational form of the more general UP concept has been shaped by all following OR
developments. The UP concept and the EP concept more generally also embed the spirits
of early- 1900s Scientific Management, later Management Science critiques, Soft Systems,
and soft OR critiques. One description of OR promoted by the OR Society which many
members endorse and find useful is ‘the science of better’, and the UP concept and EP more
generally reflect this spirit. The MS/OR roots of the UP and EP concepts are important and
useful to understand, even if you have no further interest in what they involve.
However, within the scope of the planning focus which underlies this book it is also
important to appreciate that the UP and EP concepts have provenance, scope and ambitions
which go well beyond MS/OR as these areas are usually interpreted, and well beyond the
individual contributions of the many obvious and not so obvious related adjunct subject and
discipline contributions.
All disciplines and professions contributing toolsets, skill sets and mindsets which are rel-
evant to your planning concerns can be drawn upon, and these disciplines and professions
can, in turn, freely draw on an EP perspective, including the UP concept.
Equally important is appreciating that the planning focus which underlies this book
involves a significant narrowing in focus relative to many of the ambitions of MS/OR and
the numerous other disciplines and professions drawn on within the UP.
To make the central issues of concern in this section as clear as possible,
The UP concept is emphatically not a ‘rebranded’ OR process and EP is emphatically not ‘rebranded’ MS/OR, because they draw upon and are relevant to a wide spectrum of approaches to management decision making in planning contexts requiring both formal and informal approaches, with crucial supporting craft and team- working skills.
The UP of Figure 2.1 and OR processes are very closely related, but from an EP perspec-
tive they are also quite different concepts in several crucial respects. Comparable differences
apply to the broader EP and MS/OR concepts, arguably driving the UP and OR process
differences.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
A ‘universal process’ 69
Relative to the whole spectrum of ambitions addressed by MS/OR approaches and other
related disciplines and professions, the UP of Figure 2.1 and the EP concept as a whole as
explored in this book have a very restricted focus – planning in formal terms plus underpin-
ning informal planning within commercial organisations, bearing in mind the crucial impor-
tance of informal planning and capability- culture concerns. However, within this restricted
focus a very wide range of ideas are relevant – by definition and design there are no limits.
The front- end structuring of the UP of Figure 2.1 facilitates both the efficiency gains flow-
ing from the heightened focus and the effectiveness gains flowing from the generalisation of
the areas drawn on within the planning focus. Furthermore, aspects of the capability- culture
concept change both the emphasis and the basic foundation level framing assumptions.
Some people might want to ignore this degree of difference and suggest that EP is just
marginally narrower in some senses and marginally wider in other senses than MS/OR, or
‘modestly narrower but deeper’, to simplify the differences. But in my view, this is too sim-
plistic. From an EP perspective there are too many kinds of subtle and not- so- subtle differ-
ences that matter to use this kind of analogy. The underlying features of EP which drive all
these differences as I see it is the lack of exclusive ownership or parentage of the EP concept
by any of the contributing disciplines. EP has multiple parents and guardians with many
different perspectives to draw upon, and it is worth being explicit about this characteristic,
recognising its importance.
To begin to clarify why, when discussing the deep and important two- way relationships
between empire building and science in the early days of empire building by European pow-
ers in the best- selling book Sapiens: A Brief History of Humankind, Yurval Harari (2014) points out that ‘setting up a scientific discipline was an imperial project’. There are still
very strong echoes of this imperial project flavour in most science and discipline areas. In
the 1990s during an OR Society Annual Conference public discussion, I suggested to the
soft OR advocate giving a paper on his brand of soft OR that he was misrepresenting what
traditional OR was about, and my view of the traditional OR process (a prototype UP in
my mind, but not his) could and should absorb his view of soft OR in its front end (in the
implicit select and focus the process phase of my prototype UP). He immediately accused
me of ‘intellectual imperialism’, deeply annoyed by the implications. He saw traditional OR
as a back- end process option for a set of framing assumptions which made soft OR of his
particular brand the basic framework. He was, in effect, promoting a new and competing
empire himself, an imperial enterprise playing by his rules. This illustrates the nature of terri-
torial issues which remain powerful forces within both academic disciplines and professional
societies. Some of the latter are empire builders on an epic scale, with very limited interest
in collaboration across the boundaries of professional societies, although this is not true of
MS/OR related societies I am familiar with, the OR Society perhaps being the least territo-
rial of all the societies I have come across.
If you are interested in the implications of these territorial issues in an academic context,
Lewis (2017) provides some interesting observations about the social sciences, psychology,
and economics in particular, in addition to a very useful background for understanding
decision- making bias and related planning failures.
One key foundation level framing assumption feature of the UP concept and EP more
generally is they belong to all disciplines and all professions and all professional societies and
all individuals who choose to use them – by definition and design. You and your organisa-
tion can adopt whatever features seem relevant and adapt both the UP and EP to back-
ground and context considerations regardless of the disciplines or professional perspectives
involved. The UP and EP concepts are an intellectual free- trade zone, explicitly not a basis
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
70 Foundations
for a new empire and not a territory for imperial occupation by any existing empires on a
basis which excludes free access by others.
Four particular features of the UP and EP concepts may be worth noting to help clarify
why they are conceptually different from the OR process and MS/OR perspectives.
First, the full range of arts and crafts, as well as sciences relevant to planning and associ-
ated management, needs to be drawn on, without inhibition, well beyond the range of any
discipline or profession, and this is an essential operational feature, not a marketing ploy.
Second, the role of imagination and creative aspects of planning is crucial, and imagina-
tion and creativity are clearly not the exclusive territory of any single profession, discipline,
or other grouping.
Third, making teams and wider collectives work effectively involves some team members
having important social skills plus a genuine interest in other people which again is beyond
the bounds of any professional or academic grouping.
Fourth, the capability- culture concept is a foundation level component, and the right
stuff is a key component of the capability- culture concept. This yet again takes the UP and
EP concepts beyond the reach of any single group, because no group can claim exclusive
ownership of a concept like the right stuff, and different people are free to interpret what it
means with some very different nuances.
Those with extensive MS/OR backgrounds are very much part of the target audience,
but they are not catered to in a direct way as a special group in this or following chapters
beyond helping them to see some key links which may be of limited interest to other target
readers. On the website discussed briefly at the end of this book, they will be addressed
directly as one of the groups who may be interested in using this book in specific planning
areas and what might be called ‘enlightened process planning’, building on this chapter is
currently the assumed focus.
Whatever your background and interests, I hope this chapter has provided the grounding
you need to make use of the UP concept and all the other EP tools explored in this book at a
strategic clarity level even if you have had very limited previous exposure to MS/OR. It was
written with you in mind. If you use specific EP tools at an operational level, what you call
them should be judged in your own best interests according to the context, working from
whatever professional perspective you are comfortable with. Their presentation in this book
has been framed for anyone interested in using them.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
A key component of most plans is estimates of the parameters which define the anticipated
outcome values of all relevant performance measures. Examples of relevant performance
measures include levels of future corporate revenues and costs, associated sales levels and
prices, the reliability, safety, and security of future operations, the cost of capital, the dura-
tions and costs of projects. Key parameters defining these values include expected outcomes
and outcome range measures.
Corporate goals in terms of overall performance estimates need to be linked to compo-
nent performance parameter estimates and plans via a goals– plans relationships framework,
and goals– plans relationships usually need an understanding which has been developed
bottom up as well as top down. For example, starting at the bottom, how long project
activities will take and the cost per unit of time of associated resources will define part of
the associated direct cost. Other aspects of the cost of project activities which also need
consideration include the cost of the materials used, associated equipment purchases, and
indirect costs. Corporate and operations management examples involve comparable goals–
plans relationships.
In all planning areas common practice often involves the use of ‘point’ (single- value)
estimates for performance parameters at a component level and at aggregate levels. Often,
the component estimates are combined via ‘deterministic’ planning models (models which
do not use probabilities) to yield overall performance estimates. This usually means clarity
about the implications of underlying uncertainty is negligible.
Competence in EP terms means that all performance parameter estimates at all levels of component aggregation in all plans must meet or exceed a minimum acceptable level of clar- ity. Furthermore, they must be framed by a process which can provide additional clarity as and when it may be needed in a clarity efficient manner.
Even the mandated minimum level of clarity requires a coherent view of uncertainty and
underlying complexity in terms of five key issues:
1 expected outcomes and possible associated estimation bias;
2 outcome range measures and possible associated estimation bias;
3 appropriate ‘aspirational targets’, if relevant;
4 robust associated ‘commitment targets’, if relevant; and
5 associated ‘balanced targets’, by default for a minimum clarity approach assumed equal
to expected values.
The basic ‘ABCs of targets’ embedded in the working assumptions used for issues 3 to 5 in a
minimum clarity estimate approach might be as follows, within the framing assumption that
all working assumption choices should suit the context.
3 Low to high clarity approaches and the ‘estimation- efficiency spectrum’
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
72 Foundations
‘Aspirational targets’ might be associated with a ‘plan A’ intended to aggressively exploit
any good luck that might occur.
‘Balanced targets’ might be associated with a ‘plan B’ intended to portray a plausible
mid- range outcome, assumed to be the expected outcome for convenience and simplicity in
a minimum clarity context.
‘Commitment targets’ might be associated with a ‘plan C’ intended to cope with bad
luck, including all the reasonably plausible knock- on implications of this bad luck possibility,
based on a plausible preliminary view of what a bad luck scenario might look like.
To begin to illustrate how an appropriate minimum clarity approach might compare
to common practice in a project planning context, a basic common practice Critical Path
Method (CPM) approach (Moder and Philips, 1970) often begins with direct point esti-
mates for the duration of each component activity. These estimates are then used in a deter-
ministic CPM algorithm based on a network model of activity precedence relationships to
compute a point estimate of the project duration. This defines what some people may see
as an appropriate point estimate of the project duration, perhaps assumed to be a balanced
target outcome equal to the expected outcome. However, these component activity and
overall project duration parameter estimates may be regarded by other people as commit-
ment targets, although aspirational targets may have actually been the initial intention of
those preparing the estimates. At the very least everybody involved needs a common basis
for interpreting whether point value estimates are suitable balanced targets, different unbi-
ased expectations, reasonable aspirational targets, reliable commitment targets, or ‘none of
the above’. If ‘most likely’ estimates are used as point estimates at the activity level, adding
component most likely estimates which are not additive results in a very muddled overall
estimate which is one common example of none of the above. Very often further clarity for
very little extra effort is both possible and invaluable – a ‘quick win’ which can be captured
with minimal effort by an enlightened approach.
These CPM based examples involve a project management context, the illustrative con-
text used throughout this chapter. But all the concepts developed in this chapter also apply
to operations and corporate planning contexts, and examples building on this chapter’s
ideas involving operations and corporate planning contexts, as well as project planning, are
illustrated in Part 2. For example, Chapter 5 looks at probabilistic modelling of uncertainty
about how long it might take to restock an inventory for ‘safety stock’ purposes, employing
a simple approach in a very simple operations management example context. Chapter 5 also
considers an even simpler treatment of unbiased expected value estimates for determinis-
tic optimal order quantity models when probability distributions are of limited interest.
Chapter 9 looks at the implications of the large number of people who might be killed in a
catastrophic railway incident caused by an accident or malicious terrorist attack, with cor-
related injury levels, physical damage to the system and other knock- on implications, a rela-
tively complex operations management example with direct corporate strategic management
implications. Intermediate levels of complexity are illustrated in Chapters 6 to 8.
In all contexts we want to be ‘clarity efficient’ – defining clarity as insight which can be shared, we want to achieve a maximum level of appropriate clarity for any given level of
effort/cost associated with acquiring clarity.
All clarity efficient approaches may provide more clarity for less effort than some com-
mon practice approaches for more than one reason, and high clarity approaches can offer a
very rich bundle of advantages. To some extent seeking more clarity of one kind tends to
generate more clarity of other kinds, and which kind of clarity should be seen as the primary
objective may be debatable. If half a dozen different clarity enhancement objectives are all
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 73
achieved concurrently, which is the primary objective and which are the secondary objectives
does not really matter in some ways, although motivation, presentation or storytelling con-
cerns may be important. Everybody can work in a coherent manner towards the common
goals even if their priorities are not the same, and this unity of purpose despite the pos-
sibility of different individual priorities can be a crucial aspect of teamwork and broader
collaboration.
In some contexts, we need as much clarity as we can get of all available kinds. Most of the
time a modest level of clarity will suffice, provided we are clear about what clarity charac-
teristics matter in this particular context. On the first pass of an iterative planning process,
when it is not clear which project aspects may be important, a minimum clarity approach
to estimation is usually appropriate. On later passes, once it is clear what matters and why it
matters, much more clarity of specific kinds may be essential.
To explore the whole ‘estimation- efficiency spectrum’ addressed within the focus of this
chapter, the next section considers a minimum clarity approach to estimating the duration
of one activity in a project. The following few sections explore options which provide mod-
est but useful amounts of further clarity at a relatively low cost. More following sections
gradually add further aspects of clarity by addressing the underlying complexity which may
be worthwhile, eventually considering the role of clarity efficiency, risk efficiency, and then
opportunity efficiency. This takes us well beyond just effective and efficient unbiased ‘esti-
mation’ plus recognising the role of the ABCs of targets as deliverables of clarity efficient
estimation processes. Indeed, once risk efficiency is clearly understood, it usually becomes a
central goal for most people in most organisations, with unbiased estimates plus the ABCs
of targets usefully seen as by products or component aspects of an approach designed to seek
opportunity efficiency, including component risk efficiency and clarity efficiency. However, a
need for unbiased estimates of expected outcomes and associated ranges motivated most of
the organisations involved in this chapter’s tales, so starting from an estimation perspective
with a focus on the five key concerns about estimation noted earlier is a useful way to begin.
This gives us a clear basis for all relevant estimation bias concerns plus an understanding of
the aspirational, balanced, and commitment target distinctions involved in the basic ABCs
of targets at the outset. We can then gradually add further interdependent considerations.
This approach is a useful way to begin to explore the complex relationships between
unbiased estimates, clarity efficiency, risk efficiency and opportunity efficiency. EP requires
strategic clarity about these relationships. Risk efficiency and opportunity efficiency become
the obvious central or driving concerns at the upper end of the clarity range, but they are
actually underlying concerns at all levels of clarity, including providing an underpinning
conceptual basis for a minimum clarity approach.
An important simplifying assumption is maintained throughout this chapter – we are
dealing with a single attribute at any one time, like duration or cost. Whenever multiple
attributes need consideration, this chapter assumes that either of the following:
1 All relevant attributes can be expressed in terms of a single attribute without difficulties
needing explicit attention. For example, converting uncertainty about project duration
into a suitable component of overall project cost is straightforward.
2 The pursuit of the primary attribute on its own will deliver an aligned optimality in
terms of all other attributes, so there is no need to worry about attribute trade- offs.
Problems associated with managing trade- offs between multiple attributes which are not
straightforward will be considered later, starting in Chapter 4.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
74 Foundations
The next section introduces the histogram and tree (HAT) approach to stochastic (proba-
bilistic) modelling at a minimum clarity level. A single source of uncertainty is portrayed by
a histogram with an underlying probability tree interpretation using the simplest feasible
probability distribution. This approach is gradually made more sophisticated in subsequent
sections. The HAT approach is a central EP conceptual and operational tool. It is a synthesis
of several approaches often viewed as alternatives. Its more complex relationships with these
alternatives are outlined later in this chapter in a separate section which addresses the nature
of the synthesis defining the HAT approach in a reasonably comprehensive overview manner.
The span of applications of the HAT approach discussed in this chapter ranges from very
simple examples at the outset to quite complex examples towards the end. The more com-
plex examples draw on my experience with BP North Sea project planning over an eight year
period beginning in 1976. During this period BP developed computer software to imple-
ment a HAT based approach using a probability arithmetic basis for combining probability
distributions. Most organisations now using HAT based approaches use off the shelf Monte
Carlo simulation software. The simple examples considered early in this chapter draw on a
2001 UK Ministry of Defence (MoD) consulting contract which made use of a range of
developments in the intervening decades. Some of the early discussion draws on BP practice
linked to their use of HAT based software and comparable use of simulation- based software
when this is helpful.
A minimum clarity approach to estimation
To illustrate the discussion in the first few sections of this chapter, consider a revised and
embellished version of a conversation I had with ‘William’ (not his real name) in 2001.
I was undertaking a review of UK MoD project risk management processes. My MoD con-
tract manager, who initiated the review, was responsible for some of the processes used to
approve new MoD projects and budget amendments. At his request the review was in part
based on conversations with senior MoD personnel, including the air vice- marshall respon-
sible for the MoD’s procurement budget, who we will return to later. William was the pro-
gramme manager for a major weapon programme – think of his project as a next- generation
warship – with ‘Warship William’ as a reminder of his role.
Our discussion began by William suggesting that we use an example of immediate rel-
evance to him. Say he wanted members of his warship programme team to provide an unbi-
ased estimate of how long it would take to get approval for an unanticipated design change
which had significant financial implications. Say the need for the change had just emerged.
Say he wanted this estimate in order to get his team to re plan other aspects of his warship
programme. What was the best way for his team to proceed? I responded ‘start with a mini-
mum clarity estimate’, which I would illustrate by producing one with him, where:
by definition and design a ‘minimum clarity estimate’ provides a clarity efficient estimate at the minimum acceptable level of clarity given current working assumptions.
My first question was ‘What was a plausible maximum duration, using an approximate P90
(90th percentile) value as a default approach unless some alternative percentile value (like
a P95) was preferred with good reason?’ A P90 has a 90% chance of not being exceeded.
William’s response was a P90 estimate of 18 weeks.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 75
My second question was ‘What was a plausible minimum duration, using an approximate
P10 (tenth percentile) value as a default approach or an alternative consistent with the plau-
sible maximum?’ William’s response was a P10 of 2 weeks.
These two questions are a ‘clarity efficient question set’ for an initial minimum clarity
estimate. They use a minimum level of effort to extract the information needed in the most
appropriate form and order given current working assumptions.
Figure 3.1 portrays the recommended ‘minimum clarity estimation model’ based on Wil-
liam’s answers to my clarity efficient question set.
The model used by Figure 3.1 assumes a uniform probability density function, a key
working assumption used for all minimum clarity estimates. This implies a linear cumulative
probability distribution.
Figure 3.1 portrays William’s 18 week plausible maximum as a P90 and his two week
plausible minimum estimate as a P10. The corresponding absolute minimum and maximum
values for the model are 0 and 20 weeks, and these values make it clear that there was an
assumed 10% chance of a value below 2 weeks and an assumed 10% chance of a value above
18 weeks.
Further key working assumptions that a minimum clarity estimation approach is based on
in all contexts which need emphasis now include those preparing the estimate may not know
at this stage in the process whether or not potential variability matters, but an unbiased view
of both the expected outcome and potential variability might matter, aspirational target and
commitment target distinctions might matter, and data might or might not be available.
The minimum clarity approach assumes that the balanced target is equal to the expected
outcome and defined as the plausible minimum plus the plausible maximum divided by 2,
0
probability
(density)
0 2 10 18 20
outcome value in weeks
probability
density
format
0 2 10 18 20
outcome value in weeks
0
1.0 cumulative
probability
0.5 the median (P50) is equal to the expected value and both are 10 weeks
cumulative
probability
format
0.05
0.9
0.1
Figure 3.1 A minimum clarity estimation model with the P10 = 2 weeks and the P90 = 18 weeks.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
76 Foundations
in this case (2 + 18)/2 = 10 weeks. Assuming any alternative plausible maximum and mini-
mum bounds were consistent and departed from the P10 and P90 in a symmetric manner,
any plausible range estimate would yield the same result. Given the uniform density function
assumption, the expected value is always equal to the P50.
There was no need to plot Figure 3.1 to determine the expected outcome and make direct
use of a minimum clarity approach, but Figure 3.1 is the conceptual and operational model
which underlies the use of a minimum clarity approach. If William wanted his team mem-
bers to start planning with a simple deterministic CPM model, they could immediately use
the expected value of 10 weeks as the design change approval estimated duration parameter
for a CPM model. However, William should reveal the 2– 18 weeks plausible range used to
estimate the 10 weeks expected value to anyone who would find this information useful,
unless there were good reasons for not doing so. Indeed, a usefully simple and practical
approach for communication between all parties involved would be describing this mini-
mum clarity estimate as 10+/−8 weeks, with a P10– P90 interpretation of the range and an
explicit acknowledgement that both the expected value and the range are approximate with
an underlying Figure 3.1 interpretation.
The expected value definition and the distribution shapes used in Figure 3.1 are the
simplest available. The uniform density function means that expected values are midpoints
in all relevant minimum clarity range estimates. The uniform density function also means
that the cumulative function is linear, useful when minimum clarity estimates are used for a
‘minimum clarity option choice model’ as discussed in Chapters 4 and 6.
It is always important to avoid estimating ranges in terms of absolute maximum or mini-
mum values unless there is an identifiable appropriate physical minimum or maximum
value – the P0– P100 range should normally be avoided for estimate elicitation purposes.
The P10– P90 range is usually the simplest robust interval (range- based) estimate approach to estimating probability distributions, with obvious alternatives like a P5– P95 range or a
P1– P99 range.
It is useful to recognise that most people have to think about a range to provide an
unbiased expected value estimate, and both conscious and unconscious bias is generally
decreased if they are asked to estimate a range explicitly.
If this range information is available, sharing it makes obvious sense, unless hiding it
serves some legitimate intended purpose. Insisting on sharing it helps minimise conscious bias and keep everybody honest – but there are no guarantees. It is sometimes useful to
remind ourselves that it only pays to be honest if the other people involved have a common
understanding and play by the same rules, key working assumptions relevant to all estima-
tion processes.
It is always useful to recognise that most people tend to underestimate uncertainty, and
unconscious bias effects like ‘anchoring’ need to be avoided as far as possible, the reason for
asking for first the P90 and then the P10.
It is crucial to always recognise that models which are convenient for dealing with a more
complex reality are models based on working assumptions which can be tested and changed when appropriate.
It is also crucial to understand that models should never be confused with reality. In the current context a very useful reminder of the difference between models and reality is
provided by Figure 3.2, which illustrates the assumed relationship between the working
assumptions involved in using the Figure 3.1 model and one reasonable presumed underly- ing reality, recognising that the actual underlying reality is inherently ambiguous, other presumptions are possible, and different perceptions may be plausible and relevant.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 77
probability
(density)
0 2 10 18 20
0
outcome value in weeks
one presumed reality although multiple modes may be involved and the actual reality may be very different in other respects
working assumptions used for the Figure 3.1 minimum clarity estimation model
Figure 3.2 One presumed reality underlying the approximation involved in Figure 3.1.
It is important everybody involved in using minimum clarity estimates – or any comparable simple estimates – understands the kind of approximation relationship Figure 3.2 portrays. The fact that different people may all have slightly different presumed realities is not a con-
cern so long as the model being used is a reasonable approximation to all relevant presumed
realities and everybody understands the implications. The particularly simple nature of the
minimum clarity model clarifies and emphasises this issue, often lost sight of if more sophis-
ticated alternatives are used.
A minimum clarity estimate uses a very simple and crude model which should get the
expected value and the P10– P90 range about right on average, but there should be no illu-
sions about the very different reality being approximated and the ambiguity associated with
the exact nature of that reality. Furthermore, there should be no illusions about the limited
value of a more refined smooth curve like that of Figure 3.2 if different people think it
should be in very different places, some people think two or more modes might be relevant,
and differences in opinion that matter are not explicitly addressed and resolved.
Sometimes a useful secondary interpretation of the P10 is a default value for an ‘aspira-
tional target’. If William wanted his team to be ready to get on with activities delayed while
approval was being sought, an ‘aspirational target’ might matter. A reasonable working
assumption for an ‘aspirational target’ in this case would be a ‘no problems’ target to aim
for to exploit good luck if it occurs, assuming that his team can control the plans post-
approval but not the duration of the approvals process. If the duration of the activity being
addressed was within their control, a no problems interpretation might be given a ‘stretch
target’ label and emphasis to encourage everyone to work as hard as possible in the smartest
possible manner to achieve the best performance feasible. This kind of motivation might
be important as well as understanding the assumed no problems nature of what was being
considered. If a higher or lower alternative aspirational target was preferred, it should be
stipulated if targets are relevant. In William’s context it might be very important for him
to avoid just assuming that 10 weeks was the expected delay and explicitly ask ‘What could
his team plan to do about the delay if it was just 2 weeks?’ At the very least, his team mem-
bers need to address some contingency planning because of the delay their warship project
was facing and think about contingency planning if a good luck or no problems scenario
occurred might be valuable.
Sometimes a useful secondary interpretation of the P90 is a default ‘commitment target’.
A commitment target is what is promised, or ‘counted on’, in some other sense, as distinct
from what is aspired to or expected. A commitment target can be important to manage
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
78 Foundations
possible bad luck. An assumed 90% chance of this commitment being achieved might be
reasonable, given common levels of asymmetry in the ‘penalty function’ associated with
being early or late. But 80% or 95% might be better. Making a judgement about this requires
an understanding of what the technical literature refers to as ‘the asymmetry of the penalty
function’. If a commitment target was relevant, and a higher or lower alternative was pre-
ferred, it should be stipulated.
In William’s context it might be absolutely vital for him to avoid just assuming 10 weeks
was the expected delay and ask, ‘What should his team plan to do about the delay if it was
18 weeks?’ Thinking about contingency planning if a bad luck or ‘significant problems’ sce-
nario arises might be crucial. A commitment target which was a P90 might be a convenient
bad luck outcome to be prepared for, in conjunction with a P10 good luck outcome of 2
weeks plus a mid- range average luck outcome of 10 weeks.
Planners should manage both good and bad luck, and everyone interpreting plans should understand the implications if planners do or do not do so effectively. A failure to manage
bad luck may be particularly visible. A failure to manage good luck may be less visible but
even more important. If all the good luck in a series of activities is lost because its possible
use is not planned for, and following activities cannot be started early whenever starting
early is feasible, then all the bad luck will accumulate without good luck offsets. This will
ratchet out the delay. Good luck will be wasted. There will be no ‘swings and roundabouts’
effect, cancelling out the good and bad luck variability. What some people may choose to
interpret as overall bad luck may actually be an inept accumulation of individual sources
of variability which fails to capture the good luck – a capability- culture liability which is
common, directly driven by a failure to understand very basic ‘uncertainty management’ in
terms of knowing how to manage both good and bad luck effectively.
A ‘balanced target’ is what planners should aim for when balancing the aspirational and
commitment targets for some mid- range estimation purposes. Assuming the balanced tar-
get is equal to the expected value implies a ‘symmetric penalty function’. In William’s case
a directly calculated expected outcome assumed equal to a P50 mid- range value might be
sensible for some quite sophisticated reasons considered later – a minimum clarity estimate
of 10+/−8 weeks is actually much more sophisticated than the analysis thus far may suggest.
More complex examples when balanced targets need to reflect ‘asymmetric penalty func-
tions’ are illustrated in Chapter 8.
It is always crucial to understand the potential differences between ‘expected’ values and
the ‘target’ values of all three kinds – the ABCs of targets. A lot of common dysfunctional
organisational behaviour is a direct result of a failure to recognise these distinctions, for
a number of reasons, all grounded on a failure to understand basic uncertainty manage-
ment. Some involve inappropriate decisions because the expected values which ought to be
used are biased; others involve communication failures and loss of trust concerns because
aspirations, expectations and commitments are confused; and both are frequently part of a
complex muddle. Part 2 considers a number of examples, usually involving a mix of issues.
A very important implication of the aspirational and commitment target and expected
value distinction of relevance in many contexts is that the expected value minus the aspira-
tional target defines what is usually referred to as a ‘provision’ – in this case time over and
above the aspirational target which will be needed on average if expected value estimates are unbiased.
Another very important closely coupled implication is that the commitment target minus
the expected value defines what is usually referred to as a ‘contingency’ – in this case time
over and above the expected value which will NOT be needed on average if expected value
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 79
estimates are unbiased, but will be needed to meet the commitment target with the assumed
level of probability.
Cost provisions and contingencies are usually crucial, and from an EP perspective, a point
(single- value) estimate of any performance measure attribute like duration or cost is only appropriate if the P10 and P90 are equal and the assumption that this information is correct
is shared as appropriate, even in a minimum clarity estimate context. Even if all that decision
makers want is a single value for a deterministic model, they need a range- based estimation
approach to avoid expected value bias, and the ranges used are useful information which
should always be shared unless there are legitimate reasons for not doing so. The minimum clarity estimation model is useful for getting an expected value estimate
in about the right place relative to the expected value for a more sophisticated model closer
to the underlying reality. Usually a minimum clarity estimate significantly reduces expected
value estimation bias relative to any other simple estimation approach, almost invariability
outperforming any point estimation approach. The minimum clarity estimation model is
also useful for getting the P10 and P90 in about the right place, providing a crude but
reasonably unbiased estimate of potential variability. Furthermore, these P10 and P90 val-
ues provide reasonable default aspirational and commitment target values for contingency
planning.
A minimum clarity estimate communicates sufficient understanding of uncertainty if :
1 everyone involved understands it is a minimum clarity estimate, and what that means; 2 it is obvious that there is no need to waste time thinking about the more complex underly-
ing reality or the need for more sophisticated target and commitment values on the current process pass.
A minimum clarity estimate involves a very simple and particularly convenient portrayal
of uncertainty in terms of variability for a ‘first pass’ approach to providing an estimate in
a multiple pass process. A key role for a minimum clarity approach is its use on a first pass
when it is not clear which performance measure components in the goals– plans relation-
ships structure may be important. Assuming estimation takes place within a UP or a related
iterative specific process, a first pass approach implies more effort may be appropriate on
later passes, with the level of additional clarity sought depending on the importance of the
uncertainty identified on earlier passes.
It was convenient to begin the discussion in this chapter with William using a minimum
clarity approach given considerable attention, but in my actual discussion with William it
was soon obvious he needed more clarity. William was shown Figures 3.1 and 3.2, but he
was not actually given all the information just discussed because much of it was not imme-
diately relevant. Furthermore, William actually asked a slightly different question and pro-
vided some data (about to be revealed) using numbers which have been modestly modified
to keep the example discussed as simple as possible. A ‘tale’ has just been told, and continu-
ing the tale involving William in this spirit will be useful.
When I finished my explanation, William then revealed that his estimates of 2 and 18
weeks were the two extremes of five actual observations or ‘data points’ he was aware of,
based on his knowledge of the time actually taken for comparable design change approvals
in the past. He indicated the other three data points were three, four, and six weeks. Fur-
thermore, he indicated two weeks was the current ‘corporate guideline’ for the duration
of a design change approval of this kind. He asked if this information ought to change the
minimum clarity estimate approach.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
80 Foundations
My reply took several parts. First, using the minimum of the five data points to define
the P10 and the maximum to define the P90 was entirely appropriate. Second, if a much
larger number of observations were used to define a percentile range via the minimum and
maximum observed values, a P5– P95 range basis might be more appropriate. Third, the
corporate guideline of two weeks seemed to be based on the assumption that a guideline
ought to be a no problems target – which might be explicit, but I suspected it was a ‘stealth
assumption’. Some people might be confused and think a guideline was an expectation or a
commitment, and in practice it was extremely important to make the nature of all guideline values very clear. Fourth, the potential ‘outlier’ nature of the 18 week data point suggested that using a two scenario approach might be a worthwhile way to acquire additional clarity.
A two scenario HAT example
This section considers a ‘two scenario’ HAT approach, building on the last section’s ‘single
scenario’ minimum clarity HAT approach. This section still addresses a single source of
uncertainty within a single component (activity) of the goals– plans relationships structure,
assumptions maintained for the whole of the tale associated with William.
William’s data suggested that a two scenario generalisation of the minimum clarity estima-
tion model would add useful clarity for a minimal level of additional effort. The four obser-
vations of two, three, four, and six weeks could be associated with one scenario, a ‘usual
range outcome’ scenario. The relatively isolated single observation of 18 weeks could be
associated with a second somewhat different ‘high outcome’ scenario. Some people might
see and refer to 18 weeks as an outlier observation because it looks so different relative to
2, 3, 4 and 6. This involves assumptions which William should avoid making unless he has
clear evidence.
William was asked to consider the ‘usual range outcome’ scenario first. A minimum
clarity approach to this scenario associated a plausible minimum with a P10 of 2 weeks, a
plausible maximum with a P90 of 6 weeks, and a (2 + 6)/2 = 4 weeks midpoint expected
value estimate.
The four of five data points involved suggested that the probability of the ‘usual range
outcome’ scenario was 4/5 = 0.8, and 0.8+/−0.1 provided an illustrative nominal range for
the expected value given the sample size of only five observations. Using 0.8+/−0.1 implied
that 0.2+/−0.1 was the probability of a high outcome scenario.
Using +/−0.15 instead of +/−0.1 would have had no significant implications, but +/−0.3
was clearly too big, and +/−0.0 was clearly too small. It would not have been clarity efficient
to waste time and effort worrying about a more refined approach to estimating a range for
the 0.8 and 0.2 expected values of these probabilities. But it was worth being clear that the
probability estimates were themselves uncertain, as were the ranges associated with the 0.8
and 0.2 expected values.
William was then asked to consider the ‘high outcome’ scenario. A minimum clarity
approach could obviously associate the single data point of 18 weeks with the midpoint
expected value estimate. Someone obsessed with data- based probabilities might optimis-
tically assume that 18 weeks was the only data- based estimate of both a minimum and a
maximum. They might then use (18 + 18)/2 = 18. But common sense and normal statisti-
cal practice suggest 18 weeks with an unknown variability which could be significant in the
absence of reasons to believe otherwise.
A crucial generalisation of this point is that even if William and his team had some data,
they always needed to think beyond the data, in this instance in terms of realistic P10 and
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 81
P90 values associated with the 18 weeks expected value estimate for the high outcome
scenario.
An estimate of 18+/−4 weeks in P10 to P90 range terms for an 18 weeks expected value
might have been reasonable, implying an absolute range of 18+/−5 weeks, but a bigger or smaller range might be more appropriate.
If it had been worth exploring this range, William and his team might have asked the
questions ‘Do we know why the case with an observed outcome of 18 weeks took this long,
and what other reasons for comparable high outcome value delays might be relevant?’ For
example was this project known to be in trouble? Was making the design change request a
good excuse for a review? Was it a summer vacation period? Was the end of the financial year
looming? And what other reasons of this kind might be relevant?
Now assume that William decided to assume 18+/−8 weeks in P10– P90 terms, imply-
ing an absolute range of 18+/−10, so he could use Figure 3.3, once I had explained the
implications.
Using 18+/−4 weeks in P10– P90 terms in the context of a revised version of Figure 3.3
would imply a bimodal probability distribution. A bimodal distribution assumption raises
questions about the possibility of further modes because of very different kinds of sources
of uncertainty underlying the different modes. To avoid expending time and effort on these
questions, a simple default working assumption was a single mode distribution in terms of
the current two scenario HAT structure, implying that 18 weeks was not an outlier. This
0
probability
(density)
0 2 4 6 8 18 28
outcome value in weeks
probability
density
format
using a
continuous
variable
portrayal
probability
tree
format
using a
discrete-
outcome
two-
scenario
portrayal
0.1
0.8
0.2
probability outcome value in weeks
4
18
Figure 3.3 A two scenario model with a HAT portrayal.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
82 Foundations
useful assumption implied an absolute range of 18+/−10 weeks for the high outcome sce-
nario, 18+/−8 weeks in P10– P90 terms, as illustrated in Figure 3.3. It was sensible for Wil-
liam to prefer this approach when he understood the implications.
The density format of the top half of Figure 3.3 treats the outcome value (in weeks)
in continuous variable terms which are consistent with the single- mode assumption of
Figure 3.2 and the data points, a useful basic way to interpret the available information.
The alternative bottom- half probability tree format that treats the outcomes as discrete
values can also be very useful. The joint use of both involves a HAT interpretation of Figure 3.3. This HAT interpretation implies that a default assumption associated with the
tree portrayal is discrete values are conditional expectations for an underlying continu-
ous variable distribution, rectangular in form for portrayal simplicity in density format.
That is, 4 weeks was a conditional expectation given a value in the range from 0 to 8
weeks, and 18 weeks was a conditional expectation given a value in the range from 8 to
28 weeks.
Figure 3.1 can now be interpreted as a one scenario special case of the HAT portrayal
illustrated in Figure 3.3. The underlying tree aspect of Figure 3.1 was not of immediate use,
but it was in place from the outset, and it might prove useful later if multiple sources of
uncertainty were being modelled.
The dual discrete and continuous variable interpretation of Figure 3.3 explicitly gives it a dual nature which avoids a framing assumption or a working assumption that only one of these two modelling approaches applied, common practice often locking people
into a mindset which uses only one of these perspectives and ignores valuable insights
from the other. Sometimes a working assumption that only one is needed is convenient,
but a framing assumption making both available is an important feature of the HAT tool-
set. Especially when two or more scenarios are involved in a probability density format
diagram it can be worth remembering that the usual ‘probability’ label on the y- axis is
a probability density measure, with a reminder provided in brackets for early figures in
this chapter.
The continuous and discrete value forms of the HAT portrayal are both useful, and the flexibility associated with being able to use either for different conceptual and operational
purposes is surprisingly powerful. In particular, the tree framework is the key to clarity effi-
cient ways of dealing with all forms of statistical dependence and causal dependence. Statisti-
cal dependence can be very simple, comparable to correlation, but it may be very complex,
involving non- linear cascade effects, for example. Causal dependence can require treatment
which is very simple or very complex. Part of the complexity can be portrayed by various
kinds of probability trees and decision trees. Both statistical and causal dependence need
effective and efficient treatment, which will be explored later, building on this very simple
HAT portrayal to illustrate a basis for analysis often approached by a range of alternative
methodologies, including methods based on moments, Monte Carlo simulation, and prob-
ability arithmetic as used by decision trees, event trees, and fault trees.
What the two scenario approach using the HAT framework of Figure 3.3 involves is a sim-
ple quantitative approach to a high outcome scenario which can be usefully separated from
a usual range outcome scenario without starting to explore multiple modes or contingency
planning and underlying source of uncertainty decomposition considerations. This chapter
later looks at extending this basis in various ways, using the HAT approach framework. The
HAT approach developed in this chapter underlies Chapter 4, and it is employed in all Part
2 examples using a range of variants. Part 3 revisits some issues just touched on earlier.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 83
Whatever range or model structure is associated with the 18 weeks ‘high outcome’ sce-
nario, the overall expected outcome using the two scenario structure is computed via
4 0 8 18 0 2 3 2 3 6 6 8× + × = + =( ) ( ). . . . . .
This computation just weights the conditional expectations of 4 weeks and 18 weeks by
their probabilities, estimating the probabilities for each scenario using the associated data
points. It clearly demonstrates that the clustering of the four ‘usual outcome range’ data
points in the range of 2 to 6 weeks means that they collectively make a smaller contribution
to the expected value than the single high outcome data point of 18 weeks.
Perhaps less obvious, this clustering example also demonstrates these data points make a
smaller contribution than data points uniformly distributed in the range of 2 to 18 weeks.
The increase in clarity provided by two scenarios includes a significant reduction in the
expected value – from 10 weeks to 6.8 weeks. This is a direct consequence of recognising
and modelling the low- end clustering, consistent with, although not quite the same as,
using the five individual data points to calculate the expected outcome via
2 3 4 6 18 5 6 6+ + + + =( ) / . .
A two scenario approach makes more use of the complete data set than does a one scenario
approach, formally recognising the asymmetry involved, which is clearly sensible. More clar-
ity about the high outcome value tail is also important.
Thinking about reasons for variability around the observed 18 weeks in terms of both
what actually happened in the past in the case of the one available data point and what
might happen in the future with comparable impact is valuable for better estimation. But it
also has other benefits if the approach we are discussing is going to be used to think about
contingency planning, especially if it is used earlier in the project lifecycle. Used earlier it
might lead to revising a base plan, to pre- plan likely design revision requests so that they
avoid the end of the financial year or holiday period requests if feasible, for example. Even
if for formal quantitative analysis purposes decision makers limit themselves to simple plan-
ning models which do not consider this kind of contingency planning issue in formal terms,
understanding the underlying reasons for variability provides additional qualitative informa-
tion of considerable value. Discussion of alternative reasons why a high outcome scenario
like 18 weeks might occur has direct value in terms of enhanced clarity.
The ‘additional clarity’ provided by a two scenario approach involved a model- based
formal representation which was closer to the presumed reality – that is part of what we
mean by more clarity. But the potential importance of the 18 week outcome and associ-
ated sources identified and used informally was also part of the increased clarity because the
insight it provided can be communicated without ambiguity by an analyst to everyone using
the analysis in an informed manner.
Also important was the very modest increase in required effort, whether or not data
were available, and whether reasons for high outcome scenario variability were worth
exploring.
Given these data, it was obvious that the suggested two scenario approach provides an
unequivocal improvement in clarity for modest additional effort. But even if these data had
not been available, if the presumed reality involved significant skew as shown in Figure 3.2,
then the two scenario model might be a very worthwhile clarity efficient refinement.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
84 Foundations
If William’s focus of interest was what he should do in this particular case, he might find
the simplification inherent in contingency planning based on a two scenario approach use- ful and the focus on associated sources of uncertainty that matter most particularly helpful. He could stop worrying about separate contingency plans for 2, 10 and 18 weeks based
on his single scenario model. He could focus on a base plan associated with a usual range
outcome scenario expected value of 4 weeks plus contingency plans for a high outcome
scenario expected value of 18 weeks, perhaps with more time spent on the 18 weeks sce-
nario, including starting to think about contingency responses based on what might have
led to this outcome. If contingency planning was relevant this simplification might be very
worthwhile.
If a fairly symmetric distribution was anticipated, a two scenario model might not provide
much benefit for the additional effort. The scope for more insight is driven by working
assumptions about the presumed probability distribution shape for the uncertain underlying
reality. The relevance of more insight will also be driven by the way the estimate will be used,
so those involved in preparing estimates and those involved in using estimates should have
a common understanding of these concerns.
More than two scenarios to ‘refine’ an estimate
The two scenario approach just outlined provided a useful starting point for beginning to
generalise the minimum clarity HAT approach into a conceptual and operational toolset
with very broad EP roles and capabilities.
The next step is considering more than two scenarios to ‘refine’ an estimate when ‘refine-
ment’ implies portraying a single source of uncertainty with a probability distribution shape
closer to a presumed reality, as suggested by the Figure 3.2 illustrative example.
If an n = 1 or n = 2 choice is judged not refined enough, then n = 5 as a default value in
a range like 3 to 10 is an approach relevant to William’s situation. To put this in context,
when BP developed HAT based software they provided for n = 1 to 20, but rarely used more
than 10 for a direct specification, and 3 to 10 was the usual range. Given that William had
five data points, n = 5 is a convenient choice for illustrative purposes.
Often it is convenient to use a simple rectangular density form for the histogram what-
ever value of n is being used, equating the expected outcomes for each interval to the
‘class marks’ (central values) for the histogram intervals. Sometimes it is also convenient to
assume a common interval for this histogram over part of its range or all its range. The HAT
approach to refining estimates explored in this section uses a rectangular histogram with
the expected values for each of the n = 5 histogram intervals as scenario values to define the
probability tree.
William now provided an opening for starting to explain the basis of this section’s gener-
alisation of the HAT approach by asking, ‘How could the approach of Figures 3.1 through
3.3 capture more fully the information content of all five of his data points?’
I responded by suggesting that we consider the probability tree in tabular form of
Table 3.1 plus the associated rectangular histogram of Figure 3.4. This would provide a five
scenario example of the generalisation of the HAT approach which made maximum use of
the information content of his five data points.
Consider the working assumptions used to construct the probability tree portrayed in
tabular form by Table 3.1 plus the directly dependent Figure 3.4 rectangular histogram
density function portrayal and its piecewise linear cumulative equivalent.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 85
Table 3.1 The HAT probability tree portrayal underlying Figure 3.4 using a 2- week common interval basis.
Delay scenario S (nominal range, span)
Probability P and components 0.2 per data point
Probability density P/span
Contribution to expected delay D S × P (expressed as a %)
1 (0– 2, 2) 0.2 / 2 = 0.1 0.1 / 2 = 0.05 1 × 0.1 = 0.1 (2%) 3 (2– 4, 2) 0.1 + 0.2 + 0.1 = 0.4 0.4 / 2 = 0.20 3 × 0.4 = 1.2 (18%) 5 (4– 6, 2) 0.1 + 0.1 = 0.2 0.2 / 2 = 0.10 5 × 0.2 = 1.0 (15%) 7 (6– 8, 2) 0.1 0.1 / 2 = 0.05 7 × 0.1 = 0.7 (11%) 18 (8– 28, 20) 0.2 0.2 / 20 = 0.01 18 × 0.2 = 3.6 (54%)
D = 6.6 (100%)
0
probability
(density)
0 2 4 6 8 18 28
outcome value in weeks
probability
density
format
0 2 4 6 8 18 28
outcome value in weeks
1.0
0.9
0.5
0.1
0
cumulative
probability cumulative
probability
format
0.2
expected value 6.6 weeks
expected value 6.6 weeks
Figure 3.4 A five scenario HAT model with the P10 = 2 weeks and the P90 = 18 weeks.
The first involved assuming that because five data points were involved, each contributed
a probability of 1/5 = 0.2, as indicated in the second column of Table 3.1 in the third line
of the heading.
The next involved assuming that a two week ‘common interval structure’ beginning at
zero was an appropriate basis for portraying the lowest four data points. The rationale for
this choice was William had four data points in the range 2 to 6 (2, 3, 4 and 6), so the most
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
86 Foundations
effective and plausible simple histogram portrayal of the uncertainty associated with these
four data points involved a common interval class width of two over the range from 0 to 8,
with classes representing the intervals 0 to 2, 2 to 4, 4 to 6, and 6 to 8, using midpoint class
marks defining scenario S values at 1, 3, 5 and 7.
This was linked to the assumptions that data points in the middle of any of these interval
ranges (3 in this example) should have their probability fully assigned to that range, but
data points on boundaries (2, 4 and 6 in this example) should have their probabilities split
between the two classes adjacent to the boundary.
Working through this example one data point at a time, the data point observation of 2
was on a boundary, which implied a probability component of 0.2/2 = 0.1 allocated to each
of the two intervals 0 to 2 and 2 to 4. This was shown in Table 3.1 by the only entry in the
second column of the S = 1 row and by the first entry in the second column of the S = 3 row.
The data point observation of 3 was in the middle of the interval 2– 4, so its probability
contribution of 0.2 could be allocated entirely to this class, as shown by the second entry in
the second column of the S = 3 row.
The data point observation of 4 was on a boundary, implying 0.1 allocated to each of the
two intervals 2– 4 and 4– 6, as shown in the second column of the S = 3 and 5 rows.
The data point observation of 6 was also on a boundary, which implied 0.1 allocated to
each of the intervals 4– 6 and 6– 8, as shown in the second column of the S = 5 and 7 rows.
These four data points defined the total probability P value for the scenarios S = 1, 3, 5
and 7 based on the observations of 2, 3, 4 and 6, as shown in the second column for the
S = 1, 3, 5 and 7 rows.
The observed value of 18 could have been associated with just two of the same common
interval classes, 16– 18 and 18– 20, following the common interval of 2 and boundary- value
approach used earlier. However, unless a multimodal distribution was suspected, a simpler
and more plausible interpretation consistent with a single mode distribution and Figure 3.3
was all 10 intervals of a width of two between 8 and 28 should have the one data point allo-
cation of 0.2 split between them, as shown in the second column of the S = 18 row.
The P values in the second column of Table 3.1 divided by the span of each interval deter-
mined the probability density for that interval, as shown in the third column. This, in turn,
defined the shape of both the density and the cumulative curve formats over the range 0 to
28 as shown in Figure 3.4.
The expected value of 6.6 computed in Table 3.1 can be plotted as shown in Figure 3.4.
This value was significantly bigger than the most likely value of 3 and the median of 4
because of the long right- hand distribution tail. But it was less than the 6.8 of Figure 3.3
and the 10 of Figure 3.1 because of different working assumptions about the probability
distribution shape.
I indicated to William that the approach illustrated by Table 3.1 and Figure 3.4 used a syn-
thesis of a standard basic textbook approach to histograms plus a common sense approach
to avoiding multiple mode portrayals which were not intended. It extracted the maximum
information content from the set of five data points which aligned the resulting histogram
with a simple probability tree interpretation. Its treatment as a histogram in Figure 3.4 was
directly equivalent to the simple five branch probability tree of Table 3.1, where the discrete
branch outcome values correspond to the five class mark S values of 1, 3, 5, 7, and 18, with
probabilities 0.1, 0.4, 0.2, 0.1, and 0.2. The expected value of 6.6 was exactly the same as
the sum of the data point observations divided by five used to compute the expected value
earlier.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 87
However, William and his team members were free to modify the Table 3.1 and Figure 3.4
portrayal if they believed that a multimodal distribution was appropriate, restricting the range
associated with the observation of 18 weeks. Furthermore, they could use a probability tree
based on the observed data values if they were more comfortable with this. Table 3.1 plus
Figure 3.4 was a more refined model than the Figure 3.1 minimum clarity model or the
Figure 3.3 model, but it was still a model, open to alternative working assumptions and fur-
ther refinement, including making use of any information not yet considered.
Although this example has been designed to demonstrate using n scenarios to capture the
full information content of more than two data points, William and his team did not need
any data points to make use of the representation of uncertainty provided by Table 3.1 plus
Figure 3.4 if it portrayed their subjective view of what was involved. That is, if no data were
available, they could drop the second column computations of Table 3.1 and estimate the
probabilities associated with an appropriate set of S values directly in subjective probability
terms.
The approach adopted by BP usually assumed no readily available data. The first step
was estimating a maximum plausible S value. The second step was estimating a minimum
plausible S value. The third step was rounding both onto a suitable scale to yield enough
intermediate values to provide the level of refinement deemed appropriate given the antici-
pated probability distribution shape, usually involving n scenarios in the range 1 to 10.
The fourth step, assuming an n of 3 or more, was estimating the probability of the least
attractive outcome, then the most attractive outcome, followed by allocating the residual
probability to all the other possible outcomes, ensuring that the overall distribution shape
was consistent with what those present with relevant experience believed was an unbiased
judgement. Obvious simplifications apply for n = 1 or 2, which are clearly special cases.
The fifth and final step was checking the outcome with those not present who might have
a useful contribution to make. This approach has been recommended since for all users of
a HAT approach.
In practice, if good data is readily available, and it is important to make immediate effec- tive use of it, then 10 to 100 data points is the norm, and 1 to 10 classes will usually provide
a clarity efficient portrayal of what the data is saying. The most effective portrayal of the
data will depend on the shape of the underlying probability distribution, sometimes usefully
visualised initially using a scatter diagram. But in practice good data is rarely readily avail-
able, and even if it is obtainable, acquiring and analysing it may involve significant cost, and
it will take time which may not be available. Most of the time a purely subjective first pass
estimate based on the best readily available judgement is a clarity efficient approach, with
whatever level of refinement seems to best suit the context. Further passes can address a
possible need for data, more refinement and clarification via decomposition, which will be
discussed shortly.
We can now see this section’s HAT approach to any number of scenarios worth quantita-
tive treatment as a general approach, with n = 1 to 10 the usual range whether or not data
are available, the approaches of Figures 3.1 through 3.3 as special cases for n = 1 and 2, and
Figure 3.4 a special case example for n = 5.
One or more ‘condition’ scenarios
I then suggested to William that even if he and his team members had data readily at hand, they should ALWAYS be prepared to go beyond the available data, how much depending on the
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
88 Foundations
quantity and quality of the data and the assumed gap between the data and the aspects of reality that really mattered.
To begin to address this issue, building on the approaches of Figures 3.1 through 3.4,
William was asked the question, ‘What was the chance of a ‘very high outcome’ scenario, like 52 weeks, well beyond his data set?’ The rationale behind this question was that William’s
five data points were clearly a very small sample. It might well be the case that serious prob-
lems with this project or overall MoD budget constraints and pressing reasons for spending
money differently might put this project on hold for an extended period if an expensive
design change request was made, a delay of the order of a year being a representative sce-
nario approximation to all possible outcomes involving more than the range associated with
their data set.
William replied that ‘there was a probability of about 0.02+/- 0.01 of a 52 weeks delay’,
roughly a 1 in 50 chance. Using this estimate to illustrate the implications, assuming the
0.02 probability was subtracted from the 0.2 probability of a high outcome scenario to keep
the arithmetic simple, the overall expected outcome was
4 0 8 18 0 2 0 02 52 0 02 3 16 3 42 1 04 7 48× + × + × = + + =( ) ( )( ) ( ). . . . . . . . .−
This was 13% more than the 6.6 expected value suggested by the full data set using a simple
mean or the HAT approach of Table 3.1 and Figure 3.4, a significant increase. It clearly
demonstrated that an estimate which does not go beyond the full data set might be system-
atically biased because it involved an important stealth assumption – observations beyond the range indicated by the data set are not going to happen.
This stealth assumption clearly mattered, and it demonstrated that we should always
explicitly address this potential bias by considering a suitable ‘very high outcome’ scenario,
and if relevant, a suitable ‘very low outcome’ scenario. Doing so involves two approach
options.
One option is usually referred to as a ‘conditional’ estimate approach. It involves formally
recognising one or more conditions – such as ‘this estimate assumes no possibility of an
outcome beyond x weeks, where x is defined by the assumed plausible maximum scenario’.
A second option is usually referred to as an ‘unconditional’ estimate approach. It involves
formally estimating conditional outcomes should these scenarios beyond the data set occur
plus their probability of occurrence, avoiding the bias built into the first option because
conditions are only addressed in qualitative analysis terms. For the ‘unconditional’ label to
be appropriate in a strict sense an unconditional approach must justifiably assume that all the relevant conditions have been identified and can be treated in this way.
If the first option is chosen, it is clearly important to understand the implied bias relative
to an unconditional estimate, and the unconditional/conditional estimate distinction. If
the second option is chosen, it is clearly important to be justifiably confident that all rel-
evant scenario possibilities like a 52- week delay scenario have been identified and treated
appropriately.
Furthermore, whichever of these two options is selected, arguably a key insight which
flows from understanding the possibility of a 52- week scenario was clarity about the sig-
nificant impact on the project of a possible one year delay if it happened – perhaps with knock- on effects leading to the cancellation of the project. William soon realised that he
might be well advised to put some effort into addressing the risk of this possibility in terms
of proactive (preventative) as well as reactive contingency planning without unduly worry-
ing about the exact probability of it happening or attempting to precisely measure its impact
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 89
on the expected outcome. Responding to a possible delay scenario of the order of 52 weeks
involved very different issues compared to those associated with 18 weeks.
The obvious follow- on question was, ‘What was the chance of a “project cancelled” sce-
nario, an infinite delay?’ The whole project concept might be overtaken by changes in cir-
cumstances which render it redundant, and the design change request might or might not
be the trigger for this being recognised.
William responded that ‘there was about a 0.01 chance of the project being cancelled’,
clearly revealing a further earlier stealth assumption – ‘there was NO chance of a cancelled project’.
It soon became clear that it might be very useful for the project team as a whole to be
aware of the possible cancellation of the project as an explicit formal planning issue.
At a more general level, it soon became clear that it was important to appreciate that the
duration estimate and all other aspects of their planning were now explicitly based on the
assumption or condition ‘project cancellation does not occur’. This kind of condition can be of crucial importance when relevant. Together with the very- high outcome scenario condi-
tion it can provide a starting point for generalising to an extensive list of further conditions
worth specific individual attention plus the notion of a ‘no quibbles’ provision concerned
with a comprehensive set of further conditions.
It is of course important to avoid attempting to include a possible ‘infinite delay’ in the
expected value in a manner which makes the expected value infinite, rendering it useless.
This meant that ongoing treatment of the project cancelled scenario not occurring as a con-
dition was a sensible working assumption option. Indeed, it is often important to recognise
that ‘unconditional estimates’ are not really an appropriate aspiration. The key underlying
issue is a clear understanding of the extent to which identified or unidentified conditions
are relevant.
The expected value of 10 weeks associated with the minimum clarity approach can now
be interpreted as less naïve than it might have seemed earlier in terms of a plausible provision
for a very high outcome possibility if a highly asymmetric distribution is involved. A possible very long list of other conditions associated with a no quibbles provision is discussed later in
this chapter and related Chapter 8 concerns extend and enhance this argument.
Going beyond the data as illustrated by these two further scenario examples, and rec-
ognising the implications of associated conditions, provides an obvious and unequivocal
improvement in clarity for minimal additional effort. Furthermore, even if no data had been
available as a starting point to go beyond, these possible extreme outcome concepts can
clearly be a very worthwhile enhancement of an EP perspective.
Whether or not data are available, starting by considering the chance of a cancelled pro- ject, or a very big delay, could be very useful if these possibilities were suspected to help
minimise unconscious bias, further generalising the five- step approach outlined in the previ-
ous section.
Some variant of the generalisation of Table 3.1 and Figure 3.4 provided by the last sec-
tion, plus making effective use of the conditions concept, with or without data, was the way
William and his team should plan to use this HAT approach in practice most of the time.
This would include the routine use of a first pass minimum clarity alternative when a very
simple first pass approach looked useful, seeing the approaches in Figures 3.1 and 3.3 as
special cases.
Use of these HAT variants has been a practical operational approach in a wide variety
of contexts for a wide range of organisations for many decades, and they remain a recom-
mended tool in the EP estimation toolset for both conceptual and operational purposes.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
90 Foundations
Of considerable importance is a clear understanding that in the light of these two further
scenario concepts and associated conditions, plus the general tendency of most people to be
overly optimistic, the minimum clarity approach in Figure 3.1 is actually not naïve all – it
is a sophisticated low effort but robust first pass approach designed to neutralise optimistic
bias not worth attention if it might not matter too much.
Also of importance is a clear understanding that William and his team needed to appreci-
ate that if they had the data and the refined portrayal of Table 3.1 plus Figure 3.4 but failed
to think about it in the way the usual range outcome, high outcome, very high outcome,
and project cancelled scenarios suggested, they might actually have less insight for more effort than using a minimum clarity single scenario model aware of the implications of a full range of possible outcome scenarios. This would be clarity inefficient.
William could have his team use a one week class interval to provide even more precision
if doing so looked likely to be useful in the sense of a smoother curve looking more like the
presumed underlying curve in Figure 3.2. Any desired level of precision was available. How-
ever, William and his team members should not confuse smoother looking curves with more
clarity, although they could have both smooth curves and the clarity of multiple scenario models if they clearly understood what they were looking for.
It is useful to see histogram definition of quantified uncertainty using rectangular den-
sity functions plus identification of associated conditions plus associated probability trees as
discussed so far as a starting point for further generalisation of the HAT approach. These
HAT generalisations allow a flexible approach to precision whether or not data are available.
Both scenario and histogram approaches are inherently flexible, as are associated prob-
ability tree approaches, perhaps linked to lower or higher level decision tree branches. None
of these approaches are restrictive in a framing assumption sense, provided they are seen as
special cases in a more general HAT approach framework. However, there are important
trade- offs in terms of clarity and effort/cost of analysis which need to be addressed in terms
of appropriate working assumptions for particular applications.
Any HAT approach involving a nominal common interval basis for conceptual simplic- ity (like using 2 for S = 1, 3, 5 and 7 scenarios) can be generalised using wider classes (like
using a class width of 20 for the S = 18 weeks scenario). Another approach to generalising is
replacing linear scales by logarithmic scales if this is useful, as illustrated in Chapter 9.
Generalising the minimum clarity approach of Figure 3.1 using all the HAT approaches
discussed so far is usefully thought of as employing ‘refinement options’. All these refinement options stick to one source of uncertainty within the goals– plans relationships component of interest in terms of formal modelling, which distinguishes them from ‘decomposition options’. Decomposition options involve multiple sources of uncertainty, formally separated with a
view to possible quantification, to be considered shortly.
Parametrically defined probability distribution special cases
Before moving on to decomposition options, briefly consider refinement options in terms
of an important potential option within the HAT framework which many people see as an alternative approach to defining probability distributions.
William provided an opening for considering parametrically defined probability distribu-
tions by asking ‘What if some of my team wanted to use common practice probability distri-
butions like a Normal distribution?’
Using a probability distribution function formally defined in terms of parameters, like a
Normal (Gaussian) distribution, can sometimes prove very useful. The Normal probability
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 91
distribution is defined by just two parameters, its mean and its variance, making this approach
very economical with data. Being economical in terms of data requirements is one of several
reasons parametrically defined probability distribution approaches are dearly loved by many
people with a significant knowledge of statistical methodology. Sometimes this love affair is
well founded, but sometimes it is misjudged, and likely to end in tears.
From an EP perspective it is crucial to see the use of parametrically defined probabil-
ity distribution function approaches as potentially useful special cases within a general EP
approach to probability distributions based on a HAT perspective. This allows us to test
the robustness of any parameter defined probability distribution approach within the HAT
framework. Crucially, it avoids making assumptions which are difficult to test unless we
have an unrestrictive HAT framework to fall back on. Our framing assumptions should never REQUIRE a parametrically defined approach, but parametrically defined approaches should be seen as potentially useful options within a framework which facilitates testing their working assumptions and overriding inappropriate implications.
For example, I suggested to William that after exposure to the generalisation of Fig-
ure 3.4, he should see very clearly that if anyone on his team suggested that the best practice
approach to using his five data points involved employing a Normal distribution, dropping
the 18 week data point as an outlier because it was associated with a project in trouble and
their project would not get into trouble, they should be enlightened or found another job
as quickly as possible. This was just a simple and reasonably obvious example of how some-
one unduly attached to a simple Normal distribution approach might try to get away with a
seriously inappropriate parametric probability distribution approach assumption, but it was
a parody of real people routinely encountered.
That said, BP computer software to implement a HAT based approach allowed users the
option of specifying a Normal distribution defined by P10 and P90 values or other com-
parable plausible maximum and minimum values, with the software converting this into an
n = 20 variant of Table 3.1. What proved more popular were a log- Normal equivalent and
a beta distribution equivalent, the latter allowing a hybrid with asymmetric probability dis-
tribution tails outside the P10 and P90 maximum and minimums used along with the most
likely value to provide the three defining parameters for a beta distribution.
One more complex and subtle example which I have encountered involved dealing with
all the implications of 10,000 year return period design specification assumptions for seismic
events leading to core meltdown of nuclear power stations with only a 100 years or so of
reliable data plus debatable underlying assumptions about where epicentres might occur.
Another involved missing data plus limited numbers of observations when making use of a
negative exponential distribution assumption for 100 year return period estimates of deep
ice- scour damage to proposed oil pipelines buried in deep trenches in the Beaufort Sea
(Chapman, 1988). In both these cases, extrapolating on the basis of a very limited data set is
probably the only reasonable starting point for a hybrid approach which also has to address
plausible information about extreme events from any available sources, plus relaxing naïve
assumptions like ‘provided an event is within the design specification the facility will survive
an extreme event, but as soon as an event is outside the design specification event it will
fail’. HAT frameworks provided a useful perspective for thinking about and synthesising all relevant information about both the past and the future in both these cases.
You may sometimes find it useful to see the distribution shape assumed, via a parametri-
cally defined special case or a histogram structure special case, as one of the conditions asso-
ciated with expected outcome and range estimates. They are certainly working assumptions
which condition estimates, and they sometimes matter.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
92 Foundations
This book will usually avoid discussing the details of ‘how to do it’ choices associated
with these sorts of issues. They are only mentioned very selectively or developed in outline.
In this case, the rationale is to make clear that all the issues just explored are comfortably
within the scope of what can be addressed via the HAT framework component of an EP
perspective, to overview the scope of the generality which can be provided whenever it is
worthwhile, and to clarify the rationale for using relevant data with great care to deal with
a future which may not resemble the past for reasons which may not be readily identifiable.
The synthesis underlying a HAT approach
This section provides an overview of key linkages between the very simple HAT models
used earlier in this chapter, all subsequent discussion of modelling using a generalised HAT
approach in this book, and the application of these approaches in contexts of interest to you.
It is useful to see a generalised HAT approach as a synthesis of six categories of approach
which has been designed to deliver the advantages of all six without any of the disadvantages
associated with narrower combinations:
1 probability trees with embedded decision trees using probability arithmetic,
2 continuous variable histograms,
3 probability distribution functional integration,
4 methods based on moments,
5 all other useful approaches which do not involve sampling, and
6 all useful simulation approaches employing sampling as a computation option.
Consider each in turn.
Probability trees with embedded decision trees using probability arithmetic provide
the skeleton and primary muscular structure which supports the generalised HAT approach.
The basic decision tree and probability tree graphical forms and the probability arithmetic
are as used by decision analysis (decision theory) as approached by authors such as Raiffa
(1968) and Goodwin and Wright (2014). They are also central to safety modelling using
various specialist forms of probability trees like ‘event trees’ and ‘fault trees’ – see NUREG
(1975) for example. A HAT framework assumes that the discrete outcome and probability
values employed are associated with scenarios. These outcomes and probabilities are condi-
tional expectations for associated continuous variables. Usually these conditional expecta-
tions are assumed to be central class mark values associated with rectangular histograms. The
probability arithmetic treatment of illustrative numerical examples assuming discrete values
in Part 2 have all used this approach. It keeps the examples as simple as possible while build-
ing the basis of the nuanced understanding of strategic clarity which all target readers need.
Continuous variable histograms provide an explicit generalised interpretation of proba-
bility trees plus embedded decision trees in continuous variable terms, analogous to the skin
and secondary muscularity which completes the shape and form of the HAT framework, as
illustrated in very simple basic terms earlier in this chapter. This continuous variable inter-
pretation of a discrete value underlying structure is usually implicit in conventional decision
and probability tree usage, whether or not point estimates of outcomes and probabilities are
associated with ranges. The HAT approach makes a dual discrete and continuous variable
interpretation explicit and mandatory because both discrete and continuous variable inter-
pretations may matter, and the ability to test whether they matter is an important feature of
the general framework which an EP approach provides.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 93
Probability distribution functional integration becomes a useful component HAT tool
when combining two or more sources of uncertainty. For example, if two sequential activi-
ties both had duration distributions defined by Figure 3.1, and we wanted to determine the
probability distribution for the sum of these two activity durations assuming independence, functional integration theory tells us the resulting density function will be triangular, with a minimum of 0 + 0 = 0 weeks, a maximum of 20 + 20 = 40 weeks, a mode at 10 + 10 = 20
weeks, and a median and expected value of 20 weeks. If we assumed a perfect positive cor- relation, the minimum, the maximum, the median, and the expected value would be the same, but the density function would be uniform, with no discernible mode. If we assumed perfect negative correlation, the minimum, the maximum, and central values would all be 20 weeks, with no residual uncertainty. This is a useful simple demonstration of the role of dependence.
The insight provided by this functional integration approach can also be usefully employed
to recognise the computational error implicit in any discrete probability approach to prob- ability and decision tree portrayals when an underlying continuous variable distribution is
being approximated and correct it appropriately whenever doing so is worth the effort.
For example, if two sequential activities both had their duration distributions defined by
Figure 3.4, and we added them assuming independence, the kind of probability arithme-
tic approach explored in Part 2 could use the class mark values 1, 3, 5, . . . and assign the
resulting probability products to joint distribution class mark values at 2, 4, 6, . . . with
the same interval of 2 weeks used over all the distribution ranges. It could associate a joint
duration of 1 + 1 = 2 with one possible route to this outcome and a probability of 0.1 ×
0.1 = 0.01, a joint duration of 4 with two possible routes to this outcome involving 1 + 3
and 3 + 1 = 4 with a probability obtained by 0.1 × 0.4 + 0.4 × 0.1 = 0.08, a joint duration
of 6 with 1 + 5 and 3 + 3 and 5 + 1 with a probability involving three possible routes, and
so on. This probability arithmetic result would be precise and computation error- free if we
could assume that the variables were truly discrete. But if we recognise that a duration of 1
week means 0– 2 weeks, assumed to have a uniform within class distribution, and we associ-
ate the whole of the 0.1 × 0.1 = 0.01 product with 2 weeks interpreted as 1– 3 weeks, an
error is clearly involved, with a similar error associated with all other probability product
allocations. Assuming independence, the triangular distribution actually associated with 0– 4
weeks instead of 1– 3 weeks implies one- eighth of the probability ought to be assigned to
the half interval 0– 1 weeks, another one- eighth to the bottom half of the interval 3– 5 weeks
associated with a class mark value of 4 weeks, with comparable corrections needed for all
other probability products.
As explored in Cooper and Chapman (1987), a HAT approach can build on this insight
to show that assuming independence within classes the overall error can be corrected by
moving one- eighth of the difference between adjacent class probabilities outwards from
modal values in the joint distribution. This correction approach used in conjunction with
more classes can virtually eliminate computation error and provide smooth shapes for the
resulting distributions.
In this book, numerical examples in Part 2 do not bother to make these corrections
or provide smooth curves to keep discussion of the examples as simple as possible. The
errors needing correction are demonstrably small, getting smaller as the number of classes is
increased, and the basic shape and form of the distributions computed ignoring these ‘com-
putation error’ corrections are not affected. But in practice correction is feasible if doing so
is worthwhile, and several possible approaches can be employed, including using commer-
cially available Monte Carlo simulation software as an alternative computational approach.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
94 Foundations
An approach to correcting this computation error directly based on a version of the
adjustments to a basic probability arithmetic approach just illustrated was embedded in
the HAT based software BP developed early in its evolution of the approaches discussed in
the next section. Later evaluation showed that the same level of precision using Monte Carlo
simulation approaches (subject to a comparable ‘sampling error’) involved about 1,000
times more computational effort. The drawbacks of this BP software approach are twofold:
the need for an investment in specialist computer software and the need for a ‘reducible’
sequence of combinations. Most current users of an EP approach avoid both of these prob-
lems by using readily available Monte Carlo simulation computer software within a HAT
framework, discussed briefly for those who are interested in Chapter 11.
The role of probability distribution functional integration in the HAT framework is a
contributing tool which makes the integration of the first two categories of approach more
powerful. It extends the role provided by histograms earlier in this chapter to multiple
sources of uncertainty contexts. It means that a probability arithmetic framework can con-
trol a measurable computation error to a virtually error- free level if this is worthwhile, the
computation error being part of any discrete value probability approach approximating a
continuous variable reality, whether or not this is appreciated. This verifies the conceptual
model role of a HAT approach, whether or not discrete probability arithmetic is used opera-
tionally. Simulation provides an alternative and complementary tool to be discussed shortly.
The next three categories of approach are also useful tools in a similar spirit for somewhat
different reasons.
Methods based on moments (sometimes just called ‘methods of moments’) are the gen-
eral category of approaches that a mean– variance approach falls into. Using a HAT approach
to estimating unbiased expected values and then testing dependence assumptions, followed
by using a deterministic modelling approach, can be interpreted as a very useful special case
methods based on moments approach. It is relevant when only the first moment about the
origin (expected value) matters, because variability is just noise and systemic uncertainty
issues are not a concern. This is often an extremely valuable approach because of its sim-
plicity. It can be useful to see it as drawn from the set of approaches based on methods of
moments developed by Markowitz (1959), using a Markowitz mean– variance framework at
a conceptual level to explore why systemic variability may pose problems for any determin-
istic approach and when a simple deterministic model is a sound choice.
It is also sometimes convenient to see mean– variance models as a way of thinking about
the implications of well- behaved correlation and separability assumptions in terms of a
Markowitz (1959) approach to portfolio analysis or early Program Evaluation and Review
Technique (PERT) model usage (Moder and Philips, 1970). However, variance- based
approaches and the use of further higher moments to measure ‘peakedness’ and ‘kurtosis’
have no direct operational role in the HAT approach as it is used by this book because meth-
ods that use them are too restricted or clarity inefficient in the contexts considered. If you
wish to use them, they can be formally included within the HAT framework in the sense that
associated approximation error can be tested in a more general framework.
All other useful approaches which do not involve sampling form a broad residual cat-
egory which includes all the parametric specifications of probability distributions discussed
in the previous section. This residual category also includes approaches based on under-
standing that a random arrivals process implies a Poisson distribution of arrivals over any
given time period and a negative exponential distribution of times between arrivals, drawn
on in Chapter 9. Furthermore, it includes the use of Normal probability distributions based
on the Central Limit Theorem which holds approximately even for small numbers of added
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 95
sources of uncertainty provided they are independent, plus other useful statistical ideas not
developed in this book. Basic statistics texts cover a wide range of these kinds of ideas, and
more specialist texts addressing subjects like queuing theory and safety theory cover further
comparable ideas.
All useful simulation approaches using sampling as a computation option for com-
bining multiple sources of uncertainty form another broad category which is usefully seen as
part of the generalised HAT framework. Simulation using sampling in a basic Monte Carlo
simulation sense adapted to the context modelled (e.g. see Hertz, 1964, or Grey, 1995)
is essential in some very complex ‘non- reducible’ contexts as a way of combining sources of uncertainty which have been specified using a HAT approach or any other comparable component distribution approach. These simulation approaches are also a clarity efficient
optional approach to computation associated with all the HAT models discussed in this
book if software is needed, specialist HAT software based on probability arithmetic is not
available, and developing this kind of specialist software is not clarity efficient.
All post- 1980s clients of mine who did not have access to BP discrete probability arithme-
tic based software used HAT based approaches like those discussed in this book in conjunc-
tion with readily available commercial simulation packages like ‘@ Risk’ for the probabilistic
modelling aspects of analysis requiring computer software support. My post- 1980s pub-
lications with a range of colleagues have always assumed that those using prototype EP
approaches would take the same route. Your organisation should also take this route when
software is needed unless a lot of HAT based modelling computation becomes the norm.
The key basic drawback of a simulation approach based on sampling is ‘sampling error’,
which can be controlled by just using bigger samples, or by various sampling stratifica-
tion assumption approaches. For example, to use Monte Carlo simulation to compute the
total duration distribution for two sequential activities with individual independent dura-
tion probability distributions defined by Table 3.1 and Figure 3.4, each ‘sample’ involves
extensive computational effort. A pseudorandom number generator routine has to be used
to generate a random (uniform distribution) number over the range 0– 1. This random
number is entered on the y- axis of the cumulative form of Figure 3.4 to read off a sample duration result. Separate samples for each component to be combined are required, with the
results of these samples then combined to yield one sample observation of the overall result.
Repeating this for 1,000 to 10,000 sets of samples yields a reasonably error- free overall
result, the precision required depending on the extent to which the low probability regions
of the overall probability curve matter.
The obvious advantage of using a readily available commercial package like @ Risk is low
cost software. Users of EP approaches who are not interested in the technical how to do
it issues need not concern themselves with the difference if a sampling approach is used
rather than a probability arithmetic approach provided the analysts in charge of computation know what they are doing in terms of both dealing with dependence and dealing with sensitivity analysis concerns. Some of these issues are briefly discussed near the beginning of Chapter 4 and are revisited towards the end of Chapter 11.
Simulation in a more general sense as explored by books like Tools for Thinking (Pidd, 1996) involve more general process and modelling philosophy ideas in addition to compu-
tational options. Some of these ideas can be seen as valuable additions to the set of process
styles to be considered during the select and focus the process phase of a UP. Simulation
in this broader sense might not make use of a HAT approach, but it could do so, and
simulation in this broader sense has a wide variety of application areas and exploratory roles
beyond the scope of this book.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
96 Foundations
The overall rationale for the adoption of a HAT approach as a core conceptual and
operational framework for this book includes five important arguments:
1 it does not require any restrictive framing assumptions when specifying probability
distributions;
2 it maximises the flexibility of approaches to all forms of statistical and causal depend-
ence, including embedding decision trees;
3 using a generalised probability arithmetic approach to combining multiple sources of
uncertainty as illustrated by the numerical examples in Part 2 clarifies what is being
assumed for all relevant parties – the model builders working ‘at the sharp end’, the
‘directors’ and ‘experts’ who have to approve or make all the key decisions, and every-
body else who needs to take part in the decision making or just understand the rationale
underlying decisions beyond their immediate control;
4 clarity efficient special cases of all other useful other approaches can be embedded,
be they simple shortcuts like limiting modelling to deterministic analysis using unbi-
ased expected values or Poisson process modelling assumptions for more modelling
power or the use of readily available commercial simulation software when this is
expedient; and
5 all working assumptions which have been deemed appropriate for the context are visible
and can be tested, including the use of a specific parametrically defined distribution like
a Normal (Gaussian) when this working assumption is judged useful.
HAT approaches adapted to specific contexts within an EP perspective should explicitly aim
for the support of decision- making approaches which are opportunity efficient in an overall
sense – considering decision-error costs as well as decision- making costs. HAT approaches
are a key component of the overall EP toolset which influences the EP mindset.
Any alternative to a fully generalised HAT approach which limits itself to a smaller sub-
set of the six categories of approaches just outlined is potentially less clarity efficient, with
opportunity- inefficiency implications. This is because it gives up generality which may be
useful conceptually or operationally. A key loss may be clarity about the use of sensitivity
analysis tools like those first explored at the beginning of Chapter 4 and then used through-
out that chapter and Part 2. Any loss of clarity associated with failing to undertake this kind
of effective sensitivity analysis is potentially a very serious issue.
Systematic decomposition of multiple sources of uncertainty
An important source of insight for William when the two scenario approach was introduced
was qualitative consideration of the particular source of uncertainty responsible for the data
point observation of an 18 week delay plus other potential sources of comparable delays for
his project. This qualitative analysis on an informal basis allowed William to start thinking systematically about contingency planning, including dealing with a 52 weeks scenario and
a project cancelled possibility.
We now need to develop this idea further on a formal planning basis, addressing system- atic formal quantitative analysis as well as associated supportive formal qualitative analysis. We also need to address doing so in the context of all the components of interest in a goals–
plans relationship structure – for example all the activities requiring completion to define
overall project duration and cost when planning a project. In project, operations and over-
all corporate management contexts we need to embrace a multiple sources of uncertainty
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 97
perspective with a view of the role of contingency planning which is as general, powerful
and flexible as possible.
To illustrate part of what this involves, continuing in a project planning illustrative con-
text, consider another tale. This tale is a revised and embellished version of a conversation
I had with a small group of BP International offshore North Sea oil and gas programme
managers in London in 1976. I was making a sales pitch for BP hiring me as a consultant
via Acres International Management Services at the beginning of a consultancy relationship
with BP which was very close over a period of eight years. The BP programme managers
were responsible for projects involving the construction of offshore platforms, float- out and
installation of the platforms in the North Sea, buying the pipe and arranging for the pipe
sections to be coated with concrete, laying pipelines between the platforms and their termi-
nals on the shore, and the fabrication and installation of all the ‘modules’ which packaged
the oil and gas production facilities and accommodation facilities on the platforms, plus all
associated hook- up and integration operations.
One programme manager I will call Peter started and led the discussion on behalf of
his colleagues. He indicated that BP’s current planning approaches were not letting them
deliver these projects on- time or within budget. They were under great pressure from their
board to achieve both. His wish to resolve this pressure was the reason for our conversation.
They all knew cost was heavily time- driven, but they did not know how to effectively cap-
ture the complexity this involved. They were particularly interested in understanding how
they could do better with estimating the duration and cost of laying an offshore oil pipeline.
They wanted to begin by discussing estimating how long it would take to lay a pipeline
given various assumptions about when to start and what wave height capability lay- barge to
use, computing an unbiased expected value of the duration of the pipe- laying activity. They
then wanted to use this analysis approach as a basis for an approach to obtaining an unbiased
estimate of all other activity durations plus the expected costs associated with all activities.
What was the best way to proceed?
I began by indicating that the common practice PERT probabilistic models developed in
the late 1950s which they were familiar with had been acknowledged by a growing number
of practitioners as a significant improvement on similar CPM deterministic models (Moder
and Philips, 1970). One key reason was less bias associated with activity expected duration
estimates because an explicit range- based interval estimate was used for the duration of
each activity. However, PERT model estimates of overall project duration were usually far
too optimistically biased to have any credibility, and associated cost estimates were gener-
ally even more optimistically biased. One key reason was a PERT model assumed that the
probability distribution defining the duration of each activity was independent of all other
activity durations. In practice, dependence was significant and crucial, and two quite differ-
ent aspects of dependence needed careful attention.
One aspect involved dependence in terms of the sources of uncertainty encountered by
base plans plus direct knock- on effects. PERT models assumed that if activity X was fol-
lowed by activity Y, the probabilities that activity Y will experience no problems or a lot
of problems were independent of the probabilities that activity X would experience no
problems or a lot of problems. In practice, there were usually common reasons for many of
the problems arising in different activities, like the same initial planner or contractor with
the same level of competence. In addition, in practice there were usually important knock-
on effects that implied a further significant positive correlation. These knock- on effects
might include exceedingly complex forms of positive dependence sometimes referred to as
cascade effects. Sometimes it might be convenient and practical to treat this dependence
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
98 Foundations
in statistical dependence terms via fairly simple correlation assumptions. But sometimes
an in- depth understanding of the causal dependence relationships generating this kind of
dependence was crucial. Intermediate cases best served by probability tree specification of
complex statistical dependence could also be relevant.
The other aspect involved recognising the crucial importance of causal dependence associ-
ated with reacting to delays to base plans via contingency plans. PERT models assumed that
the plan for executing activity Y after X had been completed was always created and assessed
without considering how long X might actually take. But in practice, if activity X took
significantly more or less time than expected, the plan for a following activity Y might be
altered, whether or not prepared contingency plans were in place in advance. Furthermore,
in practice planning and control should be closely coupled with this important response pos- sibility in mind, and in practice some degree of prior contingency planning should be part of most planning, associated with ‘proactive responses’ which alter base plans plus ‘reactive
responses’ which respond after the fact but may need set up in advance. PERT models
ignored consideration of prior contingency planning and unplanned later responses to delay
or earlier- than- expected completion. With or without prior contingency planning, after-the-
fact contingency responses usually reduced the duration of Y when X was delayed, inducing
negative correlation in terms of activity durations to keep the overall project duration within
the original estimate if possible. But this increased the cost, amplifying the positive depend-
ence between cost components because of common market forces, introducing a key causal
interdependency between duration and cost which was driven by contingency plans.
‘Generalised PERT’ had been developed in the early 1960s to consider prior contingency
planning. These generalised PERT models embedded decision trees at the beginning of
each activity. If prior activities were late, a contingency plan could be triggered. For exam-
ple, if activity X was on time, base plan Y1 could be used, but if activity X was late, contin-
gency plan Y2 could be triggered, and if activity X was early, contingency plan Y3 could be
used. BP needed models with this prior contingency planning capability, in addition to an
approach to correlation forms of dependency plus conditional treatment of more complex
statistical dependencies which was general and flexible. The HAT framework basis for the
Generalised PERT approach, which was explained in outline in terms of prototype vari-
ants, would provide much more realistic estimates of expected duration and cost outcomes
because it could address both statistical and causal dependence, base plans and contingency
plans.
I then suggested that BP also needed a generalised Markov process structure (dealing
with semi- Markov processes for those who want to get technical, with time- dependent
transition distributions), further extending the HAT approach within a Generalised PERT
framework. BP planners could begin with a textbook GERT (Graphical Evaluation and
Review Technique) model. Moder and Philips (1970) discussed GERT as well as more basic
CPM/PERT models, but at that time GERT had limited applications outside the US aero-
space industry to illustrate the approach. These models could decompose the uncertainty
about the answer to the question, ‘How long would it take to lay the pipe?’ into ‘How much
pipe could be laid in each successive month?’ A generalised Markov process structure was
ideal for dealing with weather windows. I explained how I had used these GERT models
successfully on an Acres Consulting Services study a year earlier of a major Canadian Arctic
oil pipeline project, when a somewhat different but equally important weather window was
involved. North Sea pipe- laying was not feasible in the winter; summer was ideal. Spring and
autumn shoulder seasons were difficult and risky but could be crucial. A generalised Markov
process structure would allow consideration of different weather expectations in each month
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 99
using appropriate weather data and reactive responses geared to progress each month, in
addition to the Generalised PERT responses at the end of each month of the pipe- laying
activity. This approach would give BP project planners the ability to test the implications of
different lay- barge capabilities and different assumed start dates making good use of avail-
able weather data.
But in order to separate weather effects from other sources of uncertainty, and use weather
data directly in an effective way, BP planners also needed a further feature, a further exten-
sion of the HAT approach within a GERT framework – later called the SCERT (Synergistic
Contingency Evaluation and Response Technique) approach (Chapman, 1979).
The basis for the SCERT approach had occurred to me when reviewing what had been
learned from the Acres Arctic pipeline study. One analysis problem associated with using the
basic GERT approach arose every time we addressed the question, ‘If we experience delays
in month X, what should we do to speed things up in the following month Y?’ The preferred
answer was almost invariably dependent on what went wrong. But the basic GERT model
does not consider that question.
Insightful analysis yielding high levels of clarity has to focus on all the questions that really matter. If questions that really matter are not being asked, something effective and efficient
needs to be done about it.
This review coincided with a separate review considering lessons learned during a separate
Acres study of the best way to approach seismic risk associated with the location and design
of nuclear power stations. I had reviewed US standards with several Acres colleagues with a
view to contributing to the drafting of Canadian standards, which Acres wanted to address
as a marketing investment.
The key to SCERT was borrowing fault tree and event tree ideas from the common
approaches to nuclear safety issues, plus basic decision tree approaches to model specific
sources of uncertainty which could be addressed by responses specific to those sources and
embedding these additional model features in GERT models without losing the general
response features of a GERT approach.
I had not yet tried out this prototype SCERT idea, but the approach looked promis-
ing. It would involve breaking down the quantified uncertainty associated with progress
in each month for an activity like pipe- laying into multiple sources of uncertainty, drawing
on the fault tree and event tree ideas used for US nuclear power station seismic risk stud-
ies (NUREG, 1975), plus basic decision tree ideas. Using this approach within the GERT
model structure, the focus for a key activity like pipe- laying might be a small number of
‘specific responses’ for very important specific sources of uncertainty, but ‘general responses’
to deal with the residual after all the specific responses for each successive month would
also be involved. Maintaining the general response concept would facilitate dealing with all
threats and opportunities not fully dealt with via specific responses, including any ‘unknown
unknowns’.
The first key source requiring separate quantitative treatment for the pipe- laying activ-
ity was ‘weather’ – more specifically wave height. One reason for the separate treatment
of weather was the significant impact of this source of uncertainty. Closely coupled further
reasons were the wave height capability specification of the chosen lay- barge was a key prior
decision, good wave height weather data was available, and given the obvious importance of
weather, it would be sensible to make effective use of the available wave height data.
Wave height data for the sea areas in question could be used to assess a probability dis-
tribution for the number of working days in each month during the working season given
the wave height capability of the assumed lay- barge. If ‘productivity variations’ implied 1 to
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
100 Foundations
5 km of pipe laid per ‘lay- day’ when the waves were below the assumed lay- barge capabil-
ity limit, with a mean of 3 km/day, then an expected value of 3 km/day could be used to
transform wave height distributions into rate of progress distributions. The initial work-
ing assumption would be that ‘productivity variations’ due to all sources of uncertainty
other than weather cancelled out, and no further specific sources of uncertainty other than
weather were involved. Markov process runs could assume a given lay- barge wave height
capability and start date and compute the probability of completion within the planned
single season, plus the chance of further seasons, with a cumulative probability distribution
for completion within each month in the feasible completion date range. BP planners could
experiment with different start dates or lay- barge wave height capabilities and get a feel for
what ‘weather’ on its own might do to their plans. They could understand the implications
of good luck, bad luck, and everything in between in terms of weather on its own. They
could then start to address any other sources of uncertainty deemed important enough to
warrant separate treatment.
The next separate source of uncertainty on this agenda might be ‘wet buckles’ – the
smooth S shape of the pipeline section from the lay- barge to the ocean bed develops a kink
which fractures the steel in the pipeline, letting in water. The pipeline then becomes too
heavy for the barge to hold, ripping itself off the barge unless it is released very quickly, sink-
ing to the ocean floor. The basic response is ‘repair’ – send down divers to cut the damaged
pipe away and ‘cap the pipe’ (sealing the end with steel cap), ‘dewater the pipe’ by sending a
‘pig’ (a torpedo- shaped object which just fits the pipe) through it under air pressure (valves
in the cap let the water out as the pig progresses), and then pick up the dewatered pipeline
and carry on. However, a rich range of alternative responses could be explored – like a
bigger lay- barge which is less prone to buckles (and faster because it can lay pipe in worse
weather).
Furthermore, secondary sources of uncertainty could be explored – like repairing a pipe
buckle might lead to the pig sent through the pipe getting stuck because of debris in the
pipe, with more than one relevant secondary response, both involving the possible loss of a
significant amount of coated pipe which might be difficult to replace without a significant
delay. Divers could be sent down to cut off the pipe behind a stuck pig and fit a new cap,
then another pig could be used to dewater the shortened pipe. Alternatively, the air pres-
sure could be turned up to try to pop the pig through the debris, with a possibility of this
response resulting in structural damage to the pipe involving even more lost pipe than cut-
ting off the pipe behind the stuck pig.
Further still, the loss of pipe involved as a consequence of a stuck pig should trigger con-
sideration of dependence with both earlier and later decisions. For example, should extra
pipe be ordered and coated for this project in advance? If it was not needed for this project
because no buckles occurred, should it be suitable for subsequent projects? And does this
imply a pipeline design strategy common to all offshore projects?
Dependence between decomposed sources of uncertainty because of common underlying
problems could be modelled directly in various ways if it mattered. For example, the com- mon interval rectangular histogram probability density function approach could be used in
a HAT framework to define anticipated lay- days and rate of progress distributions for each
month of the year when pipe- laying might be attempted. Using this starting point, progress
for each month and overall progress by the end of each month could be considered in terms
of a probability tree using class mark values. In this framework different risks of a wet buckle
could be associated with very good weather, very bad weather, and intermediate situations,
recognising that the highest risk might involve the intermediate situations. This is because
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 101
in very bad weather no pipe- laying would be attempted, and in very good weather no prob-
lems are likely, but in the mid- range problems were highly likely.
Other important decomposed separate sources to follow might be major equipment fail-
ures (perhaps also weather dependent), a late start because an earlier contract involving the
barge contracted for was delayed, coated pipe supplies delay, and so on.
A ‘dry buckle’, when a buckle occurred but the pipeline steel did not fracture, could be
resolved fairly quickly. In effect, the lay- barge could just back up and recover the damaged
pipeline, then carry on. This was a useful example of a relatively minor problem not requir-
ing separate treatment. Dry buckles were an identified example of the ‘productivity varia-
tions’ source of uncertainty, along with a list of other identified minor problem not worth
separate planning attention at this stage, plus all unidentified minor problems.
I argued that if this approach was used, along with significantly simplified versions for
other activities which might not need a Markov framework, BP planners could develop
what was later labelled the SCERT approach (Chapman, 1979). It should provide an unbi-
ased approach to activity duration estimation and associated cost estimation, because it
would address the identified flaws in current common practice approaches which gave rise
to observed bias.
Once the BP planners understood the duration of a central activity like pipe- laying
in these terms, they could build a dependent cost estimate on this framework. For example,
the cost of a lay- barge per working (lay) day and per idle (weather) day during the period
the barge was on hire could be estimated in probability distribution terms, and the cost
of the pipe, coating it and transporting it, could also be considered. Once the duration and
cost of pipe- laying were understood, earlier and later activities could be addressed, building
a model which considered dependencies between activities, including the knock- on effects
of delays to or partial incompletion of earlier activities on following activities. For example,
if delayed completion of module construction onshore threatened a missed weather window
for their installation, offshore completion of the modules could be considered, trading off
the different extra costs of the available options.
The modelling involved would address a causal understanding of some key sources of
uncertainty with a view to managing it more effectively as well as estimating its implications,
managing it more effectively ultimately proving even more valuable than unbiased estimates,
and the prime motivator for the SCERT approach.
This conversation with Peter and his colleagues, just recounted as a tale, is a modified ver-
sion of the actual conversation. In particular, it uses some ideas I did not understand until
later, and it omits a discussion of risk efficiency, part of my original sales pitch. However, as
a result of the actual conversation the BP programme managers were persuaded to try out
my suggested approach.
The SCERT label indicates a further generalisation in the tradition of PERT, General-
ised PERT and GERT, treating all these earlier approaches as special cases of SCERT, and
emphasising the broadly defined form of synthesis essential when analysis involves a formal
decomposition process of this kind. BP referred to SCERT as Probabilistic Project Planning
(PPP), but this book sticks to the SCERT label.
SCERT was the first sophisticated high clarity prototype of the specific process (SP) for
projects concept explored in Chapter 7. The SCERT approach was developed with signifi-
cant input from a number of BP staff. It was mandated by the BP board for worldwide use
for all large and sensitive projects. SCERT was used by BP for about a decade. As I under-
stand it, simpler approaches were introduced after a decade on the assumption that BP plan-
ners now understood the issues very much better, as did their contractors, and both felt they
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
102 Foundations
did not need the high clarity analysis SCERT provided. This was closely coupled to a policy
shift in terms of contracting lay- barge operators on what amounted to a fixed price basis
instead of the time-and-materials basis used earlier as part of a broader shift to outsourcing
the management of uncertainty. The eventual implications of these policy changes will be
touched on shortly and are further considered briefly in Part 2.
The first step in developing SCERT was trying out the modelling ideas and an associated
specific process using a ‘passive project’. I revisited a project recently completed by BP with
a senior BP planning engineer who had been directly involved. We asked the question, ‘If
BP were doing this project again, using a prototype of the new SCERT process and associ-
ated models, how should the planning team proceed?’
The next step was trying out the ideas just developed on a ‘live project’ – the Magnus
Field project, which came in on schedule and within budget despite some surprises.
Later statistical analysis of offshore North Sea BP projects which was undertaken to assess
bias associated with using the SCERT approach showed that projects delivered on time and
within budget was an appropriate expectation, but it was crucial to acknowledge that all
underlying assumptions ought to be understood by all the parties involved and treated as
conditions, with some nuances we had not fully anticipated in advance.
Understanding the assumptions underlying any approach to estimation is an important
clarity issue. When reviewing the outcomes for BP projects which had been planned using
SCERT to assess bias issues, we found that it was useful to discuss an explicit formal decom-
position of all relevant uncertainty into three components:
1 the ‘known knowns’ – sources of uncertainty which were ‘quantified’ (in the sense
that they were modelled probabilistically and understood appropriately in conjunction
with associated response structures using decision trees as well as statistical dependence
structures using probability trees),
2 the ‘known unknowns’ – sources of uncertainty which were identified but ‘not quanti-
fied’ (in the sense that they were treated as assumptions underlying the quantitative
analysis – usually referred to as conditions), and
3 the ‘unknown unknowns’ – sources of uncertainty which were either not identified or
misconceived.
The expected values and associated aspirational and commitment target values embodying
identified ‘provisions’ and ‘contingency’ obtained via the SCERT models addressed all the
known knowns in quantified analysis terms involving the well- founded causal and statistical
understanding of uncertainty modelled probabilistically. They also highlighted the condi-
tions these estimates assumed would hold, defined by the known unknowns. This had been
understood by everyone involved from the outset.
The BP board chose a P80 confidence level for budget determination from the outset,
and everyone involved saw the uplift from the expected outcome to the budget in terms of a
contingency which on average should not be needed. The P80 value rather than a P70 or a
P90 was a board level judgement to deal with the asymmetric implications of cost overruns
and under- runs from a board perspective.
The known unknowns were seen from the outset as identified conditions which the prob-
abilistic analysis depended up. In addition to decomposing sources of uncertainty associated
with PERT models of duration and linked quantitative analysis of cost estimates, SCERT
also decomposed sources of uncertainty which were not quantified, developing a reason-
ably rich understanding of conditions. Examples are ‘analysis assumes no more than 10 km
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 103
of pipe in excess of the planned pipeline length will be ordered and coated’ and ‘no major
equipment failures or fires’. When assessing bias by analysing completed projects we had to
recognise the implications of these assumptions for all the identified conditions which were
not met. Several assumed conditions failing to hold were common, usually without serious
implications.
The unknown unknowns were seen from the outset as significant surprises that we might
have anticipated but did not. We were surprised by each but not by the fact there would be
surprises. Their overall impact was limited and well within the contingency associated with
defining the budget in P80 terms. However, when assessing bias we had to accept that the
contingency provision used earlier was not actually set at the P80 level with the explicit
consideration of realising unknown unknowns in mind, and we really should have made an
explicit provision for the expected total impact of surprises. It is now self- evident this might
be a crucially important issue in some circumstances. It is clearly very difficult to assess the unknown unknowns because, by definition, we do not know what is involved. However, a
reasonable expectation about their collective impact is obviously greater than zero, so zero
is obviously an optimistically biased estimate.
This is a useful example of a very difficult issue which does not go away if you just put
it into a ‘too difficult’ box and forget it. From an enlightened planning perspective, those involved in planning should never be allowed to use a ‘too difficult box’ to ignore difficult issues needing explicit attention. Explicit working assumptions which may be wrong are always a preferred alternative to a ‘too difficult box’ which is ignored.
As a useful illustrative example, in retrospect we understood that it was clearly worth
making a provision for all the known unknowns that could be interpreted as ‘fine print
quibbles’ which those involved were not going to want to discuss, which together with a
possible provision for the unknown unknowns might be called a ‘no quibbles provision’.
This could be viewed as an updated and enlightened version of the traditional ‘plus 10%
(or more) for errors or omissions’ added to a deterministic estimate for centuries, perhaps
millennia, by most practical engineers and other cost estimators. It could also be viewed as
another illustration of using a closure with completeness component. In this case its purpose
was to avoid bias associated with concerns not identified and addressed individually on a
decomposed basis – because some potential concerns are simply ‘unknowable’ and some are
‘knowable’ in principle, but in practice it may not feasible to consider them on a decom-
posed basis, and some are just not worth the trouble of even considering for decomposed
treatment.
Dealing with both the known unknowns and the unknown unknowns was an issue iden-
tified as needing further sophisticated attention in each high clarity analysis context, also
needing further attention to address simple but robust approaches for all low clarity analysis
contexts.
Approaches to no quibbles provisions in all contexts remains a work in progress of impor-
tance, with no general best strategy that I am aware of. Approaches to very serious extreme
events is a separate area of ongoing concern. Approaches which are appropriate to any given
context may warrant great care or just very simple rules or something in between.
An important general implication is that what we mean by ‘unbiased estimation’ requires
clarity about the treatment of a no quibbles provision and any other relevant unknown
unknowns. It also requires clarity about who is accountable when what may have seemed
like sensible identified conditions (assumptions) do not hold.
One reason the BP project planning staff I worked with used the PPP label instead of the
‘Project Risk Management’ (PRM) label then in early adopter usage was the term ‘risk’,
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
104 Foundations
interpreted in the safety sense, was another BP group’s responsibility. Risk in a safety sense
was a rather special kind of known unknown from the perspective of the planners I was
working with, and they did not want to challenge the safety function’s ownership of the
‘risk management’ label or stray into the territory the safety department operated in. The
project planners assumed their treatment of safety and that of their contractors was reduced
to ensuring that they jointly planned to meet all standards and restrictions imposed on them
by the safety department.
The board had to look to a separate source of advice on safety. An implication I did not
consider at the time is the board also needed a separate source of advice on the linked envi-
ronmental risk issues raised by the 2010 Deepwater Horizon Gulf of Mexico disaster, plus
the complex relationships between contractual approaches and both safety and environmen-
tal concerns, both considered briefly in terms of a coordinated approach in Chapter 9.
Higher or lower levels of clarity when appropriate
Acres used the experience base our early success with BP provided with a wide range of other
clients in Canada and the US. BP collaboration with Acres included letting Acres use the
SCERT software systems which I helped BP develop for other Acres client studies in Canada
and the US, extensively adapted by my colleague Dale Cooper in some cases, as discussed
in Cooper and Chapman (1987). This HAT based software used a probability arithmetic
approach incorporating a first order arithmetic correction for the calculation error involved
of the kind discussed earlier in this chapter. Standard commercially available Monte Carlo
simulation packages which can do the same job in some circumstances became available later,
now the usual approach to implementing a HAT framework. Some of these Canadian and US
studies used lower levels of clarity, but some used very high levels of clarity for specific deci-
sions, like the study discussed in Chapman, Cooper, Debelius, and Pecora (1985).
Soon other UK based clients also followed, like National Power. National Power wanted
a simplified variant of SCERT as developed for BP, adapted to suit planning the construc-
tion of electricity generation power stations at a comparable stage in the project lifecycle to
that addressed by BP – the ‘execution and delivery strategy progress stage’ in the language
of Chapter 7. This stage followed an earlier ‘concept strategy progress stage’ and then a
‘design, operation and termination strategy progress stage’. Most of the Canadian and US
Acres studies addressed the earlier stages in project lifecycles, looking at project concept
strategy decision- making concerns.
The National Power approach was the result of the board member responsible for all
engineering projects approaching me after a public seminar on SCERT approaches. He ini-
tiated discussions which led to a combined cycle gas- fired power station being planned and
then built as a ‘demonstration project’. The demonstration project was carefully selected
as a learning context. The planning was led by a National Power planning engineer, with
my support, after internal training courses for everyone involved so they all understood in
broad terms what we were trying to achieve. The National Power variant of SCERT did
not use Markov processes, but it did use 1 to 10 separate sources of uncertainty per activity,
general as well as specific responses, and most other BP variants of SCERT characteristics. It
can be characterised as a medium clarity approach, midway between the simple low clarity
approaches discussed with William and the much more sophisticated high clarity approaches
to offshore pipe- laying discussed with Peter.
One defining characteristic of a HAT framework, as we move from the simplest of the low
clarity variants discussed with William towards the high clarity variants used by Peter and
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 105
his BP colleagues in a SCERT model, is the additional understanding of sources of uncer-
tainty. This involves causal understanding of uncertainty as well as sophisticated statistical
understanding of uncertainty. It involves being very clear about the known knowns. It also
involves very valuable insights into the known unknowns and the unknown unknowns.
While BP used some very sophisticated high clarity SCERT approaches for pipe- laying,
they also employed much simpler low clarity variants for activities which were less complex
and less crucial. When National Power introduced a simplified version of BP’s approach,
it used a portfolio of simpler approaches, depending on the complexity and importance of
each project and each activity within each project, as did other SCERT users. It soon became
clear that all organisations using SCERT variants needed a toolset of approaches which
ranged from the minimum clarity approach discussed at the beginning of the conversation
with William to the high clarity approach discussed with Peter or beyond.
Furthermore, even if they never found the level of complexity involved in the BP pipe-
laying approach directly applicable to their project’s activities, most SCERT users found the
high clarity conceptual framework employed by BP useful. It was useful as an underlying
conceptual framework for thinking about the simplifications needed for their approaches.
It was directly comparable to what some people refer to as ‘useful theory’. But it was not
theory. It was high clarity best practice, and the difference is subtle but important. A key goal for this book is providing you with an understanding of high clarity best
practice approaches to underpin your understanding of strategic clarity to help you to
understand how lower clarity approaches can be selected to best suit particular contexts that
you and your organisation have to deal with.
The defining characteristics of an EP approach to estimation include understanding in
broad terms the structure of the ambiguity uncertainty which underlies very simple mini-
mum clarity portrayals like Figure 3.1 which could be given a high clarity analysis. It is
important to understand from the outset when using a very simple approach that some
selected aspects of the underlying complexity may be worth more detailed understanding
later using suitable high clarity approaches. If the earlier simpler approach suggests that par-
ticular details matter enough to make the extra effort worthwhile, being able to apply fur-
ther effort where and how it will pay the biggest dividends is a key aspect of clarity efficiency. More generally, an EP approach is concerned about always using the simplest portrayal
of relevant uncertainty which is suitable for the immediate task, but doing so with a clear
understanding of the nature and implications of underlying simplifying assumptions to facil-
itate testing overall understanding and looping back to develop more clarity whenever it
looks as if that might be worthwhile.
Strategic clarity when choosing low clarity approaches requires an understanding of what
high clarity approaches can provide and the associated costs of additional clarity of various
kinds. The high clarity BP framework can be used as a high clarity conceptual tool to facili-
tate always asking the right questions with a view to simplifying systematically, exploiting
opportunities to simplify in the right way, avoiding risks associated with simplifying in the
wrong way.
Clarity efficient boundary concepts
This section interprets the ‘efficient frontier’ portrayal of clarity and the cost of acquir-
ing clarity portrayed by Figure 3.5 in terms of clarity efficient approaches to estimating as
discussed so far in this chapter, retaining a focus on project management contexts. Later
sections of this chapter and Chapter 4 broaden the scope of the clarity concerns addressed
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
106 Foundations
and consider corporate and operations management contexts. The rest of this book builds
on this foundation.
An efficient frontier interpretation of the clarity efficiency concept is a key conceptual tool
in the EP toolset. It is not an operational tool in a direct sense. However, understanding the
concepts which Figure 3.5 portrays makes practice much more effective.
Consider the framework of the portrayal of clarity efficiency of Figure 3.5 first and then
its interpretation.
The x- axis is not cost in £ or $ or some other currency because ‘cost’ has to be given a more general effort/cost interpretation which incorporates not only the cost of immediate
effort but also earlier investments in learning processes and formal training to develop com-
petence plus softer concerns like distinguishing between effort which is disagreeable and
effort which is satisfying. Effort which is disagreeable or awkward carries a higher oppor-
tunity cost than effort which is satisfying and straightforward, recognising that different
people may hold different views on the direct and indirect implications.
The y- axis is ‘clarity’ in the sense of ‘relevant insight which can be shared’. This is not a simple metric like ‘probability’ on the y- axis of earlier probability distributions diagrams. In
this chapter, ‘relevant clarity’ has been kept as simple as possible thus far by a focus on unbi-
ased estimation as the current dominant goal, but more than one kind of clarity is almost
always relevant – clarity is inherently a complex, multiple- attribute concept.
The minimum clarity estimate approach outlined for William is represented by the clarity
efficient point ‘c’ – less clarity is not appropriate, and using more cost (effort or time) to
acquire the same level of clarity is inefficient.
A high clarity approach like that associated with BP’s SCERT approach to pipe- laying
might be associated with point b1.
cost of acquiring clarity
a–c efficient boundary target, with
a maximum clarity approach
c minimum clarity approach
bi intermediate clarity approaches
ei inefficient approaches
c–d inappropriate clarity levels
key:
a
b3
c
d
competent management area
(usefully seen as the opportunity management area)
e1
non-feasible area
b2
b1
e2 e3
incompetent management area clarity in terms of relevant insight which can be shared
Figure 3.5 Clarity and the cost of acquiring clarity in an efficient- frontier portrayal.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 107
A maximum level of clarity, represented by the clarity efficient point ‘a’, can be assumed
to exist, but our usual concern is the intermediate clarity efficient approaches to points like
b1, b2, and b3 along with the minimum clarity point c approach.
The tale associated with William started with point ‘c’ but then explored additional clarity
associated with the clarity efficient boundary between points ‘c’ and b3. The focus of the
quantitative analysis was unbiased estimation with an appropriate understanding of associ-
ated uncertainty, but qualitative understanding of issues relevant to contingency planning
was also touched on, as were higher clarity understanding of aspirational target and com-
mitment target plus expected value and balanced target issues. This discussion addressed
what might be called ‘additional clarity without decomposition’ or ‘refinement without
restructuring’, with a concern for clarity efficient estimation of activity duration treating all
quantified uncertainty as a single source of uncertainty. This single source within a single
activity was first estimated in terms of a very simple minimum clarity approach, later in
terms of multiple scenario approaches, including the n scenario (or classes) generalisation
of the HAT approach of Table 3.1 and Figure 3.4, plus conditions associated with extreme
outcomes not embedded in the probabilistic analysis. Both simpler and more sophisticated
variants of Figure 3.4 could have been used, employing as many scenarios or classes as neces-
sary to capture the shape of the single source of uncertainty appropriately. In this sense, the
HAT approach as a whole as developed for William provided a flexible toolset for dealing
with a single source of uncertainty representation which generalised a PERT approach to
each activity, recommended as both a conceptual basis and as an operational starting point.
William’s tale also touched on additional clarity if the relevance of understanding how to
manage the underlying causes of the more extreme delay possibilities is a possible concern.
The discrete scenario interpretations explored in the HAT framework with William enrich
our interpretation of a continuous variable single source of uncertainty model. They also
hint at linked multiple source of uncertainty probability and decision tree quantitative
modelling approaches without developing these ideas.
The tale associated with Peter explores an example of seeking a point like b1 in a some-
what different and much more complex context, employing a significantly generalised vari-
ant of the HAT approach. It addresses multiple sources of uncertainty within each activity
plus multiple activities directly. The Markov process approach also breaks activity duration
into progress each month plus total progress achieved to date before decomposing sources
of uncertainty like weather and buckles. There are direct implications for contingency plan-
ning based on a comprehensive understanding of key cause– effect relationships. Decompos-
ing activity durations into rate of progress distributions for successive months was crucial
for offshore pipe- laying, but so was decomposing separate sources of uncertainty within the
monthly progress distributions. The focus of this discussion was still unbiased estimation,
but effective formal contingency planning with project implementation benefits was also
involved.
National Power’s expectations, which were met and exceeded, were analogous to seeking
overall clarity for key project activities at a point like b2, with an understanding of how to
use most of the b1– b3 range as needed for specific projects. Indeed, BP needed the whole
range, as did the MoD, recognising that warships, offshore oil projects, and electric power
generation projects are all very different, as are individual components within each project
in any context.
Whether any given organisation is currently in the competent management region at
points like e1, in the incompetent management region at points like e3, or on the boundary
between competent and incompetent management at a point like e2, is clearly a sensitive
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
108 Foundations
and potentially contentious issue. However, all organisations need to assess where they are
in relation to the competence/incompetence boundary, in terms of all relevant aspects of
clarity efficiency. This assessment needs to be understood at the level of those designing
processes, at the level of those using processes, and at the level of those responsible for asso-
ciated governance. If significantly different views amongst these players are involved, a lack
of corporate competence is probably a reasonable default assumption.
Getting onto the clarity efficient boundary involves more insight for less effort, which
can be seen as an opportunity which needs to be identified and captured. But it also means
more reward for less risk, a form of opportunity with significant potential tangible rewards
explored further in the following sections of this chapter. More generally, an important issue
to be developed later is seeing the competent management area as an opportunity manage-
ment area.
The focus of EP as a whole is opportunity management – being in the opportunity man-
agement area of Figure 3.5. This implies a constant search for clarity efficiency at the most
appropriate level of clarity. It involves understanding both the benefits and the costs of more
clarity in unbiased expected value and range terms. It also involves understanding the con-
sequences of misjudging what is needed, as a part of the more general understanding of the
consequences of incompetence.
In practice any organisation may want to use approaches over most of the c– b1 range,
perhaps even approaching point ‘a’, with choices depending on the circumstances. To do so
effectively they need to understand all the available options.
Being able to think about uncertainty over the whole of this range in flexible discrete
or continuous variable forms within a HAT framework is a useful capability, part of the
capability- culture concept an EP approach is concerned with developing to ensure organisa-
tions can operate in the opportunity management area of Figure 3.5. A HAT based approach
is clarity efficient when simple approaches are needed, and it can become crucial if complex
dependencies are involved.
Being able to think about uncertainty at various levels of decomposition, and choose
appropriate levels within appropriate structures, is also crucial. So is pursuing contingency
planning and option planning more generally, within whatever source of uncertainty struc-
ture best suits the context.
Clarity can be increased as appropriate via well- founded causal understanding of all
important sources of uncertainty whenever this is feasible and worthwhile. Within both
causal and statistical dependency approaches it is important to have appropriate clarity about
ambiguity uncertainty as well as variability uncertainty, event uncertainty, capability- culture
uncertainty and systemic uncertainty.
Having at least one person on any management decision- making team and in every associ-
ated planning team who knows how best to add clarity within this complex mix is a key EP
corporate capability, central to corporate strategic clarity.
Board level managers do not need to know how to engineer clarity efficiency – but they
do need to know enough about how it works to detect when it is missing, and they do need
to know what needs to be done about it. Others will also need to know how to do it or who
to speak to who knows how to do it.
Even if an organisation’s only concern when choosing an appropriate level of clarity for
estimation purposes is an unbiased estimate of a performance measure attribute like the
duration or cost of a project to be on the clarity efficient boundary, for this purpose alone
it is crucial to appreciate that the estimation process involves underlying plans. In a project
management context uncertainty about duration or cost is driven by uncertainty about the
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 109
underlying plans, and the assumptions those plans are based on, as well as what might or
might not happen if the plans are pursued.
Given this understanding, it should be clear that all uncertainty about the plans is rel- evant. This includes uncertainty about important strategic issues which are not yet fully
understood. It also includes uncertainty about detailed planning concerns to be addressed
later in the project lifecycle. Furthermore, it includes capability- culture issues including
capability- culture uncertainty often not addressed directly until things go wrong. To be in
the competent management area of Figure 3.5 requires a competent understanding of all potentially relevant uncertainty. Being very near the clarity efficient boundary requires a very
high degree of competence.
If we look at all uncertainty in terms of the high clarity SCERT approach developed for
BP, it should be clear that a ‘source of uncertainty’ in the EP sense illustrated by this context
may involve:
1 event uncertainty (a buckle may or may not happen), but it also involves
2 variability uncertainty (weather and its wave height metric are matters of degree), plus
3 systemic uncertainty (dependence of various kinds), plus
4 ambiguity uncertainty (what we do not understand about 1 through 3 above, as well as
all the unknown unknowns), plus
5 capability- culture uncertainty (including assumptions about all relevant capability and
culture concerns involving contractors and their subcontractors).
By 1980 the BP North Sea project planners I worked with were systematically looking at the
first three and beginning to develop a comprehensive understanding of the fourth. We did
not address the fifth, in the sense that contractors were paid on a time and materials basis to
follow BP developed plans created by those I was working with within operating standards
and rules specified by other BP players, and questioning the capability of other BP players was
not part of the agenda. With hindsight this was a potentially important weakness in the analy-
sis approach, but at that time the implicit capability- culture stealth assumptions were well
founded and held – problems associated with stealth assumptions not holding did not arise.
If we now look at all uncertainty in terms of a low clarity approach like a basic PERT
model of a pipe- laying activity and an associated cost estimate, it should be clear that dura-
tion and cost estimates treated as single sources of uncertainty at the activity component
level also involve all five of the uncertainty types listed earlier, but we are treating them
collectively as if they were a single source of variability uncertainty for each activity. Fur-
thermore, we are omitting consideration of any kind of uncertainty about relationships
between activities and associated cost estimates (systemic uncertainty), including that asso-
ciated with potential contingency planning of the SCERT variety. In a PERT framework
our ‘quantitative’ understanding of uncertainty is largely a ‘statistical understanding’ of
uncertainty which is very limited in scope and capability, with little in the way of competent
‘causal understanding’ of uncertainty beyond the role of simple precedence relationships.
The same is true of PERT models of simpler contexts, although the inherent underlying
complexity may be less.
Even if a simple minimum clarity approach is used for a single activity, as discussed with
William, the single source of uncertainty addressed has to embrace all of these kinds of
uncertainty.
There is a good case for arguing that with hindsight the basic 1960s PERT models
embraced a broad view of uncertainty comparable in many ways to a minimum clarity
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
110 Foundations
approach extended to address multiple plan components and that was the most sensible
place to start. However, the minimum clarity model nature was implicit, the 1960s PERT
model did not lend itself to generalising in an obvious way, and it did not draw attention
to the limited scope and competence of the quantification of uncertainty. What the clarity
efficient set of approaches to unbiased expected outcomes illustrated in this chapter does is
generalise PERT, via Generalised PERT, GERT, and then SCERT. This clarifies the severe
limitations of the basic 1960s PERT approach, without implying that all or some of its sim-
plifications relative to SCERT should not be used whenever this is appropriate.
In a project planning context, classic illustrations of organisations in the incompetent
management area of Figure 3.5 are those organisations which fail to consider any uncer- tainty explicitly, sticking to a CPM approach based on point estimates and comparable
deterministic costing approaches. For example, they might decompose the current plan for a
large project into 500 or even 5,000 or more activities and carefully plan and cost each indi-
vidually in point estimate terms, but omit to systematically consider any uncertainty. What
BP demonstrated convincingly in the 1970s and 1980s was that using just 20 to 50 activi-
ties, but extensively decomposing the sources of uncertainty and contingency responses to
those sources, plus secondary sources when appropriate, is usually a much better approach. More contemporary illustrations of organisations in the incompetent management area of
Figure 3.5 are all those organisations that subscribe to project risk management approaches
which involve a focus on an event or condition view of ‘risks’ that may or may not happen,
using long ‘risk lists’ or ‘risk logs’ and ‘probability impact grids’ (matrices). This is a very
widespread current practice, and a serious source of concern, addressed briefly in what needs
to be done terms in Chapter 7.
Moving beyond project planning, in numerous very important decision- making areas far
too many organisations are in this ‘incompetent management’ area. There is a tendency to
use too much of the wrong kind of detail to ask the wrong kind of questions, and a failure
to understand that the right questions addressed in a clarity efficient manner will provide
more insight for less effort.
Even minimum clarity estimates can deliver much more than unbiased estimates for a
minimal level of effort. The SCERT approach developed for BP provided the unbiased
approach to estimation being sought, and the spinoff benefit of target value, expected value,
and commitment value distinctions. But the decomposition of sources of uncertainty and
associated contingency planning also involved a very different approach to the underlying
complexity than the minimum clarity approach outlined for William, and part of the pay- off
was risk efficiency plus opportunity efficiency, not yet considered but of central importance.
More generally, clarity efficiency in terms of ‘appropriate clarity’ can involve issues which
are not limited to unbiased estimates and delivering projects on time and within cost. We
can add much more clarity in a rich variety of ways for a rich variety of reasons. We now need to start to explore how and why by addressing further issues. The first of
these additional issues is seeking risk efficiency with a view to avoiding ‘risk inefficiency’.
Risk efficiency – with avoiding risk inefficiency as the focus
Formal contingency planning implies that choices between decision options are relevant in direct and explicit formal terms. As soon as any choices between decision options become an explicit issue, in ANY context, ‘risk efficiency’ should become a central concern needing explicit attention, whether or not formal contingency planning gives rise to the choices. This atten- tion requires a clear understanding of what risk efficiency means in both conceptual and
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 111
operational terms, why risk inefficiency should be avoided, and how systematically seeking
risk efficiency should be approached in both formal and informal terms.
‘Risk efficiency’ means ‘a minimum level of risk for any given level of expected reward’. Another way of saying the same thing is ‘a maximum level of expected reward for any given
level of risk’. However, the former is operationally more useful. This is because direct met-
rics for expected reward are often feasible, but measuring risk is usually much more difficult. ‘Risk inefficiency’ involves taking more risk for less expected reward than necessary, argu-
ably a very serious form of incompetence if significant risk inefficiency is involved. It is
usually easily avoided by those who are competent in EP terms, but it is often an unknown-
unknown risk, endemic in organisations that do not understand what risk efficiency implies
or how to deal with it in a clarity efficient manner.
To discuss making use of the risk efficiency concept in practice, it is often convenient to
replace ‘reward’ by ‘revenue’, assuming costs are fixed, or by ‘cost’, assuming revenue is
fixed. In both cases, assuming that a simple metric correlated with ‘profit’ is a suitable finan-
cial measure of reward is implied. For example, in the BP offshore project planning context,
oil and gas revenue was assumed to be independent of what the BP planners I was work-
ing with were concerned about most of the time, apart from opportunity costs associated
with any delays to first oil or gas availability, which had to be estimated and used to address
unproductive capital and lost revenue issues. In this context, ‘cost risk’ becomes a way of
looking at ‘reward risk’ without addressing revenue beyond opportunity cost implications.
This section assumes that project contingency planning using a cost metric was being under-
taken, and the focus of concern was risk efficiency in terms of avoiding cost risk inefficiency.
Risk efficiency is a concept misunderstood and overlooked by many ‘experts’ in the risk
management field. Whether or not you are interested in risk management at a technical
level, understanding risk efficiency in general terms and what risk inefficiency implies is abso-
lutely essential. You also need to understand ‘enlightened caution’, ‘enlightened gambles’,
and ‘enlightened prudence’, concepts closely linked to an understanding of risk efficiency.
This set of concepts has important behavioural implications for those organisations which
understand the difference between good management and good luck, bad management and
bad luck, and for those organisations which do not. ‘Risk appetite’ is another important
concern which also needs understanding in a framework centred on risk efficiency.
Risk efficiency is a concept with origins usually associated with Harry Markowitz, who
received a Nobel Prize in Economic Sciences in 1990 for his work developing the risk
efficiency concept in the context of efficient diversification of a portfolio of financial invest-
ments. Risk efficiency as discussed by Markowitz in his seminal 1959 book, and embedded
in much of economics since, uses ‘variance’ as a surrogate measure of risk. Variance is a
basic statistical measure of spread or range. Much of the power of Markowitz’s approach
is associated with the way he deals with statistical dependence in covariance or correlation
terms. The EP approach to risk efficiency is much more general, addressing various forms of
causal dependence as well as statistical dependence. Furthermore, an EP approach explicitly
avoids the assumption that risk can be measured via variance. It uses a ‘decision diagram’
concept for each relevant attribute to avoid any surrogate measures of risk, and it offers scope for various ways of approaching complex multiple attribute trade- offs. You do not
need to understand the full details of why this EP approach is much more powerful than a
Markowitz framework if the details are not of interest, but you do need to understand the
details with sufficient depth to grasp the implications explored in this chapter and the rest
of this book, starting with the rest of this section. Risk efficiency is a core gateway concept
for any viable version of EP.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
112 Foundations
When the first BP project to use the SCERT process, the Magnus project, sought board
approval and release of funds to begin construction, SCERT analysis had been undertaken
using extensive decomposition of sources of uncertainty as described earlier, in part to give
the board confidence in the plans and associated cost estimates.
By this time BP had appointed an internal leader for the development of SCERT use
within BP, who I supported. He and all the project staff involved understood that our
primary goal was seeking risk efficiency via contingency planning using SCERT. Unbiased
estimates were usefully seen as by- products of an effective approach to seeking risk efficiency
and avoiding risk inefficiency. The central issue was risk efficiency. Obtaining an unbiased
estimate of the expected cost was just part of the process of seeking risk efficiency.
Demonstrating risk efficiency within the project management team meant demonstrating
that the team could not find a lower expected cost approach which did not involve more risk
or a lower risk approach which did not involve a higher expected cost. To seek risk efficiency,
the project team had to understand what might happen and why, and what could be done
to reduce risk or expected cost or both. This might involve after the fact reactive mitigating
contingency planning responses, which might require prior actions to make these response
options available. It might also require up- front proactive responses, which might initially
be conceived as preventative contingency planning responses for specific problems, with the
later realisation that they actually delivered even more beneficial changes in the base plans
for other reasons. Specific sources of uncertainty might be catered for directly in an exten-
sive manner, but general responses dealing with the cumulative effect of sets of sources of
uncertainty net of specific responses plus unknown unknowns might be absolutely crucial.
At the core of what we were doing was engaging in a search for opportunities to reduce
expected cost and risk at the same time – improving risk efficiency and avoiding risk inef-
ficiency. This can be seen as a core component and crucial driver of opportunity manage-
ment – a central part of the process of seeking opportunity efficiency. We were all interested
in all available opportunities to improve risk efficiency. As a consultant I was especially inter-
ested in opportunities for big improvements which might have been missed if I had not been
employed to demonstrate the value of my role. However, the project manager and all the
members of his team shared this objective and understood what was required.
One activity involved a ‘hook- up’ operation – connecting the pipeline between the shore
and the production platform to the platform. It had a target date in August. In the base
plan, a 1.6- metre barge was specified, equipment which could work in waves up to a nomi-
nal 1.6- m height. Analysis demonstrated that August was an appropriate aspirational target
date and that the use of a 1.6- m barge was appropriate in August. However, this analysis also
demonstrated there was a significant chance that the hook- up would have to be attempted
much later, in November or December for example, because the hook- up operation was
late in the overall project sequence, and there was considerable scope for an accumulation
of delays to preceding activities. Using a 1.6- m barge in winter months would be time-
consuming and might mean that the hook- up could not be completed until the following
spring, with severe opportunity cost implications associated with deferred revenue after sig-
nificant investment in the project assets.
An alternative option was available – a 3- m wave- height- capability barge. A 3- m barge
would cost more than twice as much per working day and per idle (weather) day as the
1.6- m barge, and a 3- m barge would have to be contracted in advance, before BP knew
whether preceding activities were delayed. A revised analysis assuming the use of the more
capable 3- m barge virtually eliminated the risk of going into the next season and the associ-
ated risk of a significant cost overrun, a ‘lost season scenario’ with an expected cost of the
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 113
order of £100 million. But crucial to this example, choosing the 3- m barge also reduced the expected cost of hook- up by about £5 million. Despite the significantly higher cost per unit of time because of the additional capability, the expected cost was lower because a much
shorter duration could be expected and the opportunity costs associated with a lost season
were very high.
The decision diagram used to make this decision is illustrated in Figure 3.6.
The cumulative probability distribution curves for the two barge choice options crossed
above the P50 (50 percentile value). This indicated that the 1.6- m barge was more likely
to be the cheaper option than the 3- m barge. However, the 3- m barge distribution curve
was much steeper because the outcome was less uncertain. And crucially, the 1.6- m barge
distribution had a much longer tail on the top right- hand side because of the relatively low
probability, but high cost, of a lost season. The much flatter 1.6- m barge curve in combina- tion with the significant skew underlying the long tail to the right dragged the expected cost of the 1.6- m barge option to the right of the expected cost for the 3- m barge option. And
the combination of the lower expected cost of the 3- m barge plus the lower risk was what really mattered. The analysis portrayed by Figure 3.6 indicated that the 1.6- m barge had a
better than 50:50 chance of being cheaper, but the expected cost of using the 3- m barge
was less than the expected cost of using the 1.6- m barge by about £5 million, in addition to
the 3- m barge avoiding the risk of a lost season with an additional cost expectation of about
£100 million.
Even if the long right- hand tail of the 1.6-m choice actually involved multiple modes as
a consequence of possible lost seasons, which might have been estimated and portrayed,
this graphic portrayal made it visually clear what was involved. We did not need to work
within the limitations of a mean– variance model. All probability distribution moments were
portrayed by Figure 3.6, not just variance. ‘Clarified dominance’ was involved, a form of
‘stochastic dominance’ interpretation. We were using what might be called a ‘mean plus full
cumulative distribution comparison graph’ approach with a background understanding of
‘stochastic dominance’ and visibility of all percentile values from P0, P1, . . ., P100 instead
of a mean– variance approach.
BP and other later users of variants of this decision diagram portrayal liked the term
‘enlightened caution’ associated with the 3- m barge choice in this example, because despite
cost
0.5
1
cumulative
probability
0
expected cost using the 1.6-m barge
expected cost using the 3.0-m barge
3.0-m barge curve
1.6-m barge curve
Figure 3.6 Decision diagram: one risk efficient choice example.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
114 Foundations
the obvious fact that the 1.6- m barge looked like a better choice most of the time, its
expected cost was higher, and it was also riskier in the sense it might cost a great deal more than its expected cost. The 3- m barge was actually the only risk efficient choice, and choos- ing the 3- m barge involved a subtle kind of clarified stochastic dominance referred to as
enlightened caution.
The term ‘enlightened caution’ is useful for several reasons. One of immediate concern
is emphasising that a reasonably high level of clarity is sometimes needed to identify and
then communicate the implications of a complex situation so that everybody can understand
what is involved.
Based on the discussion of Figure 3.6 the base plan was changed at the project team level.
Much later, when the project manager took his fully developed plans and budget request
to the board for release- of- funds approval for the project, he used Figure 3.6 to persuade
the board that this change was appropriate. In addition, he used Figure 3.6 to persuade
the board that this one change paid for the entire SCERT study many times over – and this change would not have been identified if they had not used the SCERT approach and Figure 3.6 format decision diagrams. The SCERT process being used could be expected to provide a massive return on the analysis investment involved. Indeed, the systematic search for oppor-
tunities to both lower expected cost and reduce risk simultaneously via effective contingency planning should be seen as the central reason for the SCERT approach. Unbiased estimates
were best seen as a useful by- product component part of risk efficiency.
The BP team planning the Magnus project used decision diagrams like Figure 3.6 to make
many risk efficient plan changes, shaping the plans to deliver less expected cost and less risk.
Once it was clear to everybody what was involved, sometimes it wasn’t even necessary to
produce a diagram to make a convincing case. Effective pursuit of risk efficiency using the
SCERT contingency planning process saved money on average, as well as reducing risk in
the sense of costs above what was expected, and that was best seen as its core role. SCERT
analysis was not an overhead cost to give the board comfort, although it also delivered
strategic plans the board could trust, and budget requests the board could understand and
approve with confidence. The overall impact was a step change reduction in both expected
costs and associated risk – an increase in risk efficiency – plus a significant increase in trust
based on much more clarity than the earlier approaches had provided and a number of fur-
ther by- product benefits.
The board approved the plan – successful despite some surprises as noted earlier. The
board also mandated the SCERT process worldwide for all large or sensitive projects because
the board was convinced that the anticipated increases in project risk efficiency would more
than pay for the costs of the process. Significant changes to BP’s approach to planning
projects and associated organisational structure flowed from these decisions, including the
relationships between planning and costing responsibilities and the way BP set up corporate
level project budgets and contingency provisions.
This Magnus example is directly based on facts, with no need for any of the modest revi-
sion of history features of the tales used earlier in this chapter.
Later estimates by a senior MoD colleague (who was responsible for key aspects of the
MoD project approvals processes and commissioned the study involving William) put the
expected return on a capably managed SCERT process investment by MoD projects at
about £100 saved for every £1 spent, implying £100 was an approximate order of magni-
tude expected outcome with an associated a range of the order of perhaps £10 to £1000.
Your organisation might get a higher or lower rate of return on an investment in EP
capability which included always delivering risk efficiency just in terms of the direct benefits
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 115
of avoiding risk inefficiency, with unbiased estimates plus further benefits to be discussed in later sections of this chapter viewed as very important indirect second order benefits which
are not included in the £100 per £1 invested return because they are very difficult to meas-
ure. The exact rate of return obtained will always be uncertain, but this uncertainty is of
secondary importance. What is crucial is seeing EP capability which delivers risk efficiency
and avoids risk inefficiency as a very sound investment, delivering a secure high yield direct
return plus important further benefits, not an overhead cost to provide comfort about unbi-
ased estimates. Unbiased estimates can be viewed as a by- product of risk efficiency provided
at no cost, just the first of a set of further spinoff benefits still to be explored.
Risk efficiency – with enlightened caution as the focus
Figure 3.6 illustrates risk efficiency in clarified dominance terms when the performance
improvement is significant if the best choice is made but the better of the two options avail-
able is a fairly close call. In a complex situation like that captured by Figure 3.6, a high clarity
approach may be needed to clarify what is involved. If the 3- m barge curve had been much
farther to the left, perhaps entirely clear of the 1.6- m barge curve, it would have been obvi-
ous that the 3- m barge was the only appropriate choice. Clarified dominance is the general
feature distinguishing a risk efficient choice from a risk inefficient choice. Enlightened cau-
tion is a special case when it is a close call and a high clarity decision diagram is needed to
clarify what is involved for all relevant players prior to making a decision. In the event, the Magnus project hook- up was actually completed in October in good
weather conditions. It was evident after the event that BP could have got away with using a 1.6- m barge. However, the prior use of Figure 3.6 demonstrated clearly to everyone
involved that the project manager had done a good job as well as making the right barge
choice and that BP had been lucky with the weather. Making the right barge choice had
involved enlightened caution, avoiding an ‘unenlightened gamble’ with the 1.6- m barge
which was not risk efficient.
Enlightened caution in this case involved a willingness to commit to a plan option which
involved highly likely additional expenditure which might not be needed because in expected
value terms (on average) it would be cost- effective to commit to this option, and this choice also involves less risk.
Had problems in the earlier parts of the Magnus project caused the hook- up to take place
in November or December, with seasonably bad weather, the change to a 3- m barge would
have been clearly justified. The wisdom of enlightened caution associated with the choice of
barge would have been verified empirically. However, given that the hook- up actually took
place in October in good weather, it was very important to be able to explain why deploying
the more expensive 3- m barge was still the most appropriate decision. It was important to clarify that both good luck plus good management were involved, and to neutralise any poten- tial negative interpretations of what had happened.
If an effective prototype EP approach using Figure 3.6 had not been followed, and the
decision to use a 3- m barge had been made on informal intuitive grounds by the project
manager, his career might have looked much less promising when it became clear he could
have got away with a 1.6- m barge. That is, the Figure 3.6 decision diagram was an EP tool
for formal planning which helped to make the best choice in advance (before the event). But it also made it very clear later (after the event) that the project manager had made the only risk efficient barge capability choice; plus the project manager and his team had done well to achieve hook- up by October, and BP had been lucky with the weather. If the project
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
116 Foundations
manager had decided to opt for the 3- m barge without the Figure 3.6 analysis output, the
project manager would have been exposed to being accused of wasting money on the more
expensive barge, overlooking completely both his more enlightened choice and his good
management of the project (getting to the hook- up by October). This might have seriously
blighted his career because his enlightenment was misunderstood. A worldly wise project
manager might explicitly recognise this possibility and might opt for the 1.6- m barge in the
absence of a decision diagram with these features. This would involve deliberately making a
bad management decision from a corporate perspective because good luck with the weather
would subsequently be confused with good management and bad luck with the weather
would subsequently just be interpreted as plain bad luck.
If an organisation cannot distinguish between good luck and good management, bad luck
and bad management, individuals will manage risk and opportunity accordingly. Without
EP support to demonstrate the rationale for their decisions, astute managers who are natu-
rally and reasonably cautious with respect to their own careers may regard risk efficient deci-
sions comparable to choosing the 3- m barge in Figure 3.6 as unwise, potentially dangerous
to their careers, seeming to demonstrate a ‘wimpish’, uncalled- for caution whenever they
actually manage the preceding work effectively. Very astute managers may avoid even look-
ing for opportunities to increase risk efficiency in this way to avoid the moral hazard of the
obvious conflict of interests, and perhaps consider looking for a more enlightened employer.
More generally, if good luck and good management or bad luck and bad management
cannot be distinguished, enlightened caution opportunities like that illustrated by Figure 3.6
will not be looked for, and there will be a tendency to pass over this kind of opportunity if
they are stumbled upon.
If everyone involved understands the after the event implications of examples like that
illustrated by Figure 3.6, the organisational culture can change. This change can be driven
by everyone looking for and choosing planning options which increase risk efficiency when
enlightened caution is involved, in addition to looking for increases in risk efficiency which
are obvious even in a simplified low clarity framework employing clarified dominance. This
means that sometimes most people will spend money on some options involving capabilities
that are not subsequently needed, even when there was a high probability they would not
be needed, because on average this will save money. Any organisation which does not always
spend some money on capabilities which may not be needed but on average will be needed
is habitually taking unenlightened risk. This is comparable to gambling in a casino where
you will always lose on average.
Enlightened caution needs to be both facilitated and demonstrated. The documentation
and making common knowledge of instances when the wisdom of enlightened caution was
not empirically verified is of particular importance.
Facilitating risk efficiency to reduce expected cost and risk simultaneously and demon- strating this was being done was soon seen as the primary goal of the SCERT approach by BP and all its other users. In the same spirit an appropriately demonstrated risk efficiency approach embracing a clarified dominance view of enlightened caution is a primary goal for
any EP approach to formal planning.
This section has been about seeking a reasonably deep understanding of enlightened cau-
tion in after the event corporate culture terms as an important addition to understanding
the broader clarified dominance basis of risk efficiency. Risk efficiency achieved in a clarity
efficient manner within the EP framework can deliver the culture change insights of enlight-
ened caution as well as freedom from risk inefficiency and unbiased estimates. They come
as a package – in marketing terms a ‘bundled product’. However, it is not ‘buy one and get
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 117
the whole bunch’; it is ‘invest in one and in addition to a big return on your investment you
will get a valuable package of further benefits’, because the achievement of risk efficiency is
self- financing with a significant return on the cost and effort invested.
We have now looked at a clarity efficiency and risk efficiency as a bundle with multiple
benefits, beginning to explore the implications of opportunity efficiency as an overall gate-
way concept. But we have not yet considered further opportunity efficiency aspects which
are key parts of the full package, all provided in one EP bundled product. Opportunity effi-
ciency can provide much more than has been revealed so far, and some of the other aspects
of opportunity efficiency need attention now.
Opportunity efficiency – with enlightened gambles as the focus
The next step towards an understanding of opportunity efficiency in strategic clarity terms
involves considering risk– reward trade- offs when taking more risk in a risk efficient manner when only one key attribute is involved. The focus of this section is an important additional
by- product of the risk efficiency basis of an EP approach – encouraging and facilitating
‘enlightened gambles’. Enlightened gambles are arguably an even more important compo-
nent of the overall cultural change EP can deliver than ‘enlightened caution’.
‘Enlightened gambles’ are defined as the selection of a potentially high return option
from a set of risk efficient options when the additional risk that comes with the high return
is considered both bearable and worthwhile by all relevant parties.
Figure 3.7 is a fabricated alternative to Figure 3.6, derived from Figure 3.6. It was initially
developed for use in conjunction with Figure 3.6 and a version of the discussion in the last
few sections using the BP examples in a culture change programme referred to by IBM UK
as their Forum 2 programme. IBM ran this two day programme at an IBM training site in
Hampshire in the early 1990s about 40 times and included all senior and middle level IBM
UK managers. The responsibilities of these managers included corporate, operations and
project management.
The opening session on the first day was an introduction by their Chief Executive Officer
(CEO), who explained in overview terms why IBM UK needed to take more risk in a risk
efficient manner to stay in business. For the rest of the morning I provided an overview of
cost
0.5
1
cumulative
probability
0
1.6-m barge curve
3.0-m barge curve
expected cost
using the 1.6-m barge
expected cost using the 3.0-m barge
Figure 3.7 Decision diagram: two risk efficient choices example.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
118 Foundations
‘what needed to be done’ to achieve the CEO’s goals. After lunch, all the participants spent
about an hour working on a bidding case study in groups of about eight people to try out
some of the key conceptual and operational tools which I had introduced. Then presenta-
tions were made by each group of ‘their bid’, and its rationale, to all the other groups. My
follow- up presentation summarised current thinking as displayed by the individual group
presentations and explored how current thinking needed to change. General discussion
then completed the first day. The second day involved those taking part in the Forum 2
event discussing how they might help IBM UK to implement what had been learned from
the first day.
Figure 3.6 was the basis for Figure 3.7, but the cumulative probability distribution for
the 3- m barge in Figure 3.7 was shifted to the right so that the £5 million expected cost
advantage associated with the 3- m barge of Figure 3.6 became a £5 million disadvantage in
Figure 3.7. This was a fabricated example. It was obtained by simply moving the Figure 3.6
curve for the 3- m barge to the right, but if the numbers in the real example had been differ-
ent, this result might have been obtained, as explained to the Forum 2 participants.
The point where the curves cross now suggested that the 1.6- m barge had about an 80%
chance of being cheaper, an increase from about 60% for Figure 3.6. However, the signifi-
cant change that really mattered was the ordering of the expected outcomes was reversed –
the expected cost of the 3- m barge was now about £5 million more than the expected cost of the 1.6- m barge. In terms of the expected value criteria, the 1.6- m barge was now preferable. This raised a conflict between the maximising expected reward criteria and the minimis-
ing risk criteria. There was no longer a single risk efficient choice. The long tail for the
1.6- m barge still implied much more risk, associated with a lost season scenario which was
assumed, for illustrative purposes, to be comparable to a 10% chance of an extra £100 mil-
lion in costs. But the expected cost of the 1.6- m barge option was now £5 million less than
the 3- m barge option.
The first key question was ‘Should this extra risk be taken by BP?’
Assuming these two options were the only risk efficient options available, the choice of
barge was now a ‘risk appetite’ concern – a matter of decision- maker preference because both
options were risk efficient. The 3- m barge option involved less risk but a higher expected
cost, while the 1.6- m barge involved more risk but a lower expected cost.
This raised a second key question, ‘Who should be the decision maker?’ A useful initial
working assumption was the board needed to make the decision based on a ‘corporate risk
appetite’ concept which included consideration of ‘corporate risk taking capability’. This
corporate risk taking capability was an important issue, determined via a corporate view of
risk efficiency which employed an appropriate risk– reward trade- off for the BP portfolio of
all projects and all other corporate operations.
For oil majors like BP involved in projects in the 1970s and 1980s with expected costs in
excess of £1,000 million, potential losses much greater than £100 to £200 million were part
of the territory. To enable them to live with this level of risk, joint ventures were common,
a way of sharing risk. Joint ventures of this kind can be viewed as a form of insurance, but
they involve sharing reward as well as risk. Over ten such projects, or over ten comparable
decisions within fewer than ten projects, taking the 1.6- m barge risk described by Figure 3.7
equates to an expected cost saving of £5 million times ten, or £50 million. Oil majors gen-
erally took the view they could not afford to pass up expected cost savings of this order to
reduce risk which did not need to be reduced. Enlightened gambles were a key part of the
culture. Organisations which do not take enlightened gambles reduce their average profit-
ability and may eventually go out of business as a direct result. A formal UP or linked SP like
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 119
the SCERT approach can facilitate, demonstrate and encourage enlightened gambles as a
basis for a direct increase in expected profit plus an important basis for engineering associ- ated organisational culture changes with linked indirect benefits.
In the context of a choice like that portrayed in Figure 3.7, if the gamble paid off, the
wisdom of the enlightened gamble would be verified empirically. However, the occasional
visible failure of such gambles can be extremely important because they demonstrate that
good decision makers who take risk efficient gambles are sometimes unlucky. If no quanti-
fied uncertainty analysis were undertaken to demonstrate an expected cost savings associated
with an enlightened gamble like that of Figure 3.7, this message would be lost, whatever the
outcome. In the absence of a demonstrated expected cost benefit plus an organisational cul-
ture which promotes enlightened gambles, astute managers might not take such gambles,
and very astute managers may not even look for them. This was another aspect of the need
for an organisational capability to distinguish between good luck and good management,
bad luck and bad management.
Further key questions which needed addressing included: when should the board want
to make this kind of decision, when should the board leave this kind of decision to project
managers, and if the board leaves this kind of decision to project managers, how should
the board ensure that the project manager’s risk appetite is aligned with the board’s prefer-
ences? For illustrative purposes in the IBM context it was useful to continue to assume a
10% chance of an extra £100 million in costs if the enlightened gamble was lost, and we can
carry on in the same way.
If an additional £100 million was going to mean the project’s budget was breached, and
if the board would prefer the £100 million × 0.1 = £10 million provision for the enlightened
gamble was not in the budget to be spent on other things if the 1.6- m barge gamble works,
the enlightened thing for the board to do was make the decision to use the 1.6- m barge and
take the risk at board level – making a £10 million board level provision against a possible
£100 million call on BP self- insurance. To give the risk to the project manager along with
a £10 million provision in this context would involve a ‘lose– lose’ situation for the board,
because if the extra £100 million was needed the budget would be breached and the board
would have to be involved in dealing with it, but if the extra £100 million was not needed the
project manager might spend the £10 million provision on something else which might not be
really needed. If the board made the decision and held the contingency, the project manager
would be responsible for managing the risk but not for taking it financially within his or her
budget. All very high impact and low probability risks warrant this treatment, with the board
deciding when the impact was too high for the gamble to be enlightened given the expected
cost and risk of an alternative ‘enlightened prudence’ option. The level of contingency in the
budget should be directly geared to the level of impact triggering board responsibility.
If an additional £100 million would not breach the project’s budget because the board
trusts the project manager with provisions plus a contingency big enough to handle prob-
lems of this order on a swings and roundabouts basis, the enlightened thing for the board
to do was leave the project manager to make the choice.
In either case, the board has to trust the project manager and vice versa, motivating all
project managers to perform by judging how well they pursue corporate best interests.
Whether or not they are lucky should not be confused with whether or not they are good
managers. At the time these interconnected capability- culture issues centred on trust and
other enlightened planning issues did not seem to be a problem for BP – the corporate ethos
meant that project manager relationships with the board were exemplary as far as I was able
to judge – amongst the very best I have ever encountered.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
120 Foundations
The key culture- change message for IBM UK staff was to recognise the corporate advan-
tages of everybody in IBM UK collectively looking for enlightened gambles, taking more
risk whenever this was appropriate within a well- understood risk efficiency framework, and
recognising that they were currently not taking many important potential enlightened gam-
bles because they were not seeing them as opportunities. This was linked to also using
enlightened caution and more obvious clarified dominance, but at the level of decision
making being addressed, a crucial additional focus of IBM UK risk management should be
taking more risk, fully understanding what was involved, not less risk. It was important for IBM staff to understand that BP, had they been faced with Figure 3.7
instead of Figure 3.6, would probably have opted for the enlightened gamble choice, and
at the IBM decision- making level being addressed, all staff involved needed to emulate this
behaviour to increase expected return for IBM, in addition to making risk efficient deci-
sions comparable to that illustrated by Figure 3.6. But IBM staff did not necessarily need
high clarity decision diagrams like Figures 3.6 and 3.7 to identify enlightened caution or
enlightened gambles – they could often get by with simple low clarity decision diagrams,
illustrated later in Chapters 4 and 6. The tale of Chapter 6 is based on this IBM culture-
change programme and demonstrates in more detail why opportunity associated with taking
more risk in a risk efficient manner can be crucial. Chapter 6 also demonstrates how all the
estimation- efficiency spectrum ideas demonstrated in a project management context in this
chapter can be applied in an operations management context.
At the level of decision making being addressed, in all operations, corporate and project
management contexts, IBM needed to achieve risk efficiency and take enlightened gambles that were not currently seen as opportunities. Often this was just a matter of maximis-
ing expected return – viewing the ‘risk’ that currently worried people inappropriately as
just ‘noise’ which should be embraced via a culture change which was based on a corpo-
rate understanding of directly relevant corporate risk appetite concerns plus the difference
between good luck and good management, bad luck and bad management. More sophisti-
cated risk appetite and related corporate capability issues will need further attention.
The key message of this section is what is involved when looking for risk efficiency plus taking more risk when appropriate is an important aspect of ‘opportunity management’, a kind of ‘treasure hunt’, where the treasure is opportunity efficient choices which deliver both
risk efficiency and the additional expected return of enlightened gambles whenever gambles are appropriate. As well as the direct impact on performance, the wider implications of the
culture change involved can be crucial, including the impact of everyone knowing they are
working for an organisation which understands the difference between good luck and good
management, bad luck and bad management, will tolerate bad luck but not bad manage-
ment, and encourages appropriate enlightened gambles as well as enlightened caution.
Opportunity efficiency – with enlightened prudence as the focus
Taking enlightened gambles needs to be coupled to and constrained by encouraging and
facilitating enlightened prudence.
‘Enlightened prudence’ is defined as the selection of a risk efficient option which reduces
return relative to a higher risk and return option because the additional risk is judged not
bearable or not worthwhile for one of the parties, and a prudent decision- making process
suggests not taking the additional risk.
Enlightened prudence means avoiding taking a gamble which is beyond the risk appetite of
any of the parties with a legitimate interest in the outcomes. Enlightened prudence was not a
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 121
serious concern for IBM at the level of decision making being addressed, because IBM was big enough to self- insure the risks involved at this level. But whenever enlightened prudence does
need attention, it is a complex concern which requires capable treatment in an appropriately
general (unrestricted) conceptual framework using viable working assumptions.
There are two key risk appetite reasons why enlightened prudence may need to constrain
gambles with a view to avoiding unenlightened gambles. Both need to be clearly understood
by all relevant people.
Most important, the additional risk may not be bearable by one or more of the parties
involved. A consensus involving all relevant parties may be crucial to avoid this kind of unenlightened gamble, and reaching a consensus may involve negotiating a change in the
risk– reward distributions involved. Usually enlightened management decision makers will
be very aware of the need to consider all relevant concerns, but some decision makers may
ignore the legitimate interests of other parties through ignorance or self- seeking behaviour
if governance is inadequate.
Also very important, but secondary, the additional risk may not be worthwhile in terms
of the operational cost implications of disrupted financial plans – suddenly having to find an
extra £100 million may involve abnormal costs even for an organisation like BP, and for very
small expected cost savings even a risk which is bearable may not be worthwhile.
All decision makers need to be clear about what is involved when either of these issues
arises, one or the other is often a concern, neither lends itself to simple rules in all poten-
tially relevant circumstances, and simplistic treatment can be counterproductive. Govern-
ance needs to address all these concerns.
For example, if very serious potential problems associated with the 1.6- m barge might cost
BP an extra £100 billion, the risk involved might be judged unbearable and certainly not
worthwhile for an expected benefit of just £5 million. Assuming the £100 billion possibil-
ity is understood, deciding to avoid this level of risk might be straightforward and obvious,
needing very little effort. But even if the additional maximum cost was just £100 million and
the £5 million expected savings might make it worth the enlightened gamble, a £0.05 mil-
lion expected savings would probably not be enough to justify the risk of an extra £100 mil-
lion, and this might also be the case for an expected saving of £0.5 million. For a significant
risk when the gamble is within the bearable range, exactly where the flip point expected cost saving is may be unclear, but it is not £0 million, and it clearly ought to vary with the
maximum amount at risk.
In effect, whenever the risk is significant the flip from enlightened gamble to enlightened prudence does not occur as the difference between riskier and safer choice passes a zero
expected cost difference point, £100 million being assumed to be bearable but significant
in this example context.
If any additional significant risk is bearable and it is avoidable, the trade- off between addi- tional expected reward and additional risk needs careful attention – risk appetite involves
trade- offs which are a function of both the risk and the reward involved. Clarity is crucial, and a high clarity decision diagram like Figure 3.7 can be an invaluable decision- making tool.
If any additional significant risk is not bearable and it is avoidable, it should not be taken. When this situation arises, clarity is also crucial. High clarity decision diagrams like Figure 3.7
can be an invaluable decision- making tool whenever a close call is involved.
Whenever the risk is not significant, there is obviously no need to consider risk appetite concerns – simply maximising expected reward is the preferred sound and simple clarity
efficient strategy. Enlightened gambles are the default option, with no need to test for the
possible need for enlightened prudence.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
122 Foundations
Insurance and other forms of partial or full risk transfer may complicate matters by intro-
ducing further new options, but the complications this raises will be deferred until later,
along with additional clarity to deal with complex enlightened prudence concerns involving
multiple attributes and multiple parties whose concerns may be very different.
Chapters 8 and 9 address contexts in which enlightened prudence requires formal plan-
ning model consideration with strategic corporate management implications. Chapters 5
and 6 emphasise a focus on maximising expected reward for most purposes. All Part 2 chap-
ters employ the same EP toolset, building on both the toolset and mindset issues addressed
in this chapter.
Interim consolidation of the opportunity efficiency concept
Decision diagrams like Figures 3.6 and 3.7 are key tools for all EP contexts, to portray
option choices which are relevant to operations planning and corporate planning as well
as project planning. So are simpler linear versions, which are illustrated later. Decision dia-
grams are the basis for opportunity efficient choices to achieve clarified dominance including
enlightened caution, plus making appropriate trade- offs when clarified trade- offs involve
distinguishing between enlightened prudence and enlightened gambles.
Figure 3.8 is a useful basis for consolidating the ideas explored in the last few sections, for
linking this consolidation to further discussion of risk appetite issues, and for building on
during the rest of this book as further complications are addressed.
The ‘efficient frontier’ view of risk efficiency provided by Figure 3.8 is a conceptual tool
which you need to understand as clearly as possible as part of the foundations of strategic
clarity. It can only be used directly as an operational tool if very strong working assump-
tions are employed, considered shortly, but it is a very general conceptual tool which needs
risk in terms of reward which fails to meet expectations
expected
reward in terms of profit or a more general composite of positive objectives
a–c efficient boundary target, with
a maximum expected reward approach
c minimum tolerable expected reward
bi intermediate expected reward approaches
ei inefficient approaches
c–d inappropriate reward levels
key:
a
b3
c
d
competent management area
(usefully interpreted as an opportunity management area)
e1
non-feasible area
b2
b1
e2 e3
incompetent management area
Figure 3.8 Expected reward and associated risk in an efficient frontier portrayal.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 123
to be understood in conjunction with Figure 3.5 to help you to interpret what the last few
sections have been about as a basis for understanding what the rest of this book is about.
Figure 3.8 is a variant of the ‘efficient frontier diagram’ which Portfolio Selection: Efficient Diversification of Investments (Markowitz, 1959) uses to discuss risk efficiency employing a mean– variance framework. Economists routinely use comparable efficient frontier diagrams
to discuss efficient choices and associated trade- offs between any two criteria. Figure 3.5 was
based on Figure 3.8, using a common format to simplify joint understanding. Figure 3.8 has
been central to my understanding of risk and uncertainty since first encountering Markowitz
and embedding his concepts in my PhD research around 1970.
Start by considering Figure 3.8 in the context of a portfolio of securities in mean– variance
terms. Markowitz could use Figure 3.8 as an operational model as well as a conceptual
framework, because he assumed variance was a suitable surrogate measure for risk.
The securities investment portfolio plans being considered by Markowitz in a mean–
variance framework were defined by the amounts of money invested in each of the possible
securities, the Xi values, with i = 1, . . . , n for n securities. These Xi variable values had to
be associated with estimates of a complete set of Ei parameter values which measured the
expected return per $ or £ or other relevant currency unit invested in security i. A complete
set of Vij 2 parameter values for j = 1, . . . , n as well as i = 1, . . . , n, which measured all the
associated variances and co- variances also required estimation, with V Vij i 2 2
= when i = j.
The y- axis was expected reward, expected return on the portfolio as a whole measured by
E, equal by definition to the sum of all the Ei Xi over i = 1, . . . , n.
The x- axis was risk associated with reward that fails to meet expectations as measured by
V2, the variance of the probability distribution defining E, a quadratic function of Vij 2 Xi Xj
terms summed for both i and j over 1, . . . , n.
Point ‘a’ was defined when the objective function E – k V2 was maximised for k = 0 (risk
was ignored – given zero weight). Graphically a horizontal line (slope zero) was pushed up
as far as it would go to still provide a feasible solution to locate point ‘a’, maximising E.
As the value of k increased the same optimisation meant that b1 was reached, then b2, then
b3, and then c. Graphically the line defined by E – k V 2 with slope k was pushed as far as it
would go upwards and to the left to still provide a feasible solution at the point of tangency
with the boundary. The c– d region involved too little reward to be of interest. A quadratic
programming approach to this model allowed the identification of a suitable selection of
portfolios on the boundary from point a to point c.
The decision maker could then select the preferred portfolio, indirectly choosing a value
for k. The value of k at the chosen optimal solution point defined the trade- off between
expected reward as measured by E and risk as measured by V2 at the optimal solution – a
special case of what is now commonly referred to as risk appetite, defined in terms of what
is now commonly known as a shadow price. In this case the shadow price could be defined
as the increase in expected reward which could be obtained if the constraint on risk were
relaxed by one unit at the currently chosen optimal solution.
The relationships between the Xi values defining the plans and the two relevant criteria
(E and V2) are the goals– plans relationships. These criteria are linear in the case of E and
quadratic in the case of V2. The model provides crucial understanding of the role of depend-
ence between different sources of uncertainty associated with plans as a central part of the
goals– plans relationships framework.
Point ‘a’ involves investing all available money in the single highest expected value outcome
security available – a classic ‘putting all your eggs in one basket’ plan. Moving down the effi-
cient frontier is about efficient diversification, changing the plan by adding more securities to
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
124 Foundations
the portfolio in an optimal manner. This involves choosing to add those securities which per
unit of expected return foregone have a maximum level of independence or, ideally, a negative
dependence ‘hedging’ effect, to provide a maximum reduction in risk for any given reduction
in expected return. A minimum sacrifice of expected return for a maximum decrease in risk is
what is being sought, over the whole of the boundary from point a to point c.
Now consider Figure 3.8 in the context of the Magnus project BP example. When BP
adopted a 3-m barge as discussed in the context of Figure 3.6, BP moved from a point like e1
to a point like b2, involving less risk and more expected reward (less expected cost). BP was
in what we judged (from a prototype EP perspective) as a ‘competent management area’,
usefully treated as an opportunity management area, looking for better ways to shape their
project plans for execution and delivery at a strategic level. The goals– plans relationships
structure associated with a causal understanding of uncertainty was more complex than the
security portfolio example in some ways, simpler in others. It was certainly different. But it
was manageable by skilled planners because activity- based plans and contingency plans with
associated duration uncertainty and linked resource choices drove overall cost uncertainty
in a way which could be defined and understood by those involved. The BP board members
could not be sure that the BP project planners had reached the efficient boundary for this
particular decision because only two points identified by their planners had been explained
using Figure 3.6. However, the board members could be very sure their planners knew what to look for, their planners had looked at other options, and they had moved their plans in
the right direction as far as they could. All the other proactive and reactive contingency plan-
ning done by the BP planning staff was part of the shaping of base and contingency plans to
move closer to the risk efficient boundary.
A choice like that discussed with IBM UK staff as portrayed by Figure 3.7 is like a choice
between b2 and b1. The additional risk levels involved in the higher risk options were bear-
able, and the complication of enlightened prudence was not seen as a concern worth empha-
sising at the level of decision making being addressed. To ‘keep it simple’ the IBM culture
change programme was about everybody in the organisation understanding risk efficiency
with sufficient strategic clarity to be prepared to use this enhanced understanding to both
seek risk efficiency and take more risk – not less risk. Most of the time this just meant using a highest expected profit approach, not even bothering to consider risk which could
be accommodated, making the working assumption they could focus on achieving point
‘a’ in Figure 3.8 terms. However, the underlying concern was using clarified dominance
to achieve enlightened caution plus enlightened trade- offs, including enlightened gambles
most of the time and enlightened prudence when relevant. The key was moving all organi-
sational decision making into the opportunity management area, whether or not formal
analysis was used to make particular decisions. Understanding the relationship between
uncertainty, opportunity and risk at a corporate level was the key to a significant corporate
culture change.
More generally, whether or not a choice like b1 or a should be taken instead of a choice like
b2 or b3 ought to depend on whether or not risk at the level being considered is just noise
or significant. If it is significant then the issue is whether it is both bearable (the absolute
level of risk and its full implications) and worthwhile (the slope of the risk efficient bound-
ary defining the trade- offs between risk and reward). Risk appetite in EP terms is about
prudence in terms of both the risk– reward trade- offs and the absolute levels of risk involved.
If the risk level involved is not significant, a focus on point ‘a’ choices without wasting any
time on risk may be sensible and clarity efficient. If the risk level involved is significant, then
proportionately more effort will be worthwhile to ensure appropriate choices are made.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 125
Risk appetite in the context of the many organisations which do not understand what this
chapter is about may be very different. Some organisations use risk appetite concepts which
from an EP perspective are a serious corporate- capability liability.
Chapter 5 elaborates on the use of expected value decision making when risk associated
with quantified uncertainty is assumed to be tolerable and not worth avoiding or even tak-
ing the trouble to think about in some cases. It does this assuming that bias in the usual
sense is not an issue provided there is a full understanding of underlying working assump-
tions, linked to several approaches to addressing risk associated with unquantified uncer-
tainty that matters. Chapter 6 explores the idea of taking more risk knowing what you are
doing in the spirit of the IBM UK culture change programme, using a tale which is, in part,
based on the IBM culture change programme and associated follow- on consultancy. Chap-
ter 7 explores the mixture of clarified dominance and trade- off issues usually encountered in
projects in more detail, using a water and sewage utility example. Chapter 8 develops some
further risk– reward trade- off concerns in a corporate planning context using an electricity
utility example. Chapter 9 considers a railway safety example when both enlightened cau-
tion and enlightened prudence involving multiple attribute issues becomes a key concern,
raising complexities resolved in a framework which generalises the interpretation of what
Figures 3.5 through 3.8 involve.
Of immediate interest, from an EP perspective some organisations are clearly in the
incompetent management area, at points like e3, because they do not even understand the
concepts of risk efficiency or opportunity efficiency. If people do not know what they should
be looking for, they are very unlikely to find it or recognise its value when they stumble on
it by accident.
People who think risk management is primarily about managing risks in the uncertain
event sense generally take this view because they do not understand risk efficiency in the
sense of Figures 3.5 through 3.8. This means that from an EP perspective people employing
a common practice approach based on an uncertain event interpretation of risks and associ-
ated overall risk are generally in the incompetent management area for at least three reasons
which have been addressed in this section and previous sections in this chapter:
1 they are not providing unbiased estimates of outcomes in a manner which facilitates
delivering projects on time and within cost and making other comparable operations
and corporate management decisions which are based on unbiased parameter estimates,
2 they are not delivering opportunity management by finding risk efficient choices,
3 they are not ensuring appropriate risk– reward trade- offs or judging prudent choices in
terms of absolute risk levels on a sound basis, because they are using an inappropriately
restricted risk appetite concept.
They may also be in the incompetent management region for a significant number of further
reasons, one addressed in the next section, others addressed in later chapters.
The importance of seeking both specific and general responses
When decision trees were embedded in PERT project planning models, and this Generalised
PERT model was further extended to incorporate Markov processes in the GERT approach,
‘general’ responses became a standard generic feature of these planning models. When I first
used the GERT approach for a practical application working with Acres in 1975 on an
Arctic pipeline project, a key lesson was the power of general responses to an accumulation
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
126 Foundations
of problems. If part way through any activity an accumulation of problems meant that a
project was falling behind schedule, a general response can sometimes deal with the effect
of all these problems, including unknown unknowns as well as an accumulation of known
unknowns and known knowns, net of responses both specific and general in earlier activities.
Sometimes proactive responses could provide this generality in advance. This involves build- ing in resilience and flexibility.
A linked but separate key lesson was the need for prior contingency planning addressing
specific responses to specific concerns in addition to general responses, a key new feature of the SCERT approach developed for BP during the late 1970s. Specific sources of uncer-
tainty and responses were not part of GERT.
The focus of much of the effort when SCERT was first used by BP was specific sources
of uncertainty and responses. For example, if a wet buckle occurs, an immediate repair was
required, unless the pipeline laid so far was abandoned, and the lay- barge went back to the
beginning to start again (because not much pipe has been laid and it was cheaper to start
again). This focus on specific responses was in part because these specific responses were
very important, in part because specific responses were a new feature of SCERT relative to
GERT. However, we soon realised that identifying key general responses remained a crucial
concern. Furthermore, we also learned that thinking about specific responses was sometimes
the catalyst for identifying very important general responses which needed to be recognised
and fully exploited as general responses. For example, when exploring how to deal with the
impact of more than one wet buckle when laying an offshore pipeline, one of the planning
engineers suggested a second lay- barge working from the other end with a submarine con-
nection when the two ends of the pipeline met. This suggestion was immediately welcomed
by everyone involved. It was soon realised that the cost of taking an option on a second
barge, probably necessary if this response was potentially important, would be too high to
justify if more than one buckle was the only relevant issue. However, it might be worth-
while paying for an option on a second barge given this response could also deal with any
combination of problems leading to pipe- laying running late, including coated pipe being
delivered late, the first barge arriving late or proving unreliable, the weather being unusually
bad, and so on. That is, a second lay- barge was initially identified as a ‘specific response’ to
a ‘specific source of risk’, but we quickly realised that it was, in fact, a particularly power-
ful general response which we would not have thought about if it had not been initially
identified in a systematic search for specific responses. This prompted the development of
a systematic search for general responses within all the sets of specific responses, as a design
feature of the SCERT process methodology. The desirability of a contractual option to use
two or three shift operations of the pipe- coating facility if needed, negotiated before the
contract was agreed, was identified as a direct consequence, one example of the generalised
search for general responses using specific responses in another context as a starting point.
When National Power introduced a simplified variant of the BP SCERT approach using a
combined cycle gas turbine electricity generation power station as a demonstration project,
we spent a lot of time looking for general responses to deal with potential delays, including
systematically reviewing specific responses for ideas which might have wider relevance. We
could not find any initially, and two sources of uncertainty were particularly worrying to the
planning team.
One of these sources of uncertainty was the potential cost of the power station being
ready to use gas later than anticipated. The owners of the gas supply had agreed to build
a pipeline from their gas field to the power station with first gas availability at a contracted
date to be specified by National Power, with heavy penalties for the gas supplier if it was late
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 127
completing the pipeline. However, National Power would have to pay for the gas from that
date, whether it used it or not. This was a ‘take- or- pay’ contract, specified by the gas supplier
to avoid getting no revenue from a completed pipeline.
The second of these sources of uncertainty was a series of permissions required from
regulators for environmental concerns like warm cooling water being put into a river – issues
involving more uncertainty than might usually be the case because this was the first power
station of this kind to be built by a newly privatised UK electricity utility, and nobody had a
clear idea about the implications of privatisation for regulatory issues. Although some sus-
pected tougher regulation was likely, others had a converse view.
The planning team eventually identified a general response strategy which proved central
to the whole plan. Initially the key general response idea was ‘start the whole project several
months early, leaving the first gas contract date in August as initially planned, to build a con-
tingency cushion’. Key further ideas building on this starting point included ‘contract for a
separate gas supply from the national grid gas supplier so that if the power station is available
early the opportunity can be used to do all the testing early’. Gas provided by the national
grid supplier could not be used to run the power station at full power for a sustained period,
but it could be used for test purposes. It was soon realised that the gas characteristics were
different, which meant that a more flexible specification of the gas to be burnt needed to be
part of the turbine contract, initially viewed as a problem and then viewed as another poten-
tial opportunity. A more flexible gas specification for the turbines would make the power
station less dependent on the main gas supplier, potentially a significant asset later.
The National Power engineering function soon became absolutely delighted with the
combined effect of these changes and the very high probability of being able to go to full
power if needed as soon as the main gas supply became available – it sold them on making
the SCERT model approach the new standard National Power practice.
The National Power finance function then became involved. It recommended that all the
general response ideas developed thus far should be retained, but the whole project, includ- ing the first gas availability date, should now be delayed so that the first full power capability could be exploited in November to give National Power a better cash flow for the project.
The National Power board was delighted by this joint engineering and finance functions
collaboration, a very high- profile example of resolving potential risk associated with sources
of uncertainty and finding a whole series of opportunities by unravelling the key complexi-
ties that really mattered. Its basis was a general response quite different and unrelated to
any responses identified earlier. It was identified by a search for general responses which was
creative as well as systematic, exploiting collaboration across as well as within functions like
engineering and finance.
The analysis was synergistic in the broad sense that it went well beyond the bounds of
the sources of uncertainty originally identified as a problem, to deal with much more funda-
mental concerns. The opportunity that really mattered was a very complex set of relation-
ships which had been disentangled by a creative team of planners working across the usual
organisational silo boundaries.
The general EP idea that these BP and National Power examples illustrate in an activity
planning context is that responses to a single specific source of uncertainty in one activity
can be crucial, but so can general responses within an activity to all sources of uncertainty
to date, and so can overall project level responses, with crucial prior proactive meas-
ures involved as well as post- event reactive contingency responses which may need prior
actions. At its simplest, what safety and reliability engineers called fault trees and event
trees deal with specific sources. At an intermediate level, decision trees involving both
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
128 Foundations
specific and general responses are embedded in network diagrams or Markov processes.
At a maximum level of complexity, underlying decision trees dealing with what some deci-
sion analysis experts refer to as a ‘bushy mess’ may need carefully picking apart because
the multiple stage tangled mess matters, and sometimes it can be tamed if those involved
know what to look for and how to deal with it, using a well- designed systematic and crea-
tive interactive process.
The broad ideas go beyond activity structures, as explored later. While contingency plan-
ning in a project planning context is a convenient illustrative framework for building on an
unbiased approach to estimation starting point for this chapter, any option choice genera- tion and selection context gives rise to the same issues, in whatever goals– plans relationship
framework is appropriate to the context, whether or not a formal approach to planning
which addresses the associated opportunities is used.
More interim consolidation of the opportunity efficiency concept
After the risk efficient frontier portrayed by Figure 3.8 was introduced and basic risk
efficiency issues consolidated, this chapter addressed opportunity efficiency in terms of risk–
reward trade- offs involving a single attribute like cost or profit. This built on earlier con-
solidation of clarity efficiency in terms of Figure 3.5. Before ending this chapter, we need to
consolidate the meaning of opportunity efficiency within the single attribute limitations of
this chapter’s focus by briefly but formally linking opportunity efficiency to underlying risk
efficiency and clarity efficiency. Later chapters will continue to develop the nature of oppor-
tunity efficiency in broader terms.
Figures 3.5 and 3.8 both have a competent management area just below the efficient
frontiers which define clarity and risk efficient approaches, also labelled as an opportunity
management area. Further examination of what this implies in terms of a synthesis of the
risk, clarity and opportunity efficiency trio is the concern of this section.
Risk efficiency in Markowitz’s mean– variance terms is a useful starting point. Given only
one measurable attribute is of interest (profit, revenue, or cost), a specific known mean value
for all components which have a known additive relationship, and an assumption that all
variability about these mean values for all components is well behaved and Normally distrib-
uted, then variance is a suitable surrogate measure for risk.
If the working assumption that a single measurable attribute is maintained, but all the
other working assumptions just noted are abandoned, then decision diagrams using the
format of Figures 3.6 and 3.7 can cope provided we assume that all the available options can
be considered two at a time in sequence. We can generalise this to cope with up to about
half a dozen options on a single diagram variant of Figures 3.6 and 3.7, but two at a time
in sequence is often easier. We can also define the ‘options’ as proportions, allowing us to
address optimal portfolios of securities or other forms of investment – for example in Chap-
ter 8 alternative sources of electric power (like nuclear, thermal of various kinds, hydro, and
other renewables) for electricity utility planning are addressed in this way. What we cannot
do is draw upon mathematical programming techniques, like the quadratic programming
tools developed for Markowitz mean– variance models, to make a variant of Figure 3.8 with
risk measured by variance an operational tool.
A generalised variant of the moment- based approach Markowitz adopted could add meas-
ures of skew and kurtosis (peakedness) to accommodate asymmetric distributions, but this
would not cope with the complex badly behaved dependence which a HAT approach can
deal with. Nor would it avoid eliminating a quadratic programming approach. An approach
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 129
based on Figures 3.6 and 3.7 can accommodate any dependence structures or distribution
shapes that the very general HAT based portrayal of uncertainty might lead to.
There is a price for this generality, in addition to not being able to use mathematical pro-
gramming. Within the contexts they are concerned about, the decision makers have to be
capable of making suitable judgements about enlightened gambles, enlightened prudence,
and enlightened caution in the framework of high clarity decision diagrams like Figures 3.6
and 3.7 or comparable variants, visualising the implications of the whole of the relevant
probability distributions as well as identified means (expected outcomes), and they have to
consider options in pairs or small groups, understanding the working assumptions implied
by the grouping structures chosen. Someone in the team involved has to have the skills and
knowledge to manage the process this involves. From the perspective of an EP framework
this is generally a price worth paying. It requires an investment in corporate capability, but
that investment will provide a significant return.
One aspect of the opportunity efficiency this involves is appropriate risk– reward trade- offs
for all relevant attributes. Assuming only one attribute, in terms of Figure 3.8 this means
choosing appropriately from among points like a, b1, b2, and b3, as addressed earlier.
A second aspect of the opportunity efficiency involved is to need to achieve risk efficiency
in the first place – getting onto or near the risk efficient boundary. BP planners understood
this was the primary goal of the SCERT approach to contingency planning – a search for
decision options which, if identified and taken, reduced expected cost and risk simultane-
ously. Indeed, for the BP planners I worked with in the 1970s and 1980s, the treasure hunt
for opportunities to reduce expected cost and risk simultaneously was the central purpose of
the whole project planning exercise – once they understood what the rich contingency plan-
ning approach SCERT provided was about. That is why the competent management area of
Figure 3.8 which they were already in was referred to as the opportunity management area.
Figure 3.8 was used to explain SCERT to all planning staff, and it went to the main board
so that all levels of management had the same understanding of what they were collectively
doing. All those involved in the National Power use of simplified SCERT variants to plan
new electricity power stations also understood this, as did those involved in the IBM UK
culture change programme, and most of the other clients I have worked with since. Indeed,
the hope has always been that all those reading How to Manage Project Opportunity and Risk (Chapman and Ward, 2011), and all earlier publications developing these ideas since the late
1970s, understood these concepts too. However, some people have put these ideas into a
too difficult box or contested them for other reasons, some addressed later.
A third aspect of the opportunity efficiency involved is the need to get on or near the clar-
ity efficient boundary with a suitable clarity and cost of clarity trade- off while pursuing risk
efficiency and appropriate risk– reward trade- offs plus absolute levels of risk for all relevant
attributes. The approach adopted has to reflect the particulars of the context to achieve
this clarity efficiency concern. That is part of the reason why the clarity efficient boundary
portrayal of Figure 3.5 also has a competent management (opportunity management) area.
This was not a problem for Peter and his colleagues, because it was clear that the SCERT
approach was delivering a significant return on the analysis investment, in addition to deliv-
ering unbiased estimates the board could trust. Peter, his project manager colleagues, all
their project management teams, and the BP board, quickly developed an understanding
of when more detailed decomposition of sources of uncertainty might or might not prove
worthwhile, and they were all convinced that a better overall framework for thinking about
all these concerns was not available. They were buying into a ‘bundled product’, but they
could vary the effort and cost associated with buying more or less clarity of particular kinds
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
130 Foundations
in an efficient manner, achieving clarity efficiency in a quite sophisticated way without actu-
ally calling it that at the time.
William and Peter both had an initial interest in estimates of expected performance param-
eter estimates and associated variability estimates which were free of bias, so they could
deliver projects within time and cost commitments. What they realised, and all those who
followed their example realised, is they could use a SCERT approach to achieve the unbi-
ased estimates they wanted as a spinoff from a search for opportunities to reduce expected
cost and cost risk simultaneously. And this process as a whole would more than pay for
itself – providing a big return on the investment, perhaps £100 in project cost savings for
every £1 spent on clarity efficient analysis. Risk efficiency achieved in a clarity efficient man-
ner provided a big return on the effort invested, as well as reliable estimates. Furthermore,
the search for risk efficiency coupled to its clarity efficient pursuit was at the centre of what
opportunity management and opportunity efficiency were about, understood at that time
in slightly different terms.
The minimum to high clarity estimation- efficiency spectrum examples explored in this
chapter are centred on a project context when a single attribute was the focus because mul-
tiple attributes like duration and cost could be integrated via goals– plans relationships which
were understood. The early emphasis with William was unbiased estimation, understanding
the distinction between what was quantified in probabilistic terms and associated condi-
tions, plus the crucial differences between aspirational targets, commitments, and expected
outcomes. The value of understanding associated scenarios in terms of underlying sources of
uncertainty for contingency planning purposes was introduced but not modelled formally.
The discussion with Peter developed the role of formal contingency planning to identify the
options whose existence defined the need for risk efficiency – because if we have choices in
terms of the plans we are making, some choices are better than others, and the risk efficiency
aspects matter. The discussions with William and Peter also begin to explore the complexity
associated with achieving clarity efficiency – because we have choices when we choose how
to do the planning. Some choices about how we do the planning are better than others –
so clarity efficiency matters. Furthermore, the discussion with Peter begins to explore the
complexity associated with achieving clarity and opportunity efficiency when duration and
cost are interdependent and their relationship needs to be understood.
One key message to take away from this chapter is a minimum clarity approach to esti-
mation involves a relatively simple but subtle set of aspirations associated with unbiased
expected value estimates, plus a clear understanding of the associated range, also estimated
without bias and interpreted in relation to the ABCs of targets when relevant.
Building on this, adding clarity may involve more refined estimates, but the key is further
clarity attributes or dimensions driven by contingency planning which increases expected
reward, including taking more risk which is bearable and worthwhile but avoiding any risk
which is not bearable, not worthwhile, or not necessary because risk inefficiency is involved.
These ideas are associated with clarity efficiency, risk efficiency and opportunity efficiency,
the last a composite concept embedding the first two.
One manifestation of corporate understanding of these concepts is a corporate under-
standing of the role of clarified dominance, enlightened caution, enlightened gambles,
enlightened prudence, and the foundations of what corporate risk appetite ought to address.
Another is a corporate understanding of the difference between good luck and good man-
agement, bad luck and bad management, and the foundations of what EP corporate culture
changes might deliver. Yet another is a corporate understanding of the need to see choosing
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Low to high clarity approaches 131
a suitable level of clarity on the estimation- efficiency spectrum for any given component of
an analysis based on opportunity efficiency.
The benefits of full corporate understanding of these issues include culture changes which
will affect informal planning and may change deep cultural issues like ethics as well as formal
planning, with widespread implications.
However, multiple attributes which are not reducible to a single attribute like cost have
not yet been addressed, and clarity itself is clearly a multiple dimension concept which is not
directly measurable in any dimension. This chapter’s initial focus on estimation, then mov-
ing into efficiency (clarity, risk and opportunity), maintaining a focus on just one directly
relevant attribute at any given time, meant that we did not have to confront a potential
conflict between more clarity in terms of contingency planning versus less clarity about
unbiased estimation or priorities which might change over time. For example, at the concept
creation stage of a project’s lifecycle some contingency planning is not relevant until we are committed to the concept strategy, but an unbiased estimate of the pay- off from the strat-
egy is crucial. Furthermore, this chapter has not confronted plans which themselves involve
multiple attributes which matter when there is no obvious well- founded way to consider the
implications.
EP, as explored in the rest of this book, is about further generalising these ideas. Chap-
ter 4 will start by confronting some of the central challenges which were ignored or navi-
gated around in this chapter, to complete the foundation level conceptual toolset needed
for Part 2.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
To keep the discussion as simple as possible, a number of challenging complexities were
avoided in Chapters 1 through 3. Those usually needing confrontation in a systematic man-
ner using the basic EP conceptual framework require overview attention now, along with
some key EP toolset components not yet discussed.
The next three sections consider challenges that Peter and his colleagues had to confront
and overcome to identify and capture the opportunities they were facing in their North Sea
projects which have direct implications for challenges needing confrontation in all Part 2
chapters. As in Chapter 3, these three sections focus on contexts when a reasonable working
assumption is we are dealing with formal planning in terms of a single measurable attribute
in the sense that multiple attributes like cost and duration can be reduced to a single metric
without difficulty.
A general view of issues associated with unbiased estimates is then considered, drawing
further on the discussions involving William as well as Peter.
A watershed following section introduces a general framework which facilitates addressing
multiple objectives involving difficulties which need explicit attention before Part 1 finishes.
This watershed section is followed by two sections addressing planning when multiple
objectives involving multiple attributes require effective explicit attention but reasonably
straightforward formal approaches may be viable, one section emphasising an important
warning against excessive focus on what can be measured, a section considering ‘planning
the planning’ in terms of key implications for some of the more difficult challenges associ-
ated with complex multiple objective concerns addressed in Part 2 chapters, and a section
on a key strategy for dealing with multiple objectives.
The penultimate section briefly reflects on the need to understand that any issues not
addressed formally always require informal treatment if they are not going to be simply ignored without understanding the implications. Informal treatment which is as effective as
possible is obviously preferable to just ignoring issues that matter, and insightful informal
treatment should be informed by our understanding of both what a more sophisticated for-
mal approach might achieve and the limitations of formal planning approaches.
A final section concludes with a ‘menu view’ of what EP decision- making processes based
on a UP can provide, with inferences about how the UP concept can be tailored to broad
context issues as illustrated in Part 2.
Sensitivity diagrams as portals to further opportunities
Sensitivity diagrams of many different kinds are key EP tools. Sensitivity diagrams of one
particular kind initially developed in a high clarity form for use in the SCERT process by BP
4 Confronting challenging complexities usually needing more clarity
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Confronting challenging complexities 133
planners are considered in this section. These sensitivity diagrams portray the way several
lower level sources of uncertainty have been combined to define a higher level composite
source of uncertainty, incorporating dependence in an underlying composition structure.
An example is the sensitivity diagram illustrated in Figure 4.1.
When using a UP or any associated SP, it is always crucial to begin the test phase by under-
standing the sensitivity of results to working assumptions about contributing components.
Different forms of analysis using different kinds of component structures often require very
different approaches to different kinds of sensitivity analysis. If the analysis involves quanti-
tative (probabilistic) treatment of multiple sources of uncertainty, then sensitivity diagrams
like Figure 4.1 are central to an EP approach to the test phase and related interpret phase
concerns. They have been used in various forms for a wide range of applications involving
EP prototypes ever since their first use by BP. In the BP context their importance extended
to use in the interpret phase to explain what really mattered to the main board. In later stud-
ies they played a similar role in most cases.
In the context of BP introducing a SCERT approach, it soon became very clear that it was
important everybody involved in using sensitivity diagrams understood how they were used
by the planners to evolve the plans, what messages the sensitivity diagrams could convey,
and what assumptions they made. This included board members. It required a widespread
Mar NovDec Jan Feb Apr May Jun Jul Aug Sep Oct
probability jacket fabrication will be complete by the dates indicated
0.2
0.4
0.6
0.8
1
0
0.3
0.5
0.7
0.9
0.1
base plan completion date
1 2
3 4
5
6
probability curves show the cumulative effect of the following sources of uncertainty:
1. yard not available or mobilisation delays 2. construction problems/adverse weather 3. subcontracted nodes delivery delays
notes: 1. the curves assume a minimum fabrication period of 20 months 2. no work is transferred offsite to improve progress 3. no major fire, explosion, or other damage
6. delayed award of fabrication contract
4. material delivery delays 5. industrial disputes
dates with relevant years indicated
Figure 4.1 Sensitivity diagram: North Sea oil project platform jacket fabrication example.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
134 Foundations
‘what needs to be done’ level of understanding grounded upon key aspects of ‘how to do
it’ understanding.
Figure 4.1 portrayed the uncertainty associated with fabricating the platform ‘jacket’
for an offshore oil production project. The jacket was the steel structure fabricated in a
‘yard’ (dry- dock facility) which was then floated out to the location of the production well,
upended and pinned to the ocean floor, fitted with all necessary modules (production facili-
ties and accommodation), and connected to the pipeline taking the oil to shore.
Curve 1 portrayed ‘source of uncertainty 1’, a quantified source (known known) defined
at a composite level by the possible effects of two underlying sources of uncertainty asso-
ciated with starting this activity – the yard might not be available on the contracted date
because another jacket being constructed in the yard had been delayed and could not be
moved out until it was completed, or the yard might not have been used for some time and
it might take time to hire and train some of the staff required. These were mutually exclusive
component sources of uncertainty – one might occur but not both. Combining them was
useful because of this mutual exclusivity. As indicated by the upper end of curve 1, a delay
of more than two months was possible, although a delay of less than one month was likely.
All these curves are cumulative format portrayals to show the probability of completion by
the date indicated.
Curve 2 portrayed source 1 plus source 2. The gap between curves 1 and 2 was defined
by the cumulative effect of source 2. Source 2 was the quantified uncertainty defined by
the combined effect of two further component sources – possible adverse weather and
possible construction problems. These two components sources were not mutually exclu-
sive. They could have a combined effect, with some construction problems being caused
or aggravated by knock- on effects of weather problems. The HAT structure within source
2 had to reflect this dependence, and curve 2 had to show the cumulative effect of source
1 plus source 2 incorporating any dependence between sources 1 and 2. Despite these
dependencies, the closeness of curves 1 and 2 made it clear that source 2 was not very
important – the least important of the six sources shown by a separate curve in Figure 4.1.
That was why a separate portrayal of adverse weather and construction problems was not
worthwhile.
Curve 3 portrayed sources 1, 2 and 3. Source 3 was the quantified uncertainty defined by
the effect of subcontracted node delivery delays. This curve showed the effect as source 3
was added to source 1 plus source 2. In this case independence was assumed, but had source
3 been dependent on sources 1 and 2, this could have been easily handled within the HAT
framework using a nested pairwise separability structure implying source 3 has a dependence
relationship with sources 1 and 2 with common features. Separate different dependence
relationships with sources 1 and 2 requiring modelling reflecting these differences would
have been more difficult but feasible, the difficulty arising because the pairwise separability
assumption which is easily handled needs relaxing, a technical issue not worth exploration
here. The closeness of curves 2 and 3 made it clear that the source 3 uncertainty about sub-
contracted nodes was not a particularly big source of concern, although it was marginally
bigger than source 2.
Curve 4 relative to curve 3 portrayed the effect of source 4, quantified uncertainty defined
by the effect of material delivery delays. Again, curve 4 shows the cumulative effect. And
again, dependence among sources 1, 2, 3 and 4 would be simple to deal with using a pair-
wise separability assumption, separate different dependency relationships requiring model-
ling involving more difficulty. The bigger spread between curves 3 and 4 makes it clear that
material delivery delays are much more important than sources 1, 2 or 3.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Confronting challenging complexities 135
Curves 5 and 6 continued in the same way. They showed the largest source was ‘indus-
trial disputes’ (source 5). But ‘delayed award of fabrication contract’ (source 6) was also
significant.
The notes identified sources of uncertainty treated as known unknowns – quantification
was conditional on these assumptions.
One purpose for Figure 4.1 was explaining to the managers responsible for letting the
fabrication contract why it was important there was no unnecessary delay. That was why
‘delayed award of fabrication contract’ was positioned last in the sequence of six sources.
Curve 6 showed that current analysis assumptions suggested a 0.15 probability of achieving
the base plan completion date at the end of March, but this increased to 0.45 if the possible
fabrication contract delay portrayed by curve 6 could be eliminated. This was part of a case
for ‘getting on with letting the contract’.
A very important outcome from early use of Figure 4.1 was triggering an extensive data
review to test the assumptions made by the planners who assessed source 5 (industrial dis-
putes). Because this source clearly mattered, it became a focus of attention. Source 5 had
been initially quantified by the planners on a purely subjective basis, drawing on their general
knowledge of offshore industry issues. What a detailed analysis of recent industrial disputes
data for all potential yards in Europe showed was the planners had sized source 5 correctly,
but they had misunderstood what was driving the potential problems and what might be
done to reduce them. None of the disputes which were explored using available data had
occurred in the first 90% of the contract duration – they were all concentrated at the end
of contracts. And they had only occurred when no more work was scheduled to come into
the yard on completion of this contract. The current assumptions that tight contract condi-
tions and penalties would contain the problems were not well founded. But if Figure 4.1
went to the board, recommending collaboration with other oil majors to get a smoother
flow of work to the yards they were all using, the board might initiate a very effective and
efficient form of collaboration which could reduce exposure to industrial disputes and get better prices for all the oil majors by avoiding the current ‘feast or famine’ situation for the yards and their employees. This could be good for all the oil majors using the yards, the yard
owners, and the yard employees. What was involved was a very useful potential win– win–
win opportunity identified by a creative planner who could see the bigger picture when he
used Figure 4.1 to trigger deeper analysis addressing data which he chose to chase because
sensitivity analysis made it clear that understanding this data might matter.
The board needed to understand what sensitivity diagrams like Figure 4.1 were saying at
various points in the analysis cycle to understand this sort of discussion. The board found
sensitivity diagrams in this format particularly useful when employed in a top- down manner
to portray the four or five key sources of overall uncertainty in the proposed budget cost,
with the option of also looking at lower level curves to further decompose these compo-
nents. A comparable breakdown of project completion date uncertainty was also useful at
the board level.
The Figure 4.1 example was memorably useful, but use of up to about half a dozen sen-
sitivity diagrams of this kind for board level top- down presentations became routine, drawn
from a comprehensive set of sensitivity diagrams which had become a key operational tool
when building the analysis from the bottom up.
Figure 4.1 is an activity level sensitivity diagram for one activity in terms of duration.
Higher and lower level diagrams were also used in terms of duration, and cost was addressed
in a comparable nested structure. The analysis as a whole involved several hundred cost
sources of uncertainty components, including associated duration source of uncertainty
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
136 Foundations
components, but the board was only presented with a limited set of sensitivity diagrams
highlighting key issues from a top- down perspective.
The planners needed to present their results from the top down, using a very selective
small set of key diagrams. But the planners had to build an extensive nested sensitivity dia-
gram structure from the bottom up as a central aspect of managing the iterative SCERT
process and understanding interdependencies. For example, the planners needed separate
the analysis of the two components of source 1 before combining them and the interde-
pendent analysis of the two sources underlying source 2.
Developing the nested sensitivity diagram structure and understanding the associated
interdependence relationships was strongly encouraged and directly facilitated by the BP
software because of its discrete probability arithmetic basis, but the same results could have
been obtained using commercially available Monte Carlo simulation software. Later users of
this approach usually used simulation- based software. The key was encouraging the analysts
to use the same underlying HAT models, understanding what was involved when making
these modelling decisions, and using the models to engage with their boards or other senior
decision makers in the same way. For example, at a modelling level to produce Figure 4.1
they would have to set up the simulation so that the sum of sources 1 and 2 was captured
to plot curve 2; the sum of sources 1, 2, and 3 was captured to plot the source 3 curve; and
so on.
Within BP and later users of related processes, the overall analysis process often involved
‘moving the bottom down’ when it became clear that more detailed understanding of a
critical area would be useful. For example, while reasonably low clarity subjective probability
analysis on a first pass minimised time wasted on sources which were relatively unimportant,
the planners had to go back to important sources, like source 5, and seek more clarity using
all their skills and creativity. That was what opportunity management in an ‘opportunity
efficient’ manner which included underlying clarity efficiency was about.
The board and other senior decision makers did not need to know how the planners
did this in detail, but they did need to understand in overview terms what was being done
and the quality of the thinking that lay behind the summary top level portrayals of direct
interest.
All contexts require their own forms of sensitivity analysis, tailored to the context. In
some cases, quite different forms of sensitivity diagrams or tables may be best suited to the
sensitivity analysis issues addressed. But everyone involved needs to understand sensitivity
analysis issues at an appropriate level in a useful manner. Figure 4.1 is just a high clarity
example when multiple sources of uncertainty assessed within a HAT structure are involved.
However, the Figure 4.1 format has some particularly useful implications with relevance to
many contexts.
An aspect of an organisation’s capability- culture assets which matters is having planners
who know how to structure the initial decomposition of analysis components (in this case,
activities, sources of uncertainty, and responses) estimate associated key parameters and
uncertainty, synthesise and interpret the results using sensitivity diagrams, and then revise
and refine their analysis where doing so will pay dividends – increasing clarity efficiency.
Acquiring and using these skills are dependent on understanding the full set of potential
roles for sensitivity diagrams within all aspects of the decision- making process as a whole in
all relevant contexts.
Another aspect of an organisation’s capability- culture assets which matters is having board
level managers who know how to use planners with these skills to best effect, encouraging
and facilitating the right stuff.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Confronting challenging complexities 137
Sensitivity diagrams like Figure 4.1 are another gateway concept in several senses, an
aspect of tactical clarity which needs overview understanding for strategic clarity.
Sensitivity diagrams have roles in the same or comparable forms in all Part 2 chapters, in
corporate and operations management contexts, as well as project management contexts.
Source- response diagrams as portals for further clarity
Some BP board members wanted to understand Figure 4.1 in terms of what underlay it in
more depth. This was often the case for senior managers using Figure 4.1 variants in many
other contexts.
BP board members found Figure 4.2 a useful portrayal of the uncertainty underlying
source 1 in Figure 4.1, clarifying the source 1 uncertainty by decomposing it in a structured
manner. Figure 4.2 provides a useful example of the deeper understanding of complexity
which an EP approach can facilitate for all those who need it, using ‘source- response’ dia-
grams in this particular context.
What Figure 4.2 portrayed was the preferred response if the yard was not available
(because another jacket under construction in the yard had not been completed) was ‘mobi-
lise and accept a short delay’ (ship steel to an adjacent site, for example). This might not
work because of a secondary source associated with the assumed primary response (accept-
ing a short delay might not work because a long delay was inevitable). The next preference
primary response was ‘find an alternative yard’ (breaking the original contract). This might
Figure 4.2 Source- response diagram: source 1 in Figure 4.1.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
138 Foundations
not work either (because an alternative yard might not be available). Accepting a long delay
was then the only response available.
Figure 4.2 also implicitly indicated that ‘start- up’ problems (associated with a yard which
had not been used for some time) had not been associated with any response option choices
involving BP actions – because it was assumed these problems would and should be dealt
with reasonably effectively by the yard and intervention by BP would not be appropriate.
This rationale might have been made explicit on the diagram.
Figure 4.2 is useful as a portal to the kind of ‘what needs to be done’ and associated
requisite ‘how to do it’ level of understanding of the contingency planning underlying the
earlier discussion.
The project manager and his team members needed the additional clarity provided by
complete sets of source- response diagrams addressing all sources of uncertainty developed
with extensive team discussion to test their understanding of the structure of uncertainty
prior to quantifying uncertainty. They could not have built the sophisticated quantitative HAT models used without this qualitative understanding because of the complexities they
had to deal with.
The board needed to understand that this was what the planners had done, with sufficient
exposure to a few examples of interest to understand broadly how it was done, but not the
details of all the diagrams used by the planning team.
This is an example of explicitly using formal qualitative analysis to facilitate bottom- up
quantitative analysis and then using the resulting quantitative and qualitative analysis dia-
grams to explain results on a top- down basis which is focused on what really matters to the
senior decision makers.
Figure 4.2 was part of a much larger diagram for this activity, which portrayed all the iden-
tified primary sources of uncertainty as circles like those for ‘yard not available’ and ‘start- up
problems’, in parallel if they were mutually exclusive, in series if they were cumulative. The
third primary source was at the end of the arrow from ‘yard not available’, and the others
further moved to the right in the same way, to show a complete set of ‘primary sources’
for this activity using as many pages as necessary. Specific reactive response options were in boxes below, ordered to reflect preferred choices, with ‘secondary sources’ as circles to the
left as shown for the first two responses. Specific proactive responses were identified and noted separately when relevant. Proactive responses involve a new base plan A2 to replace
the original base plan A now known as plan A1, while reactive responses involve a fall- back
plan B or C, perhaps dependent on prior arrangements being put in place. Diagrams for
successive activities with potential knock- on effects were linked, and other interdependences
were also shown.
The net effect of all primary sources and their specific responses was associated with a
‘residual uncertainty’ source, represented by a circle on the extreme right- hand side of the
overall diagram for each activity, and below this circle were all the general reactive responses for this activity, ordered in the preferred sequence and showing secondary risks if appropri-
ate. General proactive responses were listed separately with the specific proactive responses. These source- response diagrams were drawn and discussed to ensure that the project
team understood the relevant uncertainty in structured qualitative terms before quantifica- tion began. Identifying the conditions portrayed as ‘notes’ in Figure 4.1 was part of the dis-
cussion about what could be usefully quantified and what was better treated as a condition.
Other contexts may require other forms of qualitative structuring of the underlying com-
plexity in order to build a clarity efficient quantitative analysis, but those involved need to
understand the issues at an appropriate level for their role in the planning process, using
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Confronting challenging complexities 139
approaches designed to meet their needs in an effective and efficient manner. This is just
a high clarity example when multiple sources of uncertainty are involved in a reasonably
sophisticated HAT structure.
Chapman and Ward (2011) discuss the use of Figures 4.1 and 4.2 in a generalised version
of the SCERT process as a whole in more detail, and Chapter 7 in this book updates this
at a strategic clarity ‘what needs to be done’ level of understanding. All you need to under-
stand at present is this kind of complexity can be successfully confronted when it is relevant
if analysts who know how to do it are involved. It can be very relevant when opportunity
efficiency and underlying risk efficiency plus clarity efficiency are pursued using sensitivity
diagrams like Figure 4.1 in conjunction with decision diagrams like Figures 3.6 and 3.7.
More generally, an enlightened approach to structuring the way uncertainty sources and
associated responses are related is essential. Those shaping the analysis approach need to
understand the toolset at their disposal. They also need the skill set and mindset to use these
tools effectively, including explaining the basis of recommended plans to others. Those
involved at the board level need to understand the quality of the underlying analysis as a
whole, as well as the rationale of the approach and the results of specific interest. Further-
more, if tools like the sensitivity diagram of Figure 4.1 are never employed, board level
managers need to understand their role well enough to question why, with a view to testing
the competence of those leading the planning.
Consolidating unbiased parameter estimation issues
Chapter 3 began with a focus on unbiased estimates of performance measures like activity
duration to initiate an exploration of clarity efficiency with a focus on bias before linking
clarity efficiency, risk efficiency and opportunity efficiency. Five key bias issues were consid-
ered during the discussions involving William and Peter. These five issues are briefly struc-
tured and summarised now, followed by a brief exploration of a sixth issue.
The consolidated view of these six issues outlined in this section provides a ‘what needs
to be done’ EP framework for managing all unbiased parameter estimation concerns in a
coordinated manner. This helps to unify Part 2 discussions involving parameter estimation
bias and biased decision making in broader terms. It also confronts an important concern
avoided earlier.
Issue 1 – working assumptions about sources of uncertainty treated qualitatively
In the discussion with William, it was made clear that if feasible scenarios beyond the scope
of past experience are not quantified, biased estimates will result. Furthermore, any feasible
outcome scenarios treated qualitatively as conditions need to be understood when address-
ing bias. These ideas were later considered in various ways, including the concept of a no
quibbles provision briefly explored in Chapter 3 and the conditions illustrated by notes
addressing known unknowns in Figure 4.1.
Issue 2 – working assumptions about parametrically defined probability distributions
Parametrically defined probability distributions are common practice and often useful, but
as working assumptions they need testing for bias. Furthermore, they are never suitable
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
140 Foundations
framing assumptions because they always need testing. The HAT approach facilitates a flex-
ible approach to testing the robustness of parametrically defined probability distributions,
and its open ‘synthesis of all relevant approaches’ nature means that it will deal with all of
the issues which I have had to confront. Most readers do not need deep knowledge in this
area, but their organisations do need requisite capability when it matters, and all senior man-
agers need to understand in broad terms what this involves and why parametrically defined
probability distributions can be a source of serious risk if the associated assumptions are not
well founded.
Issue 3 – working assumptions about the role of dependence
The HAT approach facilitates a general approach to both statistical and causal dependence,
which is essential to avoid very serious bias. A failure to deal with important interdepend-
ences, including complex knock- on and feedback effects, is a common problem. Often it
renders formal quantitative (probabilistic) analysis dangerous – well beyond useless to the
extent that an organisation would be much better off if they just trusted reasonably well-
grounded informal intuition and either retrained or got rid of all ‘experts’ who cannot deal
with dependence effectively. Some of these ‘experts’ hide behind textbooks, standards and
guides which fail to clarify the implications of ineffective treatment of issues like interde-
pendences or provide adequate guidance for effectively dealing with these issues.
Issue 4 – working assumptions about minimising unconscious behavioural bias
Unconscious behavioural bias associated with estimating probabilities directly or eliciting
them from others was mentioned when discussing a minimum clarity approach and its gen-
eralisations for William in Chapter 3, and dealing effectively and efficiently with issues like
‘anchoring’ always matters. There is a large literature on unconscious behavioural bias, with
some entry points indicated in Chapter 11. Behavioural bias always needs to be understood
by all those engaged directly in estimation processes, and responsible senior managers need
to be able to judge associated competence levels. Misunderstanding the limitations of data
as well as misinterpreting data can be viewed as part of this issue. Misunderstanding the
nature of subjective probabilities and their relationship with component objective estimates
can also be viewed as a very important part of this issue.
Issue 5 – working assumptions about ‘strategic misrepresentation’
The working assumptions addressing the first four issues all involve ‘competence’ concerns
addressed in Chapter 3. A fifth issue mentioned earlier but not explored is deliberate bias
in the ‘strategic misrepresentation’ sense (Flyvbjerg, Bruzelius, and Rothengatter, 2003).
Strategic misrepresentation is usually used as a euphemism for lying – comparable to the phrase ‘being economical with the truth’. This takes us into broader capability- culture con-
cerns and decision- making bias which arguably goes beyond biased parameter estimates.
Whenever it is relevant strategic misrepresentation needs insight and care, because it matters
a great deal. It can have roots which go well beyond simple lying for personal and corporate
gain, and it may have complex roots linked to both important ethical concerns and incen-
tive issues. It is an important concern which is widely cited and addressed, but how best to
respond to its implications is not a simple matter. Significant controversy about implications
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Confronting challenging complexities 141
overlooked by many common practice responses to strategic misrepresentation which sim-
plify in the ‘wrong way’ is arguably just adding fuel to the fire and making matters worse.
Issue 6 – all other relevant working assumptions including a ‘conspiracy of optimism’
It is useful to have a residual category to synthesise all the earlier categories, make sure that
overlaps are not double- counted, and pick up any issues that were missed earlier. This is
another area which requires an approach involving a ‘closure with completeness’ component.
One set of issues only partially covered earlier needs further attention now.
At the beginning of my discussion with the air vice- marshall responsible for the Ministry
of Defence (MoD) procurement budget during the study involving William discussed in
Chapter 3, my opening question was ‘What did he think was the main problem with the
MoD approach to project risk management?’ His immediate answer, with a smile on his face
and a glint in his eyes, was a very blunt ‘everybody lies to me’.
Having captured my very close attention, he then rephrased his answer in a form that
might be used for my report, saying that the MoD suffered from what he liked to refer to
as a ‘conspiracy of optimism’. He then went on to explain in some detail what this implied.
The ‘conspiracy of optimism’ euphemism is arguably much richer than strategic misrep-
resentation, although there are important overlaps. A conspiracy of optimism is still cited
in MoD circles; it is discussed in my earlier books with Stephen Ward, and I sometimes
hear it in other contexts. It immediately became and remains one of my all- time favourite
euphemisms. Writing this book further clarified for me the considerable understanding it
can portray if explored in detail.
One key concern which a conspiracy of optimism portrays is that MoD people are very
loyal to their teams, which is a crucial corporate culture characteristic of armed forces per-
sonnel, and ‘a very good thing’ in most contexts, a capability- culture asset. However, if
this loyalty leads to conscious bias (strategic misrepresentation) in their contributions to
duration or cost or other performance estimates, the associated optimism bias needs to be
accommodated when producing a corporate estimate. Corporate clarity about aspirational
targets which are ‘stretch’ targets and their relationships with expected outcomes and com-
mitment targets can help facilitate dealing with this kind of conscious bias, but this may not
be the only response needed.
Another key concern which a conspiracy of optimism portrays is that MoD people have a
corporate culture ‘can do’ attitude linked to thinking positively which needs to be encour-
aged and preserved, up to a point, another capability- culture asset. However, this natural
and desirable positive trait also needs to be accommodated when they contribute to per-
formance estimates leading to expenditure commitments which are difficult to revise when
bias becomes evident, in the sense that if and when overconfidence which is misplaced is
involved, it needs to be neutralised. Corporate clarity about aspirational targets and their
relationships with expected outcomes and commitment targets can also help facilitate deal-
ing with this kind of unconscious bias. But dealing with it effectively may be difficult, and
again this may not be enough. ‘Commitment targets’ may involve commitments by the
MoD procurement functions to the serving forces, to politicians, and to the public, with
various shades of complex implications. Furthermore, MoD procurement function commit-
ments invariably involve interdependent commitments to the MoD from contractors who
may have serious misconceptions about what is required or agendas which are not aligned
with MoD goals, perhaps because of inappropriate contract structures.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
142 Foundations
Both these concerns were understood by the MoD people I worked with directly by
the beginning of the 1990s. So were some aspects which had no positive implications. For
example, most people do not like to reveal a lack of understanding about something they
are supposed to have deep expert knowledge about, even to themselves, and part of the
conspiracy of optimism may be attributed to deliberately hiding a lack of understanding or
overconfidence associated with an unconscious equivalent. Furthermore, a lack of compe-
tence of other kinds may be an issue needing attention.
Some conspiracy of optimism concerns overlap, at least in part, the five issues addressed
earlier in this section, but some new ‘culture’ issues are also involved. Once it is clear that
bias is a capability- culture asset and liability concern of considerable complexity, recognis-
ing the role of accommodations for capability- culture assets and liabilities presents itself as
a potentially useful tool.
For example, Chapter 6 explores the idea that a senior manager may accommodate con-
cerns about key staff being involved in a potential new contract who are already heavily
committed to other important contracts by deliberately introducing a form of bias designed
to ‘correct for’ what they might reasonably see as ‘corporate oversights’ if there is no other
way of avoiding unwise overcommitments. Providing them with a much more effective
direct way of ensuring that their concerns are effectively addressed is demonstrably better
than this kind of informal ‘accommodation’ using bias. In some cases, these ‘correction
accommodations’ needing attention may take a ‘conspiracy of pessimism’ form, generalis-
ing the conspiracy of optimism concept, as illustrated in the tale of Chapter 6. ‘Strategic
misrepresentation’ in the usual sense is not involved, although that term could be used if
you prefer it, and the corporate competence issues being corrected for are not one of the
issues identified above. The choice of label is optional, but dealing with these kinds of issues
effectively is not an option if unbiased estimates as part of a broader concern for unbiased
decision making is the goal.
As an example of a quite different kind of bias which clearly goes beyond the preced-
ing five, using an inappropriate discount rate for Net Present Values (NPV) calculations as
discussed in Chapter 7 can be construed as an issue 6 source of bias associated with biased
estimates of parameters. This example may stretch your view of what parameter estimation
bias covers too far, and you may prefer to see it as a different and separate form of biased
decision making, a more fundamental kind of decision- making error. However, it needs
dealing with however it is perceived, and in some contexts, viewing it as a particularly perni-
cious and significant form of parameter estimation bias can be useful.
The six issue framework as a comprehensive EP toolset component
All Part 2 chapters deal with biased parameter estimation issues of various forms in the
context of seeking unbiased decision making, and the six issue framework for bias just
outlined may help you to see the links. It may also serve as a useful checklist of sources
of bias you need to be aware of for all estimation tasks your organisation has to confront.
Strategic clarity about bias issues is essential for any director or expert and all supporting
staff within any organisation. At a how to do it level managing bias requires a range of
tools used in a competent manner by all those who need to know how to use them and
interpret the results.
As a final comment in this section, the competent management area of Figures 3.5 and
3.8 may be usefully thought of as an opportunity management area for a number of reasons.
One foundation level reason is a key feature of competent management is always seeking
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Confronting challenging complexities 143
better ways to understand and communicate what most needs to be understood. As part of
this, understanding why biased estimates of key parameters which decision making depends
on are so endemic and what needs to be done to eliminate bias effectively and efficiently is
a crucial opportunity.
A general conceptual framework for multiple objectives
Often the focus of decision making is one attribute (like cost) which is measurable in terms
of one criterion (like expected cost). Sometimes when one attribute (like cost) is the focus
other metrics like lost time or revenue can be addressed via ‘opportunity cost’ concepts
that allow transformations into a single metric using simple common practice approaches
without any particularly difficult challenges. For example, project duration delay is often
converted into cost by considering overhead costs and lost revenue opportunity costs which
can be estimated without serious conceptual or practical difficulties.
But frequently the effective pursuit of all relevant goals means that more than one objec-
tive is actually extremely important, and it is not straightforward to deal with multiple
objectives involving multiple attributes as well as multiple criteria for each attribute. While
suitable metrics for some of these objectives may be readily available, other objectives may
not be measurable without additional effort or cost, which may not be worthwhile, and it
may not be feasible to measure some important objectives in any useful sense.
We need a conceptual framework to understand all the potentially relevant issues in all
contexts which is as general as possible, used to design specific practical approaches which
best suit the frameworks used to address goals– plans relationships in any particular circum-
stances. That is what this section is about.
The focus of this section is an appropriate conceptual framework for considering all the measurable or indirectly/partially measurable multiple objectives issues underlying an
organisation’s goals– plans framework for decision making. The approach taken is a core
gateway concept.
This book associates the goals– plans framework with a goals– attributes– criteria hierarchy
of objectives, implying criteria are the lowest level ‘objectives’ within attributes, and attrib-
utes are an intermediate level of objectives within goals, as indicated earlier. The concern in
this section is multiple objectives at a criteria level for all relevant attributes and goals.
‘Mathematical programming’ is not used directly in operational terms anywhere in this
book, but a ‘goal programming’ variant of mathematical programming is of importance as
a central conceptual tool. This conceptual tool has direct operational implications in several
chapters, and a goal programming perspective provides a useful ‘what needs to be done’
framework for this book as a whole. Even if you have never heard of mathematical program-
ming, and you have very little inclination to confront mathematical approaches, the basic
ideas needed for strategic clarity considered briefly in this section should make reasonable
intuitive sense at an overview level.
Mathematical programming was developed initially from the input– output economic
analysis used to replan the US economy for the 1940s war effort in a linear programming
form. Linear programming evolved rapidly as computer capability increased into a well-
developed set of planning tools used worldwide by the end of the 1960s. Quadratic and
other non- linear forms soon followed.
One key idea which has been central to the use of any form of mathematical programming
models in practice since the 1950s is the shadow price or shadow cost concept. Understand-
ing the implications of this shadow price/cost concept is crucial.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
144 Foundations
Mathematical programming involves optimising a single ‘objective function’ subject to
a set of ‘constraints’. If the objective function measures ‘expected reward’ (e.g. expected
profit) and it is being maximised, at the optimal solution each constraint has a ‘shadow
price’ which defines the amount by which expected reward could be further increased if
this constraint could be relaxed by one unit. If the objective function measures ‘expected
cost’ and it is being minimised, at the optimal solution each constraint has a ‘shadow cost’
which defines the amount by which expected cost could be further reduced if this constraint
could be relaxed by one unit. Some constraints are usefully seen as ‘hard’ because of physical
reasons or because they operate as definitions that cannot be relaxed. Other constraints are
usefully seen as ‘soft’ because they can be adjusted at a cost which is beyond the bounds of
considerations embedded in the model. An immediate implication of practical importance is
the relaxation of all soft constraints needs consideration post- optimisation, as a crucial form
of sensitivity analysis. In practice, whether or not mathematical programming is employed,
it is always important to analyse the relationship between the shadow price/cost and an
assumed actual cost of relaxing all the soft constraints, adjusting constraints appropriately, as
a form of sensitivity analysis comparable in some respects to the use of Figure 4.1 in a rather
different framework for analysis.
The quadratic programming approach developed to deal with a mean– variance Markow-
itz portfolio management framework makes use of these ideas, the basis of the parameter k
defining the optimal risk– return trade- off between E and V2, discussed earlier in terms of
maximising E – kV2. In practice minimising V2 is usually treated as the quadratic objective
function, minimising V2 subject to a constraint on E, using a set of assumed minimum val-
ues for E to define a set of optimal portfolios in mean- variance terms, each associated with
a shadow price k. A classic, basic mean– variance Markowitz model involves one attribute
decomposed into two criteria, expected outcome E and variance V2.
Goal programming builds on this idea by providing a powerful general approach to mul-
tiple objective decision making. If there are n objectives, one objective is treated as the
‘primary objective’ and used to define the objective function, the choice being a matter of
convenience. The other n − 1 objectives are treated as ‘secondary objectives’ and used to
define constraints. In an iterative process, an initial ‘best guess’ is made of a suitable con-
straint level for each soft constraint associated with a secondary objective.
At the optimal solution the shadow price/cost associated with each secondary objec-
tive defines the trade- off between that objective and the primary objective. Each constraint
defined by a secondary objective can be relaxed if this trade- off value seems too high and
tightened if it seems too low. Sometimes even non- measurable objectives can be incorpo-
rated as constraints in this structure – simply constraining an option’s feasibility if it violates
an ethical objective for example. Furthermore, an iterative adjustment process can be initi-
ated and then continued until convergence is achieved with a set of shadow price/cost val-
ues which the decision makers believe represent appropriate trade- offs between all criteria.
Further still, ‘too high’ or ‘too low’ or ‘about right’ are ultimately a ‘revealed preference’
statement of appropriate trade- offs by the decision makers. Crucially:
1 Whether or not goal programming is used explicitly, this framework can be associated with multiple objective decision making, and whether or not those making choices acknowl- edge the existence of this underlying goal programming framework and all the associated revealed preference trade- offs, they may be directly relevant. If it is not practical to use goal programming directly, which is usually the case, decision makers need a reasonably simple practical process for systematically seeking an appropriate approximation to the
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Confronting challenging complexities 145
optimal trade- offs between all objectives visualised in shadow price or shadow cost terms, understanding what trade- off values are implied by the solution chosen and the relevant implications.
2 Ignorance of the existence of this underlying framework and a failure to address a suitable approximate approach does not mean that a full set of shadow prices/costs addressing trade- offs between all relevant constraints does not exist. It means that those making the decisions do not know what the trade- offs are, and they are probably using inconsistent trade- offs for successive decisions, which implies inconsistent decision making. This may be seriously inept, and it should be seen as incompetent and avoided in all circumstances where the consistent use of appropriate trade- offs really matters.
For example, in the safety and security context of Chapter 9 we know that all interested
parties want to minimise fatalities on a railway system, but they also want to minimise costs.
This implies revealed preference trade- offs between fatalities and cost generally referred to as
‘the value of an avoided fatality’ or ‘the value of a prevented fatality’. A value of an avoided
fatality concept based on a shadow cost and revealed preference interpretation should never
be confused with a direct attempt to put a monetary value on a life in the way attempted
by some early approaches to cost– benefit analysis. A failure to explicitly update the tradi-
tional cost– benefit analysis interpretations of the ‘value of an avoidable fatality’ concept
using a shadow cost basis plus a revealed preference perspective is inept. Arguably it is also
unethical, with ignorance not being an acceptable defence in a legal context or an acceptable
excuse in a more general sense. The reasons are explored in Chapter 9.
As another example, in a discounted cash- flow context using NPV, if multiple objectives
are involved, almost always the case, a goal programming perspective is imperative, for rea-
sons explored in Chapter 7.
Traditional opportunity costs, which were in widespread common use well before math-
ematical programming was developed, can be viewed as a special case of shadow prices or
costs, a useful way to embed opportunity costs in a single comprehensive framework for all
relevant multiple objective trade- offs. Before the development of shadow price/cost con-
cepts, opportunity costs were sometimes explicitly associated with constraints in a calculus
framework, implying a comparable mathematical foundation. This kind of simple traditional
opportunity cost plays a key role in Chapter 5, and variants of this general goal programming
framework at various levels of sophistication between those of Chapters 5 and 9 are used in
Chapters 6, 7 and 8.
This book will not use a goal programming approach in terms of direct operational use
of mathematical programming algorithms, but it will use a goal programming perspec-
tive in a mathematical programming conceptual framework in conjunction with a revealed
preference interpretation of shadow prices or costs at a conceptual level with operational
implications. This is an extremely valuable conceptual framework, even if some important
objectives are not measurable. It is the most general and powerful conceptual framework for thinking about and framing operational approaches to multiple objectives decision making which keep it simple systematically that I am aware of. That is the reason it is a foundation
level EP conceptual tool and a key component of the EP toolset. If you are not familiar
with mathematical programming and would like an introduction which relates mathemati-
cal programming to separability concepts without getting too technical, see Management for Engineers (Chapman, Cooper and Page, 1987), an introductory level textbook which includes an annotated list of references to many of the key classical mathematical program-
ming texts. For an overview of goal programming, see Tamiz, Jones and Romero (1998).
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
146 Foundations
Whenever the term ‘revealed preferences’ is used in this book, usually in a shadow price
or cost context, a plain English interpretation will suffice, and the implications of more
technical interpretations will be illustrated when relevant using numerical examples. How-
ever, you may find an understanding of the nature of the underlying concerns addressed
by technical meanings useful. Economists who use ‘preference’ functions (also called ‘util-
ity’ functions) to provide a theoretical basis for the ‘demand equations’ which are used to
predict consumer behaviour sometimes use the term ‘revealed preferences’. They do so to formally acknowledge that the preferences of consumers (purchasers) revealed by mar-
ketplace responses to price and income changes reflect the collective behaviour of a set of
people whose individual preference functions cannot be combined mathematically in any
meaningful way, as demonstrated by Arrow’s ‘impossibility theorem’. Chapman (1975)
discusses some of this literature, and the use of revealed preferences for decision making
in the context of a pairwise separability approach to demand analysis. When an organisa-
tion’s board members sanction any decision involving trade- offs, they always reveal collec- tive preferences (whether or not they are prepared to acknowledge this fact) which cannot be seen as a simple combination of individual preferences, and operational management
of the implications involves practical issues of importance in some contexts. For example,
in Chapter 9 the shadow costs revealed by the decisions sanctioned by the board are the
board’s collective revealed preferences, in the sense that the board has collectively agreed to the implied trade- offs between attributes like money and avoided fatalities plus avoided
injuries, but this does not mean a simple average of individual preferences is involved, with
very useful implications which are developed in Chapter 9. You may also find it useful to
understand that all preference or utility functions employed by economists in contexts like
demand theory are very different from the utility functions used in decision analysis texts
like Raiffa (1968) and Keeney and Raiffa (1976), and some of the implications are explored
in Chapman and Ward (2002).
Seeking effective low clarity views of trade- offs
The high clarity provided by the decision diagrams illustrated by Figures 3.6 and 3.7 is
often very useful. But if risk– reward criteria trade- offs in terms of attributes like cost are
not the only concern, and it may be important to consider further attributes at low levels
of clarity, using a lower clarity decision diagram like Figure 4.3 may be a more opportunity
efficient tool.
Figure 4.3 was initially developed in the mid- 1980s to explain how such situations might
be approached for public seminar purposes – courses provided by commercial professional
training organisations which I gave in London initially, internationally later. They were then
used for follow- on in- house courses for client organisations. For example, IBM made exten-
sive use of variants of Figure 4.3 in the bidding context used for its Forum 2 programme
case studies and subsequent bidding process development, and examples based on these uses
are employed in Chapter 6. Low clarity decision diagrams comparable to Figure 4.3 are now
a widely used component of the EP toolset.
When Figure 4.3 was first used, the illustrative example employed involved replacing a
photocopier. In the mid- 1980s, as a newly appointed head of a university academic depart-
ment, replacing the department’s photocopier was one of the first departmental decisions
to be made; the need for simple decision diagrams using the format of Figure 4.3 arose at
about the same time, and ‘the photocopier problem’ seemed a useful story to illustrate the
low clarity decision diagram concept.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Confronting challenging complexities 147
The photocopier story will be told as a tale – in reality, my approach was less clarity effi-
cient and more time- consuming. The version of the tale provided here builds on a version
used for The Wiley Encyclopaedia of Operations Research and Management Science (Chap-
man and Harwood, 2011) to provide a requested definition of ‘optimal risk taking and
risk- mitigation’.
The tale begins when my department’s only photocopier failed terminally. The service
engineer delivering the bad news explained that a new photocopier with appropriate features
from his firm could probably be delivered within a few days on a new five year contract. This
would involve a rental cost of £x per month plus a servicing charge of £y per copy.
The departmental secretary provided records showing copy numbers by months for
recent years. This historical record was projected forward for five years, as the basis for an
approximate P10– P90 range estimate for a nominal five year planning horizon defined by
the contract duration. The Figure 4.3 axes were constructed, with the cost scale indicat-
ing total cost over the five year contract period. The P10– P90 values were then used along
with the service engineer’s £x per month and £y per copy contract explanation to obtain
the ‘option A’ line shown in Figure 4.3. Telephoning two alternative photocopier suppliers
revealed identical contract structures for photocopiers with comparable features. But the
alternative suppliers had higher £x rates and lower £y rates. These rates were used to plot
the ‘option B’ and ‘option C’ lines in Figure 4.3.
Contract cost was assumed to be the only concern as an initial working assumption,
so option A was the only risk efficient choice, by a significant margin. This was clearly
illustrated in stochastic dominance terms in Figure 4.3, because the line portraying option
A was well to the left of the lines for options B and C. However, testing the robustness of
the working assumption that cost as currently measured was the only relevant objective was
obviously essential.
The premise that copying speed might matter was tested first. This robustness test began
by asking, ‘Was the copying speed of option A faster than that of options B and C?’ It was
faster than both – so option A dominated on this criterion too. No more effort was required
in terms of the initial working assumption that speed did not matter. Had option A been
significantly slower than B or C, then the cost of the additional lost time might have been
estimated in opportunity cost terms, but a simple ‘Was the contract cost difference worth
cost
cumulative
probability
0.5
1
0
option A option Coption B
P10
P90
P50
Figure 4.3 Decision diagram: comparison of options A, B, and C.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
148 Foundations
the speed difference?’ question and answer might have sufficed, depending on the figures
involved and the perceived importance of speed.
The second robustness test began by asking ‘Was there any evidence that options B or C
might be more reliable than A?’ There was no evidence about B or C, but the departmental
secretary assured me that the old photocopier had been very reliable for many years before
its terminal failure, so option A seemed to dominate on this criterion in terms of readily
available data with no incentive to seek more data, and no need to think about reliability–
cost trade- offs in opportunity or shadow cost terms.
The third robustness test began by asking ‘Was the departmental secretary happy with
the anticipated noise level of the option A photocopier, and any other relevant features or
aspects, including its colour, given it would be in the departmental office which was her
office?’ Enquiring about the noise levels, colour, style and any other relevant features of
the options and involving the departmental secretary in the choice were obvious matters of
courtesy. It suits the flow of the tale to introduce these issues now, but in fact, this would
have been the best place to start, including asking her at the outset if there were any issues
other than those considered earlier which she thought ought to be explored. She was better
placed to see some relevant forthcoming technology and organisational changes than I was.
She was also the most directly interested party, and her views were crucial. If colour or style
or noise or some combination of these issues mattered, Figure 4.3 was still relevant, but the
attributes of interest to the departmental secretary would need very careful explicit atten-
tion, and if they were important enough they might prove the dominant concerns in shadow
cost terms. She liked the option A features on all counts, so it remained the obvious choice,
without the need to make extensive comparisons with B and C.
The fourth robustness test began by asking ‘Could the supplier of option A actually
deliver a new photocopier in a few days?’ If several months might be involved, and option
A was now viewed as non- feasible, the choice might move to B or C. Both were risk effi-
cient assuming A was not available – B involved a lower expected outcome cost than C but
a higher level of risk, indicated by the curves crossing above the expected values (associated
with a lower £x rate but a higher £y rate, so the outcome was more variable). However, the
additional cost ‘risk’ associated with option B relative to C was really just ‘noise’ – it was not
worth worrying about.
With new supplier reliability now recognised as a potentially important issue, a valuable
new option required identification – the option A supplier making an interim machine avail-
able until the ordered one could be delivered. More generally, we are always free to think
about new options if the current set of options seems unsatisfactory, an important fall- back
contingency planning option to always keep in mind. It may help you to remember the
availability of this option if you always include a residual ‘any other options we have not yet
considered’ possibility, another closure with completeness concept.
One key point this tale should make clear in Figure 4.3 is a simple low clarity decision
diagram version of Figures 3.6 and 3.7, but it provides all the clarity needed in this rela-
tively simple context. Another important closely coupled message from this tale is if you are
used to using the low clarity decision diagram approach of Figure 4.3 and visualising range
estimates associated with P50 expected values, you do not even need to draw Figure 4.3
to visualise the expected cost and cost risk trade- offs involved. A related key issue this tale
should clarify is if risk is just noise for some attributes, then just using expected values is a
very important shortcut. Knowing that this is the case in advance improves clarity efficiency
and improves opportunity efficiency by capturing an important opportunity to achieve more
insight for less effort/cost.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Confronting challenging complexities 149
A further key learning point for you to take from this tale is that variability and risk
should never be confused – option A involves significantly greater contract cost variability than B or C but much less risk. This is clearly demonstrated in Figure 4.3 by the different slopes but the large gap between the curves for options A and B or C, because the low £x
rate for option A more than neutralises the cost variability driven by the higher £y rate for
option A. This addresses a complication with Markowitz’s mean– variance approach which
he understood very clearly, but many people using his approach do not. Variance is never an appropriate direct measure of risk. A closely coupled issue is variance is only an appropriate
surrogate measure of risk given a particular assumed expected outcome and given probability distributions with the same shape can be assumed.
Yet another key point is the potentially crucial importance of non- measurable issues,
like the implications of the photocopier’s colour or style. They need consideration in rela-
tion to what can be measured, with questions like ‘If option B is the preferred choice in
terms of a set of non- measurable advantages and disadvantages concisely expressed in verbal
terms, is the contract cost difference between options A and B worth it?’ If the choice is a
close call, this may suggest further effort and care would be worthwhile. The choice made
reveals preferences, a very simple example of a revealed preference interpretation involving
a shadow price/cost associated with insisting on the preferred colour by constraining the
option choice.
All the key learning points developed in this section were central to the IBM culture
change programme discussed in Chapter 3 which underlies the tale of Chapter 6. They also
underlie all the other Part 2 chapters. More generally, what is crucial is:
1 starting with a simple approach which can address more complexity whenever doing so
looks worthwhile;
2 comprehensive, effective and efficient assessment of robustness for the simple approach
employed initially, testing all relevant working assumptions to decide whether more
clarity is worthwhile;
3 seeking more clarity of the kind needed when and where it is needed; and
4 a possible expansion of the option set, with creativity an essential capability.
The starting point for any particular analysis will always be context specific, but the con-
ceptual framework drawn on should be general. Chapter 6 illustrates the use of very similar
variants of the Figure 4.3 approach. Chapter 7 draws in Figures 3.6 and 3.7, the high clarity
versions. Chapters 8 and 9 build on both.
A generalisation of the clarified- dominance concept introduced in Chapter 3 involves
robustness testing of the kind just discussed before asserting that ‘clarified dominance’ in terms of all relevant objectives is involved.
Simple but effective strategies for dealing with multiple objectives
The last section used one simple example to illustrate several general strategies for seeking
clarity efficient approaches to multiple objectives. In practice, those leading an EP approach
need a simple but effective operational how to do it toolset of such strategies, and at a what
needs to be done level you may find an initial feel for the nature of these strategies use-
ful now, to provide a base to build on in Part 2. For example, one useful strategy involves
starting by assuming that an analysis based on the primary objective will provide a solu-
tion which dominates in terms of all secondary objectives. Testing the robustness of this
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
150 Foundations
assumption for each relevant secondary objective has to follow, ensuring that no objectives
that matter are overlooked, even if they are difficult to measure or completely impossible
to measure.
A central concern when seeking opportunity efficiency is always testing the robustness
of a clarified dominance hypothesis by asking the question, ‘Is it worth trying to measure
anything else before we consider making a decision, bearing in mind the costs and benefits
of more measurement, and are there any concerns which need qualitative attention whether
or not they are measurable?’
If any potentially important further objective is not measurable (like photocopier colour
or style as perceived by the departmental secretary), the only choice may be to look at the
difference in cost and ask, ‘Is the difference worth it?’ in consultation with all relevant par-
ties, employing a consultation process that all relevant parties will trust.
If any potentially important further objective is measurable, it can be important to assess
whether doing so is worthwhile. An opportunity cost might be estimated (like the value
of the time wasted because of a slower photocopier) to extend what the primary objective
measures. But it may be simpler and more effective to just look at the difference (in cost
and speed/time in the photocopier case) and make a judgement without expending effort
decomposing the associated uncertainty.
We always need to make sensible judgements based on what has been measured and what
has not been measured but might be important.
This is a very short section, but the issues addressed are important, and you can build on
the examples provided during the rest of this book and via your own experience.
For those interested in the more technical concerns, applying this approach in a complex
context is central to a journal paper (Chapman, Cooper, Debelius, and Pecora, 1985) which
discusses extensive analysis of a single very important decision for the Fluor Corporation by
a team of about 30 Acres consultants over a period of several months.
A very dangerous myth and a closely coupled bias
This book is about systematic qualitative analysis and systematic quantitative analysis, pre- dominantly in terms of their joint use. Quantitative analysis starts gently in Chapter 5, but
it becomes reasonably sophisticated halfway through Chapter 6. Chapter 7 is focused on
qualitative analysis issues, but the discussion builds on the quantitative analysis discussions
of Chapters 3 and 4. Chapter 8 is more quantitative than most corporate strategic planning
before a strong emphasis on qualitative analysis issues. Chapter 9 pushes quantitative analysis
about as far as it can go in a safety context at a what needs to be done level. The discussion
throughout Part 2 is focused on what needs to be done, not tactical quantitative analysis
details that you do not need for strategic clarity. However, the value of useful quantitative
analysis underpinned by appropriate qualitative analysis is championed by this book, while
rationing the discussion of how to do it tactics to the minimum needed for strategic clarity.
From an EP perspective it is very important to appreciate that whatever the level of
sophistication of the quantitative and qualitative analysis being used, there are important
limitations on both what can be measured in quantitative analysis terms and what can even
be considered in purely qualitative terms in a systematic formal manner.
At a strategic clarity level, it is crucial that everybody in any organisation you care about understands that the mantra ‘what cannot be measured cannot be managed’ is a very dan- gerous myth with a closely coupled bias towards making decisions in terms of objectives
limited to easily measured criteria. Those who subscribe to any variant of this mantra
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Confronting challenging complexities 151
are using a framing assumption which should not be tolerated. If they hold positions of
power and influence, they may be seriously dangerous. If it is not feasible to enlighten
them, they should be contested and their influence constrained in an effective manner
before they do too much damage, if this is feasible. They are not pushing a well- founded practical mantra, they are pushing a very dangerous myth, and the difference matters
greatly. Arguably this should be self- evident in most circumstances that really matter. But
many powerful and influential people seem to buy into the myth, some because it suits
hidden agendas which they want to keep hidden, some because they genuinely seem to
have been seduced by numbers, and some because they seem to implicitly assume that
no really serious damage will be done by what they see as a helpful focus on measurable
attributes and quantitative evidence without fully understanding the implications or the
alternatives.
Such people can be an important capability- culture risk to the best interests of everyone
else. Responding to the risks they create may require enlightened planning and management
to avoid serious organisational misjudgements.
‘What cannot be measured cannot be managed’ is often promoted and sometimes even
mandated as part of a bundle of notions using the acronym SMART as a contraction of
underlying definitions like ‘Specific, Measurable, Achievable, Relevant, and Time- framed’.
There is clearly potential merit in an initial focus on SMART concerns in this sense, but at the very least there is also a clear need to recognise that any focus on measurable separate components demands a ‘closure with completeness’ component involving a synthesis of
what has been addressed within a SMART framework plus everything else which really mat- ters. If this is not done in a way which includes non- measurable concerns there will be an obvi- ous bias towards attributes which are measurable. This is likely to involve immediate errors
of judgement. Arguably even more serious, people or organisations whose performance is
judged based on a limited set of easily measurable attributes will be encouraged to change
their behaviour, ignoring attributes which cannot be measured or are just difficult to meas-
ure. There is an endemic and seriously damaging bias towards the use of attributes which
are easily measurable even when people are well aware they do not tell the whole story and
would not even consider mandating a SMART approach.
The primary reason for raising these issues now is explicitly and emphatically avoiding the
impression that EP means quantification is essential or central. Often quantification is not
worthwhile if a qualitative analysis is convincing. Frequently further quantification is not
worthwhile if partial quantification plus appropriate further qualitative analysis is convinc-
ing. Crucially, whatever the level of formal quantitative analysis, management decision mak-
ing almost always also needs further qualitative analysis plus sound well- grounded intuition
and informal planning.
The photocopier example used to explain Figure 4.3 emphasised objectives which could
not be measured as well as objectives which it was not worth measuring in a simple context.
The last section generalised these approaches.
Looking back to Chapter 2, the role in the UP of the capability- culture concept is focused
on concerns which include non- measurable or difficult to measure issues like team- working skills plus the right stuff.
Judged from an EP perspective, anyone who seriously argues ‘what cannot be measured
cannot be managed’ is using an unacceptable framing assumption because ‘what cannot be measured directly but matters actually needs VERY special care and attention to ensure effec- tive treatment’. Capability- culture issues are obvious illustrative example concerns, but there are many others.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
152 Foundations
A multiple objectives approach to planning the planning
The last four sections have addressed formal consideration of multiple objectives going
beyond a single primary attribute of interest. From the beginning of any planning process,
clarity about multiple criteria associated with a primary attribute plus further clarity about
more than one attribute is almost always needed. Sometimes these multiple attributes are
associated with very different kinds of goals, sometimes key goals have priorities for dif-
ferent players which are radically different, and sometimes one player’s gains in terms of a
key goal are another player’s loss. The general issue is ‘For each relevant multiple attribute,
should we attempt formal consideration as part of the planning process, or are informal planning judgements, perhaps coupled to formal or informal negotiations, more likely to
prove appropriate?’
The need for formality when addressing important issues is a concern requiring explicit
planning attention. It is one of several non- measurable aspects of an extensive set of higher-
order ‘planning the planning’ concerns which need explicit care and attention.
The select and focus phase of the UP of Figure 2.1 involves separation of the formal plus
informal planning the planning phases from all the other UP phases. It provides a basis for
treating the overall management of a UP as a project which needs planning in its own right.
This project planning framework for planning the planning is used for all Part 2 chapters,
with some tailoring to specific contexts. A brief overview of some implications is useful
now. It provides a conceptual background for Part 2 to be put to work from the outset of
Chapter 5.
Consider Figure 4.4, a bar chart (Gantt chart) portraying how the first complete pass
through the UP of Figure 2.1 might be visualised and planned.
Figure 4.4 portrays precedence and timing relationships in a way many project planners
would have understood a century ago, following the promotion of this bar chart format
by Henry Gantt as part of the Scientific Management movement associated with Frederick
Taylor (1911).
interpret
implement
intense effort ongoing effort no effort or intermittent effortkey:
test
phase
create & enhance
select & focus
start of the UP
end of the first complete pass of the UP
capture
shape
Figure 4.4 An example bar chart (Gantt chart) for planning the first pass of the UP shown in Figure 2.1.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Confronting challenging complexities 153
Figure 4.4 includes the current practice option of ‘linked bar chart’ dotted arrows used
to clarify precedence relationships, usually implicit a century ago. Precedence relationships
define what activities must follow or precede other activities and what can proceed in paral-
lel. For example, as suggested in Chapter 2 in slightly different terms, Figure 4.4 used as a
supplement to Figure 2.1 implies:
1 the select and focus (the process) phase cannot start until the capture (the context)
phase has started, but the first pass through the capture phase does not need to be
completed before moving on to start the select and focus phase;
2 the create and enhance (the plans) phase cannot start until the select and focus (the
process) phase has been started, but the first pass through the select and focus phase
does not need completing before moving on to start the create and enhance phase;
3 it is generally useful to take a flexible approach to proceeding with the first three phases,
employing frequent jumps among these phases and unplanned partial iterations until
initial plans have been created and enhanced, pursuing these three phases in parallel in
formal planning terms as assumed in the portrayal of Figure 4.4;
4 before starting to shape the plans, finishing a first complete pass through the capture,
select and focus, and create and enhance phases as an integrated trio is a sound practical
strategy;
5 testing the plans in the formal terms portrayed in Figure 4.4 has to follow shaping the
plans, although informal testing within each phase with potential loops back as a result
is clearly sensible;
6 interpreting the plans plus the test phase analysis of all relevant assumptions has
to be completed before looping back for a second pass through the process as a
whole; and
7 it is usually useful to plan to treat the fourth, fifth, and sixth phases in series, although exceptionally using loops back prior to the first complete pass may be convenient.
The ‘start- to- start’ and ‘finish- to- finish’ precedence relationships which Figure 4.4 uses for
the first three phases assume that these phases are pursued in parallel for formal planning
purposes, but in practice, an informal unplanned iterative structure within these three phases
is anticipated, as indicated less explicitly when discussing Figure 2.1 in Chapter 2. The
‘finish- to- start’ precedence relationships for the following phases imply a somewhat differ-
ent approach, also indicated when discussing Figure 2.1. Joint use of these different prec-
edence relationships as portrayed in Figure 4.4 is a very useful portrayal of the subtle nature
of the way the first pass of the process might be pursued which Figure 2.1 facilitates but does
not emphasise. Further passes are anticipated in Figure 4.4, building on earlier results, but
only the first pass has been formally planned.
It is very useful to see managing a UP or any SP developed via a UP based approach as
a form of project management. The seven phases of the UP of Figures 2.1 and 4.4 can be
visualised as project activities, and we can manage the UP phases as a project. If we see all
other processes in a comparable framework, as is the case in Chapters 6 and 7, we can gen-
eralise this idea. Indeed, the select and focus the process phase of the UP of Figure 2.1 was
a generalisation of the focus the process phase explored in Chapter 7.
Furthermore, it is useful to see all project management tools which have been explicitly geared to contexts involving significant uncertainty as directly relevant to process manage- ment, which means everything addressed in Chapter 7 of this book can be applied to man-
aging the UP associated with Chapter 2 in all its Part 2 roles.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
154 Foundations
One important implication of this is the seven Ws concept which is central to the Chap-
ter 7 process means we need to start the select and focus phase by asking questions about
who is involved in the analysis process, as well as who is involved in the context being
addressed, and what objectives are relevant from the point of view of all the key parties in
both groups. This has to define and shape the deliverables of the ‘project’ associated with
using a UP. Multiple objectives for a UP are inevitable, and aligning the objectives of mul-
tiple parties in both groups is crucial.
As an illustration of the subtle implications of the different interests of the team managing
the planning process and the much broader set of interests of all people involved in a planning
context, if I am hired as a consultant by a project manager to help him or her deal with a pro-
ject and their board, my mandate is clear, provided they are not engaged in strategic misrepre-
sentation. But this situation is different from being hired by a project manager’s board to deal
with the same project and its project manager, especially if the project manager is engaged in
strategic misrepresentation. And both may be very different from being hired by a regulator or
another third party concerned about what the organisation is planning to do via this particular
project. These kinds of differences may not matter in some contexts, but they can be crucial.
More generally, at the outset of the use of any UP driven process, a key issue is what are
the objectives of all the players involved in shaping the planning exercise, the wider set of
all the players involved in context of what is being analysed, and how can we best begin and
then evolve the analysis process to reflect all relevant goals with appropriate trade- offs at all
levels in all of the goals– plans relationship frameworks needing attention.
A key strategy for dealing with multiple objectives
A key strategy for dealing with multiple objectives is seeking an alignment of objectives for all
the interested parties which will avoid the need to address trade- offs in formal planning terms.
This avoids having to deal with trade- offs which are difficult to manage, and it may also help
to avoid failures of trust, gaming behaviour and serious breakdowns in relationships.
This might be seen as ‘resolving’ or ‘dissolving’ potential problems instead of ‘solving’
them, a useful aspect of some soft systems strategies. But a more helpful perspective is
avoiding creating problems in the first place. Seeking alignment of objectives at the outset
involves prevention instead of cure. Furthermore, it can be even more useful to see the
potential problems arising from a failure to align objectives as just the symptoms of a deeper
mess associated with a need for collaborative approaches as a very general aspect of seeking
opportunity efficiency.
Put simply, if crudely, if the broad basis of how a cake will be shared equitably can be
agreed on in advance, and everyone then concentrates on making the cake as big as pos-
sible, with no ‘free riders’, quibbling about the portions served later should less of an issue.
For this strategy to work, well- founded trust has to be an explicit early central goal, pursued
effectively throughout to achieve ongoing alignment of all the other goals. Variants of this
strategy are involved in all Part 2 tales.
Informal treatment of all objectives not addressed formally
Before finishing Part 1 and moving on to Part 2, it is worth briefly reflecting upon and
emphasising a particularly important implication of the EP framework noted earlier:
If any aspects of planning are not addressed in formal terms, they will need effective informal treatment;
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Confronting challenging complexities 155
otherwise, they will be overlooked, with no understanding of the implications; they will not go away, and if they are important this will matter.
One key role of the capability- culture concept is considering and dealing with the impli-
cations. Looking back to the last two sections, the way the UP of Figures 2.1 and 4.4 or
comparable equivalents are used clearly involves a lot of informal planning, and the quality
of that informal planning matters a great deal. That quality will depend on the capability and
culture of all the people involved in the process, not just those leading the process, although
the leadership provided is crucial.
Looking back to the photocopier tale, sometimes very simple formal treatment of what
can be easily structured, with some components measured, facilitates consideration of com-
plexity which matters even though that complexity is not worth measuring (like the relative
photocopier speed issue), or it is simply impossible to measure (like the photocopier style
and colour issue).
At a more general level, looking back to the multiple objectives discussion in a goal
programming framework, dealing with multiple objectives which need formal treatment can
clearly be very complex and exceedingly difficult. However, if the issues involved are not
treated formally in a planning framework that all interested parties understand and trust,
they are going to have to be treated informally by all the decision makers responsible for
both recommending and approving the choices. They are not going to go away because
they are ‘too difficult’. If they are not treated explicitly in formal or informal terms, they are
being ignored.
The nature of the ‘expertise’ required in many important decision- making areas includes
tactic knowledge and implicit decision- making skills which can be virtually impossible to
assess or test directly. However, if key issues are ‘left to the experts’, it may be very danger-
ous to assume that these experts have all the necessary expertise and their objectives are fully aligned with all other relevant parties. These assumptions certainly need testing as and when
this is feasible in whatever ways are feasible.
A ‘menu view’ of decision- making process goals
A menu view of decision- making process goals when using any process based on the UP
of Figures 2.1 and 4.4 is the concern of this final section in Chapter 4. Its role is a form of
synthesis of all of the EP foundation concepts developed in Part 1 which you may find useful
as a basis for starting Part 2.
The 1980s public seminars which I provided of the kind the director of engineering of
National Power attended to learn about SCERT based approaches concluded with a slide
which listed key ‘benefits’ if this kind of approach to planning was used. By the mid- 1990s,
based on the evolving thinking behind the first edition of Project Risk Management (Chap- man and Ward, 1997), this list of benefits was reframed as a much richer list of ‘potential
objectives’. The ‘menu of potential goals’ approach used in this section further generalises
in the EP framework of this book.
Using an EP goals– plans relationship structure as a useful starting point framework, the
potential goals of any organisation need interpretation as a menu in the sense that if an
organisation adopts the SP for projects of Chapter 7 or any of the other applications of
the UP illustrated in Part 2, it is important to consciously choose to emphasise or not the
benefits which the particular context requires on each application occasion, treating the
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
156 Foundations
benefits of particular interest as explicit key goals. The achievement of all other goals should
be treated as useful spin- offs to the extent that they come at no extra cost (they have a zero
shadow cost). They are part of the bundled product which as a whole suits the organisation
in this context. They may need to become important goals to be explicitly pursued on later
passes, but they are not an explicit priority on the early passes of an analysis of what changes
need to be made to how the organisation operates.
This is in part a multiple goals perspective on what the ‘focus phase’ prototype of the
more general UP select and focus the process phase has always been about, but greater clar-
ity and a wider scope have emerged over time.
The key potential goals for all variants of decision- making processes grounded on a
UP are:
1 clarity, risk and opportunity efficiency in the sense is developed so far;
2 associated and additional culture changes are geared to the context;
3 the ABCs of targets and expectations are clarified and extended;
4 communication in overall terms is emphasised in ways which suit the context;
5 bottom- up communication is encouraged and facilitated when appropriate;
6 creativity and satisfaction (fun) are used as effective objectives;
7 corporate learning and knowledge are made explicit;
8 better linkages exist among operations, project, and corporate management; and
9 an opportunity management appetite is stimulated via opportunity efficiency.
All Part 2 chapters involve building on the point 1 concepts outlined so far, plus all issues
associated with points 2 through 9, further enhanced by an accumulated understanding of
all aspects of opportunity efficiency.
Before we conclude Part 1, looking at points 2 through 9 on a point- by- point basis may
enhance your understanding of where this book is going.
Associated and additional culture changes are geared to the context
Moving a corporate culture towards taking more risk when that is appropriate as discussed in
Chapter 3 in terms of the IBM Forum 2 programme is a key issue addressed in Chapter 6.
But many other culture changes are addressed in Chapter 6, and all other Part 2 chapters
illustrate further corporate culture changes which may be relevant to your organisation. For
example, the culture in the organisation involved in Chapter 5 was grounded on the premise
‘we are a small manufacturing business in a particular sector’. Collective use of a UP concept
by all relevant managers tested the implicit working assumptions involved, and began to
create a new culture which generated plans based on much broader premises, like ‘we might
become a marketing organisation in a bigger sector, manufacturing only some of what we
sell, although that might involve increasing current production capacity as well as marketing
a wider range of products with some products manufactured by other organisations’.
The ABCs of targets and expectations are clarified and extended
In Chapter 3 William was encouraged to understand why it was important to understand the
basic ABCs of targets and distinguish between aspirational targets, balanced targets equal
to expected values, and commitment targets. Expected values that might or might not be
conditional were also explored. Some of these ideas were further developed in the discussion
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Confronting challenging complexities 157
with Peter, in particular, relating them to the implications of formal as well as informal con-
tingency planning.
Feedback from BP project managers I worked with in the 1970s and 1980s, and many
other client groups since, suggests that in all comparable contexts the implications can be of
enormous importance, because of the improvement in clarity of communication plus well-
founded trust between people charged with doing work at various levels. The implications
include changes in governance processes at all relevant levels built on the resulting mutual
understanding and trust. These lessons are directly embedded in the tale of Chapter 7 and
indirectly in all the other Part 2 chapters.
In Chapter 8 the balanced target and expected value distinction is explored, in terms of
when to plan to have a new electricity generation capability available if the ‘penalty function’
associated with being early or late is asymmetric. The ideas involved are often useful even in
very simple contexts, sometimes involving purely informal planning, like adding half an hour
or more to the expected time to get to an airport to catch a flight you do not want to miss.
Clarity about the ABCs of targets and expectations can be simple in other contexts in
different ways, or much more complex, and dealing effectively with both can be very impor-
tant. For example, if someone you report to asks how much something will cost and you
know that a simple one value estimate is all they will listen to, adding an appropriate level
of contingency to the expected value to define a commitment target is the only sensible
response. But in the very different and complex safety and security assessment context of
Chapter 9 a more sophisticated approach is required. Everyone involved in safety and secu-
rity management contexts involving potential low probability but high impact incidents
needs to understand that zero fatalities are always an appropriate aspirational target, the
actual level of fatalities in any given year should usually be well below the expected level of fatalities because of the asymmetric distributions involved, and the commitments made
to customers, the public, and regulators need to be very carefully thought through and
presented. Simply adding a contingency allowance to the expected outcome as a published
‘commitment target’ for safety in this kind of context is not a sensible idea, failing to be
honest or looking inept can be extremely serious, and muddled thinking about this issue can
be directly instrumental in corporate failure.
Communication in overall terms is emphasised in ways which suit the context
Communication has a crucial role in all planning, and fully developing all relevant aspects
of communication is an important EP concern which has to be at least partially driven
by the context. Communication between project managers and the board within BP was
clearly enhanced significantly by the use of tools like Figures 3.5 through 3.8, 4.1 and 4.2.
Communication facilitated by EP tools in this way will be emphasised in all Part 2 tales.
Often other organisations are involved. Usually what needs communication is reasonably
straightforward to deal with if a complete EP toolset is available, with tools discussed in
Part 1 playing a central role. But issues driven by capability- culture concerns which are
complex and somewhat different need particular attention in the later portions of the tale of
Chapter 8. Communication by the electricity utility involved with customers, contractors,
and potential collaborators aimed at transforming the fundamentals of the business model
become central. Communication about ethical issues becomes central in Chapter 9. When
communication beyond the usual bounds of corporate communication is an issue it may be
exceptionally important, in extreme cases a corporate survival issue.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
158 Foundations
Bottom- up communication is encouraged and facilitated when appropriate
In the 1980s, one Acres consultancy assignment for Gulf Canada involved being contracted
to provide advice on an appropriate process for estimating the cost of the Hibernia project.
This project involved a planned offshore ‘gravity’ platform for the production of oil off
the east coast of Canada, in a sea area frequented by icebergs. In the process of sizing the
uncertainty sources in prototype minimum clarity terms to think about how a more detailed
cost estimation process might work, I explored broader context issues by talking to a range
of Gulf personnel on an informal basis as part of a more general exploration of the context.
My conversations with oil reservoir engineers and oil market experts led me to the view
that given what all the relevant Gulf staff knew at that time about oil volume uncertainty,
plus oil price uncertainty when the oil came to market, plus cost uncertainty for the gravity
structure, Gulf really ought to look more closely at a ‘ship- shape’ alternative to the gravity
platform, or drill more exploratory wells before proceeding.
The gravity structure approach involved relatively high capital cost coupled to low oper-
ating cost, high safety, and high oil production output levels. It also needed East Coast
Canadian labour to construct it (with coupled political and other benefits). These were
key reasons why it was preferred. The ship- shape approach might use an existing oil tanker
to provide a floating oil production facility over a subsea well. Icebergs would be avoided
by removing the ship- shape, interrupting production, with revenue effects involving con-
tractual implications for prices as well as oil production volume impacts. Icebergs would
be managed via high- powered tugs when needed if the gravity structure was used, having
designed the gravity structure to cope with most of the icebergs anticipated.
The problem with proceeding with the gravity structure without drilling more explora-
tory wells was a scenario involving a plausible minimum oil volume plus a plausible mini- mum oil price when the oil came to market plus a plausible maximum capital cost for the gravity structure. Current plan assumptions meant that the financial implications of this
scenario involving three concurrent downside variability aspects needed explicit attention,
and prudence was likely to suggest reducing the risk associated with this scenario.
This was not the advice contracted for, and it was not what the senior Gulf managers
I reported to wanted to hear. But they acted on it, and their actions proved important. The
advice provided was based on letting the members of senior management know what their
own staff knew in a structured form they could act on, a form of bottom- up communica-
tion which can be invaluable. In principle, organisations should not need consultants to tell
them what they already know, although in practice they often do, sometimes simply because
communication structures linking different concerns addressed by different people are not
in place.
The key message to take away from this story is organisations need to both encourage and
facilitate structured bottom- up communication which may not take place unless a systematic
approach to this issue is an explicit part of the agenda, a capability- culture concern with vari-
ous degrees of importance in each Part 2 chapter. Its importance, and the need to address it
explicitly, was a key message I took away from many of my Acres projects.
Creativity and satisfaction (fun) are used as effective objectives
One of the reasons I greatly enjoyed working with Acres in Canada and the US was the
ethos of creativity and ‘fun’ in the sense of working with interesting people doing interest-
ing things that mattered. It was a conscious and explicit part of the Acres ‘business model’,
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Confronting challenging complexities 159
explicitly articulated by Oskar Sigvaldason, my head of department while I worked with
Acres. Its purpose was attracting and keeping good people who would help to generate
and deliver high quality work. The explicit belief was this would generate immediate profit
plus more interesting work, which, in turn, would attract more good people and interesting
work, an explicit use of the notion of seeking ‘virtuous circles’ and avoiding ‘vicious circles’.
If your organisation wants to use this approach it needs to be designed into the approaches
employed. An important agenda item for all those engaged in designing decision- making
processes, or managing the ‘select and focus the process’ phase equivalent, is keeping the
processes challenging and interesting to attract and keep good people who are capable of
doing a good job.
The quality of the people attracted and retained may matter much more than anything
else which is within the control of the organisation’s management decision makers. This idea
is in the background, not up front or explicit, in all Part 2 chapters, but it matters greatly. At
a what needs to be done level it is almost always important, and strategic clarity about the
basis of how to do it effectively is a crucial senior management capability.
Corporate learning and knowledge capture are made explicit
The first live SCERT application for a BP project took about six months, with a full- time BP
facilitator/modeller, my help about half time, plus significant input from the project team,
and this built upon an extensive earlier process development stage revisiting an earlier pro-
ject. Within about two years, comparable studies were taking about 6 weeks, with no direct
involvement from me.
Part of this change was going down a learning curve in terms of process development and
embedding process implementation skills. BP recognised the importance of this by promot-
ing the facilitator/modeller and keeping him in this role with managerial responsibility for
a growing team servicing a rapidly growing portfolio of projects and then restructuring his
team plus the estimating and planning groups as a whole.
Another part of this change was all the staff involved benefitting from the way cor-
porate knowledge about offshore oil and gas projects was captured. As one of the BP
programme managers put it, they did not have to ‘keep reinventing the wheel’. This, too,
was recognised by BP on a widespread basis and enhanced by standards associated with
the documentation of work in progress as well as reviews. For example, when a major
offshore project was decomposed into about 30 activities, like ‘laying the pipeline’, ‘build-
ing the platform’, and so on, each activity was numbered and concisely documented so
everyone involved could refer to a common understanding of what was included and what
was excluded. Each specific source of uncertainty treated separately was numbered within
the activity, as in 3.2 for the second source in the third activity, and associated with a
paragraph or more concisely indicating what was included and excluded. Each response to
these primary sources was numbered within this structure, as in 3.2.1 for the first specific
response to the second source within the third activity. Secondary sources of uncertainty
and responses continued this numbering system hierarchy. This word- processed text form
of information was kept updated and accessible. Together with the graphical associated
information using Figure 4.2 source- response diagram formats, plus later graphical sensi-
tivity diagrams and decision diagrams, it provided an effective audit trail of how and why
decisions were made. All this information also provided a valuable knowledge capture facil-
ity, useful for new staff joining an ongoing project as well as staff beginning a new project
with some common features.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
160 Foundations
Expertise and learning are very valuable assets for organisations, an important part of their
capability- culture asset set. Opportunity efficiency includes capturing this knowledge and
making ongoing use of it in the most effective and efficient manner possible.
Better linkages exist among operations, project and corporate management
Better linkages among operations management concerns, project management concerns and
corporate strategy concerns ought to be a clear and explicit goal for all aspects of enlight-
ened planning, a core aspect of the opportunity efficiency concept which opportunity man-
agement seeks. All Part 2 chapters demonstrate in outline what this implies.
An opportunity management appetite is stimulated via opportunity efficiency
The ultimate goal of this book as a whole is an enhanced appetite for effective and efficient
opportunity management. This should include aspects like encouraging creativity and the
use of all feasible approaches to a better capability- culture. Its achievement should involve
seeking opportunity efficiency in the broadest sense as a basis for best practice as a corporate
goal, defining this goal with as much clarity as possible. Each Part 2 chapter indicates in
outline what this might involve in the context of the tale being told, building on the Part
1 foundations. Opportunity efficiency is an operation form or statement of best practice,
and while opportunity efficiency may be seen as an ambiguous concept which is only slowly
beginning to acquire clarity as this book finishes Part 1, best practice without this kind of
opportunity efficiency basis is much more ambiguous and much too limited in ambition.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
5 Using a UP – an initially simple supply chain management example
Pass 1 – initiating the process
Pass 1 – creating the initial model to begin shaping the plan
Pass 1 – testing the plans and interpreting the plans
Pass 2 – refining the analysis by considering stock- outs
Pass 3 – reconsidering all relevant influences on holding costs
Pass 3 – exploiting creativity to interpret an emerging plan 3
Pass 4 – Bob’s crucial clarification of the plans/models distinction
Further creative testing of capture the context assumptions
Creative testing of all the UP front-end working assumptions
Looking wider or deeper at key strategic and tactical choices
Consolidating key concepts
Future relevance of this chapter’s tale
6 Building ‘specific processes’ – a bidding process example
Capturing some of the context information to get started
A preliminary proposal document
The initial process overview – early development Workshop presentation of the Astro SP for bidding overview
Initial observations on Martha’s rationale
Stage 1 – identifying non-starters plus ongoing concerns
Stage 2 – identifying no- hopers
The rationale for separate stages – further comments now
Stage 3 overview – bidding strategy and tactics development
Phase 3.1 – initiation of stage 3 on pass 1
Phase 3.2 – enhanced cost uncertainty analysis on pass 1
Phase 3.3 – enhanced probability of winning analysis on pass 1
Phase 3.4 – enhanced uncertainty syntheses on pass 1
Phase 3.5 – bid evaluation and development on pass 1
Trevor’s transformative insights
Trevor takes up the leadership role which Martha planned
Key steps in Martha’s approach to change management
Wider impacts on Astro Wales
The credibility of this tale and linked issues
Synthesis and reflections at an overview level
Part 2
Employing planning tools in practice – five illustrative tales
Chapters 5 to 9 contents indicating sections within chapters
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
162 Employing planning tools in practice
7 Adapting ‘generic processes’ – a project planning example
Generic processes used for project planning
The need to clarify ambiguity
The context and characters of this chapter’s tale
The role of project management within WSL
The role of the Projects Group within WSL
The new structure of the WSL Projects Group
Basic problems the new structure would address
A project lifecycle framework
The seven Ws framework
A goals– plans relationships framework
A specific process for the strategy progress stages
A specific process for the strategy gateway stages
Paul’s overall change management strategy – initial goals
The discount rate used to shape and select projects
Regulation risk – privatisation and nationalisation risk concerns
Technical workshops plus other parallel activities
Beyond the technical workshops – key corporate clarity concerns
Enlightened approaches to governance, teams and contracts
Paul’s overall change management strategy – target 1
Paul’s overall change management strategy – target 2 and beyond
Key credibility issues for this chapter
Some further implications of clarity inefficient toolsets
8 Corporate strategy formulation – an electricity utility example
Carl’s situation and all the key characters in this tale
Corporate planning horizons and associated planning categories
Long- term planning
Medium- term planning and integration upwards
Short- term planning and integration upwards
An overview of integrated short/medium/long- term planning
Futures planning of all kinds and its integration at all levels
Goals planning as a starting point for all planning
Goals planning with a capability- culture focus
Current practice compared to Carl’s proposed approach so far
Organisational separability for integrating corporate planning
Limited inferences instead of further developments for this tale
A brief exploration of the implications of context differences
9 Building well- founded trust about complex concerns – a railway safety example
Context and main characters
Generalised foundations for NER safety and security
A key board level strategic issue needing clarity
The 1970s Ford Pinto controversy in the US
Using numerical examples to clarify the board’s understanding
Building and using the rest of the base- case model set
A change in culture or ethos
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Employing planning tools in practice 163
Using risk efficiency as a basis for option choices
Enlightened prudence and the board’s leadership role
Sensitivity diagram portrayal of option costs
A process for clarifying NER revealed preferences
The core revealed preference issues for NER
Communication and capability issues for NER
Enlightened regulation and risk sharing
Ambiguity which may be resolvable at least in part
Sophie’s closing remarks
Biases, blind spots and bounds on formal analysis
Earlier examples of related safety and security analysis
Sophie’s views on politics and personal ethics
Public /private sector issues with regulator implications
Integration of safety planning and other corporate planning
Applications in other contexts when well- founded trust matters
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
This chapter’s tale begins by addressing what looks like a simple economic order quantity
analysis of a particular input component for a manufacturing process. But the initial analysis
soon evolves into the exploration of integrated supply chain management involving collabo-
rative contracts with other organisations, setting up new organisations, and then considering
integrated demand chain management and further broad corporate strategy issues. Explor-
ing complexity in a clarity efficient manner from a simple starting point can be an effective
route into complex strategic thinking, as illustrated in this chapter.
This chapter’s tale is set in a firm which manufactures lawnmowers – The Lawnmower
Company, contracted to TLC. The central characters were
George, general manager;
Nicola, a niece of George’s who had just started work at TLC;
Pete, production manager;
Ajit, accountant;
Bob, buyer;
Surinder, sales manager; and
Dave, driver of a TLC van.
Your perspective may lead you to identify with one or two particular characters more than
the others. This is not a problem, but a crucial feature of the capability culture of TLC was
the way all these individuals worked together as a team.
The historical setting is the 1980s. However, all the discussion is relevant for the 2020s
and beyond. Why this is the case is addressed towards the chapter’s end.
George initiated a simple planning process in an operations planning context by asking
Nicola to analyse the order quantity for an input component for the manufacturing process.
Nicola used a textbook variant of the OR process, but TLC’s evolving understanding of the
way this process could be used will be considered from a UP perspective throughout.
You may not be interested in manufacturing companies, or the inventory order quantity
model initiating discussion, or direct use of a UP from an analyst’s perspective. Inventories
are examples of a much more general buffer concept which most target readers need to
understand at an overview ‘how to do it’ level for strategic clarity. However, the key reason
you need to follow the general flow of the approach to analysis employed as this chapter’s
tale unfolds is to acquire strategic clarity about the way the UP concept can be used as a
basic component of an EP approach in any context.
This chapter’s tale provides a useful basis for developing aspects of tactical clarity
which need to be understood to achieve strategic clarity in all decision- making areas for
5 Using a UP – an initially simple supply chain management example
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Using a UP 165
all commercial organisations. There are several reasons which are worth understanding at
the outset.
First, it is a conveniently simple context for beginning to explore the UP concept from
an EP perspective, illustrating features of the EP perspective as well as the UP. You need
a broad appreciation of what can be achieved and what is involved in successful very basic
applications of the UP concept from an EP perspective to understand what this book is
about.
Second, the way the tale develops provides an illustration of some key features of the
UP concept which apply to its use in more sophisticated forms as illustrated in all subse-
quent chapters. In particular, it illustrates defining characteristics of the process as a whole,
the central role of its iterative nature, the role of models as formal abstractions of a more
complex reality, the need to understand broader plans which go beyond the limitations of
associated models, the need to understand and test key context assumptions, the importance
of shared understanding of goals– plans relationships, the implications of a very broad and
systematic approach to all robustness testing, the way generalising working assumptions can
be used to test robustness and create new insights, the importance of creative thinking, an
overview of UP process differences relative to more limited textbook interpretations of OR
processes, and some basic operational aspects of the nature of the relationships among clar-
ity efficiency, risk efficiency, and opportunity efficiency.
Third, the tale demonstrates the way simple common practice generic models can help
us to formalise and develop plans, and how using a UP to make these simple models more
sophisticated in a clarity efficient manner can deepen our understanding by reducing the
ambiguity uncertainty associated with assumptions, including unknown unknowns.
Fourth, the iteration issues explored include using successive passes through a UP to seek
a progressively wider or deeper perspective, depending on which is most relevant at that
point in the analysis. This ‘looking wider or deeper’ when managing the iteration process,
sometimes usefully characterised as ‘expansive strategic exploration of complexity or deeper
tactical exploration of complexity’, involves an important change in direction. Recognis-
ing when analysis priorities change in a way requiring a change in analysis direction can be
crucial.
Early sections of this chapter follow the multiple pass iteration structure of the UP out-
lined in Chapters 2 and revisited in Chapter 4 using Figures 2.1 and 4.4. Initiating the first
pass of the process is discussed in the next section, which covers concerns addressed by the
first three phases in an indirect manner. The fourth phase, the ‘shape the plans’ phase, is
addressed in the following section. ‘Test the plans’ and ‘interpret the plans’ resulting from
this initial first pass analysis suggest a need to refine the analysis by generalising key assump-
tions, leading to a second pass. This second pass then prompts a third pass, which further
redefines some of the initial modelling assumptions in a significant manner. Creative inter-
pretation of the implications, followed by a fourth pass, then leads to a much broader view
of the plans involved. Subsequent sections extend the discussion to consider wider perspec-
tives before drawing together what all readers should find useful about this chapter’s tale.
The crucial working assumptions which must hold for this chapter’s tale to work as
intended include your finding it credible that an organisation like TLC could have the
requisite relevant capabilities for effective use of UP variants after Nicola’s arrival without
it having occurred to anyone to identify the potential opportunities explored by this tale
prior to Nicola’s arrival; that Nicola was capable enough to transform her understanding of
textbook concepts into a practical approach and lucky enough for a good fit between the
theory she happens to be aware of and the practice requirements of the context; that Bob,
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
166 Employing planning tools in practice
George, Surinder, and Pete were all individually capable enough to use Nicola’s explanation
of her first three passes to make immediate use of the UP concept in a manner appropriate
to their TLC roles; and Ajit was capable enough to understand what he needed to know
about a UP from Nicola’s explanation of her first three passes to make immediate use of it
in terms of somewhat different top- down thinking about corporate planning in a way which
complemented the focus of the others. If these working assumptions stretch your credibil-
ity you are simply being reasonable and realistic. However, this stretch is in a good cause
because it maximises the scope to communicate what effective use of a UP can achieve.
Please approach the credibility of the tale with this in mind. The closing sections discuss
plausibility issues and the future relevance of this chapter’s ideas in other organisations.
For clarity the language of EP developed in Part 1 is used whenever this is helpful in this
and all other Part 2 tales, translating from the language usage you might expect in the actual
context of the tales as well as during asides linking back to Part 1 discussion. The past tense
is used for describing the evolving scene for all the tales and observations by the characters
within the context of the tales, with the present tense used to distinguish asides and interjec-
tions related to wider implications and applications in other contexts.
Pass 1 – initiating the process
Nicola was a central character in this chapter’s tale. Nicola completed a degree in mechanical
engineering in the 1980s. She undertook a year of travel between university and beginning
her career to see the world while she still had the freedom from responsibilities. She was
not lacking in ambition or talent, but she was uncertain about her career plans and wanted
to think about them carefully after her year of travelling. When Nicola returned home, she
took her mother’s long- standing advice and asked her uncle George for a job, making her
current career indecision clear to George to avoid misleading him.
Nicola’s uncle George was the general manager of TLC, a traditional family business of
modest size and success which manufactured a limited range of lawnmowers in the UK’s
industrial Midlands, in Birmingham. George had run TLC since his father retired, having
previously worked in various roles in TLC and in other similar manufacturing companies.
George hired Nicola because she was his niece and his sister was a silent partner in the busi-
ness. During the first three days, Nicola worked for George without a clearly defined role,
‘learning the ropes’ and doing odd jobs.
On day four of her new job, Nicola was accompanying her uncle as he walked around
the premises. George was observing what was going on and discussing issues which people
wanted to raise. George was approached by Dave, who appeared somewhat agitated. Dave
used a large van to pick up input components for the production process from local sup-
pliers. Dave also made some of the local deliveries of finished lawnmowers. Most finished
lawnmowers were shipped to customers via a range of delivery services, according to order
size and destination location. Inputs which were bulky, or came from a distance, were deliv-
ered by the supplier.
Dave was a valued TLC employee for several good reasons, but he was not very socially
adept. About a month before Nicola’s arrival, Dave had offended the dispatch staff of the
supplier of engine speed controls which were used for all the firm’s lawnmowers by com-
plaining about the quality of the tea and biscuits they offered him. Their response when
Dave returned for another order about the time of Nicola’s arrival was to stop giving him
any refreshment. After brooding about this, Dave resolved to try to persuade George that
less frequent trips to the engine speed control supplier would be a good idea.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Using a UP 167
With Nicola listening in, Dave suggested to George that it was silly for him to pick
up the standard order of 2,000 controls about once a month. His van could cope with
about 12,000 every six months. The cost each trip incurred in terms of his time plus wear
and tear on the van and fuel could be reduced if he only went once every six months,
and some of the time saved might be used to deliver more of the finished lawnmowers,
saving external delivery service charges which were expensive compared to the cost of
sending him.
George was aware that Pete, his production manager, had determined the order size of
2,000 in consultation with Bob. Bob was his ‘buyer’, responsible for the supply of all pro-
duction components, along with other duties. George was also aware that Pete would prob-
ably not mind larger orders of these relatively small components, and could probably find
space for them. However, George knew that Ajit, his accountant, would object to the extra
funds tied up in larger inventories unless it could be demonstrated that overall costs would
be reduced, with good reason from George’s perspective. George explained these concerns
to Dave and Nicola and suggested that it might be useful if Nicola undertook an analysis
to test Dave’s suggestion. Dave had made an issue of the order size which needed resolu-
tion. Although George’s intuition suggested that 2,000 was about right, and he knew Dave
well enough to suspect the kind of ulterior motives that lay behind Dave’s suggestion; his
approach to all TLC employees meant that if he did not take Dave’s advice he wanted Dave
to understand why. He also thought this might be an interesting challenge for Nicola. He
knew she had studied some management as part of her engineering degree, and he wanted
to test her capabilities. He also wanted to stretch her and give her an opportunity to feel she
might be a useful contributor to TLC.
At this stage nobody questioned the assumption ‘the plan we are interested in is an opti-
mal order quantity’, and there was no obvious need to involve Ajit, Bob or Pete beyond
letting them know what Nicola had been asked to do, indicating they would be involved if
anything interesting emerged, and asking them to help Nicola if called upon. This can be
interpreted as an implicit treatment of all three front- end UP phases in terms of accepting
the status quo, assumptions questioned later but not initially.
As part of her engineering degree Nicola had taken an introductory management course.
The material covered by this course later led to the textbook Management for Engineers (Chapman, Cooper and Page, 1987), which provides a broad coverage of a range of topics,
including an introduction to uncertainty and complexity management processes using basic
inventory generic models and a textbook 1980s OR process variant of the UP concept.
Nicola was aware that these common practice generic inventory models were deliberate
abstractions of reality which could be used to capture the core issues of the relevant deci-
sions in a useful manner, analogous to a map being useful to plan a walk or a car journey.
She also understood the role of OR processes comparable to a UP at the ‘hard tactical
analysis’ end of the spectrum of possible UP approaches in an operations planning tactical
decision- making context. But she had never used this kind of model or process, she had no
understanding of alternatives, and she had a very limited understanding of the more strate-
gic aspects of using bottom- up or top- down approaches to planning.
Nicola spent several days talking with people, observing what went on, and developing
her ideas about modelling the issues associated with an optimal order quantity model. This
confirmed the desirability of a multiple pass iterative analysis which would consider the costs
associated with reordering and holding inventory given different order levels, starting with
the simplest available model and adding more complexity in the process of testing earlier
working assumptions by exploring alternatives.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
168 Employing planning tools in practice
Nicola then prepared a presentation for George which is described in the following sec-
tions. She made the presentation in a manner which reflected the iterative nature of her
deliberations, as if she were going through the process with George, eliminating the unpro-
ductive blind alleys she had explored along the way and glossing over the incidentals but
incorporating educational initial oversimplifications. She acknowledged her management
course sources, but she presented the analysis as ‘her creation shared with George’.
Pass 1 – creating the initial model to begin shaping the plan
Nicola’s chosen formal modelling approach started with the simplest inventory control
model she was aware of, the most basic form of the Economic Order Quantity (EOQ)
model, with closely linked Periodic Review variants, described in most management text-
books that deal with modelling inventory issues.
Constructing her model began by identifying a variable which would be used to define
her first pass ‘base plan’, referred to as a ‘decision variable’ in the language of this kind of OR
modelling. A ‘primary decision variable’ was the OR modelling terminology for what had to
be decided in a direct manner, the ‘order quantity’ Qo. In EP terms, Qo can be viewed as a
core component of a very simple goals– plans relationships framework which TLC needed to
explore to determine the base plan defined by Q0*, the ‘optimal value of the order quantity’.
Nicola was aware that looking for Q0* involved an important basic working assumption –
it was reasonable to assume that asking ‘What should the optimal order quantity be?’ was
a sensible question to begin with. However, she could see no reason not to start on this
basis. George had asked her to question the size of the current fixed order quantity of 2,000
controls, but not the fixed order quantity assumption itself.
Key parameters that would influence the optimal order quantity were identified next.
One obviously important parameter was the ‘demand rate’ for controls, Rd. As a simplify-
ing working assumption, Rd was assumed to be a constant, implying a stable monthly total
output of lawnmowers of all types (they all used the same control). Nicola estimated Rd at
about 2,000 controls per month, as indicated by Dave and confirmed by Pete.
The decision of direct interest was the optimal order quantity, but given a constant
demand rate assumption, a Qo value implied indirect determination of an ‘order interval’ To
value, a ‘secondary decision variable’ which had an implied optimal value linked to Qo * . To
was defined in terms of Ro and Qo by the ‘identity’ constraint
T = Q / R .o o d
Nicola then focused on the identification of an ‘objective function’. The objective function
had to define what was to be maximised or minimised in terms of the decision variables
and associated parameters subject to any relevant constraints. To start to identify a suit-
able objective function, Nicola defined the ‘order cost’ parameter Co as ‘all costs which are
incurred on a per order basis’.
Co posed problems in terms of measurement precision. When Nicola consulted Ajit he
suggested that Co could be as low as £20 during the mid- winter period: about £15 for
Dave’s time, vehicle wear and tear plus fuel, and a further £5 to cover the cost of processing
the associated invoice. During the winter Dave and his van had periods of idle time, staff
processing the associated invoices also had spare time, and only basic salary and marginal
vehicle operating costs needed consideration. But Co could be as much as £40 in the spring
when significant opportunity costs or overtime/extra staff costs were involved, say, an extra
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Using a UP 169
£18 opportunity cost for Dave and his van, plus an extra £2 to cover the cost of processing
the associated invoice using agency staff. In the spring the upturn in demand for gardening
equipment meant Dave working overtime and TLC still having to use expensive external
delivery services for low volume local deliveries which Dave could deal with given more
available hours. In addition, the spring rush of orders meant inexperienced agency account-
ing staff had to be taken on and managed carefully to avoid mistakes.
For her first- pass analysis Nicola decided to use (£20 + £40)/2 = £30, an average value Co
estimate. Ajit’s £20 and £40 estimates were treated as plausible minimum and maximum values
and a mid- range £30 was equated to a Figure 3.1 minimum clarity estimate of the expected
value. However, systemic uncertainty associated with an annual cycle was clearly involved in
addition to other variability uncertainty which might not be systematic. This systematic uncer-
tainty had implications which might be exploited by a more sophisticated analysis. Nicola
planned to return to the implications of this systematic uncertainty and associated extremes
later, if these variations seemed relevant, and she pointed this out to George when explaining
her thinking. The seasonal variations underlying the Co = £30 value were a form of variability
uncertainty which was largely predictable, but that predictability was not being addressed
by TLC’s current fixed order quantity approach, and it was not addressed by Nicola’s pass 1
model. Its presence suggested that any optimal order quantity should perhaps be a function of
the time of year, with bigger orders in the spring and smaller orders in the winter. This kind of
policy would clearly introduce a significant complication involving complexity which Nicola
prudently wanted to avoid for the time being, but it was important to be aware of the issue.
Nicola next defined the ‘holding cost’ parameter Ch as ‘all costs per control (item of
inventory) per unit time (using months)’. Ajit had suggested Ch was about £0.06 per con-
trol per month, involving two components. Assuming a value per control of about £3 and
an average cost of capital of about 20% per annum, £0.05 per control per month was the
approximate cost of capital component. Assuming other holding costs of the order of £0.01
per control per month (an allowance for space, security, insurance, and pilfering) seemed
about right to Ajit. Pete had confirmed that space was not a problem, and he could think of
no other holding costs. Nicola indicated that Ch was subject to the cost of capital variations
which Ajit associated with pressure on working capital in the spring when lawnmower inven-
tories peaked and at other times of the year if new equipment had to be purchased. They
were less important than seasonal variations associated with Co because they were smaller
and less systematic, so she had relied on Ajit’s 20% estimate basis for Ch and not bothered to
use a range- based approach to estimating the cost of capital component.
Having explained these variables and parameters, Nicola then showed George Figure 5.1.
inventory
level in
controls (items)
time in months
slope - Rd Qo
To
Figure 5.1 Order cycle relationships.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
170 Employing planning tools in practice
Nicola explained that Figure 5.1 was a simple graphical portrayal of a typical order cycle.
Figure 5.1 indicated working assumptions already noted, like a constant demand rate Rd,
and further working assumptions, like no planned shortages.
In addition to assuming no planned shortages, she was currently assuming no unplanned shortages, both assumptions she would return to later. Pete had indicated that Dave usually
responded to an order for more speed controls the next day, and never took more than three
days, so Pete used a ‘safety stock’ of three days of production, ordering more controls three
days before he ran out of stock. This safety stock was not part of her pass 1 model, but it was
part of Pete’s plan, and it made the ‘no unplanned stock- outs’ assumptions reasonably robust.
Nicola then defined the composite parameter Cc as the ‘joint cost per order cycle’, where
C = C + C T Q / 2c o h o o ,
and used Figure 5.1 to explain the obvious Co term (one order cost per order cycle) and
the less obvious Ch term. Ch was the cost per control per month. Ch had to be multi-
plied by the area of the triangle (the base times the height divided by 2), because this area
defined the number of ‘control months’ (one engine speed control held in stock for one
month = one control month).
Nicola explained that the joint cost per cycle defined by Cc was a convenient place to start looking at joint order cost and inventory holding cost because it related clearly to the cycle
portrayed by Figure 5.1 as just explained. However, of more interest was Ct, the ‘joint cost
per unit time (month)’, defined by dividing her expression for Cc by To to obtain
C = C / T + C Q / 2t o o h o .
To reduce the two interdependent decision variables to one independent decision variable,
Nicola then substituted the identity defining To into the Ct expression to obtain the form of
the Ct expression used as an objective function to define the EOQ model,
C = C R / Q + C Q / 2t o d o h o .
The Periodic Review variant of the EOQ model uses To instead of Qo as the primary deci-
sion variable, reversing the substitution. In either case the only constraint is employed to
transform the objective function into a function of only one variable, facilitating a direct and
simple approach to optimisation using a table, graph or calculus.
Nicola then explained that the EOQ model assumed they wanted to minimise Ct, and
they could visualise doing this using the portrayals of Table 5.1 and Figure 5.2.
Table 5.1 and Figure 5.2 were complementary sensitivity analysis portrayals, showing
Ct*and Q0* plus the components driving the overall sensitivity of Ct as Qo changes. They
both clearly indicated the overall relationships involved, and the role of the two component
contributions to the EOQ objective function defined by Ct as Qo varies above and below Q0*.
The holding cost component of Ct was a linear function of Qo. It was defined by Ch Qo /2,
with Ch = 0.06. If Qo was equal to 500 controls, then Ch Qo /2 = £0.06 × 500/2 = £15. If
Qo approached zero, then ChQo /2 approached zero. As Qo increased, ChQo /2 increased
by £0.06 × 500/2 = £15 for each successive 500 controls. Table 5.1 and Figure 5.2 both
showed this clearly, and employing them together provided useful clarity.
The order cost component of Ct was a non- linear function of Qo, defined by CoRd /Qo,
with Co = 30 and Rd = 2000. Each time Qo was doubled, the order cost component was
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Using a UP 171
Table 5.1 Joint cost per month and its components as the order quantity changes.
Order quantity Q o 500 1000 1500 2000 2500 3000
Order cost component Co Rd /Qo 120 60 40 30 24 20 Holding cost component Ch Qo /2 15 30 45 60 75 90 Joint cost per month Ct 135 90 85 90 99 110
cost in £
joint cost Ct
Ct
Qo order quantity in controls
1,000 2,000 3,000
0
50
100
Qo*
* holding cost Ch Qo/2
order cost Co Rd/Qo
0
Figure 5.2 Joint cost per month and its components as the order quantity changes.
halved, and each time Qo was halved, the order cost component was doubled. Table 5.1 and
Figure 5.2 both showed this clearly.
In addition, Table 5.1 and Figure 5.2 showed why Ct had an optimal value, Ct* in the
region of £85 per month, associated with an optimal order quantity, Q0* in the region of
1,500 controls (shown in bold in Table 5.1).
Table 5.1 and Figure 5.2 showed Ct* was not very sensitive to rounding Q0*. For exam-
ple, rounding Q0* from 1,500 to 1,000 or 2,000 (plus or minus 500) increased Ct by
90 – 85 = £5 per month. However, they also showed that sensitivity associated with round-
ing errors was not symmetric: increasing Qo from 1,500 to 2,500 (plus 1,000) increased Ct
by 99 – 85 = £14, while decreasing Qo from 1,500 to 500 (minus 1,000) increased Ct by
135 – 85 = £50. This asymmetry suggested ‘round up the order quantity when in doubt
about the robustness of the assumptions used’. This can be viewed as a very simple example
of using an understanding of an asymmetric ‘penalty function’ to choose a ‘balanced target’.
The optimal values Ct* and Q0* were defined by the point of zero slope of the Ct curve
portrayed by Figure 5.2, in turn defined by the point where the negative slope of the order
cost component curve became equal in absolute terms to the positive slope of the hold-
ing cost component curve. Calculus could be used to determine the classic EOQ formula
for Q0*:
Q = 2 R C / C ,o *
d o h( ) ½
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
172 Employing planning tools in practice
also known as the economic lot size or Wilson formula (after one of its earliest [1920s] pro-
ponents). The square root nature of this formula also led to a ‘square root rule’ terminology
some people preferred, and it was a notable feature of the relationship because it dampened
the sensitivity of the optimal order quantity to parameter estimate errors.
In this case setting Rd = 2000, Co = £30, and Ch = £0.06 implied Q0* = 1414, but it was
convenient and sensible to round this up to 1,500, to avoid implying spurious precision.
Because the sensitivity in Ct* terms was low, this rounding involved an increase of only
£1.56 per month.
Pass 1 – testing the plans and interpreting the plans
To begin to explain the idea of testing the robustness of the plans, Nicola developed the
spurious precision issue further. She began by pointing out that rounding the order quantity
from 1414 to 1500 involved an extra £1.56 assuming:
1 that all the current model parameter estimates were correct, AND 2 the model itself was correct.
‘Correct’ in the context of the model being correct meant ‘the joint effect of all the assump-
tions involved reflect the reality being modelled in an unbiased manner’. The second set of
assumptions, about the model itself, was much more important than the first because the
second set of assumptions were about asking the right questions, as opposed to the precision
of the answers to the questions posed. A clarity efficient approach at an appropriate level of
clarity requires everyone understanding what this implies.
At this stage in her analysis the ‘correctness’ of this simple model was inherently limited
and highly uncertain in all aspects of its assessments because its relationship with the full
complexities of reality was highly uncertain. One simple implication was they might con-
sider more significant rounding to reflect the relative sensitivity of rounding up or down
in relation to gaps between the model and the reality of interest. For example, rounding
the answer to 2,000 might be appropriate. As Table 5.1 showed, Ct increased by £5 if Qo
was increased from 1,500 to 2,000, but an important merit of 2,000 was that it confirmed
the status quo. This confirmed Pete’s judgement in choosing 2,000 some time ago, and it
avoided making matters worse for Dave. It might be very important to avoid challenging the status quo for potential cost improvements which were both uncertain and small, especially
if doing so could make matters worse for other reasons. What was very clear was the way
Table 5.1 and Figure 5.2 jointly illustrated the lack of sensitivity to this kind of adjustment.
This sensitivity analysis was a very useful feature of the model for its creators and users.
Nicola emphasised that it was much more important to understand the asymmetry driven by the non- linearity of one component this sensitivity analysis portrayed than the calculus-
derived optimum.
Developing this sensitivity analysis issue further, Nicola explained that uncertainty associ-
ated with estimates of Co, Ch, and Rd could be explored in the same framework. For example,
as the EOQ formula suggested, overestimating the value of Co would involve overestimating
the value of Qo*, which would involve a lower cost penalty than underestimating Co. How-
ever, errors in both directions had an impact on cost which was relatively small because of
the square root relationship involved. For example, assuming Co was £45 when, in fact, it
was £30, a +50% error, could be shown to involve a 2% increase in cost, while assuming Co
was £15 when in fact it was £30, a −50% error, involved a 6% increase. This demonstrated a
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Using a UP 173
remarkable robustness to parameter errors which was a very useful feature of EOQ models.
It also demonstrated the advantages of erring on the high side when estimating Co and Rd
but erring on the low side when estimating Ch.
Nicola went on to point out that sensitivity analysis was an important component of
robustness analysis, and arguably it was a useful place to start because it was the simplest
aspect of robustness analysis. However, she made it very clear to George that robustness test-
ing should never be limited to sensitivity testing. It was much more important to understand the implications of inappropriate assumptions which shaped the form of the models being
used plus more fundamental assumptions about the underlying plans and context, argu-
ably starting with the most vulnerable looking assumptions. This was most easily explained
and explored by looking at less restrictive working assumptions while exploring linked pos-
sibilities for more refined analysis using a sequence which starts with the relatively simple
changes.
Put slightly differently, Nicola had to assess all relevant assumptions for robustness, and in
order to develop some feel for which assumptions mattered the most, she needed to explore
their implications by pursuing generalisations which relaxed them in a sequence suitable for
building her understanding and George’s. The second and further passes explained now are
the explorations of generalisations built on earlier hunches about the need to relax specific
earlier assumptions which proved well- founded, ordered to achieve a systematic building of
understanding.
Pass 2 – refining the analysis by considering stock- outs
Nicola indicated that she had started pass 2 by exploring the implications of possible
planned stock- outs, but her main concern was building on this to consider unplanned stock- outs.
Planned stock- outs (shortages) in the context of the optimal order size component of the
plan were equivalent to deliberately planning to run out of the controls needed to produce
lawnmowers for finished goods inventory, implying putting lawnmowers into ‘finished’
goods inventory without controls fitted. Planned stock- outs could be visualised in terms of
a variant of Figure 5.1 as a negative inventory of controls which equated to a positive inven-
tory of lawnmowers without controls fitted, requiring post- production rectification. Nicola
suggested that the cost of a controls stock- out per unit of inventory per unit time might be
about 100 times the cost of holding controls inventory per unit of inventory per unit time,
because a lawnmower without a control ties up about 50 times as much capital, and it would
probably be worth at least doubling this figure to reflect the inconvenience for Pete and his
production staff. She had assumed a factor of 100 for illustrative purposes, but Pete would
have to confirm or revise this before her analysis results were used, if the analysis suggested
changes. Extending the model to allow for planned shortages this way, the new optimal
order quantity was 1,428 (compared with 1,414), with a ‘planned shortage’ of 14. Taking
the model literally, it was worth being out of controls for about (14/1,428) × 100 = 1% of
the time production takes place, about 2 hours each month, to avoid carrying the extra 14
controls for the rest of the cycle.
Nicola emphasised that it would be silly to take this pass 2 extended model too literally,
and both versions provided an answer which needed rounding to Qo* = 1500. However, the
idea that limited planned stock outage was not a disaster and might be viewed as an oppor- tunity rather than a problem was a useful insight, and even more useful was the implications
in terms of a starting point for modelling the possibility of unplanned stock- outs.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
174 Employing planning tools in practice
For those who are interested in the modelling details, the equivalent of Figure 5.1 involves
two degrees of freedom instead of one (the maximum stock- out as well as Qo), so the math-
ematics becomes marginally more complex, as explored by Chapman, Cooper, and Page
(1987). Nicola shrewdly avoided troubling George with this.
Nicola then used this planned stock- out model discussion as a basis to consider unplanned stock- outs, suggesting that it would prove much more useful to refine this part of their pass
1 plan assumptions.
During her investigations Nicola observed that Pete kept a ‘safety stock’ as noted earlier,
a second part of his overall plan. Pete’s ‘plan part 1’ was order 2,000 at a time. His ‘plan
part 2’ was a form of pre- emptive contingency plan for delivery which might not be imme-
diate. It used a safety stock of 300 controls, placing an order for another 2,000 when there
were still about 300 left. Pete’s rationale for 300 was that most of the time, Dave delivered
an order almost immediately (the same day), but sometimes other jobs took priority and it
took a day; occasionally it took two or three days. TCL produced lawnmowers for about 20
‘production’ (working) days a month, about 100 lawnmowers per production day, so Pete’s
safety stock equated to about three days of production. Pete currently used a ‘two bin’ safety
stock system, initiating an order when the main bin was empty, but he was planning to move
to a computer- based equivalent shortly.
Nicola suggested that the chances associated with Dave taking zero to three days to
replenish the stock could be defined using Table 5.2, which treats ‘days’ as a discrete variable.
She had estimated the P(n) probability values herself, based on a conversation with Pete,
and she had not yet confirmed them with Pete, but she believed they were the right order of
magnitude. She had then calculated the CP(n) cumulative probability values, which define
the probability of n days or more, convenient for the next step.
Nicola then used Table 5.2 in conjunction with the assumption that a mower without a
control held in inventory cost about 100 times as much as a control held in inventory to
produce the analysis of four discrete alternative policies portrayed by Table 5.3.
Pete’s current policy of ‘a safety stock of 300 controls’ (three days of production) was
selected by Nicola as the ‘base case’ for policy 1 of Pete’s plan part 2, with three obvious
alternatives working to the nearest 100 controls. For those familiar with decision analysis,
Table 5.3 is a decision tree in tabular form. For all readers, this is a very simple and conveni-
ent HAT (histogram and tree) approach.
All the costs addressed by this safety stock model were over and above the EOQ model
costs, which ignored safety stock costs and the cost of stock- outs.
To appreciate the modelling choice Nicola made with her approach based on Tables 5.2
and 5.3, you need to know there are alternative common practice ‘prescriptive models’
which, like the EOQ model, provide optimal solutions. The ‘descriptive model’ HAT
approach in Tables 5.2 and 5.3 was used by Nicola’s textbook because it is more general (less
restrictive) and more insightful. The approach used in Tables 5.2 and 5.3 is characterised
as a ‘descriptive model’ because it describes the implications of alternative options without
Table 5.2 Days to replenish stock: probability and cumulative probability distributions.
n, number of days 3 2 1 0
P(n), probability CP(n), cumulative probability
0.01 0.01
0.04 0.05
0.15 0.20
0.80 1.00
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Using a UP 175
Table 5.3 Safety stock policy – analysis of four alternative policy options for Pete’s ‘plan part 2’.
Policy 1: Pete’s current policy as a base case • Safety stock of 300, 3 days production. • No chance of a stock- out, no stock- out costs. • Total safety stock cost per month = extra holding cost per month = 300 × Ch = £18. • Total rounded cost £18.
Policy 2: modest reduction in safety stock • Safety stock of 200, 2 days production. • Probability of a stock- out of 100 for 1 day = 0.01. • Expected cost of stock- outs per month = 0.01 × 100 Ch × 100/30 = £0.20 (the probability P(n) times 100 Ch times the 100 controls times 1/30 of a month).
• Extra holding costs per month = 200 × Ch = £12. • Total expected safety stock cost per month = £0.20 + £12 = £12.20. • Total rounded cost £ 12.
Policy 3: aggressive reduction in safety stock • Safety stock of 100, 1 day of production. • Probability of a stock- out of 100 for 1 day = 0.05,
200 for 1 day = 0.01. • Expect cost of stock- outs per month = (0.05 × 100 Ch × 100/30) + (0.01 × 100 Ch × 100/30) = 1 + 0.40 = £1.40. • Extra holding cost per month = 100 × Ch = £6. • Total expected safety stock cost per month = £1.40 + £6 = £7.40. • Total rounded cost £7.
Policy 4: elimination of safety stock • Zero safety stock – reorder when stock- out occurs. • Probability of a stock- out of 100 for 1 day = 0.20,
200 for 1 day = 0.05, 300 for 1 day = 0.01.
• Expected cost of stock- outs per month = (0.20 × 100 Ch × 100/30) + (0.05 × 100 Ch × 100/30) + (0.01 × 100 Ch × 100/30) = £4 + £2 + £0.6 = £6.60. • Extra holding cost per month = zero. • Total expected safety stock cost per month = £6.60. • Total rounded cost £7.
using calculus or any other direct form of optimisation to choose an option in the manner
of a ‘prescriptive model’ like the EOQ approach, but it clearly provides relevant information
for making choices, including sensitivity analysis. George did not need to know this, but
understanding the need to make modelling choices like this is part of strategic clarity.
George was taken through this analysis of the four policies fairly quickly. He wanted to
understand the analysis implications in broad ‘what needs to be done’ terms, but not the
modelling choices or the calculation details. His focus was the final line for each policy in
Table 5.3, rounded to the nearest pound as a reminder to avoid attributing spurious accuracy
to the results. It was the £18 to £12 to £7 and then another £7 that really interested George.
Nicola might have prepared a decision tree equivalent to Table 5.3 and used it to explain
the arithmetic. She did not bother because the implications which concerned George were
clear from the table, and it would not have been clarity efficient to expend her effort or
George’s time in this way.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
176 Employing planning tools in practice
Furthermore, Nicola might have produced a decision diagram comparable to Figure 4.3
or Figures 3.6 and 3.7 to relate the expected outcomes of Table 5.3 to potential variability
to help make the case for risk efficient choices which did not accept more cost to reduce risk.
It is useful to understand this possibility is available, but it would not be clarity efficient for anyone to expend their effort in this way in this context.
It was important for Nicola to acknowledge that stock- outs have an important nuisance
value for Pete and his staff which might go significantly beyond the stock- out penalty cost
factor of 100 relative to Ch which this model assumes. Pete would have to be consulted
before using this analysis, and even if the factor of 100 and all the other parameters were
acceptable to Pete, a zero safety stock would probably not be a good idea, despite its appar-
ent slight advantage prior to rounding. She provisionally recommended a safety stock of
100, policy 3, most of the year, noting the special problems during the spring rush might
deserve a special spring policy, like a safety stock of 300 plus bigger orders, a point to be
revisited later.
Nicola suggested that given her current assumptions, from a corporate perspective Pete
might be adopting a slightly cautious ‘zero risk of stock- outs’ policy that might not be a
good idea from a broad TLC perspective except when the spring rush was involved, although it made obvious sense from Pete’s perspective all year round. Policy 3 involved 0.01 prob-
ability of a two- day stock- out and 0.04 probability of a one- day stock- out. This might
sound (and feel) significant to Pete in the context of a stock- out cost 100 times Ch. But in
the context of a monthly order of 2,000, a safety stock of 200 involved 100 controls which
were only needed for one day every 1/0.01 = 100 months and another 100 which were
only needed for one day every 1/0.05 = 20 months, on average. The consequences of a
stock- out were irritating for Pete and potentially serious for TLC during the spring rush but
not catastrophic for either. In addition, effective routine communication between Pete and
Dave about priorities could make the recommended safety stock of 100 fairly free of hassle
as well as cost- effective, provided Nicola’s ongoing concerns about seasonal swings were
addressed and her factor of 100 assumption was confirmed.
Given her current assumptions Pete’s current policy was probably costing TLC about
£10 per month more than a safety stock of 100 = one days production policy, which she
characterised as ‘an aggressive reduction in safety stock’ policy. Adjusting Qo from 1,500 to
2,000 might increase Ct by about £5 per month, which might (or might not) be overlooked
because of other considerations. However, regardless of whether the order quantity was
1,500 or 2,000, adjusting the safety stock reorder rule from three days production to one
days production might save about £10 per month (£18 – 7.40 = £10.60 rounded).
Nicola finished this part of her presentation by indicating that if both changes suggested
by her pass 1 and 2 analysis were implemented, a Qo of 1,500 and a one- day reorder rule,
about £15 per month could be saved given her current assumptions. This would involve an EOQ based monthly order and holding cost of about £85, plus a safety stock holding and
storage cost component of about £10, a total cost of about £95 per month. The major
saving would be associated with a more aggressive approach to shortages with a modest
downside which would be minimalised by Pete simply letting Dave know when more con-
trols were needed.
Pass 3 – reconsidering all relevant influences on holding costs
By this stage George was seriously impressed by Nicola’s presentation. He could see clearly
that the issues raised went beyond controls and savings of £15 per month. Nicola then
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Using a UP 177
explained to George that she had needed the first two passes at modelling the decision
context to gain experience and confidence, but she also needed to take George on the same
learning curve she had travelled because they both needed a shared basis for understanding
what was to follow. Building on George’s growing confidence and respect, Nicola suggested
they needed to further refine the plans portrayed by her pass 2 models to incorporate addi-
tional aspects of the decision context and a broader approach to TLC plans.
During her investigations, Nicola had observed that invoices for controls orders take
about 25 days to arrive, payment falls due in 30 days, and there were usually payment
transfer delays of about 5 days. This implied a ‘free credit period’, Tc, of about two months.
She pointed out that during this period Ch was actually only £0.01, not the £0.06 assumed earlier, because the £0.05 cost of capital component did not apply. Furthermore, Nicola
observed that inflation on components like controls was currently running at about 16%
per annum, implying an inflation- driven ‘capital appreciation’ on inventory holdings, Ca, of
about £0.04 per control per month.
Nicola explained that the EOQ models used earlier assumed that inventory holding costs
applied from the moment inventory arrived, and inflation- driven price increases she associ-
ated with ‘appreciation’ on inventory holdings was not relevant – both important assump-
tions to test for robustness. Indeed, her university ‘management for engineers’ course had
addressed how the basic EOQ model could be further extended to consider the impact of
both the inflation- driven capital appreciation and the free credit period. The mathemat-
ics of this model were considerably more complex than that used earlier, but the basic
approach was the same, and all George needed to understand was Figure 5.3, Table 5.4, and
Figure 5.4, which she now showed him and explained.
Nicola first explained that Figure 5.3 was a generalisation of Figure 5.1, with Tc defining
the free credit period previously assumed to be zero. That is, Figure 5.1 was just a special
case of Figure 5.3 which assumed no free credit period and ignored inflation, both stealth assumptions of importance if economic lot size formula users are not aware of them. The
area of the shape abcd could be associated with the £0.01 holding cost component less inflation- driven appreciation of £0.04 but no cost of capital. This involved a negative cost
(positive appreciation net of holding cost) of £0.01 – £0.04 = - £0.03. The rest of the usual
holding cost triangle could be associated with £0.06 holding cost component less apprecia-
tion of £0.04, implying a £0.06 – £0.04 = £0.02 positive cost.
Helping George to interpret Table 5.4 and the associated Figure 5.4, Nicola explained that
what was involved was a generalisation of Table 5.1 and Figure 5.2, with the same special case
relationship just considered. The order cost component of Table 5.4 was the same as that in
Table 5.1. A more complex expression was needed to derive the holding cost component,
with negative costs (net appreciation savings) which reached a minimum between Qo = 6,000
and 7,000, when the rate of growth of further appreciation costs was overtaken by the rate of
growth of further holding costs. This effect together with the continually declining order cost
component as Qo increased meant that the joint cost per month, Ct, reached its minimum of
about −£64 per month at a Qo of about 7,000, highlighted in Table 5.4 for easy identifica-
tion. In terms of this generalised model, a Qo of 2,000 implied a Ct of zero, and a Qo of 1,500
implied a Ct of a little more than +£20 per month – her pass 3 result was about £64 per month
cheaper than the current situation and about £85 per month cheaper than her pass 1 sugges-
tion because of crucial changes in her working assumptions.
Helping George understand why the order quantity was so much bigger, Nicola
explained that two systematic effects were working together, both missed completely by
her earlier models. Inflation- driven appreciation meant a lower holding cost, in general,
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
178 Employing planning tools in practice
time in months
inventory
level In
controls
slope - RdQo
To
Qc
d
ea
cb
Tc
Figure 5.3 Order cycle relationships: credit period allowed.
Table 5.4 Joint cost per month and its components as the order quantity changes: two months free credit period.
Order quantity Qo 1000 2000 3000 4000 5000 6000 7000 8000
Order cost component Holding cost component Joint cost Ct
60 – 15 45
30 – 30 0
20 – 45 – 25
15 – 60 – 45
12 – 70 – 58
10 – 73 – 63
9 – 73 – 64
8 – 70 – 63
cost in £
joint cost Ct
Qo order quantity in controls 0
4,000 8,000
–50
100
Qo *
holding cost component
order cost component
50
–100
Figure 5.4 Joint cost per month and its components as the order quantity changes: two months free credit period.
potentially a negative holding cost (positive return on investment). Furthermore, appre-
ciation in conjunction with somebody else’s money being used as working capital meant
a profit on somebody else’s money – ‘gearing’ or ‘leverage’ in financial parlance. She had
pointed out earlier that the optimal order quantity was more sensitive to overestimating
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Using a UP 179
Ch than it was to underestimating, and ignoring both appreciation driven by inflation and
a free credit period was a form of very serious overestimation of Ch which went beyond
biased parameter estimation in the usual sense because it was the pass 2 decision- making
model that was biased.
In part to make George more comfortable with the subtle but massive implications of this
new pass 3 model, she pointed out that many people using EOQ models were not aware
that ‘no appreciation and immediate payment for stock on arrival’ were underlying assump-
tions, although the details were developed in her course materials (the basis of Chapman,
Cooper and Page [1987] chapter 12).
As an aside addressing a point Nicola did not make, I did not realise the importance of
the ‘cash on delivery’ and ‘no inflation’ assumptions implicit in basic EOQ models until the
early 1980s, 50 years after these models were first used and 20 years after studying them
as part of my first and second degrees. The resulting paper (Chapman, Ward, Cooper, and
Page, 1984) revealed I was not alone. It flushed out some related earlier papers, generated
considerable discussion, and led to a stream of subsequent papers addressing the impact of
inflation and free credit periods. However, for half a century most people who wrote about
or used a very popular basic inventory model were not aware of key implicit assumptions, an
important form of ‘stealth assumption’. Many remain unaware. Yet the consequences in this
chapter’s tale are orders of magnitude more important than Nicola’s pass 1 and pass 2 mod-
els, and they take the solution in a different direction. Using OPM (other people’s money)
is not a new idea – this was a published term in the US more than a century ago and used
internationally as a core concept for many centuries in many industries, but its implications
in EOQ models had been largely overlooked for a very long time.
The free credit period and inflation- driven appreciation modelling developed in this
chapter’s tale can be interpreted as both a useful illustration of the importance of stealth
assumptions and the related importance of using framing assumptions which are as general (unrestricted) as we can make them. The way Nicola was potentially moving in the wrong
direction in terms of order quantity size on pass 1 and pass 2 as identified by pass 3 illustrates
the importance of both issues in a reasonably simple way, providing a basis for understand-
ing the implications in more complex contexts.
Stealth assumptions are comparable to and can be usefully viewed as unknown unknowns
in the sense that they are misconceptions associated with ambiguities beyond the scope
of our framing assumptions. These unknown unknowns are very important aspects of
ambiguity relative to the more obvious known unknown aspects of uncertainty – what
we know is an issue even if we cannot size or shape it. The no free credit period or appre-
ciation stealth assumptions implicit in all early EOQ models illustrate unknown framing
assumptions which once recognised in this chapter’s tale transformed the analysis and all
follow- on planning.
Unknown unknowns are often associated with events which are not predictable, some-
times using the term ‘black swans’ (Taleb, 2007). However, stealth assumptions involving
misconceptions can be even more important than unpredictable events, and they can be
minimised by ensuring that we use the widest feasible set of framing assumptions as a basis
for explicitly testing working assumptions. This is much easier said than done. However, the difficulty involved does not make it less important – it makes it more important to know
how to approach the key issues.
Linking this to the discussion of Chapter 3, this kind of potential decision- making error
is not usually considered ‘bias’, but its basis and implications are comparable – omitting to
address important sources of underlying uncertainty. Furthermore, this important source of
uncertainty and bias is an unknown unknown for everyone using the stealth assumptions
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
180 Employing planning tools in practice
underlying the basic EOQ model. It can be viewed as a ‘risk’ associated with a capability-
culture source of uncertainty – leading to missed opportunities if those involved fail to
recognise these important issues.
Returning to the tale, Nicola pointed out that direct use of this pass 3 model required
prediction of average appreciation per month over the order interval To. In practice, the implications of general inflation for the cost of controls would not be a small amount of
appreciation per month, but a significant amount linked to price changes every 6 to 12
months, recent and more distant past appreciation being no more than a guide. However,
a large recent price increase for controls might suggest no further price increases for some
time. Assuming they were aware of the inflation issue, one important residual risk (or mis-
take) was a relatively large order based on assumed appreciation when none took place or
an economic slump led to price decreases. A much more important risk (or mistake) was a
relatively small order based on assumed stable prices when large price increases took place
regularly. But by far the biggest risk (or mistake) of all was simply ignoring inflation driven
appreciation and associated gearing because of a free credit period – because the associ-
ated stealth assumptions had not been identified and tested. Using a simple expected value
approach was not ideal, but it was much better than just ignoring the issue.
Nicola finished this part of her presentation by indicating that these results might be
implemented directly with a safety stock = one days production rule. Relative to the cur-
rent standard order quantity of 2,000, in terms of order, holding, and appreciation costs as
captured by the revised economic order quantity model, Table 5.4 indicates that they could
achieve a saving of about £64 per month with an order quantity of 7,000. And the safety
stock implications (risk of stock- out every 3– 4 months instead of every month) should save
a further £16 per month relative to the three days production rule with Qo = 2,000, about
£80 per month in total.
Pass 3 – exploiting creativity to interpret an emerging plan 3
By this stage George was glowing with family pride. Nicola, who was beginning to enjoy
herself, and to see why her mother’s advice to join the family business was a lot better than
she had previously thought, now built on the firm foundations had laid so far. She had
already approached the boundaries of conventional textbook interpretation of traditional
OR process use. She now drew on her innate creativity, crucial to really effective use of a UP,
to begin a transition into more general UP territory.
Nicola observed that a working assumption underlying Table 5.4 and Figure 5.4 was
inventory was paid for when it was used during the free credit period. Nicola then observed
there would be no point in offering to pay for inventory when it was used (not the case now)
unless this benefit for the firm supplying controls could be balanced by a reciprocal benefit
for TLC. The obvious reciprocal TLC benefit in her current view was a longer free credit
period. A longer free credit period would result in an even larger order quantity, delivering
no inflation- driven price increases during an even longer period between orders. She had
undertaken sensitivity analysis on the effect of paying for stock as it is used in conjunction
with a longer free credit period which suggested that if an extra four months of free credit
on unused stock was on offer, paying for stock as it was used, this justified TLC using an order of about 12,000 (rounded from a calculated value of 12,884), a six- month supply. This was
coincidentally what Dave had suggested – but for rather different reasons. If 12,000 was
the TLC order on offer, the control supplier might be persuaded to offer an inducement
comparable to four months’ free credit on unused stock. This would generate a further
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Using a UP 181
cost saving of about £80 per month. TLC might not be able to secure agreement to this
much free credit from the controls firm, but it would be a good stretch target to negotiate
towards. As a linked suggestion, Nicola observed that they did not need to see this target
‘plan 3’ as an optimal fixed order quantity Q0* of 12,000 controls roughly every six months.
They might find it useful to see it as an optimal fixed order interval T0* of six months for the
roughly 12,000 controls required every six months, using a Periodic Review interpretation
of the EOQ model.
George immediately recognised this creative interpretation of her pass 3 results involved
important insights which he and Nicola needed to share with Bob, TLC’s buyer.
Pass 4 – Bob’s crucial clarification of the plans/models distinction
After Bob was fully briefed by George and Nicola, taking him through the same learning
process to reach Nicola’s target plan 3, Bob spent some time thinking about the implica-
tions. During this period he had several discussions with Nicola to clarify his evolving ideas.
He then approached the managing director of the firm supplying controls with what we will
call ‘Bob’s plan 4’, built on a transformation of ‘Nicola’s plan 3’ by Bob and Nicola work-
ing together. Bob referred to his plan 4 when talking to Nicola and George as ‘their TLC
plan 4’.
Bob started his discussion with the managing director of the controls firm by indicating
that TLC would be interested in replacing their current orders for 2,000 controls as needed
by contracting for a six- month supply of about 12,000 controls at a discounted price to be agreed. The contract would involve TLC paying for controls without the need for invoices at a uniform rate over the six- month period at the end of each month, roughly equivalent to paying for them as they were used. The controls firm would deliver them at the control firm’s convenience without letting TLC run out of inventory, with stock- level information available for the controls supplier as needed. He suggested this might be in the context of a one or
two year rolling contract, rolled forward every six months, with the price for the next six
months negotiated well in advance.
From a TLC perspective what was desirable about this transformation of Nicola’s plan 3
was the discounted price for six months negotiated in advance, plus zero TLC investment in inventory, plus delivery of controls by the controls supplier, plus no stock- outs, plus not having to cope with a very large inventory every six months, plus no invoices to process monthly because automatic payment by the bank could be arranged, plus some longer price and relationship stability built into a rolling contract agreement, plus a new collaborative relationship with a supplier of a key component which increased TLC market power as a
supply- side buyer.
From the control supplier’s perspective what was desirable about this transformation of
Nicola’s plan 3 was regular payment monthly without invoices requiring generation, plus less free credit for TLC on average than the present case, plus the freedom to deliver when it suited the controls firm. The controls firm might or might not see freedom to deliver when
it suited them as beneficial, but the other features of the collaboration with a customer pro-
posed by Bob ought to be attractive.
Bob was clear before he approached the controls firm that there was no need to even men-
tion the extended free credit or fixed order quantity aspects of Nicola’s plan. Bob was also
aware, before he visited the controls firm, that freedom to deliver when they wanted might
not be highly valued, but it would be very valuable if the controls firm could exploit it by
negotiating similar arrangements with other firms they supplied with the same component,
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
182 Employing planning tools in practice
and when he briefly explained why the managing director of the controls supplier became
very interested.
A similar arrangement with the other firms supplied with controls would allow the con-
trols firm to significantly reduce their finished goods inventories and use longer production runs. Producing an inventory of controls on equipment used to produce a wide variety of
other components involved a ‘set- up cost’. This setup cost was equivalent to TLC’s order
cost, and directly comparable holding costs were involved, with no free credit period pos-
sibilities. This implied an Economic Lot Size model approach might be employed, directly
comparable to an EOQ approach. However, because the controls firm primarily supplied a
small number of local firms that all purchased their controls when it suited them, demand
for the controls firm was not smooth – unlike TLC’s demand for controls, it was decidedly
lumpy and unpredictable. This implied a relatively large safety stock for the controls firm,
assuming orders were unpredictable and could not be kept waiting, in conjunction with
short production runs. An arrangement with the other firms supplied with controls compa-
rable to Bob’s plan 4 would allow the controls firm to produce a large batch, deliver it to
all its customers on planned routes to top up current stocks, and then repeat this cycle on a
fairly stable basis. It could size its production runs based on its distribution plans, with most
inventories being held by its customers.
Bob knew that the joint implications of a more flexible and cost- effective approach to pro-
duction of controls and associated inventories plus the stable revenue based on pre- agreed
prices should be worth a significant discount from the control firm’s perspective if they
could make effective use of TLC’s plan 4 with other customers, in effect sharing the benefit
with TLC of an improved operating position for both firms which the controls firm could
exploit with its other customers. What was involved was a form of ‘supply chain manage-
ment’, taking a partnership view of supply chain linkages to share the benefits of integrated
practices to some extent but looking after TLC’s interests as far as possible. Bob had limited
interest in the deeper implications for TLC suppliers, but he understood very clearly why it
would be useful to know about and be sensitive to their concerns.
Even if the controls firm could not negotiate similar arrangements with its other custom-
ers, it should be in their interests to undertake delivery for no more than the cost to TLC
of using Dave, with some discounting of the current standard price to share a more limited
set of other benefits.
Bob had recognised that Nicola’s plan 3 based on her creative interpretation of the model
developed during pass 3 provided crucial insights. However, he had used those insights and
his understanding of Nicola’s UP- based planning process to build a pass 4 plan which was
much better. It replaced the fixed order quantity model or its fixed order interval equivalent with a negotiated contract for a fixed period, at discounted prices, with flexible deliveries by
the supplier at the supplier’s convenience, plus other suitable contract provisions, plus less
formal but important understandings and agreements.
When Bob reported back to George, he started by indicating that the negotiated contract
meant 12,000 would not have to be made available for collection by Dave every six months.
All the controls would be delivered to TLC by the controls firm. It would do so as and
when it wished, in whatever delivery size it wanted, but would guarantee TLC would not
have a stock- out. Crucially, there would be a significant discount on the current standard
price fixed for six months at a time on a rolling basis. Bob explained that he had bargained
hard to negotiate what he believed was a fair discount, suitable arrangements to ensure
both companies worked together to avoid stock- outs, and a rolling review process to ensure
contract renewal was always timely. Letting the controls supplier make the deliveries when
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Using a UP 183
it suited them was a key step in generalising Nicola’s invaluable plan 3 basis for their TLC
plan 4 because TLC did not want the order costs or the physical inventory with its hold-
ing costs. TLC wanted a reliable supply at a low price which offered protection from rising
prices and no capital tied up in inventories. Indeed, having a large physical inventory at
the start of each cycle would be distinctly unhelpful and might require rethinking using an
opportunity cost approach to storage costs.
When George, Bob and Nicola then got together to review the implications of Nicola’s
plan 3 and Bob’s plan 4 they jointly agreed on a number of observations:
1 Nicola’s concerns about seasonal swings in cost parameters driven by the upswing in
demand for lawnmowers in the spring plus complications associated with stock- outs in
the spring raised a very important set of concerns, but they would all be resolved with-
out further effort, because Bob had arranged for delivery by the controls supplier. It
was agreed that the resolution of these concerns was very important. Nicola had prob- ably understated the impact during the spring rush of any component supply stock-
outs, and the way Bob’s plan reduced their problems during the spring rush would be
very useful.
2 It had been worthwhile delaying exploring the complexities of this spring rush compli-
cation, because the effort would have been wasted at best, reducing the clarity efficiency
of the overall analysis and synthesis process. It would also have risked diverting analysis
into a complex mess which might have led to them missing the real opportunities. If
Bob’s plan 4 insights had not eliminated this concern, it would have deserved atten-
tion, perhaps warranting an approach like a once a year using an order of 6,000 or more
controls just before the spring rush.
3 It had also been worthwhile delaying worrying about Nicola’s factor of 100 assump-
tion about the difference between stock- out costs and Ch, but suitable stock- out cost
penalties could be built into Bob’s plan 4 contract with the controls firm if this was
appropriate.
4 In system monitoring terms, Pete would have to keep an eye on his stock of controls,
but life should be easier for him.
5 Ajit and the accounting staff who paid the bills should be pleased with the outcome
because their workload would be lower and smoother.
6 Dave would never have to visit the controls firm again, and he would have more time
for delivering finished goods inventory or other tasks throughout the year.
7 TLC would no longer have to be a ‘price taker’ with respect to the price of con-
trols because Bob had established a new negotiating position which pleased him and
strengthened TLC’s position.
8 Had they simply implement Nicola’s pass 1 or pass 2 plans, they would have moved in
the wrong direction, arguably making things worse, not better. This was an ‘enlight-
ened retrospective view’ from the position of insight they had gained via Nicola’s plan
3 plus Bob’s plan 4. They did not have this insight when they started, or at the end of
passes 1 and 2. They had acquired it on the journey during passes 3 and 4, a corporate
learning process as well as a corporate planning process.
9 It would be appropriate and useful to acknowledge the value of the textbook OR pro-
cess contribution in terms of the generic model and process ideas which Nicola had con-
tributed to the overall TLC learning process, and the crucial impact of the free credit
period and inflation model. However, the TLC corporate learning process also needed
to recognise the importance of Nicola’s insight and creativity, George’s judgement, and
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
184 Employing planning tools in practice
Bob’s absolutely crucial role – transforming Nicola’s plan 3 into a supply chain plan 4
which included a much more sophisticated approach to shared benefits from collabora-
tion with the control supplier firm than Nicola’s.
All three agreed there were remaining sources of uncertainty, risks, and potential problems
which would need addressing. For example, if the controls firm suddenly went bankrupt, or
tried to drive up prices unreasonably, TLC would need another supplier quickly. As a sec-
ond example, lawnmower sales were reasonably stable but inherently uncertain, and if the
demand for TLC lawnmowers suddenly slumped, TLC might find itself locked into a supply
contract for more controls than they wanted. However, both of these issues needed separate
management anyway, perhaps with more attention than they currently received.
From an EP perspective, Dave’s grumpy reaction to no tea and biscuits provided an
opportunity to put the new skills Nicola brought to TLC to use – pure serendipity in a
sense – but arguably a minor problem which George responded to wisely, transforming a
problem into an opportunity because of the way Nicola and Bob worked together as a very
effective team.
TLC’s primary concern from Nicola’s initial perspective was dealing with the ambiguity
uncertainty and underlying complexity associated with 2,000 being the right answer to the
question, ‘What should the order quantity be?’ The underlying, long- standing assumption that
an order quantity approach to supply- side management for input components like controls
was appropriate was not questioned. Both the deterministic prescriptive EOQ models and the
probabilistic descriptive safety stock models were tools for understanding this complexity and
managing its implications, resolving the associated uncertainty. The EOQ models provided a
deterministic model understanding with no associated stochastic model understanding. The
safety stock models provided associated stochastic model understanding. The EOQ and safety
stock models together with a creative interpretation of the implications provided the basis for
qualitative and quantitative analysis for a supply chain plan which went well beyond both the
initial models in terms of its scope and the opportunities addressed.
Bob’s contribution in terms of planning a more general supply chain contracting approach,
using his creativity and insight into their supplier’s objectives as well as TLC’s real con-
cerns, was about further exploring the complexity and ambiguity uncertainty addressed by
Nicola. Bob saw beyond the assumption that an optimal fixed order quantity or an optimal
fixed order interval was appropriate. He addressed a higher- order level of uncertainty and
underlying complexity. He envisaged and implemented plans which separated the strategy
associated with the contracting period from the tactics of delivery. In this sense Bob’s plan
4 involved a level of both strategic vision and strategic clarity which was completely missing
in Nicola’s more tactically focused plan 3 and the earlier TLC fixed order quantity plan, and
it clearly distinguished between his plan 4 as a whole and the models underlying its shaping
while Nicola was leading the analysis development.
Discussing his emerging approach at length with Nicola before his visit to the controls
supplier was important for Bob, Nicola, and TLC. It was during these discussions that
Nicola made the connection between her EOQ approach for TLC and an Economic Lot
Size approach for the controls firm for its stock of controls. Her sharing of this connec-
tion with Bob triggered the insights behind a key aspect of Bob’s plan 4. Bob discussed his
approach at length with Nicola after his visit to the controls supplier, before he reported
back to George, and at Nicola’s suggestion they jointly labelled Bob’s plan 4 approach
‘My- Place- Your- Timing’ – as a catchy variation on the currently fashionable ‘Just- In- Time’
label. One potential problem with this label, which Bob observed, was the perspective basis.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Using a UP 185
My- Place- Your- Timing from a TLC perspective was Your- Place- My- Timing from their sup-
plier’s perspective. However, Bob wanted Nicola to feel that she had an acknowledged
ownership stake in ‘their TLC plan 4 approach’, and he did not see any need to quibble
about a label which he thought some suppliers might find attractive, using ‘their version’ of
the language of the label. These labels and their discussion with George triggered some of
George’s subsequent ideas.
The overall teamwork was crucial. Nicola had used a traditional OR process, with par-
ticularly effective robustness testing to pick up the free credit period and effects of inflation
on the cost of capital, plus a particularly creative interpretation of her results. George wisely
saw the need to involve Bob at this point. Bob then recognised the need to move from
EOQ models or Periodic Review equivalents to a complete supply chain plan for controls
with a negotiated rolling contract based on supplier delivery at the supplier’s convenience
plus other terms and conditions, supported by Nicola in terms of knowledge about generic
inventory models for supply- side or demand- side analysis plus the role of a systematic
approach to testing all of the assumptions associated with a plan.
Put slightly differently, Nicola had been testing assumptions systematically to resolve
uncertainty by understanding complexity within the assumed question, ‘What is the optimal
order quantity?’ George sensed the strategic implications of Nicola’s answers. Bob then
tested the assumption this was the right question and had the insight to see that it wasn’t.
Bob then initiated a higher- order exploration of uncertainty and underlying complexity,
guided by what they had both learned from the lower- order analysis. In UP terms Nicola was
looking at assumptions within the bounds of a fixed order quantity or interval model, but
Bob addressed relevant ‘create and enhance the plan’ assumptions which could be relaxed.
To replicate the effectiveness of their joint efforts in other EP contexts, the UP’s separa-
tion of create and enhance the plans to address all relevant concerns and ‘shape plans by
modelling core issues’ which are a subset of the broader plans is a very helpful distinction.
It facilitates understanding how to make the transition from Nicola’s plan 3 to Bob’s plan
4. It also stimulated George to go right back to the capture the context phase of the UP,
the next issue in our tale. Bob built on Nicola’s pass 3 models, but he couched his plans in
spreadsheet terms with his altered assumptions, using price as a variable to be negotiated
below current standard price based on the attractiveness of the other contract features to
the controls firm. Bob’s ‘test the plans’ involved testing all his assumptions, and he inter-
preted his plans creatively throughout the process of their development. Whether or not Bob thought of ‘Bob’s plan 4’ as the result of using a UP pass 4 approach building on Nicola’s
passes 1 through 3 does not really matter in some ways, but ongoing joint planning does
make corporate use of an explicit UP concept perspective useful for everyone involved.
Furthermore, it is the team use of the UP concept which needs a collective understand-
ing. The UP front- end phase structure clarifies the difference between a model of some core
issues, the plans as a whole, based on insights provided by models, the context the plans are
set in, and the effective implementation of those plans. Different people may have differ-
ent contributing roles to play in effective corporate use of a UP approach, but they need to
work together, moving towards a common vision of where they want to go. They may not
share this common vision of where they want to go at the outset, and finding an enlightened
shared vision may be a particularly important part of the journey.
Some OR textbooks, and some courses based on them, present EOQ models as tools for
direct use – once you understand the formula you can plug in your parameters and get the
optimal answer. Some practice- focused OR experts are highly critical of their use in any way.
Gene (Robert E. D.) Woolsey provides a good example of the latter school – see Real World
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
186 Employing planning tools in practice
Operations Research: The Woolsey Papers (Woolsey, 2003, chapter 4.4) for example. The EP perspective advocated by this book assumes that EOQ models are an excellent conceptual
framework for developing an initial understanding of what matters a lot, and what does not matter quite so much, provided all the working assumptions that matter are tested very carefully, even when that leads to abandoning an order quantity concept altogether. This makes EOQ models particularly good examples for explaining how to use a UP concept. But an EOQ initial model approach can demonstrably move suggested changes in plans in the wrong
direction and provide seriously unhelpful advice if the robustness of the assumptions is not
tested carefully. Initial and later models of all kinds always need very rigorous testing in the
general UP sense.
To link this point to the anti- aircraft guns on merchant ships example of 1930s- 40s OR
discussed in Chapter 2, Nicola’s pass 1 and pass 2 analysis is comparable to a statistical
analysis of the relative ability of ship versus shore- based anti- aircraft guns to destroy enemy
aircraft without looking at the extent to which anti- aircraft guns mounted on merchant
ships increases the survival capability of the ships and the morale of their crew. Nicola’s pass
3 followed by Bob’s plan 4 were comparable to beginning to see the bigger picture more
clearly in the sense that it was not losing the war that really mattered.
Nicola, Bob and George were all delighted with their new approach to controls, and all
three appreciated there was clearly scope for extending this kind of joint bottom- up corpo-
rate planning process to other input components for the firm’s lawnmowers. Some suppliers
might not be interested in anything other than a conventional order quantity relationship,
but some might welcome a variant of the controls supplier approach, and some might war-
rant very different approaches developed in a related manner. Bob’s plan 4 soon became
known as ‘Bob’s proposition’, and the further propositions which it seeded, to be discussed
in the next section, are a much smaller scale variant of the rapid spread of OR use through-
out the Allied war effort. They each went well beyond the range of the initial TLC concerns,
and collectively they embraced extensive use of a UP concept which went well beyond the
scope of the UP variant Nicola had started with to address a bottom- up approach to corpo-
rate strategy development.
Further creative testing of capture the context assumptions
George, Bob and Nicola working together soon came up with a very different proposition.
This time George provided the leadership, Bob and Nicola the support, and the result
became known as ‘George’s proposition’. Surinder, the sales manager for TLC’s lawnmow-
ers, then led the development of ‘Surinder’s proposition’. Lunchtime TLC discussion of a
proposition developed predominantly by Pete was then followed by a proposition from Ajit.
Consider George’s proposition first, then the others. Bear in mind each ‘proposition’
was the result of very preliminary pass 1 through a UP equivalent without any formal use
of models by a group of people focused on ‘create and enhance plans to address all relevant
concerns’, in each case starting in a slightly different context but learning from Nicola’s use
of a UP and her resulting plan 3, plus Bob’s use of the UP concept and his resulting plan 4,
plus the immediately preceding propositions and the way they were generated.
George’s proposition – an engine wholesaler
George began his first discussion of this proposition with Bob and Nicola by reminding
Bob of the discussions they had about a year ago involving switching the TLC lawnmower
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Using a UP 187
engine supplier from a British based supplier used for decades to a Japanese supplier. TLC
customers liked the new engines, and all relevant TLC staff liked the new arrangement.
However, George felt TLC was vulnerable to the long supply route and possible price
increases if a dominant supplier situation developed. He was also concerned about the space
requirements for large order sizes because the engines were bulky. He observed that Bob
had recently commented that buyers for several other local firms producing pumps, genera-
tors and other equipment had switched to using the same small engine supplier, and while
they were equally pleased, they shared the same concerns. George then indicated the lease
for the warehouse across the road from TLC was currently available.
Putting these issues together, George proposed a new company – a small engine whole-
saler. He suggested calling the new company The Engine Company – contracted to TEC.
He suggested TEC should be controlled and probably wholly owned by TLC, but some
funding involving local users of TEC’s services might be useful. TEC’s initial core mandate
would be contracting with TLC’s Japanese engine supplier as a wholesale distributor for
their small engines for the UK Midlands. TEC’s initial core market would be TLC and all
the other local users of these engines. The core service offered to TLC and other local users
by TEC would be a Your- Place- My- Timing supply of engines on a rolling horizon contract
arrangement to provide secure supply and stable prices without the current space concerns
for TEC customers, with alternative contracts negotiable. This could approach a Just- In-
Time position from a TEC customer perspective if a Your- Place- Your- Time variant was
what the customer wanted, with the customer specifying delivery times. However, the TEC
approach would be better conceived as daily or weekly deliveries to top up safety stocks with
Just- In- Time special case (Your- Place- Your- Time) options TEC would not recommend. If a
TEC customer had lots of storage space, a special case Your- Place- My- Timing variant might
be offered with the customer storing engines well beyond their immediate needs, with dis-
counts, to employ this space in a jointly useful manner, but this could be seen as an option
to be recommended only when relevant.
These TEC customer arrangements would require coordination with TEC contracts
and delivery arrangements with their Japanese supplier. A My- Place- Your- Timing contract
could be offered to the engine supplier. This might facilitate quite large deliveries to TEC
if this suited the Japanese supplier. Recognising the more complex demand forecasting now
involved would be important. However, this issue could be addressed effectively, and a
reduced risk of engine stock- outs should be feasible for TLC and all other TEC customers.
A number of important developments from this basic concept would need careful con-
sideration. TEC might diversify its engine suppliers and provide a range of engine choices
for TEC customers, perhaps with related design advice. TEC might expand its geographic
remit. All these options would depend upon the market opportunities and threats.
The core concept driving the creation of TEC was making a successful business proposi-
tion out of a gap in the market for an organisation which would resolve difficulties for TLC
plus comparable difficulties for other firms using the same engine supplier. It might also
resolve difficulties for their engine supplier, but it could be viewed as creating difficulties
for their engine supplier if their engine supplier attempted to exploit their market position,
operating as ‘a tasty carrot with big stick potential’. TEC’s basis was a buffer between one or
more small engine suppliers and users of those suppliers, ideally developing a scale of busi-
ness which would provide market power plus a broader understanding of the small engine
market for TLC.
With important stimulation, support, and help from Nicola and Bob, what George had
created was a significant generalisation of the My- Place- Your- Timing approach to the
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
188 Employing planning tools in practice
controls supplier, creating a new intermediate engine supplier operation with a new profit
stream in addition to other direct and indirect benefits for TLC.
George’s proposition was eventually explained to all interested parties in an ordered fash-
ion as outlined earlier, but it had not suddenly occurred to George in this orderly manner –
it had emerged from iterations of his UP equivalent, building on what Nicola and Bob had
demonstrated and interactive discussions with both of them.
The crucial lesson George had picked up was testing all his planning assumptions as he experimented with new planning ideas, being especially sensitive to long- standing context assumptions. In UP concept terms he had started to test ‘capture the context’ assumptions, asking even more fundamental questions than Bob or Nicola did. The key new idea was
TLC did not need to stick to just manufacturing lawnmowers if an interesting upstream sup-
ply chain opportunity presented itself. George had tested and relaxed an assumption which
he had inherited from his father which had become central to the nature of TLC’s business
and culture. Until now they had been assuming that TLC’s only business was manufacturing lawnmowers. Now their eyes were opened to other opportunities.
George’s proposition demonstrates a whole new level of creativity with strategic vision
and strategic clarity implications for any organisation which might be triggered by the wide- spread use of a UP which targets the testing of context assumptions. In particular, once
George understood the benefits to both TLC and their controls supplier of supply chain
management based on Bob’s plan 4, he could see a rich set of related contract variants as a
basis for a new company which could make a profit out of a gap in the small engine supply
market as well as resolving TLC engine supply problems. His starting place for TEC was
trying to resolve the obvious difficulties associated with simply replicating Bob’s approach to
controls when the items purchased from a supplier were relatively large, relatively expensive,
and shipped a long distance, by a company which was becoming very popular with a wide
range of customers in the UK and worldwide.
Once George had the basis of his proposition clear in his own mind, he began shaping the
plan and further testing of associated assumptions with Nicola and Bob. They did not take
it as far as using formal models to shape the emerging plans, but they did test and develop
some of the key planning assumptions necessary for a very preliminary business case model.
George was very keen to discuss his TEC proposition with Ajit as soon as possible. He
knew he might be getting out of his depth with a whole new business model involving TLC
and TEC. George relied on Ajit’s council on all strategic choices. However, at this point Ajit
was taking a well- deserved holiday.
Surinder’s proposition – a machinery company
Surinder and George had always struggled with the very large inventories of lawnmowers
which accumulated during the autumn and winter, prior to the spring rush and the reasonably
steady summer demand rate. George had used local warehouses to provide space on an ad hoc
basis, partially linked to comparable arrangements for large deliveries of engines, and on an ad
hoc basis Surinder had negotiated some large winter period sales of TLC lawnmowers to large
retailers with price reductions they passed on to customers, plus various other tactics of this
kind. These ad hoc approaches had kept TLC finished goods lawnmower inventories within
manageable bounds, but both felt a more systematic approach was overdue.
A related but separate issue was George and Surinder had always agreed ‘standard prices’
for TLC lawnmowers on a basis largely driven by costs and set with limited pricing expertise
within the upper ‘quality’ end of prices charged by competitors. Surinder had always seen
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Using a UP 189
TLC as a ‘production driven’ organisation, with her sales role as primarily an order proces-
sor for the retail outlets and garden machinery servicing companies selling the lawnmowers
which TLC produced. She had always wondered if a stronger marketing lead from her might
be better for TLC as well as more stimulating for her.
Once George was comfortable with the way his TEC discussions with Bob and Nicola
were progressing, he began related discussions with Surinder. Surinder was particularly
interested in Bob’s increased market power in the controls supplier relationship, plus the
broader market power issues raised by George’s proposition for TEC. In addition to her
conversations with George, Surinder started talking to Bob and Nicola, and the seeds of her
own proposition began to take root.
The assumption that TLC was just in the lawnmower manufacturing business had been
questioned and revoked by George’s TEC concept, a change of working assumptions with
some quite fundamental further corporate change implications. Surinder and George soon
saw some of the implications in terms of a further new company which Surinder developed
preliminary plans for, calling it The Machines Company – contracted to TMC.
TMC could share warehouse space with TEC. Its initial core business could be systematic
marketing of TLC lawnmowers to current sales outlets, using Your- Place- My- Time options
comparable to TLC and TEC. But these options could be generalised to include delivery
of lawnmowers direct to customers when ordered in a shop which just kept a limited range
of show models, or a garden machinery servicing company which just kept a range of bro-
chures. It could be further extended to direct marketing to customers via a mail- order busi-
ness prototype of post- millennium online web- based direct sales. Managing TLC’s finished
goods inventory would be a by- product of this TMC core business.
However, the really big growth in business which TMC might be set up to exploit could
be using TMC direct retail and wholesale capability to market machinery manufactured by
other companies as well as TLC lawnmowers. As Surinder saw it, this was revolutionary
because it moved TCL’s focus from producing lawnmowers to marketing machines.
Two very different and probably incompatible directions for a single new company might
be pursued. One was selling generators, pumps, and other equipment, probably manu-
factured in the UK by companies with sales capabilities which were limited in a manner
comparable to TLC, perhaps some of the initial partners in TEC, or other TEC customers.
The other was selling machinery manufactured by other companies which was all garden
machinery, possibly manufactured in other countries with lower costs, probably limited to
high- quality up- market products like TLCs current mower range. Surinder proposed a focus
on the second option.
Surinder’s proposition was initiated by applying the supply- side (upstream) thinking
behind TEC to the demand- side (downstream) management of lawnmower sales, initially
transforming George’s proposition into a generalised form of ‘vertical integration’. How-
ever, the potential ramifications of a fully developed TMC concept were very much broader
and more complex, taking the TLC/TEC/TMC group into a whole new corporate strategy.
George soon began to see the wider implications and started an ongoing conversation
about some of the more important possible consequences with Pete. George greatly valued
Surinder’s new ideas and enthusiasm, but he had well- founded concerns about Surinder’s
limited experience base for managing a venture like TMC, and even more concerns about
how he might manage the important implications involving Pete and TLC’s manufacturing
operations. George was now becoming very anxious about the need to talk to Ajit. But he
was far too enthusiastic about the possibilities emerging to slow down his discussions with
Surinder and Pete until Ajit’s return.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
190 Employing planning tools in practice
Pete’s proposition – formally testing design, production and product range assumptions
George’s father had set up and nourished a long- standing TLC tradition of companywide
morning coffee and afternoon tea breaks. They were usually completely informal. Football
and similar topics or the latest gossip usually dominated discussion in small groups, but
sometimes company business was discussed.
After extensive discussions between Pete and George, and considerable thinking about the
issues by both, George used a longer than usual coffee break to outline ‘Pete’s proposition’.
Pete was personable, thoughtful, and respected by all, but slightly shy, and he preferred
George presenting his proposition. George was very pleased to do so. Pete’s proposition
raised some sensitive issues which George wanted addressed along the lines Pete proposed.
However, George believed it would be useful if he was the one explaining Pete’s proposition.
George began by first outlining Dave’s suggestion, Nicola’s plan 3, Bob’s plan 4 or prop-
osition, George’s own TEC proposition, and Surinder’s TMC proposition. He suggested
most TLC employees had no doubt heard about these propositions via the usual gossip, but
he wanted them all to hear the same story as he saw it in its current form. He nodded to Ajit,
who had returned from holiday the previous evening, and indicated that now that Ajit was
back, he and Ajit had a lot of further work to do, with Ajit taking the lead in some key areas,
but he wanted to bring them all up to date now. He then explained that Pete had built on
Surinder’s TMC ideas to develop several suggestions, in effect dealing directly and explicitly
with two sensitive issues which Surinder’s proposition had exposed but avoided discussing.
The first was the advantages and disadvantages of their current product range relative
to competitors. The second was the possible advantages and disadvantages of outsourcing
some of their current production and design processes.
In effect, a questioning of long- standing working assumptions had been kicked off by
Dave. The basis of a new approach to all supply- side issues had been triggered by Nicola.
This development had been very ably transformed into a more general supply- side plan-
ning approach by Bob. George had then taken the supply side ideas further with his TEC
proposition. Surinder had taken George’s ideas further still, into demand side issues includ-
ing sales of machines produced by other manufacturers. But Pete had started to integrate
Surinder’s concerns with some very fundamental core issues for TLC. George and Pete had
spent some time developing some preliminary plans for an internal review of key working
assumptions underlying their current manufacturing operations. There was a lot more work
to do, but he and Pete had provisionally agreed to a four stage process, subject to discus-
sions he now wanted to initiate with Ajit, and further ideas any TLC staff might like to
suggest.
First, Pete and Surinder should test the quality of TLC’s lawnmower designs in terms of
purchaser preferences, benchmarking the TLC designs against all relevant competitors. Pete
thought Surinder might lead, perhaps with support from Nicola, but he also thought buying
in some marketing expertise would be wise. In the past George and Pete had evolved TLC
designs based on feedback from TLC lawnmower owners and garden machinery servicing
companies plus their use of the products in their own gardens. They had never undertaken
a systematic direct investigation of all relevant competitive machines. TLC needed a step
change increase in its marketing management and planning capabilities.
Second, Pete should test the quality of TLC’s current and evolving lawnmower designs
in terms of economic production in a more formal way than they had in the past. Pete
thought he could probably manage this on his own, but they might consider outsourcing
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Using a UP 191
some aspects of design to get the best possible overall approach once he and George and
Surinder were clear about what they thought mattered and wanted to test their prelimi-
nary plans.
Third, Pete should test the efficiency of TLC’s production processes more formally than
he had in the past, perhaps benchmarking it against outsourcing production to a more
efficient supplier of complete or partially assembled lawnmowers with a view to ensuring all
current TLC production processes remained competitive. Pete thought Bob could probably
help with this, and possibly lead on some aspects. Pete suggested visits to carefully selected
other production facilities would be a crucial aspect of this, and George might want to be
involved before it progressed very far.
Fourth, George, Pete, Bob and Surinder, perhaps with help from Nicola, and certainly
with the enhanced marketing capability noted above, should test the effectiveness and effi-
ciency of TLC’s range of products in portfolio terms at an overall corporate strategy level,
with a view to a wider range of complementary equipment, some produced by them, some
supplied by others. Their intention would be preserving a core of high- quality British prod-
ucts built in their current Birmingham facilities, perhaps as only part of a much wider set
of complementary products bearing the TLC brand. Expansion of their current production
with bigger premises and a wider range of products was George’s aim, but this ambition
would need to be tested against the obvious alternatives.
In terms of all four stages just carrying on as at present was an option, but assuming
this was the best choice needed a systematic approach to assumption testing. A fixed order
quantity for controls had needed testing – but it was pure serendipity that Dave had raised
the issue and Nicola had been there to kick off a very enlightening chain of developments.
Pete and George were now convinced that a systematic approach to well- considered change
as and when appropriate throughout TLC was probably crucial to a healthy future for TLC,
starting with core business assumptions.
George thanked Dave for raising the need to test the fixed order quantity assumption for
controls, briefly but sincerely. He then thanked Nicola, Bob, Surinder and Pete for their
subsequent efforts, much more fully as was clearly appropriate, with special emphasis on the
crucial roles played by each. He suggested all TLC employees help to carry these proposi-
tions forward when opportunities to make suggestions occurred to them. TLC did not have
a ‘suggestions box’, but the employees all knew that he was open to suggestions anytime.
He now wanted to spend a significant amount of time discussing all these propositions with
Ajit before they went any further.
George was a natural optimist, and he liked having other natural optimists around him.
However, he was well aware that this sometimes left him exposed to misguided enthusiasms.
George was also aware of two other relevant issues: he knew he was not as systematic as he
might be, and he recognised he was not the natural top- down thinker Ajit was. Nicola had
motivated him to start testing all key assumptions in a more systematic manner, as well as
triggering creativity which led to the emergence of a number of potentially very important
possible opportunities. But he relied on Ajit to keep his more dangerous ideas in check and
to impose some systematic top- down strategic planning on TLC.
Ajit was a natural top- down strategic thinker, who liked to start with questions like ‘Where
do we think we are trying to get to?’ and ‘Why?’ before getting into the details of ‘How are
we going to get there?’
In addition to insisting that George address these top- down concerns, Ajit played the
insightful natural pessimist role in the context of discussing George’s latest enthusiasm – he
saw this as part of his job description. However, Ajit reverted to an optimist out of the office
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
192 Employing planning tools in practice
and in the office once he believed George had a reasonable grip on the threats as well as the
opportunities associated with his current passions.
George and Ajit were a leadership partnership, and both understood the complementary
nature of their roles.
Ajit’s proposition – the TXC group mission, ownership and financing structure
After extensive discussions between Ajit and George, George asked Ajit to outline his propo-
sition at a TLC coffee break.
Ajit started by indicating that George thinking of TEC as a separate company was useful
in several ways, as was Surinder thinking of TMC as a separate company, but Pete’s propo-
sition made it very clear that TLC, TEC, and TMC had to be operated within a coher-
ent overall corporate strategy, and further separate companies needing a collective strategy
might be required over time.
Ajit suggested they use TXC with X = L + E + M + other company identifiers in the
group which might emerge over time as a working label for the new group of companies.
TXC might evolve in ways they could not foresee now, but an initial view of what direction
they wanted TXC to move in and how it should decide its strategy was an urgent priority
concern.
A particular component concern for Ajit was financing for the kind of expansion being
discussed. He did not think simply going to the bank for further finance was viable, and he
did not think doing so would be sensible even if it did prove viable.
This raised quite fundamental concerns – the desirable future of TXC in terms of its own-
ership structure and its basic mission as seen by its owners.
One option which he and George had discussed was George and other members of his
family, Nicola’s mother for example, taking big mortgages on their houses and pulling
together other assets to transform TLC into TXC. One variant of this approach was TXC
becoming listed on the stock exchange as a public limited company after five or ten years.
This could keep the TXC cost of capital reasonably low in the crucial early years and provide
a long- term return for George and his family when TXC was floated in a form they might
find attractive. However, it would concentrate the risk on George and his family in the
short- term and the medium- term, and the current TLC friendly and collaborative family
company feel – very much enjoyed by all TLC employees including him – would probably
be lost if and when TXC became a public limited company.
Ajit emphasised the obvious fact that shareholders in public or private companies of
any kind wanted a return on their financial investment, and George was not an exception.
George was very keen to make long- , medium- , and short- term profits. But George saw
the corporate culture issues associated with his approach to collaboration and respect for all
employees as an enlightened way to achieving lower costs and higher productivity as well as
a matter of social responsibility and enjoying working with people who took pride in what
they did and how they did it. George understood that the TLC employees he wanted to
hire and retain wanted to work in an atmosphere where their contributions were valued and
they could work with people they respected in a collaborative manner. This was a central
aspect of the current TLC corporate culture. It had begun with George’s father, and it could
be lost when George retired if TLC had become a public limited company in the interim.
For example, even if TXC thrived in all respects, if it were a public company a majority
shareholding might be purchased by a non- UK manufacturing company which valued the
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Using a UP 193
TLC brand plus TEC and TMC operations but not manufacturing in the UK, with massive
implications.
George and his father had been long- time fans of the Cadbury family approach, the
locally based (Bournville) world- famous chocolate company founders who began with 19th-
century Quaker zeal to build an ethically driven form of capitalism. The Cadbury’s approach
was respected worldwide, inspiring Milton Hershey for example, America’s ‘chocolate king’.
It would be 25 years before Kraft buying Cadbury- Schweppes in 2010 seriously damaged
Anglo– US relationships from the perspective of some people in the UK. The results received
very hostile UK national press attention immediately and later because of the behaviour of
Mondelez International, the company Cadbury was spun into by Kraft, and ongoing con-
cerns show no current sign of going away. George and Ajit did not have the foresight to
anticipate this outcome, but in general terms, both understood this kind of threat.
George knew TLC needed to raise more capital, but his current preference, which Ajit
strongly supported, was an employee partnership form of corporate structure using features
of the John Lewis Partnership and other similar approaches. Ajit was letting everybody
know this now, with George’s approval, because they both wanted to test this proposition in
terms of everyone’s reaction. If George decided to go this way, it did not mean that all TLC
employees would be required to invest financially in TXC themselves. But it did mean that
part of their income would be dividends on an ‘employee partnership shares’ basis, a form of
bonus linked to profitability. Furthermore, they might opt to take all or some of this bonus
in the form of ‘investment partnership shares’, which would earn a return on the capital they
choose to invest. Further financial investment might be feasible for those who wished and
encouraged for those with senior management responsibilities.
Ajit emphasised that making this kind of employee partnership organisation work effec-
tively and efficiently over the long-term was much more complex than at first sight it might seem. The John Lewis Partnership had a long and highly respected history in the UK retail
trade, with department stores and Waitrose supermarkets, but it had a carefully crafted
structure suited to its particular context. To develop a variant suited to TXC, George would
have to test all key assumptions very carefully, with reference to what worked well and what
failed elsewhere, from all relevant perspectives.
Ajit did not think it was appropriate to share with all TLC employees his view that the
most appropriate ownership structure for TXC needed careful consideration in the context
of an assessment by George of the extent to which his ambitions for TXC were motivated by
his relative priorities in terms of a balanced mix of a desire to:
1 provide a sustainable and secure income for himself and his family,
2 provide his employees with sustainable and rewarding jobs, and
3 provide his customers with high- quality products at a fair price,
4 plus being socially useful in other ways important to George.
However, he did share this view with George. Ajit knew the fourth item probably included
continuing to support George’s favourite football team as a corporate sponsor plus his TLC
apprentice training programme, but he did not say so or elaborate further on any of the
other issues. Indeed, he explicitly omitted even thinking about the considerable complexi-
ties involved, because doing so would not be productive.
Although Ajit did not think it relevant to take this further with George at this stage in
their discussion, you may find it useful to briefly explore the basis of Ajit’s understanding of
what shapes corporate ethos in a very simple organisation like TLC, linking this discussion
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
194 Employing planning tools in practice
to the capability- culture concept and the underlying corporate structure and values concepts
using one of many possible sets of simple common language. Doing so has immediate value,
but it also allows comparisons in the contexts of the tales in Chapters 6 to 9.
Ajit included personal ‘passions’, ‘responsibilities’ and a wide range of ‘issues in between’
in his views of how he, George, and other people might approach all relevant corporate
objectives. He used a common working definition of passions – ‘what normally gets you up in the morning and keeps you going all day’. He used a compatible definition of responsibili- ties (that need to be met) – ‘what normally keeps you fully engaged all day in an appropriate manner even if you are not really enjoying meeting the responsibilities involved’.
From Ajit’s perspective everyone had their own individual views about what were pre-
ferred ‘pleasurable passions’, ‘painful responsibilities’ that needed to be met, and everything
in between. He saw corporate motivation as a very complex set of issues not immediately
relevant to his discussion with George, but he knew George’s pleasurable passions included
high- quality lawnmowers at a fair price and football, his painful responsibilities included pay-
ing all relevant taxes at a fair level, and his issues in between included providing a sustainable
and secure income for himself and his family, providing his employees with sustainable and
rewarding jobs, and training apprentices who might or might not stay on as TLC employees.
Pleasurable passions for Ajit included his family and doing his job as well as he could, and
he was not much interested in either lawnmowers or football, but he strongly approved of
George’s underlying values defined in broad terms to include his generosity of spirit, his
open approach to everyone, his ethical values, and their effect on the ethos of TLC.
Ajit greatly enjoyed working with George and the opportunity provided by his role in
TLC, and he believed that most other TLC employees shared his appreciation of the oppor-
tunity provided by their role in TLC, with a wide mix of pleasurable passions and views on
painful responsibilities. He believed that the TLC corporate ethos was shaped by George’s
own version or variant of Ajit’s passion- responsibility concept and underlying values plus the
reactions of others to George’s actions, and the family firm feel about TLC was part of this.
It was also shaped by the social and political culture which TLC operated in. Ajit thought
corporate ethos was an important issue of considerable complexity, even in the context of
this relatively simple family business. It mattered greatly, and it had an impact on corporate
capability as well as corporate culture, although exactly how corporate ethos operated was
ambiguous and not a matter of immediate concern. It would become a matter of great con-
cern if George allowed the management of TLC to become controlled by an organisation
which wanted the brand but not the manufacturing organisation.
Returning to Ajit’s discussion with George, having raised the issue of ownership and
financing structure options, Ajit then suggested that whatever approach TLC adopted, TXC
corporate governance arrangements would have to be more formal than the structure cur-
rently used by George as TLC’s ‘general manager’. He assumed George would become
Chief Executive Officer (CEO) and Chairman of the Board for TXC if and when it emerged
as a separate company. He assumed Bob, Pete, Surinder and he would all become executive
directors with ‘operations director’ roles. He assumed an employee partnership organisation
would require at least one non- executive ‘employee director’ representative of all employ-
ees, if they went that route. He assumed at least one non- executive director representative
of all investors who were not employees, like Nicola’s mother. Furthermore, and crucially,
he thought two new executive director positions were probably essential – defined by sig-
nificant TXC roles which were currently less significant TLC roles.
One of these new executive director roles was leading corporate planning in top- down
strategic terms, coordinating this with the bottom- up strategic implication of all the current
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Using a UP 195
propositions and those to come. For example, it seemed to him that Pete’s proposition
should be their first priority, then Surinder’s, then George’s, the reverse of the order in
which they had emerged, because Pete’s proposition was at the very core of TLC’s current
operations and Surinder’s TMC was more closely coupled to this core than was George’s
TEC proposition. TXC needed a more ordered and experienced approach to this kind of
top- down strategic thinking than he or George could currently provide, and in his view,
early recruitment of someone who could bring the needed new capabilities but fit well with
existing TLC capabilities and culture was an important consideration.
Ajit believed a wide range of backgrounds might be suitable for the new corporate plan-
ning director TXC needed, but someone with extensive marketing experience who could
mentor Surinder and help shift the corporate culture towards marketing would be particu-
larly attractive. TLC needed this culture shift, support for Surinder was needed, and meet-
ing both via the new corporate planning director with marketing expertise seemed a useful
possibility.
The other new executive director role was leading all the change- management ‘project
planning’ – the portfolio of projects and programmes which would achieve the envisaged
corporate changes. For example, Pete needed someone to help him lead the change pro-
gramme he was suggesting they should embark on who could later help Surinder and George
with the coordinated change- management programmes associated with TMC and TEC.
A relatively stable past had allowed George to deal with the top- down strategic corporate
planning and the project (change- management) planning leadership roles plus the CEO and
chairman roles, with support from Ajit. However, the proposed transition from the current
TLC to TXC would need careful preparation and testing of emergent corporate structures.
This ought to involve at least two more people with additional skill sets who could also help
them to lead the ongoing changes.
Ajit’s natural capabilities had been marginally strengthened by using some features of the
UP demonstrated by Nicola and the resulting development by Bob plus the propositions by
George, Surinder and Pete in a somewhat different manner. However, he was not a fluent
user of process ideas, and he recognised that he and everyone else needed support if the
degree of change envisaged was going to be pursued effectively and efficiently. He found
it difficult to translate his colleagues’ use of the UP concept in simple bottom- up analysis
terms into the forms needed, as discussed in Chapters 6, 7, 8 and 9 of this book, because
there was a big difference between bottom- up and top- down thinking, and he lacked the
requisite experience base. Bottom- up thinking was not the focus of his normal approach to
the issues just discussed. Indeed, his natural top- down thinking approach with a direct con-
cern for strategy was crucial to balance the bottom- up creativity of the others.
Testing assumptions about how top- down and bottom- up thinking meshed was crucial.
Ajit did not have formal training in this area, but his significant natural ability included the
wisdom to see when more capability was needed because of his own limitations.
Ajit might have noted that Nicola was very lucky that her knowledge of inventory models
and a UP concept fitted the controls context so well – with a corporate need for wider access
to relevant generic models and specific processes as well as other variants of the universal
process which Nicola could contribute to but not necessarily lead. Furthermore, he might
have noted that Bob was clearly able and experienced and very insightful to have seen the
advantages associated with transforming Nicola’s plan 3 into Bob’s plan 4, but he might
benefit from more exposure to wider supply chain management experience via professional
courses. Pete was a very capable operations manager, but some more exposure to formal
concepts in this area would probably be useful. And as he indicated earlier, Surinder could
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
196 Employing planning tools in practice
definitely do with some additional marketing expertise and experienced support to exploit
her obvious potential.
Creative testing of all the UP front- end working assumptions
If he had been familiar with the UP concept and the rest of this book, Ajit might have
understood and mentioned several additional points.
First, the testing of working assumptions associated with models used to shape plans is the
traditional focus of OR process explanations, but the importance of testing working assump-
tions is often overlooked by those who use models from other professional perspectives. It
is crucially important to test underlying assumptions for all relevant models, including those
borrowed from behavioural and physical sciences not always seen as cognate sciences, what-
ever the professional decision- making perspective involved.
Second, extending the testing of working assumptions to a view of plans which is much
broader than just the models used to shape aspects of the plans is an important issue
highlighted by adding the ‘create and enhance the plans’ phase to the UP concept. It
significantly broadens the scope of the UP concept, as demonstrated by Bob’s plan 4. It
acts as a gateway concept which facilitates bottom- up strategic thinking instead of just
tactical thinking.
Third, extending the testing of working assumptions to ‘capture the context’ assump-
tions is highlighted by adding this phase to the UP, and it further broadens the scope of
the UP. This was demonstrated by Bob seeing the need to shift to a much broader supply
chain perspective which triggered George thinking about a new business opportunity and
then the successive propositions generated by Surinder, Pete, and Ajit. Creative testing of
the assumptions associated with ‘capture the context with appropriate clarity’ ultimately had
to address the implicit TLC assumption ‘TLC is a family owned lawnmower manufacturing
business’. George was the most appropriate person to lead on this, but Surinder, Pete, and
Ajit built on George’s leadership in the way good leaders encourage. The UP concept should
encourage this kind of thinking and interaction. Understanding the relationship between
assumptions and reality is the key to creative leaps. Creative leaps are central to opportunity
efficiency. The higher the strategic level of the assumptions, the bigger the prizes for creative
leaps. The early analysis by Nicola was, in effect, an investment in insight and understanding,
a learning process with returns which could be leveraged by an understanding of all relevant
issues in top- down terms. To fail to look for and make associated creative leaps is to waste
this investment. Creative leaps are central to clarifying opportunity, complexity, uncertainty,
and risk. Careful analysis without inspired synthesis can be a serious waste of opportunities.
Creative leaps and inspired synthesis may be led by specific individuals, but corporate groups
who enjoy the kind of teamwork- based capability- culture assets demonstrated by TLC are
particularly well placed to make effective use of their opportunities.
Fourth, creative testing of ‘select and focus the process for appropriate clarity’ is also
important, to complete the testing of UP front- end assumptions. Adding the ‘select and
focus the process’ phase to the UP has yet to play a demonstrable role, but Chapter 6 starts
to do so, and Chapter 7 does so more extensively. Chapter 7 also outlines a project planning
process which Ajit’s suggested project’s director might bring to TXC, Chapter 8 outlines
a corporate planning approach which Ajit’s suggested corporate planning director might
bring to TXC, and Chapter 9 extends the capabilities of the Chapter 8 framework.
Fifth, this chapter demonstrates the effective use of a simple basic form of the UP con-
cept as understood by Nicola and built on by the others, but it does not even attempt to
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Using a UP 197
demonstrate how best to turn George’s proposition and all subsequent propositions into
operational plans as explored in following Part 2 chapters.
Ajit’s crucial observations about the need for top- down thinking plus necessary addi-
tions to TXC competence before they go very far could be extended to include experience
with unfamiliar UP and SP concepts as well as the knowledge and experience specific to
wholesale operations and marketing which TLC did not currently have. Put a bit differ-
ently, if TLC wants to build on George’s proposition and any of the other subsequent
propositions in an effective manner, TLC would be well advised to acquire all the plan-
ning capabilities discussed later in this book. The UP concept employed by this chapter
achieved a great deal for TLC, in part because the tale of this chapter was based upon
working assumptions explicitly chosen to maximise what might be achieved. However,
while Nicola’s approach in passes 1, 2 and 3 generalises in a fairly obvious way to Bob’s
plan 4, and use of the front- end phases of a UP is clearly linked to the propositions gener-
ated, it is not obvious how the UP can be used to develop these propositions. The basic
difficulty is that the very general (unrestricted) nature of the UP concept is valuable when
starting with a simple and specific ‘problem’ like ‘What order quantity is appropriate?’ But
the model focused interpretation of the UP Nicola understood is not broad enough or
capable enough when we start with a much more complex context like the need to con-
sider how to develop TXC as a whole or as specific new companies in its portfolio, issues
addressed in Chapters 7 and 8. We need a more developed set of capabilities for the UP
which recognises the need to address these issues, and seeing this need as a key capability-
culture concept issue is useful.
Finally, it might seem that the capability- culture concept as elaborated in Chapter 2 has
yet to play a demonstrably important role, but Nicola’s arrival would arguably have had
very little immediate impact if TLC did not have the capability- culture asset characteristics
portrayed in this tale. TLC had a capability-culture with deeply embedded natural compe-
tence and teamwork, and when TLC acquired the modest UP concept process knowledge
Nicola brought, it was used very effectively. This role for capability- culture assets was shaped
by the underlying corporate structure and the way George interpreted and developed that
structure. The capability- culture concept lies behind and reinforces Ajit’s case for a sig-
nificant strengthening of TLC capabilities before making radical changes in terms of TXC
development. It should also help to clarify the ‘learning process’ ideas introduced in this
chapter and later tales and the capability- culture interactions. This capability- culture assets
and corporate structure relationship are also used in a range of quite different contexts in
following chapters.
Looking wider or deeper at key strategic and tactical choices
The iterative process illustrated by this chapter’s tale may suggest a smooth succession of
iterations, gradually seeing a bigger picture emerge as ‘layers of the onion’ are peeled away,
and gradually developing more definition detail of what really matters as what matters most
gradually emerges from the initial fog of ambiguity. The TLC tale was designed to build
understanding in layers which is bound to give this impression. However, in practice itera-
tions do not usually proceed as smoothly as this chapter’s tale suggests. Even if we have
useful prior views about the scope for bigger picture or ‘greater depth’ perspectives, we may
need to overcome some significant surprises and obstacles to get there.
Nicola’s discussion of her three pass approach explicitly acknowledged that she simply
ignored dead ends she had explored in her discussion with George. Bob’s pass 4 discussion
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
198 Employing planning tools in practice
also indicated that the development of his proposition was not entirely straightforward. This
was implicit in all the later proposition development discussion when not mentioned.
Successive testing of working assumptions and creative interpretation of results is the basis
of the iterative nature of the UP. Sometimes the consideration of further possible iterations
suggests that enough has been done and the process should stop. Sometimes it suggests a
need to ‘look deeper’ – as when planned and unplanned stock- outs were addressed and a
free credit period plus inflation- driven appreciation was identified as a key issue. Sometimes
looking deeper suggests big opportunities may be available if the process changes direc-
tion to ‘look wider’ – as when the much bigger order suggested by the free credit period
in conjunction with inflation identified by Nicola triggered Bob’s move into a supply- side
management strategy which meant delivery by the controls supplier when it suited the sup-
plier based on a collaborative supply contract. George’s proposition and the subsequent
TXC generalisations kept looking wider, but each required later attention to more depth.
Iteration management involves some subtle issues which significantly benefit from craft
and experience, associated judicious anticipation in particular. Effective and efficient itera-
tion management also benefits from remaining flexible and open to surprises.
Any of the characters in this tale might have had less luck than they all enjoyed. The tale
of this chapter involves an explicit assumption that good luck does not always happen, but
good luck needs to be appreciated and exploited when it does happen, and bad luck needs
to be overcome too.
It was convenient starting our series of chapter length tales with a story which illustrates
an optimistic perspective by people who on this occasion are lucky. Later chapters give
more emphasis to bad luck. Sometimes bad luck needs to be overcome when quite serious
unanticipated outcomes occur. Sometimes bad luck needs to be avoided in advance. Often
a combination of both is the best planning basis.
This is a very brief section. However, the subtle nature of iteration control is of crucial
importance in terms of strategic vision and strategic clarity plus underlying tactical clarity.
It involves a key capability- culture asset. That is why a separate section has been used. It is
central to the significant discussion in all following chapters of this book, although drawing
attention to its importance each and every time is avoided because doing so would soon get
tedious.
Consolidating key concepts
To help you synthesise the strategic clarity roles of some of the ideas explored in this chapter,
this section provides some consolidating comments.
Goals– plans relationships development
A lack of clarity about all relevant objectives is a fundamental problem for any kind of plan- ning. The obvious follow- on problem is a failure to understand how plans should lead to
achieving those objectives, including the use of appropriate contingency plans. This is what
goals– plans relationships development is about, and clarity about the relationship between
what is planned and the outcomes anticipated in terms of all relevant objectives is important.
What Nicola’s pass 1 demonstrated was the development of a very simple starting point
set of assumptions about goals– plans relationships for that part of Pete’s current plan which
had been questioned – the optimal order quantity. This starting point addressed the trade-
off between the cost per order and the cost of holding inventory per item per unit of time
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Using a UP 199
in the simplest terms feasible given a competent level of clarity about the working assump-
tions employed. Uncertainty about the expected value of both these costs was addressed in
terms of an interval estimate of the expected values, with an explicit focus on the wide range
associated with Co and the seasonal variations suggesting bigger orders in the spring period
clearly flagged. A simple understanding of the relationship between stochastic (probabilistic)
and deterministic (non- probabilistic) models to explore goals– plans relationships was evi-
dent from the outset. But the emphasis in pass 1 was a simple view of what both order costs
and holding costs involved, initially used in a prescriptive deterministic model of a central
causal relationship in the optimal order quantity concept. This model suggested plan revi-
sion from 2,000 controls per order to 1,500 controls.
Nicola’s pass 2 identified and clarified a starting point set of assumptions about goals–
plans relationships for the safety stock part of Pete’s current plan which had not been ques-
tioned in pass 1. However, it was worth attention in an integrated way to review Pete’s
current plans for the supply of controls as a whole. Furthermore, the stochastic (probabilis-
tic) modelling approach illustrates the way a statistically based understanding of uncertainty
in a HAT framework can be integrated with a very different kind of model, tailoring a hybrid
approach to what is needed.
Nicola’s pass 3 and her creative plan 3 interpretation demonstrated the value of a deeper
look at these relationships in terms of assumptions about free credit periods and related price
inflation. In this case the aspects of uncertainty addressed, completely overlooked earlier,
were rather different, and the plan revision from 1,500 to 7,000 involved a massive move
in the opposite direction. This illustrated how important what we overlook can be, a way of
looking at unknown unknowns driven by stealth assumptions.
Bob’s pass 4 to create plan 4 broadened the scope of the plans, dropping the assump-
tion of a fixed order quantity determined by TLC, but retaining the testing of all relevant
assumptions to shape the plan. In this case understanding the value of going beyond the
original plan assumption that an optimal order quantity was appropriate was very different
again, and even more important in terms of cost savings. This further demonstrated how
important testing this kind of assumption can be in relation to clarifying goals– plans rela-
tionships. We need to test all assumptions associated with plans, not just those associated
with aspects which have been modelled.
The basis of the resulting plans and their relationship with associated objectives remained
fairly clear for Bob’s proposition. However, the plans necessary to implement the following
propositions started to become relatively ambiguous.
All the subsequent proposition developments involved further questioning of earlier TLC
corporate assumptions associated with goals– plans relationships, clearly a beneficial process.
George’s proposition was a key one for unlocking further creativity. It questioned the basis
of the family business in a way Surinder could build on, which, in turn, exposed the need for
Pete to develop his proposition.
Ajit introduced some basic top- down goal planning ideas. All organisations need a top-
down approach which asks ‘Where do we think we want to get to?’ and ‘Why?’ Goals– plans
relationships need top- down assessment. Ajit made it clear that this top- down approach
should lead to a coordinated view of all corporate objectives, addressing how all aspects of operations management and project management contribute to overall performance, to
plan how to get there. But Ajit did not address how to undertake any of this planning, even
in outline what needs to be done terms. These issues are explored in Chapters 6, 7, 8 and 9.
One of this book’s framing assumptions is all target readers need the broad understanding of the role of processes and models in helping planners to achieve clarity about goals– plans
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
200 Employing planning tools in practice
relationships. This chapter provides a foundation for this aspect of strategic clarity, exploring
its tactical clarity basis in a relatively simple context.
Useful perspectives on the opportunities which underlie unexplored uncertainty
A starting point for several linked issues briefly explored in this subsection is a quotation
often attributed to Sir Winston Churchill:
A pessimist sees the difficulty in every opportunity; an optimist sees the opportunity in every difficulty.
Churchill may have borrowed the underlying idea from someone else, but whoever first cap-
tured this idea, Churchill’s version is a very useful reminder to stay optimistic when appropri- ate even when confronted with great difficulties, interpreting difficulties in generalised terms as unwelcome constraints, problems, challenges, risks or threats.
During my involvement with the IBM UK Forum 2 culture change programme in the
1990s, an often- quoted mantra embedded in IBM culture, which can be seen as a variant of
the Churchill quote, was:
There is no such thing as problems – only opportunities,
with some of the older and wiser hands muttering under their breath the proviso:
although sometimes opportunities are insurmountable!
This IBM mantra can be seen as a very useful reminder to focus on the opportunities when appropriate, remembering that exploiting some opportunities may not be feasible. In a worst- case scenario trying to capture ‘insurmountable opportunities’ can serve as a self-
inflicted fatal wound.
It is obviously important to understand that not all opportunities present as unwelcome
constraints, problems, challenges, risks or threats. Some opportunities present as ‘gifts’.
Gifts may be just variants of fate or good luck, and gifts may be the straightforward genuine
benevolence of family or close friends, but especially in a commercial context, gifts may be
conditioned by an expectation of reciprocation. An obvious and possibly relevant saying
here is perhaps:
Never look a gift horse in the mouth,
with constructive tension provided by the proviso,
but sometimes assumptions about ‘gifts’ need a ‘there is no such thing as a free lunch’ test.
Opportunities of many different kinds have been considered in this chapter’s tale, with-
out always using the term opportunity. The starting point was Dave attempting to find an
opportunity to reduce the number of times he had to go to the control supplier by testing
the current TLC order quantity assumption. George, Surinder, Pete and Ajit’s propositions
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Using a UP 201
to test current TLC corporate boundaries were end points for this chapter but opportunities
which are starting points for much bigger journeys to a TXC company group. The central
part of the journey in this chapter involved Nicola looking for useful ways to test and relax
key assumptions underlying her EOQ model, followed by Bob seeking a much better supply
plan and supplier relationship with the benefits of Nicola’s ideas plus the further benefits of
generalisation of the assumptions underlying the EOQ concept.
When Nicola began developing the basic EOQ model in pass 1, the basic form of the TLC
plan was predetermined – assumed to be a fixed order quantity. The uncertainty addressed
was limited to ‘What value should this fixed order quantity take, and why?’ The underlying
complexity first addressed was the relationship between holding cost and order cost per unit
time for different order quantities. Both Nicola and George needed to understand how this
relationship worked, and the asymmetry associated with rounding the order quantity for
convenience and errors in their parameter estimates. The basic EOQ model should be seen
as a way of educating people about this relationship, not a simple golden rule with a formula
for inventory management to crank out optimal solutions.
When Nicola began pass 2 by demonstrating planned stock- outs altered the result so lit-
tle that rounding produced exactly the same answer, she was illustrating the robustness of
her pass 1 result to the implicit assumption that planned stock- outs would not be consid-
ered. This was complementary to demonstrating that unplanned stock- outs were important
but arguably overplayed by Pete’s three day (300 controls) safety stock. Reduction to one
day (100 controls) plus an informal contingency plan involving enhanced communication
between Pete and Dave might save significant sums. Resolution of additional uncertainty
by exploring underlying complexity that might matter, with everyone understanding why it
did or did not matter, was the purpose of the exercise, not just a more sophisticated model.
When Nicola began pass 3 by exploring the implications of a free credit period and simul-
taneous significant inflation, the optimal order quantity which pass 1 moved from 2,000
down to 1,500 shot up to 7,000, making it very clear that passes 1 and 2 were taking TLC in
the wrong direction. The 7,000 figure in itself was not all that important. What was crucial
was understanding that if you can enjoy the upside of inflation by buying large quantities of
inputs with someone else’s money, you are onto a winner – an opportunity which passes 1
and 2 had completely missed.
Bob’s further developments in his pass 4 plan built upon Nicola’s pass 3 to grasp what
really mattered. They did not actually want to order a large amount every few months
for immediate delivery. What they wanted was a fixed price for six months, orderly price
increases within a rolling annual contract, and no stock- outs. It was better for TLC and for its controls supplier to let the controls supplier deliver smaller amounts when it suited
the controls supplier so long as TLC did not run out, paying for the controls according to
an assumed uniform rate of usage over the six- month period. Some of the benefits to the
controls supplier could be leveraged into lower prices. This could produce a win– win situa-
tion, with Bob making sure that TLC got as much of the joint pot of winnings as he could.
Getting beyond the EOQ model to an understanding of the My- Place- Your- Timing
approach, and an ability to sell it to its supplier, is a demonstration of clarity efficient resolu-
tion of uncertainty and underlying complexity to harvest an opportunity that was just wait-
ing to be found. Nicola and Bob were on a treasure hunt, looking for opportunities amongst
the uncertainty previously unexplored. Generally speaking, people do not find treasure if
they do not know both what to look for and how to search for it in a systematic manner.
The tale ensured that Nicola and Bob jointly had the toolset, skill set, mindset, and good
luck to find treasure.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
202 Employing planning tools in practice
What they found, having understood its nature and implications, then inspired George,
Surinder and Pete to undertake similar searches for comparable opportunities in other areas
of the business. In each case there were associated risks which needed attention. But it was
the opportunities that mattered most, and responses to uncertainties which jointly captured
opportunities while limiting risks were central to all the propositions. Indeed, there is an
important argument that the biggest single risk for TLC was failing to capture sufficient
opportunities, leading to corporate failure because the competition was better at finding
and capturing opportunities.
However, as Ajit pointed out, before they set out to make too many decisions they needed
to enhance their capabilities without weakening their considerable current capability- culture
assets. They had been very good at operating a stable business model which had changed
very little over a long time. Radical change required some new toolsets and skill sets as well
as the new mindsets already emerging.
Nicola had started with a very self- contained and limited amount of uncertainty – ‘What
should the order quantity for controls be, and why?’ But TLC was now addressing very
large and complex sets of uncertainty, needing tools beyond the collective capabilities of the
current TLC team.
A common thread in the contributions of all this chapter’s characters, starting with
Dave, was:
Seek opportunities by testing all working assumptions, including those perceived as difficulties for current plans, by creative exploration of alternatives using a generalisation of those assumptions with a view to discovering possible new opportunities, and then testing working assumptions about these possible new opportunities in an iterative systematic process which keeps it simple throughout.
The starting point for a generalised view of what opportunity efficiency is about is a synthe-
sis of this common thread throughout this chapter, the Churchill quote, the IBM mantra,
and the gift- horse sayings. You may look at this starting point in many ways, and choose to
emphasise various features.
One aspect is the recognition that a well- managed unwelcome constraint, problem, chal-
lenge, risk or threat – or any other variant of these terms – can become a very big oppor-
tunity. Constraints, problems, challenges, risks or threats are usefully viewed as difficulties
waiting to be transformed into opportunities. This involves testing all relevant current
assumptions and resolving ambiguity and complexity in a clarity efficient manner with a
view to opportunity efficiency in a general sense whenever this is feasible. Most of this chapter has this emphasis.
Another aspect is the recognition that a badly managed opportunity can become a very
serious difficulty. All the TXC propositions have this possibility if they are not individually
and collectively managed in an opportunity efficient manner, as Ajit observed.
A further aspect is the importance of starting with a systematic search for opportunities
before getting too bogged down with the difficulties. This is, in part, linked to the com-
monly observed notion that most people cannot work effectively in ‘optimist mode’ and
‘pessimist mode’ at the same time – because we cannot multitask effectively in this way.
Indeed, some people by nature work in one mode or the other virtually all the time. This
implies ordering these two modes of operation when designing processes is essential, and
having a well- balanced and mutually supportive team is crucial.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Using a UP 203
Further still, making sure that all the opportunities which present themselves as gifts
are also identified and addressed effectively and efficiently is equally important, testing all
assumptions. A top- down corporate planning view of this ‘futures planning’ issue which
has to work alongside ‘goals planning’ is discussed in Chapter 8, along with long- term,
medium- term and short- term planning frameworks, all separable but interconnected aspects
of an enlightened approach to planning.
Ajit’s recommended new corporate planning director would need to develop a top- down
approach to both opportunities and threats which encouraged ongoing bottom- up develop-
ment of strategy propositions, integrating all relevant bottom- up strategic corporate think-
ing with systematic top- down processes.
Achieving clarity efficiency by starting with simple plans and basic models
Simple plans and basic models are very valuable, as illustrated by the TLC fixed order quantity
plan for controls and basic EOQ model used in Nicola’s pass 1 analysis. The basic structure
and guidance such simple models provide can be a crucial starting point for examining and
understanding key features of a decision situation and guiding further analysis. In particu-
lar, they can identify basic goals– plans relationships which need to be understood to shape
plans and suggest appropriate choices. The ‘prescriptive’ (yielding an optimum solution) and
‘deterministic’ (non- probabilistic) aspects of an EOQ model can be an asset, not a limitation
or liability, because the deterministic aspect facilitates a clear focus on particular aspects of the
ambiguity being addressed without the complexity of associated variability, and it facilitates
a prescriptive modelling approach which yields a recommended optimum plan given the
assumptions, sometimes the ‘right kind’ of initial simplification of a complex situation. As
Stephen Ward (1989) observed when describing the ‘constructive simplicity’ concept which
underlies both enlightened simplicity and clarity efficiency, constructive simplicity designated
and described in various ways is not a new or revolutionary idea. For example, the EOQ or
‘Wilson’ rule has been around for about a century, and it has been widely used, with good
reason. But it is not a tool to be used in a mechanistic manner to crank out ‘optimal order
quantities’, as very vocal critics of simplistic EOQ use have argued from the outset.
An EOQ model is comparable to the minimum clarity estimate approach outlined earlier
in Chapter 2 in the sense that both involve a minimum level of complexity to get started.
But there are important relevant differences. The minimum clarity estimation approach of
Chapter 3 was designed to be robust in relation to a very broad set of working assumptions
and provided unbiased first pass estimation is all that is sought, and fairly conservative esti-
mates of both the expected outcome and the associated range posed no concerns, deeper
analysis may reasonably be assumed to be less likely to prove productive than more obvious
sources of concern revealed by the first pass. The EOQ model was not developed with this
kind of robustness in mind, and the underlying issues are different. It is worth constant
reminders to start with a simple capture of the key features of a context, but it is equally
important to appreciate that further complexity may need to be recognised and explored,
and knowing when this is the case can be important. Nicola’s pass 3 illustrates this.
Simple probabilistic models can also be very useful, even if they are descriptive in their
basic form (not yielding a recommended optimum plan), as illustrated by the stochastic
safety stock model introduced by Nicola’s pass 2. Complexity is added, but in the spirit of
clarity efficiency, it more than pays its way.
All passes through the UP of Figure 2.1 should involve keeping the models as simple as
possible, preserving clarity efficiency but facilitating the introduction of further clarity as and
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
204 Employing planning tools in practice
when appropriate. This was illustrated by Nicola’s pass 2, further illustrated by the consid-
eration of inflation and free credit in Nicola’s pass 3 and her plan 3. It was illustrated again
in Bob’s My- Place- Your- Timing plan 4. Yet more follow- on illustrations build on Bob’s plan
4 in a series of TXC propositions. Central to this incremental addition of more clarity in an
efficient manner was the transparency of assumptions, the rigorous testing of assumptions,
the highly iterative nature of the process, and the concern for a form of synthesis which
extends to ‘interpret the results creatively’. Enlightened simplicity as a basis for clarity effi-
ciency involves many features not yet touched on, which subsequent chapters will illustrate.
However, this chapter’s tale provides a useful starting point.
Developing a ‘feel’ for when deterministic models will be useful is an important capability,
central to the EP skill sets which are needed for corporate clarity efficiency and corporate
tactical clarity which corporate strategic clarity has to address. If you do not have this feel
yourself, you should be aware of the need to work with those in your organisation who do
when it could be useful – otherwise, you are going to lack strategic clarity and miss oppor-
tunities. If nobody in your organisation has these capabilities, this is a capability- culture
liability worth attention by someone, even if you are not directly interested yourself.
When parameter values are uncertain it is always important to understand sensitivity
issues, including asymmetries associated with the implications. Beginning with a clear
understanding of plausible ranges to estimate expected values is always good practice, but
sometimes the additional effort of a stochastic model is not clarity efficient. One reason,
illustrated by this chapter’s tale, is systemic uncertainty (such as seasonal variations) may
be involved, and if this is the case, some aspects of it may need to be clearly understood
and treated effectively. But this may not involve useful modelling. Furthermore, if the
outcome of a modelling approach will be a very complex plan, it may be useful to consider
an altogether different approach, a very different approach may be the consequence of
other concerns, and it may be useful to delay any complex modelling and planning until it
is clearly needed.
The example model this chapter is centred on is a prescriptive calculus- based model, a
particularly powerful exceptionally simple model in terms of insight provided the assump- tions hold. Such models predate the mathematical programming models which began their extensive development and use in the 1950s as computer capability increased, and they can
be seen as a very simple special case of a mathematical optimisation approach which facili-
tates a very traditional prescriptive interpretation even when great complexity is involved.
Some people who are particularly fond of ‘optimising’ things are very keen to use prescrip-
tive models whenever feasible, but often it is wiser to settle for a less ambitious descriptive
model, used within a less direct optimisation process framework. Sometimes it is useful to
stick to a qualitative analysis ‘exploratory model’ of the soft OR variety, which simply helps
with exploring the relevant issues. Sometimes a quantitative exploratory model based on
simulation as explored by Pidd (1996) can be very useful. Decision trees in a HAT frame-
work as used for safety stock modelling in this chapter are just a very simple example of a
descriptive model approach. Prescriptive models are more useful if suitable models are avail- able, but the assumptions required may not be tenable, and if the assumptions required are
not obvious, these models can be extremely misleading.
When Nicola addressed safety stock sizing, uncertainty associated with stock- outs had
to be addressed in stochastic terms to estimate the expected costs involved. Avoiding
specific probability distribution assumptions led to a descriptive basis used within a tabu-
lar framework based on an underlying decision tree to seek optimality less formally – a
simple HAT approach. This is an example of a descriptive model used within a less formal
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Using a UP 205
process- based approach to optimisation. Prescriptive safety stock models are available, but
they make strong assumptions about the shape of the probability distribution defining
uncertainty associated with demand – a Normal (Gaussian) distribution for example – in
addition to assuming expected values for holding and stock- out costs. They provide no
way of testing this kind of assumption, and they can also encourage less concern about
approaches to testing other assumptions, including sensitivity analysis about all parameter
assumptions.
Developing an understanding of when the complexity of stochastic modelling using an
informal process- based approach to optimality is needed for even a basic level of clarity, and
how to couple simple stochastic models plus linked deterministic modelling is part of the
EP toolset and skill set craft which Nicola brought to the family firm. If you do not need to
have these skills yourself, you may still need to know when they could be useful and how to
get access to people who can apply them for you when this would be useful.
Appreciating trade- offs when seeking opportunity efficiency
An important aspect of shaping plans via models is the role of trade- offs between perfor-
mance measures. In this chapter’s tale trade- offs between order costs and holding costs was
a central issue. Simple deterministic models can be particularly useful in clarifying the nature
of appropriate trade- offs when different parties have different priorities.
Informal approaches are usually biased by the provider of the intuition. This can shape
informal plans to a greater extent than related formal processes which make a point of mini-
mising bias. For example, Pete’s intuitive approach led to a slightly larger order quantity
than the pass 1 analysis. This was marginally more expensive for TLC if the pass 1 working assumptions were accepted as valid, and marginally more convenient for Pete. This result should not have been a surprise. Pete did not need to be familiar with EOQ models to have
an intuitive grasp of the trade- offs involved, for himself and for TLC. On a learning- by-
doing basis, Pete would tend to settle on a reasonably appropriate solution in these terms.
However, even a very simple formal model may provide the insight for a modest improve-
ment, in terms of separating out what is in the best interests of the individuals and what is
in the best interests of the organisation, as well as in terms of classifying what key trade- offs
need attention, provided the assumptions used are robust. What pass 3 and 4 suggested was Pete was actually closer to the true optimum than Nicola’s pass 1 plan, albeit perhaps for
the wrong reasons. That is, if Pete had not thought about inflation and free credit period
assumptions, bias might be the explanation.
Explicit treatment of uncertainty associated with Pete’s safety stock was useful to capture
further trade- offs, in this case between holding costs and stock- out costs. This provided
much more fertile grounds for improvements in cost, despite the very low level of uncer-
tainty involved, in terms of zero, one, or two days needed to replenish stock and the small
size of the safety stock relative to the order quantity. Again the validity of Pete’s intuitive
solution was debatable, but in the context of Bob’s plan 4 level of clarity, both bias issues
no longer mattered.
Seeking opportunity efficiency including clarity efficiency in the most general sense feasi-
ble within the EOQ framework was clearly the thrust of pass 3, and the ‘interpret the plans
to exploit creativity’ phase developments by Nicola for plan 3. Bob builds on this for plan
4 but moved beyond the EOQ formula. George, Surinder, Pete, and Ajit then each built
their propositions on the basis of the clarity provided by the preceding stages of the plan-
ning process.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
206 Employing planning tools in practice
Risk efficiency in a basic expected value form as a basis for opportunity efficiency
A minimum level of expected cost without any direct concern for the level of associated
cost risk was the objective pursued by Nicola on behalf of TLC in terms of her formal
EOQ models. She recommended taking account of asymmetric penalty costs associated
with departures from expectations when dealing with the cost of stock- outs, the basis of her
rounding- up recommendation for Q0*. But she did not recommend any formal trade- off
processes involving expected cost and associated cost risk in the sense associated with the
decision diagrams of Figures 3.6 and 3.7 and the efficient frontier of Figure 3.8. This was
appropriate because cost risk in these terms was not relevant. When the level of risk involved
means that risk does not need attention in the sense of Figure 3.8, ignoring risk is sensible.
That said, to understand her working assumptions within broader framing assumptions,
from a broader perspective on relevant costs each time a successive pass identified an expected
cost reduction, this was an improvement in terms of risk efficiency, and it may be useful to
see these improvements in terms of finding opportunities by working in the opportunity
management area of Figure 3.8, moving from a point not on the efficient frontier towards
point a without being concerned about either the absolute level of risk or risk– reward trade-
offs at the boundary point being sought.
If you are familiar with modern decision analysis and use its concepts as part of your perspec-
tive on decision making, it is important to understand that the equivalent of ‘risk neutrality’
in modern decision analysis terms is a common and useful mode of seeking risk efficiency in
EP terms, directly equivalent to aiming for point ‘a’ in Figure 3.8. However you prefer to
look at this issue, a focus on point ‘a’ does not make the risk efficiency concept less useful. It
simply demonstrates a common and very important special case. If risk, in the sense of down-
side departures from expectations, is not important in terms of objectives directly relevant to
decisions being addressed, we can concentrate on the opportunity implications of uncertainty,
and ignore expected reward trade- offs with the associated risk of downside departures from
expectations. We can aim for point ‘a’ in Figure 3.8. Nine times out of ten this may be clarity
efficient and effective in an opportunity efficient sense – the best way to proceed.
It is also important to understand that the opportunity/risk datum concept discussed
earlier involves a degree of subjective convenience. Relative to George’s starting posi-
tion, Nicola uncovered enormous opportunities. Relative to where Nicola, Bob, George,
Surinder, Pete and Ajit working together with other TLC staff could take TLC, prior to
Nicola’s arrival TLC was arguably risk and opportunity inefficient to a degree which could
lead to bankruptcy if TLC competitors were to prove better at opportunity management.
Risk efficiency is always a basis for finding opportunities in the process of seeking oppor-
tunity efficiency, but when risk, as modelled directly in terms of downside variations from
expected outcomes, is not a useful issue to talk about in terms of risk– reward trade- offs, a
focus on all relevant uncertainty from an opportunity management perspective can be more
useful. This does not mean we can or should forget about risk in a more general sense.
Nicola was careful to consider some of this more general risk. Bob’s plan did so too. George,
Surinder, Pete and Ajit’s propositions would require significant attention to risk.
Risk and risk appetite as a multiple attribute concern
Throughout this chapter, minimising expected cost has been the modelling focus, as pointed
out in the last subsection and earlier. However, risk in a general sense has been recognised
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Using a UP 207
as important and addressed, sometimes without emphasising that risk was involved or using
risk management language.
For example, George did not risk offending Dave by failing to take his suggestion seri-
ously, and Nicola explicitly avoided the risk of offending Pete for negligible gain because
that kind of risk often matters a great deal in practice.
Furthermore, Nicola suggested it might (or might not) be worth forgoing a change in
the status quo with implied criticism if only £5 per month was saved, but a larger saving
might make the trade- offs worthwhile, even if inconvenience for Pete and other staff were
involved. A moderately aggressive approach to cost risk is not incompatible with a cautious
approach to offending staff. A sizable cost improvement may make inconvenience as well
as sensitivity to changes worth tolerating, but a diplomatic approach to managing this kind
of change may be essential to avoid the risk of permanently damaged relationships. What
some people call risk appetite is a multiple attribute concern, and some attributes which are
important are not measurable.
Furthermore, there are some serious sources of supply chain risk which Nicola drew
George’s attention to. For example, if the firm producing controls went into liquidation,
or suffered a major industrial dispute, or dramatically increased prices, and there were no
alternative suppliers, TLC might be in significant difficulty. George’s proposition could be
seen as building on this in the engine supply context, when it mattered even more.
All the TXC propositions involved significant risk that was not even considered. These
opportunities require significant further attention beyond the scope of this chapter.
Put slightly differently, a clear understanding of all the relevant issues can lead to a suc-
cessful transformation of difficulties into opportunities, but requisite knowledge and skill are
important, and a limited understanding of what matters most can be dangerous, transform-
ing an opportunity pursued into a risk or a problem.
In some ways even more important is what might be called ‘missed opportunity risk’. If
the opportunities identified by TLC had been missed, much higher expected costs would be
endured, perhaps ultimately leading to company failure. A broad view of what really matters
is crucial. The risk our conceptual and operational models are missing relevant concerns and
our creative interpretation of results does not address all important relevant concerns is the
really difficult risk to manage and often the most important. Anyone who argues that it is
always easy to keep it simple using simple methods usually does not have even the vaguest
idea about what is involved. In my experience they often subscribe to the common initial
interpretation of KISS – keep it simple, stupid. The point of the revised definition – keep it
simple systematically – is the right kind of simplicity for any given context needs systematic
attention to avoid being simplistic, and the wrong kind of simplicity is a basis for both seri-
ous risks and important lost opportunities.
A somewhat different example of the right and wrong kinds of simplicity with immediate
and later direct relevance is worth brief consideration now. At a time when mathematical
programming in general and linear programming in particular was generating a great deal
of interest, Dasgupta and Dasgupta (1966) used crop yield and related market price data in
a very interesting manner. They demonstrated that a linear programming approach to max-
imising income for typical farmers in India would double their income, as also demonstrated earlier by Indian agricultural economists who were urging the government to spend more
money on ‘educating’ the farmers to change from current traditional crop patterns to the
agricultural economists recommended crop patterns based on linear programming mod-
els. The Dasgupta paper then demonstrated that the same data employed in a Markowitz
(1959) quadratic programming model, which used the same expected outcomes in addition
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
208 Employing planning tools in practice
to considering risk associated with income variability measured by variance, led to several
very different inferences. The Dasgupta quadratic programming model made it clear that
the farmers were already using an optimal solution which maximised expected income given
a minimum level of risk. The farmers were intuitively using a Markowitz mean– variance
portfolio management model. The farmers needed to minimise risk because given their
lack of financial assets, without any form of government insurance scheme to protect them,
they and their families could starve in a bad year. It was the agricultural experts who needed
‘educating’, not the farmers, because the agricultural experts were using crucial stealth
assumptions with myopic models. The government also needed educating, but not by the
agricultural ‘experts’ who believed linear programming was the way forward.
Over many generations the Indian farmers had developed a highly diversified crop pat-
tern which protected them as far as possible from drought, crop diseases, and market price
fluctuations, maximising expected income as a secondary objective within a minimum risk
primary objective, an approach common to most subsistence farmers. The trend towards
monoculture prevalent in Western economies was the result of a combination of wealth
accumulation by farmers plus government price intervention and support when disease or
other reasons for crop failure would otherwise devastate the industry. Governments in both
developed and developing economies needed to understand the insurance role they could
play. If they did so there were very beneficial behavioural implications, provided the farmers knew they had an explicit insurance contract with their governments.
One general point which follows from the Dasgupta paper, as well as this chapter’s tale,
is that following simplistic models when they fail to capture key issues can do more harm
than good. If Nicola had not progressed beyond her pass 2 analysis, it might be reasonable
to judge her in the same light as the Indian agricultural experts limited by a linear program-
ming perspective. Later chapters will explore further examples linked in a fairly direct way
to the Dasgupta paper.
A second general point is the importance of not underrating the intuitive expertise of
someone like Pete or the Indian farmers, and the importance of being very careful before you suspect bias or enforce changes their intuition is suspicious about, testing the assump-
tions responsible for differences in views.
A third general point is that in top- down strategic corporate management terms, cost varia-
tions relative to unbiased expected outcomes were just noise, not risk worth the analysis effort,
but seriously biased expected values and biased decision making more generally were risks.
Furthermore, while the set of propositions considered could lead to some very useful TLC
corporate risk reduction via diversification, a more general approach to the key risks which
could put TLC out of business was needed, drawing on ideas developed in later chapters.
The most difficult risk to manage is failing to understand what really matters – what a
management consultant friend of mine has called ‘stupidity risk’ in its most serious forms,
although this is clearly not a label any sensible consultant would use with his or her clients
unless the consultant was a very good judge of the client’s sense of humour. ‘Strategic myo-
pia’ might be a better label for discussions within organisations, whether or not consultants
are involved, comparable to the use of strategic misrepresentation as a euphemism for lying,
but my inclination is using the phrase ‘lack of strategic clarity’. Whatever you chose to call
your equivalent of strategic myopia or lack of strategic clarity, the EP approach attempts to
manage it in various ways. The central unifying theme is trying to persuade all relevant deci-
sion makers to adopt a systematic approach to using identified and tested working assump-
tions within framing assumptions which are as general (unrestricted) as possible to achieve
strategic clarity plus requisite linked tactical clarity in a clarity efficient manner.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Using a UP 209
The key role of the capability- culture concept and revisiting the tale’s credibility
TLC seems to have had no basis for getting beyond a very simple optimal order size and
safety stock approach, intuitively sized by those directly involved, until Nicola arrived. In
stealth assumption terms, George and his colleagues seem to have used an implicit approach
to planning which accommodated very limited capability in terms of testing inventory policy
and broader supply chain assumptions. Nicola brought the capability needed to explore the
sizing of order quantities and safety stocks. But much more important, she also prompted
the development of a corporate curiosity to test assumptions and explore more sophisticated
policies with implications for purchasing and marketing strategy, pushing back the ambigu-
ity associated with many TLC working assumptions via the use of explicit UP variants, trig-
gering an ambition to evolve the capability to do this more effectively.
This may seem implausibly flattering to Nicola and unreasonably unfair to all other TLC
employees. But the tale is trying to demonstrate what adding the capabilities associated
with a basic understanding of the UP concept and generic models can do, even if they are
brought into an organisation by someone like Nicola with very limited experience.
What is crucial to add is with the apparent exception of prior skills with the formal process
aspects of the UP concept and related generic models, TLC was a very capable organisation
with a very sound corporate culture. If George, Bob, Surinder, Pete and Ajit had not played
their roles as they did, as very effective members of an unusually capable team, no significant
progress would have been made.
More generally, Nicola’s management course material was important to the progress
made by TLC – but so was her intelligence and skill. Furthermore, even more important was
the inherent planning skills of all this tale’s characters. In particular, their collective ability to
synthesise what they learned from analysis using models with the complexities of the ‘real
world’ was central to the inspiration behind the My- Place- Your- Timing strategy variants
developed by Bob and George and then built on by all the others.
In terms of the Figure 2.1 portrayal of the UP concept, TLC had the right stuff in ele-
ment 11 terms plus elements 1 through 10, with the exception of UP/SP process skills and
associated generic models plus the EP tools to be developed in the rest of Part 2.
As an observation on the credibility of the tale, a crucial working assumption when draft-
ing this chapter was that Nicola was bright enough and lucky enough to reach her plan 3
and not insightful enough to get beyond her plan 3, but Bob was insightful enough to trans-
form her plan 3 into his plan 4. This implies the person in the organisation in the supply side
operations director role has to be insightful enough to see the benefits of what was clearly a
better approach, or a more experienced variant of Nicola is needed to help them. Further-
more, each of those in all the other operations management roles discussed have to have the
insight to play their roles – otherwise, the organisation is clearly going to be opportunity
inefficient. Finally, George in his CEO role, and Ajit in his supporting leadership role, clearly
need to acquire further capability for TLC before proceeding with the propositions gener-
ated, perhaps in the form of two further directors as well as further training and support for
all existing TLC staff. Your credibility may have been stretched by assuming all these people
were enlightened enough to quickly pick up and employ Nicola’s UP concepts after her
arrival, without indirectly having come to similar conclusions informally on their own, but
this stretch was in a good cause. What is particularly stretching is the importance of Nicola
as a catalyst, especially given other TLC staff were so adept at communicating and working
as a team as well as being creative on their own.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
210 Employing planning tools in practice
Despite their assumed lack of earlier exposure to generic models, UPs and SPs, in practice
we could expect any organisation with this level of capability and cultural maturity to find
some of the opportunities discussed in this tale on an ad hoc basis. Many small organisations
like TLC make negligible use of UP concepts or generic models, but few have the capabili-
ties attributed to TLC, and any that did would probably generate some of the improvements
suggested by Bob, or the propositions suggested by the others, without the equivalent of a
contribution from a Nicola.
Future relevance of this chapter’s tale
My most recent involvement in the practice of managing a manufacturing business was in the
early 2010s, providing in- house training on managing uncertainty in change management
projects for Volvo staff in Sweden via Chalmers University of Technology in a series of
courses over several years. The change management issues they had to deal with included
Volvo trucks manufacturing MAC trucks (a US brand) and Renault trucks (a French brand)
in addition to its own Volvo brands, Nova buses, and Terex construction equipment, while
Volvo cars initially became part of Ford and then part of Geely. Geely, originally in the refrig-
erator business in India, also acquired the British manufacturer of London Taxi Cabs, and
in late 2017 Geely bought an 8.3% stake in the Volvo Group, ‘putting them in the driving
seat’ (Lea, 2017a) of a reunited version of the 100 year- old Swedish Volvo company with
90,000 employees. The Volvo Group’s annual revenues as of 2017 were £27 billion, and
Geely is China’s most international automotive group, with a Hong Kong stock market
value of £22 billion. Volvo had earlier bought a large minority shareholding in Dongfeng,
China’s largest lorry maker and a rival to Geely in the car market.
Simply being aware of the world around us makes it very clear that globalisation in the
manufacturing industry has changed everything beyond recognition relative to the 1980s
setting of this chapter’s tale. The full implications of many other issues are still emerging,
like the ongoing development of artificial intelligence. There is no reason not to expect new
currently unpredictable fundamental changes. However, while current and future levels of
complexity may be orders of magnitude greater than that faced by TLC, most of the basic
concerns TCL faced as addressed in this chapter have not gone away. They remain a useful
starting point for Part 2 of this book in terms of the basic broad questions addressed.
The pace of change is unlikely to slow for any organisation in the manufacturing business,
and the same is true of most other businesses in most countries. Some widely acclaimed
authors saw this coming some time ago. For example, Alvin Toffler (1970) in Future Shock famously said, ‘The illiterate of the 21st century will not be those who cannot read and
write, but those who cannot learn, unlearn and relearn’. This first book in a series of three
by Toffler sold 17 million copies and became mandatory White House reading. Current
eagerness to learn about what may be involved is illustrated by the popularity of books like
Beyond Competitive Advantage (Zenger, 2016), a useful overview of what sustains corporate success in a rapidly changing world drawing upon a range of illustrative examples.
An interesting contemporary book looking at a ‘new’ basis for sustainability which accords
with the ideas cherished by George and Ajit is The Conscience Economy: How a Mass Move- ment for Good is Great for Business (Overman, 2014). The basic ideas explored are not new in the sense that the Cadbury’s approach to looking after its employees which George and
his father admired has a very long history, which Steven Overman discusses. However, some
aspects are new in the sense that Overman argues that the current new generation of young
adults (millennials) have a basic inbuilt wish to do ‘good’ as a way of doing ‘well’ which he
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Using a UP 211
believes was not observable in the preceding generation, and they want to do so through
their businesses as entrepreneurs, their work as employees, and their purchases as consum-
ers. Overman argues that this is being driven by post- millennium technological, economic,
social, and political forces which imply it is not a passing phase, and it is actually a threat to
organisations which fail to see the opportunities associated with a significant mass move-
ment. The threatened organisations may be in the private sector with any form of ownership
or in the public sector.
This contention is controversial and what will actually happen is highly uncertain, but
what is certain is this issue matters, and it is now receiving significant attention. I did not
read Overman’s book until August 2017, but within a few weeks I observed a lead article
in a feature on franchising with the title ‘When Doing Good Is Great for Business: Socially
Responsible Franchising Is a Growing Trend Which Can Benefit Business and the Commu-
nity’ (Coleman, 2017), followed later by another titled ‘Tech Kings Must Learn from Our
Great Entrepreneurs’ (Byrne, 2017), attacking organisations like Google and explaining the
scope as well as extolling the efforts of George Cadbury, William Lever, and John Lewis.
Coleman’s article warns there are difficulties which need care, comparable to Ajit’s warning
about getting a model like the John Lewis Partnership approach to work effectively. Fran-
chising is a controversial form of business operation for many people who are interested in
the kinds of operations which ‘doing good’ implies, but franchising operations with closely
coupled community and charitable goals which are consistently operated within good, sus-
tained values are a development which is clearly a plausible and potentially optimistic pos-
sibility given the many other aspects of our post- 2020 future.
The relevant and interesting short article ‘Bosses Discover it Pays Dividends to Put Work-
ers First’ in Luke Johnson’s regular Enterprise column in the Times (Johnson, 2018) argues there is a strong case for medium- sized family businesses using this strategy, based on the
recommendations of highly successful entrepreneurs like Julian Richer and his book The Ethical Capitalist. He suggests the complexities involved in large public companies, or those owned by institutional private equity, will raise additional significant challenges, but
he points to General Electric as a current example of why they need to move forward too.
Whatever the ultimate directions of the key aspects of change, the current pace of changes
implies a growing need to understand how to make effective use of the most sophisticated
UP concept equivalents available, avoiding the limitations of simplistic interpretations of
models like Nicola’s pass 1 and 2, and achieving the much more useful results of Nicola’s
plan 3 and Bob’s plan 4, plus the follow- on TLC propositions. Furthermore, the role of
capability- culture concepts in the UP concept matter. The current pace of changes and
uncertainty about the direction of key aspects of these changes imply a growing need to
generate propositions suited to a frequently disrupted future and then deal with them col-
lectively in an opportunity efficient manner in all organisations. How to deal with them in
an opportunity efficient manner for a very unpredictable future is directly relevant to all the
tales to follow in Part 2, the basis of strategic clarity.
The specific example context used in this chapter is clearly dated, but not the messages,
and most of the messages are relevant to all organisations. The same comment applies to all
other Part 2 chapters, for different but related reasons.
The very simple context used in this chapter was a convenient starting point for Part 2,
but much more complexity now needs addressing.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
The experience base which underlies this chapter’s tale about ‘Astro’ includes a case study
initially developed in the 1990s with an IBM UK senior executive for a culture change pro-
gramme, its use with groups of IBM staff about 40 times, plus follow- on consulting which
included helping IBM to develop and use related bidding processes. You may find it helpful
to see Astro as an organisation comparable to IBM.
The scope of this chapter is much broader than bidding might suggest for several reasons
worth understanding before beginning the tale which is central to this chapter.
First, bidding is a particular approach to selling which is an aspect of the much broader
field of marketing which, in turn, is a central aspect of operations management with cor-
porate strategic management roles. In this chapter, bidding and associated marketing are
assumed to include underlying aspects of managing internal market structures plus strategic
relationships with subcontractors, with important implications for approaches to both the
outsourcing of currently internal corporate operations and the ‘friendly takeover’ purchases
of valued subcontractors who might otherwise be bought by competitors. The implica-
tions are relevant to many aspects of demand chain management, supply chain manage-
ment, interdependent product design, and broader operations and corporate management
concerns.
Second, much of the tale of this chapter is also relevant to organisations concerned with
managing purchases because to manage their side of what can be a complex relationship,
‘customers’ often need to understand their ‘suppliers’ concerns and perspective.
Third, some aspects of the approaches discussed in this chapter are of interest in all
decision- making areas. As one example, this chapter’s tale illustrates one key reason why
organisations may spend 80% to 90% of their effort using formal planning approaches to
clarifying aspects of a situation which they understand fairly well, only 10% to 20% of their
effort using formal approaches to providing more clarity for aspects they do not under-
stand as well as they need to, when reversing these effort allocations would be far more
opportunity efficient. This switch in effort allocation may be difficult, but it may be crucial,
and it may require new conceptual and operational tools to achieve a new and significantly
enhanced level of clarity efficiency.
Fourth, even if you are not currently interested in any of these particular areas or aspects
of decision making, the fundamental reason you need to understand the basic ideas explored
in this chapter is because several of the conceptual and operational tools discussed will be
used as basic building blocks in later chapters. A key example is the ‘specific planning uncer- tainty and complexity management process’ concept, contracted to ‘specific process’ or ‘SP’.
An SP is a ‘process model’ – a model which is itself a process involving a designed sequence
for using one or more lower- order models in a specific corporate decision- making context.
6 Building ‘specific processes’ – a bidding process example
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 213
The relationship between a UP and the generic models of Chapter 5 is extended – the UP
is used to create, enhance, shape, test and implement an SP.
As in Chapter 5, some aspects of this chapter’s tale may stretch your credibility, but this
stretch is in a good cause, so please suspend your concerns about credibility until closing
sections of this chapter with this ‘good cause’ in mind. The set of credibility issues involved
are more usefully discussed later, in conjunction with the relevance of a 1990s tale to post-
2020 contexts and the role of other real cases underlying the tale. The credibility issues are
not the same as those of Chapter 5. In this case they maximise the scope to communicate
what effective use of a UP to create SPs can achieve. The tales of following chapters also all
delay the consideration of plausibility issues until the end, but the issues will be different in
each case. By way of a reminder linked to these credibility issues, the language of EP will
be used throughout, as in all Part 2 tales, translating from the context language usage you
might expect within the tales.
Capturing some of the context information to get started
The tale of this chapter is set in the 1990s in the Astro Wales regional head office of Astro
UK. Astro UK was part of Astro Inc, an international shareholder owned firm in the com-
puter information systems business based in the US. The key characters were
Martha, marketing manager for Astro Wales, recently appointed;
Ben, bidding process facilitator reporting to Martha;
Trevor, transport-sector lead sales representative for Astro Wales;
Sian, Systems manager for Astro Wales;
Sam, a Systems group member reporting to Sian;
Rhys, regional manager for Astro Wales; and
Steve, Service Bureau manager reporting directly to Rhys.
As with the last chapter’s tale, it is not a problem if you identify with some of these characters
more than the others, but seeing all of them as a team who worked together exceptionally
well is important. The nature of the teams’ working relationships was different from those in
TLC, to be expected when moving from a small family owned business which was manufac-
turing driven to an international shareholder owned organisation which was market driven.
Martha was a central character in this tale. Martha studied philosophy at university because
ideas about different ways of thinking particularly interested her at that stage in her educa-
tion, and she was uncertain about what kind of ideas interested her most. Martha’s first job
was in computer sales for Astro Scotland, part of Astro UK. She made very rapid progress,
in part, because a wide range of ideas associated with having a successful career seemed very
interesting to Martha. She was still in her 20s when she was promoted to marketing man-
ager (Head of Sales) for Astro Wales.
Ben was also a central character in this tale. Immediately after completing a BSc degree in
economics and management science Ben started work as a Sales trainee at the Cardiff office
of Astro Computer Systems, head office for Astro Wales, reporting to Martha.
Trevor was the third central character in this tale. He was the transportation industry lead
sales representative Martha assigned to the bid which forms the basis of this chapter.
Sian was a further key character in this tale. Sian was Systems manager (Head of Systems)
for Astro Wales – Martha’s opposite number in the Sales– Systems silo structure used within
Astro Wales. Sam, a Systems group member reporting to Sian, had a minor role.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
214 Employing planning tools in practice
Rhys was the regional manager for Astro Wales who Martha and Sian both reported to.
Rhys was also responsible for the Astro Cardiff Service Bureau managed by Steve, operat-
ing outside the direct remits of Martha and Sian in a separate operations management silo.
Martha was in ‘respond- to- crisis’ mode the morning Ben started working for Astro.
Astro Wales had just lost the third of three unsuccessful bids that month for major ‘systems
integration’ projects. Systems integration was Astro jargon for a complete computer- based
information system, including hardware, software, physical facilities, and training, normally
incorporating the functionality of ‘heritage systems’ (existing hardware and software) in
addition to providing new functionality, usually tendered for on a fixed- price basis with
performance penalties. Martha knew that the failure of three major systems integration bids
in a row was not just bad luck. There were other clear symptoms of underlying problems.
For example, sales representatives were complaining to Martha that the ‘Systems people’
were taking far too long to cost things, and their solutions to problems were often far too con-
servative (safe and expensive), resulting in lost sales that should have been won. In addition,
the Systems personnel were complaining to Martha that the ‘Sales people’ were putting unrea-
sonable pressure on them to provide quick responses to difficult questions and to cut corners,
that most of their efforts were wasted because of lost bids, and that the bids that were won
were usually the ‘wrong ones’. Both groups were implying that the problems were serious,
and it was up to Martha to resolve them. Neither group seemed to understand why their own
approaches might be part of the reasons Astro Wales was so unsuccessful. The ‘panic level’ was
rising. It was clearly up to Martha to clarify the situation and resolve the underlying problems.
Martha knew that Astro Inc was experiencing similar difficulties worldwide. She also knew
that a common perception within and outside Astro was that ‘Astro Inc had lost its way and
was in danger of losing the plot altogether’. Martha’s view was that ‘the technology and the
markets had changed, the competition had become a lot tougher, and Astro Inc had not yet
adjusted in terms of organisational structures, products, processes and culture’.
Since her promotion Martha had focused on the ‘strategic’ or ‘macro management’
aspects of her new job, which were unfamiliar. She had deliberately avoided getting involved
in the details of individual bids or bid processes. However, Martha was aware that she did
not have sufficient clarity about Astro’s many problems, and more specifically, she knew
she needed to get a grip on the details of Astro Wales difficulties with bidding for systems
integration projects.
The present systems integration sales crisis was viewed by Martha as a catalyst for sorting
out bidding processes more generally in Astro Wales and starting to address the difficult
organisational, process, and cultural issues within her remit. She subscribed to the notion
‘never waste a good crisis’. If she could help to sort out Astro Wales, starting with systems
integration bids, she might get the opportunity to help sort out Astro UK on her next step
up the Astro corporate ladder.
The nature of Ben’s degree was understood by Martha, and she had checked that he had the
background to understand bidding models and processes, the insight to critically appraise and
learn from the relevant literature, and the interpersonal skills to make a competent facilitator.
Martha asked Ben to act as coordinator and facilitator for a new bidding process, an ‘SP for
bidding’. Martha and Ben would design this process drawing on other Astro expertise.
Martha had taken numerous advanced- level professional development short courses
which gave her exposure to both general and specific process concepts, using variants of
the EP, UP, and SP concepts discussed here. Martha’s academic training, especially the way
she saw philosophy as a kind of ‘glue’ among behavioural issues, politics, economics, and
everything else that mattered, underlay a very structured approach to business issues, with a
strong concern for the internal consistency of sets of assumptions, and clear logic.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 215
An invitation to tender had been received from Transcon the day before. Transcon was a
containerised freight road transport company recently formed from the merger of two suc-
cessful firms, one based in Wales and one based in Hampshire, with a view to attracting an
increased share of the growing Euro container freight market. Martha had decided to use a
bid for Transcon as the basis for developing a new prototype SP for bidding for Astro Wales
which they could then continue to develop for use with all similar future bids.
Astro had a tradition of ‘throwing new staff into the deep end of the swimming pool’,
with a view to stretching them and rewarding them if they learned to swim the requisite
strokes quickly. Martha made this clear to Ben, and she made it clear that she expected
him to take the lead with the details of the bidding model and some of the related pro-
cess design issues. However, she reassured him that she would provide leadership on
the overall approach framework, all his ideas could be tried out on her first, she would
contribute as required to their development, and her authority would be behind him
when they had an agreed- on approach. Martha let it be known that she was marketing
manager for Wales while she was still in her 20s because she had been given similar tests
in Astro Scotland. Martha was now on a fast track, and if Ben succeeded, he could expect
similar recognition.
It had been obvious to Martha that Trevor was the Astro sales representative to help with
the development of the bidding process. This was not just because Trevor was the transport
sector lead sales representative when the Transcon opportunity arose. He had particularly
enlightened intuition based on extensive experience which Martha wanted to capture for
all sales staff to learn from. In addition, Trevor was respected by all the other Sales staff –
Martha fully appreciated that having Trevor help her to change local processes would make
change easier. To a significant extent, Martha had seized the opportunity to use Transcon as
the basis for developing an SP for bidding because of Trevor.
Martha introduced Ben to Trevor. Martha indicated that she wanted Trevor to help
her and Ben develop a new systems integration bidding process for Astro Wales using
the Astro bid for Transcon as a prototype development project. She emphasised that she
believed Trevor’s input could help to shape the process as a whole in a crucial manner.
Furthermore, she wanted Trevor to take on some key leadership roles when the roll- out
of the prototype became a follow- on project. She wanted Trevor to use Ben as a bid coor-
dinator and facilitator for the Transcon bid. Trevor knew Martha well enough to know
that responding as constructively as possible to a nuanced interpretation of her requests
was the only sensible option.
Martha also introduced Ben to Sian. Sian was responsible for the Systems people who
would cost the Transcon systems integration project and the Systems people who would
implement some of it if the bid was won, other aspects being outsourced to subcontractors.
Martha made Ben’s and Trevor’s role clear to Sian. She then emphasised the importance of
Sian’s input to shaping the process as a whole and making it work on an ongoing basis. In
some ways Sian knew Martha even better than Trevor. Sian and Martha were technically on
the same management level, both reporting to Rhys, regional manager for Astro Wales, but
Martha had a significant edge because Astro was a market- driven organisation. Sian under-
stood that Martha was inviting her to join in an Astro Wales Sales– led Sales/Systems process
development project as a partner, and she was very happy to do so.
Martha indicated that she wanted Trevor, Sian, and Ben to prepare a document for her
that summarised what this bid was about, in the sense that everyone involved could agree
that the document they had prepared succinctly captured the current best practice starting
point for formal analysis of the Transcon bid. In UP terminology this was a ‘capture the
context’ phase document.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
216 Employing planning tools in practice
Ben was provided with a desk and Martha left him to it. It wasn’t even time for lunch on
his first day with Astro Wales.
To help you follow the way Martha and her colleagues developed a bidding process for
Astro as outlined in this chapter, Figure 6.1 is provided now.
Although it was not formalised until later, Martha did anticipate some of the key charac-
teristics of Figure 6.1 from the outset, and seeing it now in its final form will help you to
better understand how the SP for bidding takes shape. The overall rationale for the structure
portrayed by Figure 6.1, plus some immediately relevant aspects of its details, will now be
explained in terms of the story of their development. Martha’s initial vision of the Figure 6.1
overview plan at this stage in its evolution is referred to as the ‘initial process overview’.
A preliminary proposal document
Trevor and Ben had a brief initial discussion right after Martha’s briefing, and Trevor sug-
gested they start by visiting Transcon together. Trevor would do most of the talking, but
Ben could take notes. Extended and reshaped by Trevor, these notes would form the basis
of the documentation which ‘Sales’ passed to ‘Systems’ to facilitate a technical specification
Stage 1: Eliminate non-starters
Capture basic context information and key associated assumptions. Eliminate all contracts that are not likely to be worthwhile for reasons other than the margin.
Stage 2: Minimise no-hopers
Preliminary quantification of expected margin and the probability of winning. Eliminate most contracts that are unlikely to provide much margin.
Stage 3: Bid strategy and price
Analysis of opportunities which look worthwhile, developing a detailed bidding strategy in terms of both price and non-price advantages and disadvantages.
Cumulative Proposal
Document (CPD)
Proposal Document 1
(PD1)
Basic context information and key
associated assumptions
Proposal Document 2
(PD2)
Additions to the PD1
Proposal Document 3
(PD3)
Additions to the PD2
very early
no bid
early
no bid
late
no bid
bid
Figure 6.1 The overview plan of Astro’s SP for bidding portrayed as a flow chart.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 217
and costing. Trevor was very happy to start this way. Ben would be making himself useful
as Trevor’s assistant, and Trevor could control the shape of the document Martha wanted.
Ben was also pleased to start this way. By lunchtime Trevor had arranged an appointment
that afternoon with key Transcon staff.
Following the meeting with Transcon, Ben edited his notes and passed them to Trevor.
Trevor added detail, shaped the terminology as well as the structure, and appended the
‘invitation to tender’, which he always referred to as an ‘ITT’.
Trevor loved acronyms and similar abbreviations. Although most Astro staff winced every
time a new one was invented, they usually adopted it, for convenience as well as to humour
Trevor, whom everybody liked. Trevor referred to Ben’s reshaped document based on the
meeting and other information as the sales draft of the ‘PPD’, a contraction of ‘Preliminary
Proposal Document’.
This draft PPD began with some general background information. For example, Transcon
had been formed eight months earlier by the merger of Eurofleet Containers and Continental
Haulage. The new company would be operating from 15 depots, all in the UK. The head
office was in Cardiff. The managing director of Transcon had commissioned an assessment of
Transcon’s future information system’s needs by a leading firm of consultants. The consultants
had reported that about half the systems being used in the former companies would remain
useful, but they should be converted for a single new- generation mainframe computer. The
rest of the systems should be scrapped and replaced with proprietary software packages, and
a major new distributed scheduling system would be required, with capability distributed to
the 15 depots. The consultants went on to recommend that Transcon should go out to tender
for a new computer system capable of taking over the existing work and providing additional
capacity to cope with the new scheduling system. Their offer to develop the invitation to ten-
der was duly accepted, executed, and sent to Martha at Astro Wales amongst others.
The PPD provided a brief summary of the technical details in terms of Trevor’s views
on what should be provided. This involved a simple five item breakdown of the work to
be included in the bid. The PPD also provided a brief summary of the marketing issues in
terms of information gleaned by Trevor, influenced to a very modest extent by Ben’s input.
Following their meeting with Transcon, Ben had pushed Trevor to give him what he referred
to as ‘a preliminary maximum bid estimate’ and Astro’s chance of winning at this bid price level
to size the uncertainty about winning at the upper end of the bid range. He was anticipating
what he believed would be needed for the most basic bidding model he could envisage and
thought this was a good time to elicit this information. Trevor’s preliminary maximum bid
estimate was £20 million, with a 10% chance of winning. Trevor reminded Ben that a senior
manager in Transcon had let slip (probably deliberately) that Transcon’s budget for this contact
was a maximum of £20 million. He also observed that no information on competitive bids was
available as yet. Ben had also pushed Trevor to give him what he called ‘a preliminary minimum
bid estimate’ and Astro’s chance of winning at this level to size the lower end of the bid range.
Trevor’s preliminary minimum bid was £15 million, with an 80% chance of winning. The pre-
liminary maximum and minimum bid information was noted in Ben’s first draft of the PPD.
As soon as Trevor had a draft PPD that he was happy with, he asked Ben to discuss it
with Sian and indicate that it was clearly a preliminary document which did not require any
commitments on Sian’s part, but it needed completing by Systems at the level of detail set
out in Trevor’s input. Sian agreed that Sam, a member of her Systems staff, would do this.
Table 6.1 summarises core aspects of Ben’s and Trevor’s input in Trevor’s five item break-
down structure, plus Sam’s added elaborations, leaving out details which you do not need
to see to get the flavour of what was involved.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Table 6.1 Preliminary Proposal Document (PPD)
Item 1, Central Site Equipment One new mainframe computer was required to replace two existing mainframe computers.
It would be located in Transcon’s Cardiff head office. It must run the converted and new software. The proposed equipment should have spare memory capacity of at least 50% in terms of the converted and new software as installed, to allow for post- installation growth. Contractual guarantees on this were required. An Astro mainframe computer and peripherals were specified and the direct cost to Astro Wales for this equipment (from Astro Inc) was £3.6m. The usual selling price for this equipment was £5.0m.
Sam noted a high degree of confidence in the utilisation estimates for the proposed new equipment, based on extensive experience of similar bids. However, there was a source of risk for this component linked to item 4 implying a possible need for additional memory.
Item 2, Computer Suite Modifications The new mainframe computer would be installed in a room adjacent to the existing mainframe
computer room at Transcon’s head office, to avoid disturbing existing operations. The existing mainframe computer rooms would be emptied and released for alternative use as soon as parallel operation was no longer required. Astro would design the new facility and provide a project manager to supervise the work. A subcontractor, Zenith Controls, could undertake the installation work for a fixed price.
Sam noted that an Astro direct cost estimate of £0.30m had been prepared based on the information provided by Trevor plus further enquiries. Zenith had been approached based on Trevor’s information and further enquiries, and they provided a preliminary quote of £1.00m. Sam had a high degree of confidence in the Astro component estimate and suggested treating it as a fixed price. However, actual prices contracted with Zenith were usually between 10% and 30% higher than preliminary quotations.
Item 3, Initial Operation and Training Operators would be provided to run the new systems in parallel with the old systems initially.
Transcon’s staff would then be trained to take over, and supported during the takeover process. The transition should be complete within 6 months from the installation of the new system. Trevor indicated two options were available.
(a) Zoro Computing could undertake this task on the Transcon site for a fixed fee. (b) An existing Astro computer site, an Astro Service Bureau close to Transcon’s head office,
could be used for initial operation, with Service Bureau staff training the customer’s operators on Astro premises.
Sam noted that Zoro had been approached based on Trevor’s information and further enquires and a £1m provisional bid estimate obtained. However, experience with Zoro on previous contracts suggested that fixed prices actually contracted were between 20% and 40% higher than preliminary quotations. Zoro had a very good track record of success on Astro and other contracts. However, they were currently the subject of a hostile takeover bid.
Sam noted that Astro had two contracts in progress similar to option (b), both going reasonably to plan, but no other similar contract experience. In this case the Astro estimate of direct cost was £1.14m, based on estimated hours of effort by Astro Service Bureau staff at standard rates with backup from Astro’s training staff. Both existing contracts were close to completion. One was currently 10% below budget, the other 40% over budget.
Sam recommended option (b).
Item 4, Convert Existing Programmes A subset of Transcon’s existing financial programmes would be converted and associated
documentation would be updated. Run- time savings of 25% would be required. The ‘invitation to tender’ (ITT) stated that current documentation was ‘believed to be reasonable – 95% accurate’. Trevor indicated two options here.
(a) Datapol Systems could use new generation languages to rewrite all the software. (b) Sysdoc Autocode could use an automatic converter to translate about 90% of the code,
manual patching to deal with the balance of about 10% which proved non- viable for automatic translation.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 219
Sian and Sam were both pleased with the form and content of the PPD. Sian had ensured
that the quality of the drafting matched Trevor’s, and she was confident that the level of
the detail provided would be seen as useful. When Ben received the PPD with the Systems’
contribution incorporated and edited, he was very impressed. He showed it to Trevor, who
interpreted it as the best practice he expected of his Sales team plus the best practice he
expected of Sian’s Systems’ staff. Ben then showed it to Martha, who said she would start to
think about it that evening and discuss their next steps with Ben first thing the next morn-
ing. In the meantime, she would like him to start a systematic review of relevant bidding
Sam noted that Datapol were a large well- established company. Using Trevor’s information and further enquiries Datapol was willing to commit now to the required run- time savings and a ‘firm fixed price’ of £1.2m.
Sam noted that Sysdoc was a relatively new company, with only 100 employees, but a good record of success on smaller projects. Sysdoc would not commit to a fixed price but would commit to a measured day- rate contract. They had provided an initial estimate of £0.5m based on a fixed day- rate. Cost variability was associated with the proportion of the translation achieved automatically, programme documentation, house standards, and so on.
Sam also noted that either approach could give rise to a requirement for more memory than that assumed in relation to the central site equipment (item 1 above). Sam believed there was a 50% chance that additional memory would be needed to meet the capacity and performance guarantees. Sam estimated that if sufficient additional memory to meet this contingency was specified and installed at the outset, the additional cost would be £0.50m. If installed subsequently, Sam estimated the additional cost would be £1.0m.
Sam recommended the option (a) Datapol approach and pre- installing the additional memory.
Item 5, Distributed Scheduling System A distributed scheduling system would have to be developed, to operate in each of the 15
depots, linked to the mainframe. Sub- second response time was required. Trevor indicated two options in this case as well. Both involved 15 Astro workstations with a direct cost of £3.00m (usual selling price £4.20m) However, the development and installation of the system involved a choice between
(a) using Astro Systems staff and (b) a subcontract with Zoro.
Sam noted that information provided by Trevor plus further enquiries suggested the Astro approach would involve a day- rate cost of £1.10m plus a 20% contingency for a total of £1.32m. Astro had completed three successful projects very similar to option (a).
Sam noted a preliminary estimate from Zoro of £1.00m, and drew attention to the comments on Zoro associated with ‘initial operation and training’ (item 3 above), in terms of high- quality deliverables but a potential hostile takeover, and 20% to 40% uplifts from preliminary quotes to fixed prices.
Sam also noted that either approach might give rise to a need for more powerful distributed computing equipment. Sam believed there was a 50% chance that additional performance would be required to meet response time performance guarantees. Sam estimated that if higher performance equipment to meet this contingency was specified and installed at the outset, the additional cost would be £1.00m. If installed subsequently, Sam estimated that the additional cost would be between £1.5m and £2.5m.
Sam recommended the Astro Systems staff option (a) and pre- installing the additional performance.
Preliminary maximum and minimum bids estimates Preliminary maximum bid estimate £20m, with a chance of winning of 10%. Preliminary minimum bid estimate £15m, with a chance of winning of 80%.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
220 Employing planning tools in practice
literature. He should be aware from the outset that she had a multiple stage process in
mind, which she explained in terms of the ‘initial process overview’ which would evolve to
become Figure 6.1.
The initial process overview – early development
Martha had at least two phases in mind for her process overview plan from the outset, with
front- end filtering to ensure that the time wasted on bids which Astro did not want or could
not win on a profitable basis was minimised. She was not surprised by the PPD information,
but reflecting on it confirmed her preference for three stages, for reasons explained shortly.
When discussing her evolving thinking with Ben the next day, she asked him to test her
three stage choice in his review of the literature, as well as a list of other concerns. This
was an initial aspect of drawing Ben into the process of jointly evolving her initial process
overview, the basis of Figure 6.1. Ben would test her strategic clarity primarily by helping to
develop tactical clarity, ensuring that there were effective and efficient how to do it ways to
achieve what Martha currently thought needed to be done.
As part of a process of familiarising Ben with Astro practice and terminology, Martha
explained that current Astro UK policy was a ‘normal’ overhead recovery rate of 30% (for
indirect costs) on all Astro direct costs. Direct costs included all internal Astro transfer costs
plus subcontractor costs. Current Astro UK policy also specified a further 10% profit uplift
on total (direct and indirect) cost for all contracts, implying a cumulative ‘normal uplift’ of
43% which was usually referred to as the ‘margin’. Margin was the common usage Astro term for selling price less direct costs, a contribution to profit measure, profit being defined
by margin less overhead costs.
Martha indicated that she had observed some people interpreting the ‘normal’ uplift or
margin of 43% and its components as a corporate standard which should be achieved on every
potential contract whenever that was a convenient interpretation for their current purposes,
these same people being ambiguous about corporate norms whenever they did not suit their
purposes, while other people were consistently ambiguous. Under her watch, using their new
SP for bidding departures from ‘normal’ mark- ups would not just be routine; they would also
be expected, but so would a clearly stated and documented case for departures from the norm,
a formal tracking of margin achieved in relation to margin expected, and a formal tracking of
cumulative achieved margin relative to cumulative normal margin. Her light touch approach
to enlightened governance required everyone being clear about how all component aspects of
Sales performance contributed to corporate judgement of their overall performance.
As part of her coaching on process development planning for Ben, Martha explained
that at a later stage in the development of their bidding process a useful alternative to the
30% overhead uplift plus a further 10% profit uplift (43% cumulative uplift for both) might
be a ‘nominal 60% expected margin’ uplift. This 60% might include the current 43% plus
a 17% provision for errors and omissions in the traditional sense plus further a no quibbles
provision, in part, to keep it simple and, in part, to remind everyone that errors and omis-
sions plus a no quibbles provision would usually be involved. However, she did not want to combine these separate reasons for the uplifts embedded in margin estimates until later,
primarily because she did not want to distract attention from the more basic changes she
wanted addressed first. But she also suspected that 50% or 70% might be more appropri-
ate than 60%, and she thought it would be useful to acquire some data before suggesting
a starting point estimate which could evolve over time. She explained her reasoning to
Ben, emphasising that they would have to plan which changes to implement in what order,
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 221
informally and intuitively most of the time, but explicitly some of the time. She was just
using a simple example to illustrate what was involved in early planning for the evolution of
her initial process overview.
Martha and Ben spent some time working through what they anticipated would be the
formal analysis structure within each of the stages in Martha’s initial process overview, build-
ing on the PPD starting point plus Ben’s literature review. They also discussed how they
should set about prototype development and testing. Martha ensured that Ben’s initial liter-
ature review was focused, and any interesting ideas were explored further and incorporated
in their emerging plans when relevant. On the basis of this initial process overview develop-
ment, Martha briefed Rhys on their progress, tested all her key ideas on him, and sought his
support for any controversial issues. Martha and Ben then planned a workshop with Trevor
and Sian which would take a full day and arranged a date.
Workshop presentation of the Astro SP for bidding overview
Martha opened the workshop by suggesting that bid preparation for systems integration
contracts ought to involve a multiple stage process. Based on the PPD prepared for Transcon
and her understanding of current Astro problems and processes, plus analysis of a literature
review by Ben, she was proposing Astro Wales move to a bidding process with three stages.
Stage 1 should be concerned with eliminating from further consideration all contracts
which were ‘non- starters’ in terms of issues other than expected margin as a basis for a very
early no-bid decision. This would involve identifying all the important positive and negative
issues over and above margin. Stage 1 would also ensure the residual ‘starters’ with impor-
tant positive and negative issues had these issues flagged for attention during the decision
making of later stages. Two related but different stage 1 concerns were implied.
One was making some no-bid decisions as early as possible before wasting time addressing
margin concerns, whenever this was sensible. For example, the potential customer might
involve significant credit risk or other potential payment difficulties, the potential customer
might involve significant reputation risk, or the potential project might mean Astro effort
investment in activities which did not involve synergies with ongoing Astro corporate goals.
The second was identifying all key non- margin issues which needed ongoing considera-
tion if the process continued to stage 2. These might include positive issues like knock- on
sales opportunities, a particularly good fit with new Astro sales expansion plans, and other
positive corporate synergies. They might also include negative issues which were not serious
enough to eliminate the possibility of a bid in stage 1, like an indifferent fit with sales expan-
sion plans which would be tolerable if the profitability was good.
Direct cost, selling price and the margin defined by the difference would not be stage 1
issues, but an extensive set of other concerns would be.
Stage 2 should have the goal of eliminating most of the ‘no- hopers’ – contracts which pass
the stage 1 hurdle but did not look promising enough to analyse any further because of the
likely potential margin given ongoing considerations identified in stage 1. The assessment
of likely margin would be preliminary and simple, with a balanced approach to likely cost if
the contract was won and the likelihood of winning it.
Stage 3 should be concerned with developing a bidding strategy and tactics including
deciding a bid price for the remaining candidates, with a flexibly controlled low level of late
or very late no-bid decisions.
She thought the outline information provided by Trevor in the PPD was perfect as the
basis for stage 2, aimed at identifying and minimising no- hopers, but it did not address the
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
222 Employing planning tools in practice
issues she had associated with stage 1. She thought the more detailed information provided
by Sian and Sam in the PPD was perfect as the starting point for stage 3, aimed at develop-
ing a bidding strategy and deciding a bid price, given a potential contract had passed the
stage 2 hurdle. However, she thought Sian’s input was too good for stage 2. Furthermore,
like Trevor’s input it did not address the issues she had associated with stage 1. Two con-
cerns needed addressing immediately.
First, there was a mismatch between the levels of effort provided by Trevor and Sian. She
appreciated that Trevor may have expected the level of detail provided by Sian from the out-
set in the past, and Sian may have felt a professional approach required it, but Astro could
not afford to invest this much effort in a bid proposal which might be revealed as a no- hoper
with relatively simple input from Sian’s staff.
Second, the PPD was missing a suitable basis for a stage 1 process, concerned with iden-
tifying and eliminating non- starters plus effective and efficient ongoing consideration of
relevant non- margin issues.
She might have raised further concerns – like Trevor’s input was good enough for stage 2
but only because he had provided answers to Ben’s questions about preliminary minimum
and maximum bids, and a much more sophisticated approach to match Sian and Sam’s
efforts was required for stage 3. However, she saw deferring these concerns until later and
then dealing with them indirectly as a more useful approach.
Martha suggested that Transcon might not warrant any significant attention in this stage
1 context, so it was perhaps entirely reasonable that neither Trevor nor Sian had noted any
relevant issues. Furthermore, she had not encouraged a direct conversation between Sian
and Trevor or other Sales and Systems staff about balancing their efforts in a multiple stage
process earlier, so it was not unreasonable this mismatch had arisen. However, everyone involved in Astro bidding would benefit from a simple non- starters test followed by a simple
no- hopers test, as well as a more orderly ongoing treatment of non- margin issues, and she should have thought about this sooner.
Sian and Trevor were both pleased by the way Martha put these observations, a situation
Martha had planned for to encourage acceptance of both criticisms and change. Trevor and
Sian both realised that either might have been held at least partially accountable and appreci-
ated Martha’s deliberately accepting the oversight as hers, setting a tone which minimised
recriminations and encouraged candid critiques of current practice.
Martha then suggested that the time Sales and Systems staff would save by filtering out
the non- starters and no- hopers in stages 1 and 2 should be used to provide more clearly
communicated depth of understanding in stage 3 than she had seen in past bids.
In summary, using EP language, Martha wanted a step change in opportunity effi-
ciency to be achieved by everyone using their time and effort in a new formal planning
framework which would be designed and then developed via prototype applications with
clarity efficiency in mind, achieving different levels and kinds of clarity in each successive
stage.
Martha indicated that the exercise to produce the current PPD as portrayed in summary
form by Table 6.1 was in no sense a waste of time – it was a crucial first step. What Trevor
and Sian had produced was of good quality, perfect in the right context. Also, it was impor-
tant to begin with a clarification of the best of current practice before considering change.
Martha regarded the PPD material of Table 6.1 as an approximation to the best of current
practice which they could continue to use when appropriately placed within a more clarity efficient overall process. However, they now had to address change, starting with the basis for a suitable stage 1 and a plan to develop that basis in stages 2 and 3.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 223
To preserve the spirit of Trevor’s PPD concept – but clarify what was different – she
would refer to the replacement for the PPD as the ‘Cumulative Proposal Document’ (CPD),
with a Proposal Document 1 (PD1) portion of the CPD generated by stage 1, a Proposal
Document 2 (PD2) portion of the CPD generated by stage 2, and a Proposal Document
3 (PD3) generated by stage 3. She then showed them an early version of Figure 6.1 using
this notation.
When she was sure everyone was comfortable with the basis of Figure 6.1, Martha sug-
gested they began consideration of stage 1. You should now see how the process portrayed
by the overview plan of Figure 6.1 was starting to emerge.
Initial observations on Martha’s rationale
Martha wanted to take all the relevant people with her through the ‘capture the context’ phase
and into the ‘create and enhance the plans’ phase of what we can interpret as Martha’s implicit
UP. She was using her own intuitive approximation to the UP concept, but we can interpret
what she was doing using the formal UP framework of Chapter 2. Creating the PPD had been
a part of her ‘capture the context’ phase which explicitly involved Trevor and Sian. But all her
Astro Wales involvement to date was also crucial, as was her wider Astro experience.
Martha had already concluded that the ‘select and focus the process’ phase of her UP
required the development of an SP for bidding. She had an outline of the SP for bidding
of Figure 6.1 in mind already, but she wanted her initial plan to evolve with all the relevant
people playing a part in shaping it, so they would ‘buy in’ and ‘own it jointly’, the position
eventually achieved with Figure 6.1 and the component processes for each stage.
In terms of her implicit UP, Martha was beginning with a top- down vision of what she
wanted the SP to achieve. In terms of immediately obvious objectives within her emerging
goals– plans relationships framework, stage 1 and stage 2 were designed to filter out the
customers Astro did not want and the projects Astro did not want. This would allow Astro
to use the time wasted by their current single stage process to much better effect in stage 3.
But better focus on different issues which needed a different kind of treatment at different
stages in the process was also part of the rationale, for greater clarity efficiency. New features
in her approach within each stage would address root causes of her concerns with the cur-
rent bidding process and underlying corporate problems.
The overall uncertainty which had to be addressed was visualised by Martha in terms of
seven partially separable component sources of uncertainty about:
1 the desirability of doing business with the potential customer,
2 synergy between the potential project and Astro corporate goals,
3 what package of deliverables to commit to,
4 the cost of meeting these commitments if the bid was successful,
5 how the bid price and contractual terms should be influenced by uncertainty about the
cost estimate and the desirability of winning the contract,
6 non- price advantages relative to different potential competitors which might be influ-
enced by the package of deliverables committed to and other aspects of the approach
taken to the bid, and
7 how best to approach the first six sources of uncertainty.
Martha had developed a structured appreciation of this complexity working in sales in
Astro Scotland, ultimately in a role like Trevor’s. She knew Trevor understood these issues
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
224 Employing planning tools in practice
intuitively, possibly better than she did, but probably without this clarity of structure. She
was less certain about Sian’s understanding.
It is fairly obvious that increasing a bid price increases the chance of making a profit if
the bid is accepted, but it also reduces the chance of the bid being accepted. Conversely,
decreasing a bid price increases the chance of winning work but also decreases the chance
of the work being profitable. What is much less obvious is how to develop a clarity efficient
approach to all seven of the components of uncertainty which Martha was aware of and
achieve an opportunity efficient level of clarity for each.
What Martha was doing with her initial overview plan structure was suggesting that they
focus stage 1 on sources of uncertainty 1 and 2 – the uncertainty about the desirability of
doing business with the potential customer, if a contract were won, in terms of customer
characteristics and project issues which were not contribution to profit related in a direct
manner. Part of the rationale for this focus was filtering out undesirable customers and pro-
jects for non- margin reasons. But also important was making sure that they did not overlook
any big opportunities in terms of non- margin issues, and they did not lose sight of ongoing
positive issues as well as ongoing negative issues. Indeed, early attention to upside non-
margin issues which might need ongoing attention was crucial before getting too immersed
in downside concerns or costing and pricing detail.
Her stage 2 was focused on potential profitability in terms of margin if they won the bid
and associated unconditional expected margin incorporating the probability of them winning
the bid. She wanted stage 2 to be an effective minimum clarity approach to estimating both
the probability of winning and the margin if they did win, without getting into the details
of sources 3 to 7 of her seven sources of uncertainty decomposition. In terms of the cost of
delivery if Astro got the job, to some extent this view of stage 2 is analogous to the minimum
clarity approach provided for William in Chapter 3, but the probability of winning issue in the
bidding context added some new and rather different forms of complexity.
Her stage 3 was focused on addressing sources 1 through 7 in more detail in a holistic
manner, building on stage 1 and 2 information and making effective use of the time saved
by eliminating wasted effort using stages 1 and 2.
Stage 1 – identifying non- starters plus ongoing concerns
When opening her discussion of stage 1, Martha reminded them that non- starters meant
potential contracts that Astro did not want to get involved with for reasons which were not
contribution to profit related in a simple or direct sense. One key role for stage 1 was pro-
viding a screening process designed to filter out non- starters because of negative concerns
which are fairly obvious if the usual rush to look at direct cost and margin is put aside for
the moment. But a second key role was documenting important positive synergies and any
other non- margin issues needing ongoing attention in later stages. Indeed, this second role
was arguably much more important than the filtering role because it was about effectiveness
while filtering was indirectly about effectiveness but focused on efficiency.
There were a number of quite different kinds of uncertainty and complexity associated
with capturing issues of ongoing concern and filtering out non- starters, but she wanted all
the relevant stage 1 uncertainty and complexity issues addressed in a very simple qualitative form. In principle this might just involve qualified ‘yes’ or ‘no’ answers to a series of basic
questions about key assumptions, but she believed they would find ‘traffic light’ models
useful for communicating answers to an initial overview question, subsequent more detailed
questions and a final summary overview question, developed in a manner which she would
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 225
illustrate shortly. A final summary traffic light rating explained briefly by a verbal ‘balanced
view’ would provide the opening executive summary for the PD1.
Visiting the customer should not be necessary most of the time, but exceptions could be
made if the lead salesperson involved thought this was appropriate.
The starting point for preparing the PD1 would be the basic context information provided
by the potential customer’s invitation to tender (ITT) information plus a covering note which
summarised relevant context facts and issues known to the person drafting the PD1 which
were not included in the ITT. This would be followed by an initial overall traffic light response
to the question, ‘Can we assume this customer and this project bring no special opportunities
or difficulties?’ Her traffic light interpretation suggestions for this first traffic light were:
red – unambiguous significant difficulties,
amber – possibly important difficulties,
green – a neutral situation with a clear balance of modest opportunities and difficulties,
yellow – possibly important opportunities, and
blue – unambiguous significant opportunities.
She had put the red light first, the blue light last because the sequence red– amber– green
was the common practice ‘traffic light’ starting point, but in practice she would like them to
use the reverse sequence and start by looking for the unambiguous significant opportunities.
The blue light was associated with unambiguous significant opportunities. This might be
interpreted in ‘blue light emergency service’ terms, giving crucial opportunities the priority
treatment accorded to ambulances, fire engines, and police cars, even at junctions involving
red lights, provided due care and attention was involved.
More than one traffic light colour could be used. For example, the joint use of red and
blue lights would indicate significant difficulties plus significant opportunities which need much more detailed attention if on balance proceeding to stage 2 and then perhaps stage 3
was appropriate. Astro staff would need to discuss and jointly agree on interpretations like
this, perhaps modifying them as experience was gained. The role of the traffic light model
she was suggesting was simply a way of providing clear and concise qualitative communica-
tion in an agreed- on ‘traffic light’ format at a summary level.
An initial overall traffic light assessment resulting in a blue light plus an amber light
might suggest a complex set of issues warranting decomposition immediately or in a later
stage. But a green light with no other colours might indicate ‘a minimum effort stage 1 was
appropriate’ which only required the need to address two further closely coupled questions
of detail using a second traffic light:
1 Is this potential customer a good credit risk?
2 Does this customer have a track record for paying their bills in full without difficulty?
To answer both of these questions within stage 1, simple tests of creditworthiness and atti-
tudes to valid payment demands should be tried first. If these simple tests proved inconclu-
sive, more onerous tests should be used next. Very expensive tests might be left until stage
3, to ensure a bid looked promising enough to warrant the cost. If simple tests revealed a
clear lack of creditworthiness or other serious payment difficulties, the proposal should be
given a red light and dropped at the outset of stage 1, before any serious analysis effort was
wasted. But if the position was ambiguous, progress to stage 2 or 3 noting the concerns and
using an amber light might be sensible.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
226 Employing planning tools in practice
She would assume for present purposes that Transcon had an initial overall green light fol-
lowed by a ‘creditworthy customer who pays without difficulties’ green light which implied
a final summary green light and the end of stage 1 analysis. However, this assurance was
important, and a simple ‘triple green’ stage 1 outcome should not be presumed to be the
norm or assumed to apply to individual cases without asking the relevant questions. Trevor
needed to confirm this position using the accounting department staff.
To give them a few illustrative examples of departures from this minimum effort triple
green stage 1 outcome, a blue light might be associated with a customer who wanted a
sporting events audience information system for international level events which would be
central to television coverage and give Astro valuable public exposure, or a business system
in a rapidly expanding new area Astro Inc wanted to break into or a system requiring spe-
cial Astro skills which Astro was currently underusing and wanted to keep fully employed.
A red light might be associated with a customer who was known to use exploitive labour,
or dishonest customer relationship practices, or could seriously damage Astro’s reputation
for any other reasons.
The more important the ambiguity posed by blue– yellow plus amber– red combinations,
the greater the need for more detailed questions to explore underlying issues, but in stage
1 the focus was stopping further effort if a bid did not look appropriate without trying to
resolve ambiguity better left until stage 2 or 3.
Important ambiguity involving multiple issues should be clearly flagged by multiple traffic
light ratings – a multiple light cluster with different colours visible as separate lights rather
than a single ‘white light blur’. A linked key issue was always using short and concise notes
to explain the nature of each judgement, with no attempt to use simple scoring schemes
which ‘add apples, pears and oranges and equate the sum to bananas or banana equivalents’.
That is, a red light plus a blue light is not equal to a green light, and there is no meaningful
way to just average the implications of different colours within this framework.
One detailed question revealing a lack of synergy about the project or the customer might
suggest an amber light but not define a non- starter, but two or more might make a non-
starter designation likely. Questions which reveal very strong synergy would be useful for
passing the stage 1 or 2 test and for opportunity management purposes in stage 3, but their
immediate purpose would be counterbalancing negative answers when taking and explain-
ing a balanced view, and it would be important to identify them as part of stage 1 for all
these reasons.
To help clarify the boundaries when ethical judgements might be involved, and indicate
the possible need to flag some stage 1 end- point ethical issues, Martha suggested they use
a variant of a ‘red face test’ once suggested by her uncle Paul: if you would go even slightly
red in the face explaining to a valued customer, your boss, or a close family member, why
you did something because of the ethical/moral issues involved – give it a red light. If you
would not go red but had to think about it carefully – use an amber light to communicate
your concern.
It was important to understand clearly that such judgements were complex and debatable,
but Astro needed a simple corporate system to ensure that the obvious was not overlooked
and important concerns were treated consistently even if the issues involved were debatable.
Martha suggested it would clarify matters if she gave them a simple illustrative example
of the overall approach which Astro needed in her view – a view endorsed by Rhys. If Astro
had a potential customer who sold very high interest rate ‘payday’ loans – widely but not
universally judged to be extortionate – she would personally be inclined to give a bid for
them a red light because in her view there was a case for rejecting this kind of customer for
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 227
very good business reasons as well as for important underlying moral reasons. However, she
could clearly understand why Rhys or those further up the Astro corporate ladder might feel
there was a case for charging the upper limit of Astro’s mark- ups or perhaps not even trying
to make a moral judgement of this kind. She had never encountered this particular kind of
case and would not waste time attempting to make a judgement until she did, but when
such a case arose, if she had any doubts, she would seek advice from Rhys and expect him to
seek advice from his boss if need be. Furthermore, she thought all Astro Wales staff should
see Rhys as their local gatekeeper for these issues, working up the corporate ladder to the
Astro UK board as and when necessary.
The key message was she did not want individual Sales or Systems people feeling uncom-
fortable about making these ethical judgements on their own, and she did not want them
making them inappropriately. If Astro Wales staff flagged this kind of issue to her and Sian via
the formal bidding process traffic light approach as soon as possible, they would take advice
when in doubt, and the necessary communication channels would be used effectively and
efficiently. She wanted a consistent and open policy which the Astro UK board ultimately
took full responsibility for when critical choices were involved, defined by potential issues
as they arose, unless Rhys or the board or intermediate levels were alerted to issues they
wanted to cascade down. She and everyone else would then be obliged to conform to an
internally consistent board- led Astro position all staff understood or find other employment.
A reasonably obvious flip side generalisation of the red face test was a ‘halo test’. For
example, a blue or yellow light might be associated with a deserving charity with good busi-
ness case reasons for a low margin (they could project a favourable public image and perhaps
Astro could sell similar systems to other customers), as well as ethical reasons the whole
office and the Astro UK board might support.
A less obvious generalisation of the red face test for ethical concerns was a ‘frowns test’
to deal with all non- ethical concerns associated with any other red or amber light issues,
like key staff needed for a potential customer were already fully stretched on other highly
profitable and sensitive work. She did not want to overdo the idea of carefully chosen labels
for all tests associated with qualitative issues leading to traffic light colours other than green.
However, raising the possibility of such labels emphasised the need for clear recognition of
all relevant qualitative issues needing ongoing consideration. She did not want such issues
addressed via bias of the strategic misrepresentation variety when it came to the estimates of
quantitative measures in stages 2 and 3. For example, later in this tale Sian will be seen to
want higher cost estimates than Trevor for what might be usefully treated as a frowns test
amber light result, explained by the stage 1 documentation using a brief note from Sian
which Trevor and Martha would find useful.
The flip side generalisation of a frowns test was a ‘smiles test’, associated with all yellow
or blue light triggers which did not involve halo test ethical issues, like key staff needed for
a potential customer were underused and needed this kind of opportunity, or Astro Inc was
encouraging a move into this new business area.
The overall implication was four basic kinds of tests leading to all the coloured lights: a
smiles test and a frowns test for positive or negative issues needing qualitative attention, with
halo test and a red face test for special cases for ethical concerns.
As an aside, Martha suggested that Trevor or others might prefer alternative labels for the
tests of their own choosing, and she was always open to this kind of modification.
Martha indicated that she wanted the completed PD1, however complex or simple, to
involve a final overall summary traffic light assessment with a concise verbal statement of the
recommendation. A triple green might not warrant anything more than a standard ‘proceed
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
228 Employing planning tools in practice
to stage 2’. However, one or more other colours would warrant an ‘executive summary’
which argued a view for or against proceeding with a bid ‘on balance’. If moving to stage 2
was recommended, preliminary guidance should be provided on any recommendations for
higher or lower than normal margin, with the detailed questions and answers plus the rest
of the PD1 serving as backup.
She would expect Sales to operate this stage 1 process, with support from Systems on
issues like which Astro skills were currently underused or overstretched, the Accounting
Department regarding credit checks, and other departments as appropriate. Stage 1 should
be formal, focused, carefully designed, evolved and updated in the light of experience, and
documented on each application. All aspects of it should be both efficient and effective. She
would take responsibility for the nature of the process, but she wanted Rhys to take overall
responsibility for the way it was implemented, which he had agreed to. This was because
Rhys had to be the conduit for red face tests and halo tests, and he was best placed to take
responsibility for ensuring that the more general frowns tests and smiles tests took place,
with top- down input when relevant about key new red or blue light concerns. However, for
most purposes oversight from both Rhys and Martha would involve very light touch audits.
They would all need to work together to evolve the operational details, but neither Martha
nor Rhys would expect to be consulted about individual bid PD1’s unless asked.
As guidance to be used to develop early versions of the stage 1 analysis, Martha suggested
Sales and others involved should aim to apply about 1% to 5% of the effort required for a
complete three stage bid process. The 1% estimate was her current best guess of a plausi-
ble minimum to use as an aspirational stretch target if a triple green was involved. The 5%
estimate was her current best guess of a plausible maximum to use as a target if a joint blue
and red was involved. A 2% anticipated average outcome was her current best estimate –
the skew implying most outcomes would be closer to 1% than 5%. These estimates were
intended to illustrate roughly what she was anticipating, but the reality might prove quite
different.
Furthermore, they should expect to reject about 10% to 50% of the potential contracts
at this stage, with 50% as a stretch target when lots of work was available. She would be
surprised if they found less than 10% or more than 50% appropriate, but a 30% midpoint
of the 10% to 50% range was a very nominal rough guess at the expected outcome without
ruling out surprises.
These anticipated ranges and associated expected outcomes should be subjected to care-
ful review over time to see if adjusting their expectations would improve the efficiency and
effectiveness of the bidding process as a whole.
Ben would coordinate the development of this stage 1 process, working with Trevor
and his sales team plus Sian to start with, then other sales staff, and then other Astro staff,
reporting to her and Rhys as appropriate.
Martha observed that the computer- based information systems nature of Astro’s business
might suggest to some Astro staff that ‘the deliverable’ documentation for stage 1 should
be electronically based. If and when this development looked effective and efficient at a later
date, she would obviously consider this possibility. That is, at some time in the future the
questions could be presented to the sales staff responsible for each bid electronically, and
their responses would trigger the follow- on questions. This would be convenient for staff
executing stage 1 analysis. It would also ensure that she had a good database for Ben to use
for post- outcome analysis of bids continuing past stage 1, bringing important special cases
to her attention and consideration by Rhys when appropriate. However, it was premature to
waste time even thinking about this possible later development now, and her personal view
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 229
was the complexity involved and the flexibility needed implied that they should not rush a
computer- based stage 1 process.
Martha emphasised she expected enlightened judgements reflecting the context and
asymmetric penalties. For example, if winning new contracts was especially difficult, then
risking early rejection of what could prove to be a big opportunity would not be worth a
minor saving in effort. But if Astro sales were at a very healthy level, and Systems staff in a
crucial area were already seriously overstretched, then a potentially difficult customer might
warrant early rejection without too much probing of the opportunity potential.
She also emphasised that it was separability she wanted to see between the different stages,
not independence. The objectives and the resulting processes and outputs were highly inter-
connected in very important ways, despite important reasons for the separable structure. In
particular, stage 1 would be qualitative in nature and quite different in character from most
aspects of stage 2, but stage 2 would draw upon stage 1, stage 3 might add to the qualitative
knowledge generated by stage 1 as well as drawing upon it, and the stage 2 process was a
special case of the more general stage 3 process.
She would not expect to see most PD1s, Rhys would be involved only on an exception basis
unless he wanted to feed down Astro UK concerns, and sales personnel would normally move
directly to stage 2 if a potential contract survived stage 1. But a formal hurdle test administered
by the responsible sales executive would be involved, and subsequent audits might include
non- starters as well as those potential contracts which progressed to stages 2 and 3.
The focus of this monitoring would be the system and the players in the system rather
than individual decisions, but seriously inappropriate individual decisions would receive
attention, as part of an overall concern for learning from experience.
She wanted Ben to work with Trevor and Sian and other Astro staff as appropriate to
develop a simple PD1 example for Transcon, in parallel with their follow- on development
for stages 2 and 3. However, for that day’s purposes they would just assume a prototype
stage 1 process would lead to a final overall summary green light for Transcon with a ‘pro-
ceed to stage 2’ recommendation, subject to confirmation from Trevor that Transcon had
a triple green light. She would finish her presentation on stage 1 now, taking feedback after
they had time for further reflection whenever issues occurred to them, unless they had
immediate questions. There were no immediate questions, so Martha suggested a coffee
break, to be followed by addressing stage 2.
Stage 2 – identifying no- hopers
Martha began the stage 2 discussion by indicating that the preliminary view of profitability required by the test for ‘no- hopers’ of stage 2 justified the working assumption that Astro
could use a simple single- pass approach to unbiased estimation of:
1 the expected direct cost of meeting bid commitments,
2 the expected value of the probability of winning for a plausible set of margins on the
expected cost just estimated.
The working assumptions employed to estimate the expected cost associated with the first
part would be used to assess the probabilities in the second part. However, further inter-
dependence between assumptions underlying expected cost estimates and assumptions
underlying expected probability of winning estimates at any given level of margin would be
deferred until stage 3.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
230 Employing planning tools in practice
After a single pass through sequential analysis of these two components, synthesis which
also addressed stage 1 results would be needed to filter out the stage 2 no bid decisions – the
no- hopers. As a rough guide, Martha was anticipating rejecting 10% to 50% of the potential
bids which survived the non- starter tests, with 50% as an aspirational stretch target if a lot
of work was available, 10% as a comparable lower bound, and 30% as a mid- range expected
value. She was anticipating about 5% of the effort required for a complete three stage bid
process would be needed in stage 2 for all potential customers, centrally located in a 1% to
10% range. The 5% was an expected outcome estimate which should be achievable on aver-
age, although she would not rule out surprises.
Stage 2 was similar to stage 1 in terms of the anticipated rejection rate, the use of a single
pass process, and the need for ongoing review and evolution. But stage 2 would be very
different from stage 1 in most other respects.
Martha indicated that PD2 would be produced by a five phase process. She then showed
them an early draft version of Figure 6.2, emphasising that all the ideas and associated
assumptions might evolve.
You should find it helpful to see the final form of Figure 6.2 now and use it to fol-
low the way Martha and her colleagues developed it. The five phases imply five process
phase 2.1:
initiation of stage
phase 2.2:
cost uncertainty analysis
phase 2.3:
probability of winning analysis
phase 2.4:
uncertainty synthesis
phase 2.5:
margin evaluation
to
stage 3
no bid
Figure 6.2 The stage 2 plan of Astro’s SP for bidding.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 231
components for stage 2, referred to as phases 2.1 to 2.5, producing five components for the
PD2, referred to as the PD2.1 to PD2.5.
Phase 2.1 – stage 2 initiation to produce a PD2.1
Martha indicated that the information provided by Trevor in the draft PPD of Table 6.1 was
a good basis for what she would call ‘phase 2.1’ – phase 1 of stage 2. If a member of Sales
wanted to take a bid beyond a PD1, they should proceed as Trevor had, visit the customer,
and make summary notes. The resulting document, a PD2.1, could then be sent by Sales to
Systems for cost input. It should have the completed PD1 as an annex.
A triple green or a complex multiple colour rating would clearly be of direct interest for
decision making in stage 2, but it should not significantly affect the effort required in stage
2. If stage 3 was reached, then a multiple colour rating would have an impact on the stage
3 effort required, especially if a blue light signalling an important opportunity was involved.
Martha might have added that preliminary draft preparation of part of the PD2.1 based
on the PD1 prior to visiting the customer might provide a plan which would help to focus
the discussion in a useful manner in areas of uncertainty flagged by stage 1. However, she
knew Trevor and other sales staff would make good use of the PD1 in this way and did not
want to labour the obvious.
Phase 2.2 – stage 2 cost uncertainty analysis to produce a PD2.2
Martha indicated that if a PD2.1 as just described was sent to Systems in the future, in her
view Sian or a senior member of the Systems group ought to be able to suggest an unbiased
expected direct cost estimate of about £12m associated with a preliminary range of actual
outcomes of about £10m to £20m without expending the considerable effort Sam had
put into the PPD. By ‘unbiased’ she meant ‘an estimate which was not systematically too
high or too low – about right on average’. The £10 million lower end of the range would
be a ‘nominal minimum’ which was in the P1– P10 range, with allowable ambiguity about
which percentile value was involved, although she would recommend targeting a P10. The
£20 million upper end of the range would be a ‘nominal maximum’ which was in the P90–
P99 range, with allowable ambiguity, although she would recommend targeting a P90, and
consistent upper and lower nominal range boundaries were essential.
Crucially, it should be feasible to do this on the basis of about an hour’s consideration at
a simple overview level, without referring it to Sam or other staff for the development of the
considerable level of detail which Sam had contributed to the PPD of Table 6.1. Whoever
did it would need an overview perspective based on experience, like Sian, but it might be
someone reporting to Sian trained for this task. This was the level and kind of input from
Systems which Martha wanted for stage 2. Could Sian provide this sort of precision with
that level of effort?
Sian had not actually read the draft PPD prepared by Trevor or the invitation to tender
before passing it on to Sam, but her subsequent close reading of Sam’s detailed estimates
suggested an expected cost easily exceeding £13 million could have been estimated without
all the detail Sam had provided. She did not want to seem too pessimistic relative to Mar-
tha, and the nominal range was wide, so she said she was happy to agree that an estimate of
expected direct cost of £13 million associated with a £10 million to £20 million anticipated
outcome range was the kind of ‘preliminary’ estimate that she or one of her staff with wide experience could produce in an hour or so. Sian liked the way the discussion was going, and
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
232 Employing planning tools in practice
she believed that with suitable mentoring Sam could meet the requirements Martha was
suggesting.
Martha smiled when Sian increased the expected outcome estimate from £12 million to
£13 million but did not comment. She emphasised that they were now assuming £13 mil-
lion was an expected outcome in a nominal uncertainty range of the order £10 million to
£20 million and indicated that adding this preliminary cost estimate to the PD2.1 would
make it a PD2.2, which could then be sent to Ben. She then asked Ben to explain what he
would do for phase 2.3, phase 3 of stage 2.
Phase 2.3 – stage 2 probability of winning analysis to produce a PD2.3
Ben started by reminding everyone that Trevor had provided a preliminary maximum bid
estimate and a preliminary minimum bid estimate with associated estimates of the chance
of winning in the draft PPD. He indicated that in terms of future practice the working
assumption was Trevor would produce these estimates for a PD2.3 in the light of the
PD2.2, making use of the initial Systems view of expected cost in the PD2.2, in this case,
the £13 million estimate just discussed. For today’s purposes he would assume Trevor had
guessed correctly that £13 million was the kind of estimate Systems would provide later.
Trevor’s estimates would be used to produce a ‘preliminary probability of winning curve’,
illustrated in Figure 6.3.
Ben began explaining Figure 6.3 by pointing out that the version of the preliminary prob-
ability of winning curve he recommended for use today was the solid straight line between
points ‘a’ and ‘b’, the simplest approach feasible. However, it would be useful to spend a few
minutes discussing what using the straight line between points ‘a’ and ‘b’ in Figure 6.3
implied in terms of underlying assumptions and associated more complex possibilities. Trevor
had provided a preliminary maximum bid estimate of £20 million, with a chance of winning at
P(B)
probability of
winning
key: example discrete values
some discrete values of particular interest
preliminary probability of winning curve
assumed underlying curve
possible extrapolations
10 15 20 25
B bid in £m
0.2
0.4
0.6
0.8
1
0
c1
c2
c3
b
a
d1
d2
d3
d4
Figure 6.3 Preliminary probability of winning curve.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 233
this level of about 10%. This information had been used to plot point a in Figure 6.3. Point
‘a’ was linked to the information slipped, deliberately or otherwise, by a Transcon senior
manager, that £20 million was their budget figure. The £20 million figure was presumably
based on a recommendation from the consultants who prepared the invitation to tender, and
most of their competitors probably had the same information. It was sensible to use such
a figure as a preliminary maximum bid estimate, with an associated probability of winning
estimated approximately to the nearest 10%, as Trevor had done. Trevor had also provided a
‘preliminary minimum bid estimate’ of £15 million, with a chance of winning at this bid level
of about 80%. This information had been used to plot point ‘b’ in Figure 6.3. Ben indicated
it had not occurred to him at the time to ask Trevor what the basis of point ‘b’ was, and the
rationale involved might be important. However, when drawing Figure 6.3 he assumed that
Trevor had chosen £15 million because he had a Systems preliminary cost estimate of about
£13 million in mind, he was unsure about the shape of the probability of winning curve below
this value, and he did not think it was worth worrying about the curve below this value given the £13 million expected cost. By implication, he was reasonably confident about the shape of
the curve for bids above £15 million.
Trevor indicated that Ben was quite right in assuming he was reasonably confident over
the range £15 million to £20 million but not below £15 million, and he doubted that a bid
below £15 million would be of interest, given the information available. Trevor had smiled
approvingly at Ben’s inferences on his behalf, particularly appreciated because he too had
not thought this through earlier. He had certainly not explicitly thought about the role of the £13 million expected cost estimate in the way it was assumed he had in mind, but he
liked the idea of formally doing so. He also liked the idea of avoiding considering a curve
range that did not matter. Martha had carefully coached Ben to make sure he was aware of
the need for this sort of facilitation skill, helping him to acquire this particular insight, and
Ben was a fast learner.
Ben thanked Trevor for confirming his interpretation of point b. He then indicated that
he had joined point ‘a’ and ‘b’ in Figure 6.3 with a solid straight line to define a preliminary
probability of winning curve over the ‘preliminary bid range’ £15 million to £20 million,
plotting intermediate points d1 to d4 at £16 million to £19 million.
Ben had added the curve shown by dashes between ‘a’ and ‘b’ to indicate the assumed
nature of the underlying curve. He highly recommended using a ‘best estimate’ of this dashed
curve in future stage 2 analysis if they found this option helpful, but he would avoid doing
so today to ‘keep it simple’. He had also added the dotted curve beyond the preliminary bid
range at the top end, and he had added three dotted curves at the bottom end for bid values
below £15 million to indicate some of the possible scenarios which might be involved. For
example, the curve to c1 implied a bid of about £12 million would win for sure. This might be
interpreted in terms of three linked working assumptions: Transcon could be convinced that
such a bid was the result of a decision by Astro ‘to buy the work to keep Astro staff busy’, Astro
was the highest quality bidder in the contest, and a £12 million bid by Astro represented out-
standing value for Transcon which no competitor could touch. As another example, the curve
through c3 implied a bid below about £14 million would lead to a decrease in the probability
of winning. This might imply that a bid this low would suggest to Transcon that Astro lacked
credibility, Astro was desperate for work because it was inferior to others in the marketplace,
and Astro did not understand what the job was about. The curve through c2 was an intermedi-
ate case, with a peak probability of winning of about 0.9 at a bid of about £13 million.
Ben encouraged a limited discussion about these dashed and dotted curves, to gain under-
standing of (and approval for) his simple linear probability of winning curve by exploring
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
234 Employing planning tools in practice
the potential alternative natures of the underlying ‘true’ curve it approximated, both within
and beyond the presumed range of interest.
When Ben was pushed about the obvious lack of data associated with this preliminary
curve, Martha immediately interjected that it should become clear shortly that such a curve
was always used implicitly if it was not made explicit, and crude but simple explicit use was much better than ambiguous implicit use for a number of reasons. One of the key reasons of immediate relevance was the basis it provided for gathering appropriate data later, and
ultimately building a data basis for probabilities which had to be purely subjective initially,
but could be given an objective data- driven basis if they started on a purely subjective basis.
Another immediately relevant key reason was testing for alternative views, by revealing the
views to be used if testing for alternative views did not suggest changes.
Martha observed that Astro staff currently always made a note of the expected cost estimate
used for bidding, £13 million in this case. They discussed any relevant differences of opinion
which were generated by the process of committing to the figure generated. Furthermore,
they compared each outcome with the estimate, testing for bias over sets of completed projects
for bids won. From now on they were also going to make a note of the assumed probability of
winning in stages 2 and 3, testing both these probability estimates for bias immediately when
relevant differences of opinion surfaced, and later using outcomes to test over sets of bids won
and lost. For example, if over ten projects which made stage 3 the average stage 3 probability
of winning was 0.6, and bidding with an expectation of winning six, they actually won five, six,
or seven of the ten potential contracts, it would look as if they were reasonably competent at
estimating probabilities in stage 3. However, no wins or ten wins would indicate very serious
stage 3 probability of winning estimation problems.
Martha then suggested they leave the question of how data might be accumulated and
used for the moment, pressing on with the stage 2 analysis, assuming that the preliminary
linear curve between a and b provided sufficient precision over a suitable range for the pre-
liminary quantitative analysis involved in stage 2 to define the PD2.3. Trevor indicated he
was happy with this, although he was not too sure where it was leading. Sian looked puzzled
but intrigued. There were no more questions, so Ben pressed on with his presentation.
Phase 2.4 – stage 2 uncertainty syntheses to produce a PD2.4
Ben indicated that the preliminary cost estimate of £13m provided by the PD2.2, plus the
preliminary probability of winning curve illustrated by Figure 6.3 which defined the PD2.3,
needed integration. This integration would be provided by a ‘preliminary margin evaluation
table’ illustrated by Table 6.2, which would define the PD2.4. He would now explain how this
table was produced, and how they could interpret it, as a basis for using it in the next phase.
The first row in this table introduced the E(C) = £13m notation for the expected cost
estimate provided by Systems, which he had put into his spreadsheet as soon as Sian con-
firmed the £13 million estimate. His working assumption earlier had been £12 million, but
a spreadsheet basis for Table 6.2 made this kind of real- time adjustment simple.
The second row headings for the columns defined further notation which clarified the
calculations. The concepts used and the associated notation choices were all straightforward,
but he would take them through each entry in the first option row, so they had a clear view
of what was involved. He would then consider relationships between option rows.
The first three entries of the B = £15m row of Table 6.2 indicated that the corresponding
point in Figure 6.3 was ‘b’ and the associated probability of winning was P(B) = 0.8. The
fourth entry indicated a corresponding conditional margin of M = £15m – £13m = £2m.
This M value was the contribution to overhead and profit if the bid of £15 million was
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 235
Table 6.2 Preliminary margin evaluation table (the PD2.4 produced in stage 2 phase 4).
Expected direct cost E(C) = £13m working to the nearest £m with an outcome in the range of £10m to £20m anticipated
Bid values B (£m)
Points on curve used
Corresponding probabilities P(B)
Margin (conditional) M = B – E(C)
Expected margin E(M) = P(B) × M
Comments
15 b 0.8 15 – 13 = 2 2 × 0.8 = 1.60 Minimum bid 16 d1 0.66 3 1.98 Buying work worth it? 17 d2 0.52 4 2.08 Maximum E(M) 18 d3 0.38 5 1.90 Approaching normal M 19 d4 0.24 6 1.44 Normal 20 a 0.1 7 0.70 range for M
A decision tree with six options and row values indicated in bold for the maximum E(M) option
successful – it was conditional on winning. Multiplying M by P(B) yielded E(M) = £2m ×
0.8 = £1.60m, the expected margin indicated by the fifth entry.
The B = 16, 17, 18, 19, and 20 rows of Table 6.2 completed the underlying the deci-
sion tree. There were six bid value choices in the first column, working to the nearest £m.
About half a dozen alternative bid scenarios was a suitable level of detail for preliminary
margin evaluation, but more detail could be provided if deemed useful. The second column
indicated the corresponding points in Figure 6.3 and the third column indicated the cor-
responding probabilities of winning. The interpolated points d1 to d4 taken from the linear
version of the curve of Figure 6.3 used a second decimal place to aid interpretation.
The final column provided a set of notes indicative of the discussion which needed to take
place when exploring the basis for selecting a preliminary bid and using possible choices to
assess whether Transcon was a no- hoper. For example, the ‘maximum E(M) bid’ of £17 mil-
lion involved an M of £4 million, a 31% uplift on E(C) = £13 million, well below the 30%
plus 10% ‘normal M’. The £18 million bid increased the uplift to 38%, while the £19 million
bid increased it to 46%, above the ‘normal M’. These uplifts could be included in the table.
Table 6.2 used a HAT approach which is comparable to but marginally more complex
than Nicola’s safety stock model, a relationship illustrating the portability of HAT models
you may find useful.
There were no questions about how Table 6.2 was generated or the basics of its interpre-
tation, so Ben pressed on to discussing how they could use it.
Phase 2.5 – stage 2 margin evaluations to produce a PD2.5
Ben indicated that stage 2 phase 5 involved the use of the PD2.4 plus the PD1 to discuss
possible bid choices and evaluate the key associated trade- offs given the information avail-
able to date. If Transcon had a stage 1 traffic light rating that was anything other than the
triple green that they were assuming, the PD1 would need careful attention when interpret-
ing the right- hand column of Table 6.2, synthesising the quantitative basis of Table 6.2
with the qualitative basis of the PD1 executive summary to prepare these notes. They did
not need to determine a proposed bid value – they just needed to decide whether to make a
no bid decision or move on to stage 3. The potential profitability of this contract given any
PD1 issues was the core concern in stage 2, and a specific bid price was not required. The
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
236 Employing planning tools in practice
PD1 information about the desirability of the customer and Astro’s appetite for more work
of this kind might be highly relevant, including complex circumstances like the need for a
balanced view based on a mixture of blue, yellow, green, amber, and red lights.
A summary of this discussion added to the PD2.4 would define the PD2.5. Ben suggested
they have such a discussion with an assumed triple green stage 1 result to illustrate what
was involved in a simple basic case. For example, if the preliminary estimate of the expected
direct cost was £13 million, and the simple linear ‘curve’ of Figure 6.3 was approximately
correct, then a bid of £17 million maximised expected margin at a value of £2.08 million,
with a conditional margin of £4 million and a chance of winning of about 0.52 (shown in
bold in Table 6.2 to assist identification). They would have to increase their bid to £19m to
achieve a bid above the ‘normal M’, which reduced the probability of winning to 0.24, but
this might be acceptable if M was more important than P(B) on this occasion. The key was
Table 6.2 gave them a basis for this kind of discussion, without any need to determine their
bid at this stage. All they had to decide was whether to carry on to stage 3. Ben suggested
that Table 6.2 clearly indicated to him that the Transcon bid was not a no- hoper. Transcon was a preliminary ‘goer’, requiring a more detailed stage 3.
At this point Sian suggested that the Transcon contract would be a useful one to win. How-
ever, in her view it was ‘nice to have’ rather than critical. She suggested that a bid of £19 million
with an assumed margin of £5 million on a £14 million cost looked about right to her. She
would be happy to get the work, but with her £13m cost estimate increased by a further £1m
provision to allow for unknown unknowns and ensure they did not overrun on cost, playing by
the rules on margin. She had been tempted to increase Martha’s £12 million initial suggestion
to £14 million earlier but did not want to look too pessimistic at that stage of their discussion.
Trevor’s rejoinder was he thought a bid of £15 million with an assumed margin of £3 mil-
lion on a £12 million cost looked about right to him. He believed that £13 million might be a
reasonable estimate by Sian for the expected cost given her current assumptions, and a further
£1 million provision for the reasons Sian suggested might seem sensible from a Systems per-
spective, but they could reduce the expected cost if both Systems and Sales worked at it. Fur-
thermore, there was no point in risking the loss of this important contract in what was clearly
a tough market. He did not say so, but he was not keen to chance four lost systems integration
contracts in a row because it would seriously damage his credibility and the commission- based
part of his income, and playing by the rules on margin was not a big concern for Trevor.
Martha then suggested that Sian and Trevor had both raised extremely important issues
which needed addressing. However, before they were considered, along with a number of
other concerns, like how do these issues relate to PD1 multiple coloured lights, it would be
worth beginning a stage 3 analysis. It would be helpful to clarify both the expected direct cost
and the shape of the probability of winning curve in the process of addressing Sian and Tre-
vor’s concerns, in addition to reasons possibly underlying Sian’s comment about this contract
being just a ‘nice to have’ with a much higher expected or commitment value cost than Trevor
was suggesting. Sian and Sam had already done what was needed as a starting point basis for
the stage 3 expected direct cost estimate, but some additional effort was needed by Systems,
and much more needed to be done by Sales to underpin the probability of winning curve.
Martha did not raise the issues associated with Martha and Trevor having different incen-
tives because Trevor was on commission and Sian was not, but this was part of the con-
siderable complexity stage 3 had to deal with and part of the cost estimation process bias
concerns (by no means the whole story).
Martha indicated that Rhys was fully briefed and happy with the stage 2 process as just
outlined by Ben, and she would take ongoing responsibility for the stage 2 process, with
important input and support from Trevor, Sian, and Ben.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 237
To illustrate how she wanted the formal stage 2 process to work in the future, Martha sug-
gested they finished stage 2 now by agreeing that Transcon was a ‘preliminary goer’, progress-
ing to stage 3 was appropriate for the Transcon bid, and it would be reasonable for sales staff
to make this judgement subject to later audit by her. Everyone readily complied. She then
indicated that this decision noted on the PD2.5 should be taken to define the end of stage 2.
The rationale for separate stages – further comments now
Reflecting on the details of stages 1 and 2 as explored so far, one of Martha’s reasons for
separating stages 1 and 2 was using stage 2 to concentrate its focus on a minimum clarity
quantitative approach while stage 1 provided a concentrated focus on a minimum clarity qualitative approach, both minimum clarity approaches, but very different in nature. Mar- tha was aware that experience in many contexts over many years suggests that people can
do a much better job of two different kinds of tasks at any given level of total effort if they
do one at a time in a suitable sequence. In this case, stage 2 clearly had to follow stage 1
to avoid rejecting blue and yellow light projects with low contributions to profit in margin
terms, the rationale for the ordering.
Assuming what needed to be done in stages 1 and 2 could be separated and ordered in
this way, with greater efficiency and effectiveness as a result of exploiting the separability
involved, this implied the effort saved by no bid decisions after stage 1 can be viewed as a
bonus even if the proportion rejected was quite low – a clarity efficient bundled package of
benefits was involved. Despite the obviously important value of screening out non- starters
in stage 1, the more fundamental reason for separating stages 1 and 2 might actually be seen
as more effective analysis in both stages with less effort. This was a clarity efficiency process
design issue which is useful to understand, even if designing processes is not your direct
responsibility or concern. For example, if you are a board level manager you might want to
ask why other comparable processes do not have this kind of feature.
Separating stage 3 from stages 1 and 2 involves a further bundle of benefits. However, this
time stage 3 builds on two earlier stages using a significant increase in sophistication relative
to both, and you need a working understanding of what stage 3 involves before exploring
these benefits. As a hint about what is involved, we can exploit the benefits of focusing on
one task at a time in a suitable sequence, and deal with complex interdependencies, if we plan on using an iterative approach, the basis of key differences in stage 3.
Martha briefly summarised these observations for Sian and Trevor so they had some feel for
her overall process design rationale and then finished the morning’s discussion by suggesting
they now enjoy a well- earned lunch, with a view to dealing with stage 3 suitably refreshed.
Stage 3 overview – bidding strategy and tactics development
After lunch Martha prefaced their workshop resumption by reminding Sian and Trevor that
stage 3 was about developing a bidding strategy and tactics involving non- price features of
their bid and then making a pricing decision to bid for potential contracts Astro wanted to
win, with a small chance that a late no bid situation might emerge.
Her first key point of substance was introduced by indicating that she would maintain
oversight of the form of the stage 3 process as well as an audit of outcomes, as for stage
2. However, with Ben’s support and her assistance when needed, Trevor would be for-
mally responsible for individual stage 3 execution exercises involving Transcon and all other
transport sector bids, in conjunction with Sian whenever joint working was essential, and Trevor’s role would go beyond just being responsible for the outputs of individual bid
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
238 Employing planning tools in practice
exercises as in stages 1 and 2. He would be leading bidding process development on behalf
of Sales. Sian’s role would also extend beyond her Systems Manager roles in stages 1 and
2 to joint leadership of bidding process development. Trevor’s leadership in the shaping of
stage 3 process developments in partnership with Sian would be a key feature of the way
Astro Wales would address stage 3. Both Martha and Ben would be behind Trevor and
Sian’s development of the generalisation of the starting point prototype process outlined
today, but Trevor and Sian would take leadership of bidding process development as far as
possible, with Trevor leading the Sales aspect of what had to be viewed as a Sales/Systems
partnership. Furthermore, once Trevor and Sian were comfortable with the evolution of
a suitable version of today’s prototype process, and she was also happy with it, she would
expect Trevor to help other sales staff responsible for all other areas of Astro Wales business
to develop variants suitable for their areas.
Martha then moved on to the specifics of the stage 3 process, emphasising that stage 3
was a multiple phase iterative process by design. Stages 1 and 2 were both single pass non-
iterative processes by design. This multiple pass aspect of stage 3 was an important differ-
ence which would significantly affect the tasks facing the process developers as well as all the
users of this stage 3 process.
She then showed them an early draft of Figure 6.4. Like Figures 6.1 and 6.2, the final
form of Figure 6.4 has been provided now so you can use it to follow the way Martha and
phase 3.1:
initiation of pass
phase 3.2:
cost uncertainty analysis
phase 3.3:
probability of winning analysis
phase 3.4:
uncertainty synthesis
phase 3.5:
bid evaluation & development
no bid
bid
Figure 6.4 The stage 3 plan of Astro’s SP for bidding.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 239
her colleagues developed it. She indicated that after a few overview comments they would
consider phases 3.1 to 3.5 in turn, producing PD3 components 1 to 5, referred to as PD3.1
to PD3.5.
Pass 1 of stage 3 would be centred on higher clarity quantitative analysis address-
ing uncertainty about the expected direct cost of meeting bid commitments, followed
by higher clarity quantitative analysis of uncertainty about the probability of winning
for an assumed margin, separating the analysis of these two components. But follow-
ing this separate higher clarity analysis of these two components, an iterative approach
which addressed interdependences would be needed to explore planning decisions about
what the customer is being offered which affect both the expected cost and the expected
probability of winning. Iteration was required because of these interdependencies.
Eventually detailed bidding tactics within an evolving bidding strategy would become
the focus.
In outline the stage 2 and stage 3 processes had exactly the same basic phase structure,
but because a multiple pass approach was involved in stage 3, each stage 3 phase was more
complex than the stage 2 equivalent. The stage 2 approach could be viewed as a special case
of the more general stage 3 approach – a minimum clarity variant they might see as ‘pass
zero’ version, not to be confused with ‘pass 1’ of stage 3, which would seek a much higher
level of clarity than stage 2.
Phase 3.1 – initiation of stage 3 on pass 1
Martha indicated that initiation of stage 3 on pass 1 would begin in phase 3.1 by planning
the remaining four phases for pass 1, bearing in mind what had been achieved in stages 1
and 2, and anticipating further stage 3 passes. This kind of planning the planning was not
part of the stripped- down more basic phase 2.1 variant used in stage 2.
As part of these plans for the development of the Transcon bid, Sian and Ben would
be asked to plan the production of an ‘enhanced cost estimate’ to form the PD3.2, the
deliverable of phase 3.2. This would be an enhanced version of the PD2.2 preliminary
cost estimate.
Trevor and Ben would be asked to plan the production of the PD3.3, the phase 3.3 deliv-
erable. The PD3.3 involved an ‘enhanced probability of winning curve’, a more developed
version of the simple Figure 6.3 approach used for the PD2.3.
The PD3.2 and PD3.3 would then be combined by Ben to produce an ‘enhanced margin
evaluation table’, defining the PD3.4, an enhanced version of Table 6.2.
PD3.5 development would follow, analogous to but generalising phase 2.5 production
of the PD2.5. The phase 3.5 process had to consider a possible no bid decision, but its
focus would be refining earlier analysis in an iterative manner. Further analysis passes would
address selected non- price aspects of their bidding approach, initially at a broad strategic
level and then in terms of tactical detail within an emerging strategy.
Sian and Trevor both agreed that this outline made sense, but Martha had primed this
reaction by suggesting that she was sure they would not yet be very clear about where the
discussion was going and that they would have to explore what followed in some detail
before they could really understand this initiation phase.
Martha now indicated that Sian and Sam had already provided a starting point basis
for the stage 3 cost estimates, in terms of the details in the PPD of Table 6.1 discussed
earlier, which she had excluded from the PD1 and PD2 documents. Ben had already used
this material to complete a draft of the PD3.2 deliverable, with her support and input.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
240 Employing planning tools in practice
Ben would present this draft now. Martha observed that in practice Sian or her delegated
Systems person and Ben should discuss Ben’s contribution to this document before it was
released to others, and Martha would become involved only when requested. Ben’s current
example was purely illustrative. Martha also observed that Sian or other Systems people and
Ben or Trevor did not need to agree on all issues, but it would very be important to high- light any differences in view.
Phase 3.2 – enhanced cost uncertainty analysis on pass 1
Ben began his presentation of the PD3.2 by revealing Table 6.3.
Ben indicated that he would explain the Table 6.3 item entries based on the PPD infor-
mation line by line, using backup diagrams to elaborate the summary provided by Table 6.3
in some cases, as soon as he had explained the four individual direct cost column headings.
Table 6.3 would be spreadsheet based to facilitate evolution, and together with associated
backup diagrams Table 6.3 would be his part of the PD3.2.
When explaining the four individual direct cost column headings, Ben indicated that:
1 the ‘base’ estimate was the starting point estimate provided by Sam,
2 he and Martha had agreed to use P10 to P90 ranges based on the information provided
by Sam,
3 they also agreed to assume a uniform probability distribution model,
4 this meant that the nominal expected outcome would be a simple average of the nomi-
nal minimum and maximum, and
5 a more complex assumed underlying reality needed general understanding.
What each of these five points implied was then outlined, with a view to discussing alterna-
tives later. Its basis was discussed with William in Chapter 3 in the context of minimum-
clarity estimates, not repeated in this chapter.
Ben then made a series of observations for each of the five items listed in Table 6.3, start-
ing with item 1.
Item 1 – central site equipment
The base cost for item 1 was the direct cost to Astro Wales of £3.6 million. This could be
taken as the nominal minimum, maximum, and expected cost. There was no uncertainty
of interest because determining Astro Inc to Astro Wales transfer costs was well beyond
Martha’s remit.
For Table 6.3 purposes direct cost was the issue, not selling price, and it was important
to maintain this focus. Margin (overall overhead cost recovery and profit) would be deter-
mined later. After the bid price was decided, the bid presentation package for the customer
would be drafted using standard list prices and suitable discounts.
There was no choice to be made for item 1, as noted in the right- hand column. This
item provided the simplest possible example of the cost analysis required in the stage 3 cost
estimate format of Table 6.3.
The issue of possible additional memory could be addressed as part of item 1, but he
would deal with it as part of item 4, following the approach to the descriptive material taken
by Sam.
Everyone nodded approval, so Ben moved on to item 2.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
T a
b le
6 .3
E
n h
an ce
d c
o st
e st
im at
e s
u m
m ar
y t
ab le
( th
e P
D 3
.2 p
ro d
u ce
d b
y p
as s
1 i n
p h
as e 2
o f
st ag
e 3
).
It em
s It
em c
om p on
en ts
a n
d
op ti
on s
D ir
ec t
co st
s (£
m )
C h oi
ce s
a n
d a
ss u
m p ti
on s
b a
se n
om in
a l
m in
im u
m n
om in
a l
m a
x im
u m
n om
in a
l ex
p ec
te d
1 C
e n
tr al
s it
e e
q u
ip m
e n t
A st
ro m
ai n f
ra m
e p
lu s
p e ri
p h e ra
ls 3
.6 3 .6
3 .6
3 .6
N o
c h
o ic
e
2 C
o m
p u
te r
su it
e m
o d
ifi ca
ti o n s
A st
ro Z
e n it
h T
o ta
l
0 .3
0 1
.0 0
1 .3
0
0 .3
1 .1
1 .4
0 .3
1 .3
1 .6
0 .3
1 .2
1 .5
N o
c h
o ic
e 3
I n
it ia
l o
p e ra
ti o
n a
n d
t ra
in in
g (a
) Z
o ro
(b )
A st
ro 1
A st
ro 2
1 .0
0 1
.1 4
0 .8
1 .2
1 .0
0 .7
1 .4
1 .6
1 .1
1 .3
1 .3
0 .9
N o
h o
st il e t
ak e o
v e r
e ff
e ct
S am
’s p
re fe
rr e d
o p ti
o n
M ar
th a’
s re
v is
e d o
p ti
o n
4 C
o n
v e rt
e x is
ti n
g p
ro g ra
m m
e s
(a )
D at
ap o l
(b )
S y sd
o c
1 .2
0 .5
1 .2
0 .5
1 .2
1 .5
1 .2
1 .0
S am
’s p
re fe
rr e d
o p ti
o n N
o n
o n - c
o m
p li an
ce e
ff e ct
s M
ar th
a’ s
p re
fe rr
e d o
p ti
o n a
ss u m
in g S
y sd
o c
2 A
d d
it io
n al
m e m
o ry
(a )
P re
- i n st
al l
(b )
P o st
- i n st
al l if n
e ce
ss ar
y 0
.5 1
.0 0
.5 0
0 .5
1 .0
0 .5
0 .1
S am
’s p
re fe
rr e d
o p ti
o n
M ar
th a’
s p re
fe rr
e d o
p ti
o n
5 D
is tr
ib u
te d
s ch
e d
u li n
g s
y st
e m
A st
ro e
q u ip
m e n t
(a )
A st
ro s
o ft
w ar
e (b
) Z
o ro
s o ft
w ar
e T
o ta
l (a
) o r
(b )
A d ju
st m
e n t
to (
a)
3 .0
1 .1
1 .0
+ 1
3 .0
1 .2
1 .2
4 .2
3 .0
1 .4
1 .4
4 .4
3 .0
1 .3
1 .3
4 .3
S am
’s p
re fe
rr e d
o p ti
o n
N o
h o
st il e t
ak e o
v e r
e ff
e ct
C le
ar s
w it
ch o
f p re
fe re
n ce
t o (
b )
w it
h n
o o
th e r
im p ac
t A
d d
it io
n al
p e rf
o rm
an ce
(a )
P re
- i n st
al l
(b )
P o st
- i n st
al l if n
e ce
ss ar
y 1
.0 1
.5 – 2
.5 1
.0 0
1 .0
2 .5
1 .0
0 .4
S am
’s p
re fe
rr e d
o p ti
o n
M ar
th a’
s p re
fe rr
e d o
p ti
o n
C o n
tr a ct
n o m
in a l d
ir e ct
c o st
O v e rh
e ad
a t
3 0
% N
o m
in al
t o
ta l co
st P
ro fi t
at 1
0 %
N o m
in a l p ri
ce N
o m
in a l m
a rg
in
1 0 .4
1 5 .7
1 1 .8
3 .5
1 4
.3 1
.4 1
6 .7
4 .9
R o u n d t
o £
1 2 m
i n
t h
e r
a n
g e £
1 0
m t
o £
1 6
m
R o u n d t
o £
1 7 m
R o u n d t
o £
5 m
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
242 Employing planning tools in practice
Item 2 – computer suite modification
The item 2 base cost estimate of £0.30 million for Astro’s direct costs was the £0.30 million
provided by Sam, as was the £0.3 million nominal maximum, minimum, and expected value,
apart from rounding to the nearest tenth of a million. They could treat this as a fixed price
estimate as suggested by Sam, with the P10 = P90. However, spurious precision wastes time
and, more importantly, can be misleading. Working to the nearest £0.1 million (about 0.5%
of total expected cost) would be useful for most purposes in the Transcon case, with a view
to overall clarity efficiency.
The item 2 base cost of £1.00 million for Zenith was as provided by Sam, but in this
case the nominal minimum, maximum, and expected values were different. The range of
10% to 30% uplift from preliminary quotes to actual prices for Zenith as reported by Sam
had been used to determine a P10 nominal minimum value of £1.1 million (£1.0 million ×
(10 + 100)/100) and a P90 nominal maximum value of £1.3 million (£1.0 million × (30 +
100)/100). Ben suggested values outside this range were clearly possible, but in the present
context it was not worth being too concerned about the chances of this.
The expected value of £1.2 million defined by (1.1 + 1.3)/2 could be associated with a
uniform probability distribution density function defined by the P10 nominal minimum and
P90 nominal maximum values, with an associated underlying assumed reality as discussed in
terms of Figures 3.1 and 3.2 in Chapter 3. These Chapter 3 figures using an MoD context
were not used here, but variants based on this item could have been employed to explain
minimum clarity estimates using an Astro example.
Table 6.3 shows a single total for the two components of item 2. The no choice note in
the right- hand column is applied to the total, and by implication both components. The
highlighting (in bold) makes later interpretation of totals easier.
Ben’s presentation on item 2 finished with a request for questions or comments, but none
were forthcoming, so he moved on to item 3.
Item 3 – initial operation and training
Ben began his discussion of item 3 by indicating that when two or more choices involved
probability distributions for a continuous variable, a useful conceptual and operational
framework which was widely employed involved plotting cumulative probability distribu-
tions for both choices on the same graph, what he called a ‘decision diagram’. He then
showed them Figure 6.5, which used simple linear cumulative curves for three options.
All three curves in Figure 6.5 were plotted using the Table 6.3 item 3 nominal minimum
and maximum values interpreted as P10 and P90 values.
The Zoro curve nominal minimum P10 = £1.2 million used the Zoro choice (a) base value
of £1.0 million plus 20%. The corresponding Zoro curve nominal maximum P90 = £1.4 mil-
lion used the same base value of £1.0 million plus 40%. The base value of £1.0 million was
provided by Sam, as were the +20% and +40% estimates for the P10– P90 range. The expected
value was the midpoint, £1.3 million. This Zoro curve ignored the possible implications of the
hostile takeover bid, as noted in the right- hand column of Table 6.3.
The Astro 1 (before adjustment) curve nominal minimum P10 = £1.0 million used the
Astro choice (b) base value of £1.14 million less 10%. The corresponding P90 = £1.6m used
the same base value of £1.14m plus 40%. The base value of £1.14 million was provided by
Sam, as were the −10% and +40% estimates for the P10– P90 range. The expected value was
£1.3 million, the same expected value as the Zoro option.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 243
In terms of ‘risk’ associated with direct cost variability as portrayed by Figure 6.5, Astro
1 (before adjustment) was clearly riskier than Zoro. Sam’s preference for Astro, noted in
the right- hand column of Table 6.3, could be assumed to be based on the hostile takeover
threat, which Sam had associated with ambiguous implications and uncertain consequences.
Ben indicated that Martha had taken the view that Sam’s preference was both reasonable
and representative of current best practice, but it raised three issues which she had discussed
with him that they all needed to discuss now.
First, a hostile takeover of Zoro was a risk of strategic concern for Astro. Astro routinely
used Zoro to good effect, and it was important to preserve the possibility of doing so in
the future because Zoro was a key subcontractor for Astro. To facilitate this, Martha had
approached Rhys about a possible friendly takeover approach by Astro Wales. Rhys had acted
immediately. Zoro would shortly become a ‘strategic partner’, owned by Astro, but run by
its existing management on terms that the Zoro management team were very pleased with.
Second, given the elimination of takeover risk, the Zoro option clearly became the pre-
ferred option if risk associated with direct cost variability was considered, no other objectives
apart from expected direct cost were relevant, and the Astro 1 (before adjustment) option
was the only alternative. However, Martha had approached Steve, the Service Bureau man-
ager, and with approval from Rhys, they had agreed to a 30% discount on the charge rates
relative to the standard rates assumed by Sam, the basis of the Astro 2 (after adjustment)
curve. The reason Martha had taken this action was the Service Bureau was very short of
work, with very low marginal costs and a risk of redundancies if their business did not turn
around soon. Steve and Rhys both preferred a 30% discount on standard rates to the Astro
Service Bureau option being dropped from item 3 in the Transcon bid – a smaller margin for
the Service Bureau was better than no margin at all. Ben did not say that Martha achieved
the 30% discount in a hard- nosed bargaining discussion, but this was understood without
being explicit. This point raised a number of questions because it introduced the concept
of different parts of Astro Wales bidding internally for work in a way not previously enter-
tained. But these questions were not asked or addressed immediately.
The expected cost of the Astro 2 (after adjustment) option was lower than the Zoro option
(£0.9 million compared to £1.3 million, a difference of £0.4 million), and the Figure 6.5
curves did not cross. This meant that the Astro 2 (after adjustment) choice was risk efficient
direct cost in £m
Astro 2 (after adjustment) Zoro
0.7 0.9 1.1 1.3
0.2
0.4
0.6
0.8
1cumulative
probability
0
1.5
Indicates expected cost (also median)
Astro 1 (before adjustment)
Figure 6.5 Decision diagram: item 3 (initial operation and training) example.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
244 Employing planning tools in practice
with a clear basis for clarified dominance. Sticking to cost for the moment, Astro 2 (after
adjustment) had a lower expected cost, and the probability of exceeding any given cost was
less. The Astro 2 (after adjustment) choice involved more variability, but it was less risky.
Unless ‘qualitative benefits’ (non- quantified benefits) worth £0.4 million could be associ-
ated with the Zoro choice, Astro 2 (after adjustment) was Martha’s preferred choice, as
indicated in the right- hand column of Table 6.3.
Third, there was a difference between the Zoro approach and the Astro 2 (after adjust-
ment) approach which Astro needed to sell to Transcon as an important non- price advan-
tage for Astro relative to the competition. This was Transcon being able to use the Astro
Service Bureau site before the new Transcon site was ready. The supporting case Astro
should provide ought to argue that this Astro choice would allow the training to start well
in advance of the Transcon facility being ready, spreading it out for the convenience of
Transcon staff (and Astro staff), and it would decouple training from the risk of any delays
with the new computer suite in addition to saving £0.4m. None of their competitors could
offer this non- price advantage. There might be important linked ongoing non- price advan-
tages which also needed effective marketing, like backup services in the event of a major
fire in the Transcon office, because all Transcon software would be up and running on the
Service Bureau hardware and Transcon staff would be familiar with its use at the nearby Ser-
vice Bureau site. To win bids they needed to systematically search for non- price- advantage
opportunities like this. They also needed to try to avoid letting their competitors convince
Transcon that training they would provide on the Transcon site was better than off- site
training by Astro. Furthermore, they needed to offer a Zoro option on the Transcon site
for an extra £0.4 million if there was any risk that they might fail to convince Transcon that
the off- site Astro training option was a good one, preferably linked to an approach to item 5
which demonstrated the value to Zoro of their new strategic partnership with Astro.
Trevor and Sian were both comfortable with the obvious implications of this part of the
presentation. Neither had come across the risk efficiency or clarified dominance concepts,
but both seemed clear enough in this context. At this point Sian asked Martha for clarifica-
tion about the internal market issue.
Martha responded by first indicating that the Astro Inc ‘normal’ 30% and 10% mark- ups
for overhead and profit were what she would call ‘nominal’, for use as a guideline to indicate
what was needed as a minimum on average. In very tough market conditions, they would
have to accept less or go out of business. Conversely, in very good market conditions, they
would have to take much more, in order to achieve at least the company norms on average.
They needed to be explicit about departures from nominal values, to avoid implicit pressure
to bias direct cost estimates and distort associated option choices. Trevor and Sian were well
aware of the need to respond to market conditions in these terms assuming that standard
internal transfer prices were correct and competition between different parts of Astro and
external contractors was not an issue, but she suspected that they had not thought about the
biased cost estimates and related distorted choice issues which could arise if these assump-
tions were inappropriate.
What was now self- evident to her, although she was sure it was less obvious to both of
them, so Martha elaborated the point briefly, was the process of choosing between options
(a) and (b) for this item involved a lower level bidding process with one part of Astro in
competition with Zoro. The same basic principles applied to all levels of bidding processes.
The Astro Service Bureau site associated with option (b) was truly desperate for the work, in
a very tough market for them, perhaps the beginning of a long- term decline. Overall opti-
misation of the Astro Wales position required the use of adjusted costs whenever this kind of
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 245
consideration was relevant. Whether or not it was seen as a formal part of her brief in terms
of her responsibilities to Rhys, Martha had agreed with Rhys and Steve that it was in the best
interests of Astro Wales to ensure that underused capacity in the Astro Wales Service Bureau
was pushed with the support of Sales and Systems. The Service Bureau Manager reported
directly to Rhys, outside the Sales and Systems structure, but when she could Martha needed
to push their sales as well as the sale of Systems peoples’ time and Astro Inc hardware and
software. This kind of adjustment was the instrument they all needed to use to help push the
marketing of underused Astro capacity. Of course, Astro could not sustain consistent nega-
tive adjustments of this kind in any resource area and would look to the elimination of such
areas in the medium term. However, in the short term it was important to make explicit neg-
ative adjustments which made direct cost as computed using internal accounting conventions
more accurately reflect actual marginal cost to align short- term choices with longer- term
corporate goals. The flip side of this, protecting high- demand resource areas from overuse via
positive adjustments to reflect opportunity costs, would arise in item 5.
Trevor and Sian were a bit stunned by this revelation, but they were both smiling, and
Sian could already anticipate part of what was coming in item 5.
Ben then indicated that the Figure 6.5 decision diagram format was a crucially important
basic tool for considering choices. He would organise some computer software to make
their production simple. But once Astro staff became comfortable with the operational use
and implications of decision diagrams, Table 6.3 could be used on its own. That is, the risk
efficiency and clarified dominance of the Astro 2 (after adjustment) option for item 3 were
indicated by a lower expected value and a lower nominal maximum value on Table 6.3.
Once everyone involved became familiar with the use of backup figures like Figure 6.5,
there was no need to actually produce a Figure 6.5 equivalent each time. The Figure 6.5
decision diagram could become a background conceptual device which only required opera-
tional use when someone did not understand direct use of Table 6.3 to make risk efficient
choices or it was a close call or controversial for other reasons. If it was a close call, they
might use a higher clarity version which he would explain later.
Trevor and Sian both nodded again, and Sian smiled, indicating she appreciated the effi-
ciency of the simple procedure which was emerging to justify clear- cut decisions by experi-
enced team members.
Ben then moved on to item 4.
Item 4 – convert existing programmes
Ben began by showing them Figure 6.6.
The ‘firm fixed price’ of £1.2 million for the Datapol option implied a vertical cumula-
tive probability curve in Figure 6.6, assuming P10 = P90 = £1.2 million as indicated in
Table 6.3. If a failure to comply with contract conditions or successful additional cost claims
by Datapol against Astro which could not be passed on to Transcon were a significant risk,
the actual Datapol curve should have a tail moving to the right at the top before the cumula-
tive probability reaches 1.0. The possibility of Datapol going broke could imply a very long
tail for the underlying actual curve as it moved to right at the top, perhaps going off the
page. Figure 6.6 assumed a P90 nominal maximum for Datapol of £1.2 million in the same
spirit of ignored tails as earlier nominal values, although in this case the P100 equal to the
P90 was a very extreme special case.
Ben indicated that Sysdoc’s P10 nominal minimum of £0.5 million was derived from
Sam’s base value and discussion in a straightforward way. Although the chance of a lower
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
246 Employing planning tools in practice
cost than this P10 = £0.5 million might be negligible in practice, this did not matter. The
obvious significant difficulty with Sysdoc was a basis for defining a complementary P90
nominal maximum. Sam had not made any suggestions, and the unknown nature of this
nominal maximum cost was an understandable basis for Sam’s suggested choice of Datapol.
However, Martha had taken the view that although Sam’s preference was reasonable, and
it was probably typical of current Astro practice, it was not best practice. It was actually bad
practice which highlighted several issues which she wanted everyone involved to under-
stand, and he would explain them now.
First, the unknown nature of such a maximum cost should not induce automatic prefer-
ence for a known cost alternative like Datapol. Whenever there was significant uncertainty
about a nominal maximum cost, different ‘scenarios’ might be worth considering, without
any immediate consideration of associated probabilities, to explore a plausible range of pos-
sible values. For example, to choose between Datapol and Sysdoc by considering a plausi-
ble range for possible P90 Sysdoc values, Figure 6.6 portrayed three illustrative alternative
nominal maximum value scenarios which provided a useful basis for discussion:
Sysdoc 1, with a P90 maximum of £1.0m, 100% more than £0.5m (minimum × 2),
Sysdoc 2, with a P90 maximum of £1.5m, 200% more than £0.5m (minimum × 3),
Sysdoc 3, with a P90 maximum of £2.0m, 300% more than £0.5m (minimum × 4).
Putting aside ‘non- price advantages’ qualitative concerns for the moment, assuming that
both Astro and the customer perceive no ‘quality’ difference in the two approaches, a Sys-
doc choice in terms of each of these three scenarios could be compared with the Datapol
choice in risk efficiency terms.
If the Sysdoc 1 curve applied, then Sysdoc was a clearly preferred choice in terms of risk
efficiency and clarified dominance. Sysdoc cost was more variable, but it is always cheaper,
and the expected cost saving was £0.45 million (1.2 – (0.5 + 1.0)/2), using a second deci-
mal place at this stage to clarify comparisons.
0.5 1.0 1.5 2.0
0.2
0.4
0.6
0.8
1
direct cost in £m
cumulative
probability
0
key:
expected cost (also median)
Sysdoc 1
Sysdoc 2
Sysdoc 3
Datapol
Figure 6.6 Decision diagram: item 4 (convert existing programmes) example.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 247
If the Sysdoc 2 curve applied, the expected cost saving was £0.20 million. This was still
a significant saving, but Sysdoc was not the only risk efficient choice, because the Sysdoc 2
curve involved a risk of costs in the range £1.2 million to £1.5 million not associated with
Datapol. A trade- off was involved in terms of this particular decision, clarified by the Fig-
ure 6.6 portrayal.
At this point Ben made the point that Astro Wales had to make choices involving trade-
offs like that portrayed for the Sysdoc 2 versus Datapol in terms of minimising expected cost,
accepting the risk involved to do so, on a routine basis in many contexts. That is, all decisions should always be made on the basis of minimising expected cost unless unacceptable risks were involved, in order to aggressively minimise costs and maximise profits over time on average. Achieving risk efficiency in the Figure 6.5 sense was important, but so was taking acceptable
risk in the sense of using Figure 6.6 assuming Sysdoc 2 was the relevant curve. If Sysdoc 2 was
the appropriate curve, there was a clear case for Sysdoc. Put another way, the possibility of an additional cost up to that they might reasonably associate with the true curves underlying the
Datapol and Sysdoc 2 curves in Figure 6.6 was an acceptable risk in the context of an expected
saving of £0.20 million on this particular decision. In the context of a £15 million to £20 mil-
lion contract, this was not really risk at all – it was just variability which was ‘noise’. Hence,
Astro should choose Sysdoc if the Sysdoc 2 curve was believed to be appropriate.
If the Sysdoc 3 curve applied, then Datapol became the only risk efficient choice, with an
expected cost advantage of £0.05 million ((0.5 + 2.0)/2 – 1.2), and clarified dominance was
portrayed in Figure 6.6.
In summary, if Sysdoc 2 was a ‘best estimate’, or a curve in the Sysdoc 1 to Sysdoc 2 range
was thought to be appropriate, then Sysdoc should be the initial choice prior to considering
qualitative benefits. If Datapol were to be selected without reference to qualitative benefits,
then it would be important to verify that the Sysdoc 3 curve or a curve even further to the
right was a best estimate of the reality, with a flip point between the Sysdoc 2 and Sysdoc 3
curves when the expected Sysdoc and Datapol costs coincide.
Ben then indicated that Martha had suggested using the Sysdoc 2 curve as an illustrative
best estimate for illustrative purposes, with the Sysdoc 1 and 3 curves interpreted as ‘nominal
bounds’, in a second- order probability model sense which could be explored further later.
As an aside, this three scenario approach has some of the characteristics of a ‘parametric
analysis’: finding the value of an unknown parameter which ‘flips’ the decision, in this case, a
‘nominal maximum’ of £1.9 million or more. However, it places this flip value in the context
of a best estimate. Even if you are not interested in techniques like this form of parametric
analysis at a detailed how to do it level, you should have a what needs to be done level of
strategic clarity which includes understanding the value of analysts who can implement this
kind of technique whenever they become relevant.
You may find it useful to understand that this part of the analysis was dealing with a
higher- order level of uncertainty which matters. It could be dealt with by parametric analysis
or modelled directly, the choice depending upon a number of issues.
You should certainly be aware that more sophisticated curves for Sysdoc (tilted S shapes)
and a tail for the Datapol distribution would enhance the case for Sysdoc, and a skilled mod-
eller using an EP perspective should understand this in all comparable contexts. The simple
modelling used here is robust, and simple follow- on testing for robustness would probably
support the choice of Sysdoc if Sysdoc 2 looks appropriate using the simple model and the
underlying judgements are reasonable.
Trevor and Sian were both impressed by this part of the presentation. They were both
comfortable with the Sysdoc choice given the analysis while recognising that Datapol, as
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
248 Employing planning tools in practice
recommended by Sam, was typical of past Astro decision making. Everyone understood the
new bidding process should lead to taking more risk to reduce expected costs and increase
expected profits as demonstrated by choosing Sysdoc rather than Datapol. Sian did not feel
threatened by Sam’s Datapol recommendation being reversed or by the modification of
Sam’s other recommendations.
However, both Trevor and Sian wanted to talk about the Datapol approach in terms of
what Trevor called ‘non- price advantages’ and Sian referred to as ‘qualitative benefits for the
customer’. This was, in essence, the same concept from two different perspectives, although
there are nuanced distinctions in some cases, and value added or added value are further terms which Trevor and Sian both used for the same basic concept. Ben let them ‘have the
floor’ to do so. In brief, the Datapol new generation language rewrite approach would be
a much better basis for future changes, and as such from a Transcon perspective it would
provide value added relative to a Sysdoc approach.
Martha then made the point that all these arguments were sound, and they should be
documented as part of the PD3.2 in order to make them clear to the customer later, using
clearly explained and consistent common language. However, it would be a serious mistake
to add cost to the bid by making this kind of choice for the customer, behaving like a ‘nanny
contractor’. It was crucially important to bid on the basis of the lowest expected cost choices
which are compliant with the customer’s stated requirements and offer options at extra cost in cases like this. That is, if Sysdoc 2 is the accepted basis in terms of minimum expected cost, Transcon should be told that the Astro bid minimised the cost to Transcon by using Sys-
doc, but for the difference in expected costs, currently estimated at £0.2 million, Transcon
could have Datapol as a ‘costed option’, with a clear case for why Transcon should seriously
consider this option.
Martha then made the additional point that costed options like this should be individually
and collectively sold to customers as ‘bonus opportunity options’ – a set of opportunities for
the customer to get a better system than that required for compliance with their contractual
terms, in this case, a ‘bonus opportunity option’ associated with the deliverable provided
by Datapol being better suited to future changes than the Sysdoc deliverable. Part of the
rationale was the competitive advantage this could give them in both non- price and price
terms. But also important was looking professional, guiding the customer towards better
choices for a range of reasons without being critical of the customer’s specification or pushy.
Trevor pointedly indicated he would normally expect to do this, but Sian’s uncomfort-
able look suggested he was not always given the opportunity to do so. Martha picked up the
message clearly, and without making Sian more uncomfortable than necessary made sure
that Ben, Trevor, and Sian had understood and resolved to make good use of this insight –
the new Astro bidding process would improve communication on this kind of issue – and
better communication in this sense should be a built- in feature of the process, worth explicit
emphasise when other Astro staff were later introduced to the SP for bidding.
Ben then suggested they move on to the ‘possible additional memory’ issue.
Item 4 – the possible additional memory issue
Ben began by showing them Figure 6.7, which used a simple conventional textbook deci-
sion tree format to portray the uncertainties associated with the two choices available to
address this issue as initially specified by Sam (pre- install the additional memory or post-
install if necessary). Figure 6.7 also included one key chance node and associated possible
outcomes identified by Martha which Sam had overlooked (if the extra memory is needed
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 249
Astro might have to pay for it, but this might not be the case, and Martha was assuming
an 80% chance Transcon would have to pay for it). The version of Figure 6.7 which Ben
was using at this point was without the text in brackets on the ‘pre- install’ or ‘post- install if
necessary’ options. Ben did not want to show this text until he used it to discuss the basis of
an option for Transcon which Martha had suggested, but you might like to see it now, and
using just one version of Figure 6.7 is simpler for our purposes.
The top branch from the square decision node associated with choosing the ‘pre- install’
option shows the £0.5 million ‘cost to Astro’ outcome assumed by Sam. No chance nodes
were associated with this branch, so the expected Astro cost outcome shown on this option
branch was obviously £0.5 million without any need to formally compute it. When prepar-
ing Figure 6.7, Ben had assumed that Astro would pay this £0.5 million if the pre- install option was chosen, but Martha’s bonus memory option for Transcon to be discussed shortly
assumes a £0.4 million contribution from Transcon if Transcon decides to take the bonus memory option, an opportunity for Transcon and Astro not revealed at this point, with formal modelling of associated higher- order uncertainty not even considered for reasons
which will be clarified shortly.
Two chance nodes were associated with the ‘post- install if necessary’ branch. The first
addressed the probability extra memory would be needed or not. The second addressed the
probability Astro would have to pay or not if additional memory was required. Ben indi-
cated that Sam had estimated the cost to Astro if it paid at £1.0 million, used as shown. Sam
had also estimated the 0.5 probability extra memory would be needed, as also shown. Sam
had not addressed the possibility Astro might not have to pay for a post- installed memory,
but Martha’s reading of Sam’s documentation in the PD3.2 suggested a probability that
Astro would have to pay of the order of 0.2, not the 1.0 implicitly assumed by Sam. Sam’s
0.5
Key:
decision node choices available indicated above ‘choice branches’ expected values indicated below ‘choice branches’
chance node alternative outcomes indicated above ‘chance branches’ probabilities indicated below ‘chance branches’
cost to Astro
£0.5m
£1.0m
£0m
£0m
pre-install
expected Astro cost = 1.0 × £0.5m = £0.5m (expected Transcon cost = 0.0 × £0.5m = £0.0m)
post-install if necessary
expected Astro cost = 0.5 × 0.2 × £1.0m = £0.1m (expected Transcon cost = 0.5 × 0.8 × £1.0m = £0.4m)
extra memory needed
0.5
Astro pay
0.2
Transcon pay
0.8
no extra memory needed
Figure 6.7 Decision tree: possible additional memory example.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
250 Employing planning tools in practice
oversight of this issue might reveal a preference for pre- installing. Ben emphasised that
Martha had no wish to embarrass Sam or Sian about this issue – but Martha’s belief was
Sam’s oversight here involved missing a key first order probability issue which was typical of
current Astro common practice, a result of a lack of uncertainty management competence.
One consequence was a tendency to favour ‘safe options’, which Martha was determined to
change immediately, starting today.
This probability estimate of 0.2 was clearly linked to the choice of option for item 4 (con-
vert existing programmes) and less strongly linked to the option choice for item 5 (distrib-
uted scheduling system). It should be seen as conditional on the choice associated with these
items, yet to be confirmed. However, assuming for the time being that the probabilities of
0.5 and 0.2 were both ‘correct’ in the sense that they were unbiased expected values, there
was a 0.1 probability (0.5 × 0.2) that ‘post- install if necessary’ would cost Astro £1 million,
but a 0.9 probability (1 – 0.1) of no additional memory cost for Astro. The Astro expected
cost, a best estimate of what it should cost Astro on average, was £0.1 million. This was
the provision Martha and Ben suggested they add to the cost estimate when ‘post- install
if necessary’ was the option selected. This was £0.4 million less than the expected cost of
£0.5 million if the ‘pre- install’ option was selected. Martha and Ben recommended the
‘post- install if necessary’ option on the grounds that the expected cost difference between
the options was likely to favour this choice, and currently the best estimate of the cost dif-
ference was £0.4 million. There were some clear qualitative benefits to Astro associated with
pre- installation, which should not be overlooked, like making it easier for those developing
the new systems associated with items 4 and 5. But at present Martha took the view that the
sum of all such benefits could not be assessed in terms of a value exceeding the £0.4 million
difference in expected costs. There were also clear and important implications for Transcon
which this analysis had ignored thus far.
Sian immediately indicated that she was not comfortable with this approach. Sian’s chal-
lenge was that Ben and Martha’s approach put only £0.1 million into the budget as a pro-
vision for an event which might cost £0 but might cost £1 million – and would never cost £0.1 million. Surely, she argued, the ‘long- term average’ cost might be £0.1 million if this
was done time after time, but this was a ‘one- off ’ project. Putting £0.5 million in the budget
for the ‘pre- install’ option was a prudent (safe) policy for this project, while a provision of
only £0.1 million with a one in ten chance of a £1 million ‘hit’ was a high risk gamble. Surely
playing it safe was sensible in a situation like this.
Martha responded by saying that if extra memory was needed which Astro had to pay for,
the Transcon project would have to take a £1 million ‘hit’ against a budget with a provision
of only £0.1 million in it for this eventuality. But this should average out with other reasons
for cost variations on this and other projects. Her policy, approved by Rhys and as of that day Astro Wales policy in any comparable bidding context, was ALWAYS avoid increasing expected cost to reduce direct cost risk when making option choices like this. Astro could afford to take the risk some projects would involve a direct cost outcome which was often more than expected,
occasionally a lot more than estimated. It was not opportunity efficient to increase expected
cost to avoid risk Astro could live with. She was determined that Astro would ALWAYS avoid increasing Astro expected direct cost by reducing risk which did not need reducing.
Reducing risk which did not need reducing by increasing the expected direct cost involved
what she would call ‘unenlightened caution’. Taking risk which reduced expected direct
cost and could be afforded involved what she would call ‘enlightened gambles’. The post-
install if necessary choice, putting a £0.1 million provision in the cost estimate assuming a
one in ten chance £1 million will be needed, was a good example of an enlightened gamble.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 251
As of that day enlightened gambles were mandatory, and unenlightened gambles involving unenlightened caution would not be tolerated. The pre- install choice was a good example of unenlightened caution because it involved an expected cost of £0.4 million more than post-
install if necessary choice. A comparable issue had been central to their earlier discussions
involving Figures 6.5 and 6.6.
Martha elaborated briefly, explaining why the need for change arose. She explained that
for cultural reasons many Astro staff, not just Sian and Sam, currently took what they saw
as ‘low risk’ decisions, like pre- installing the extra memory in this case, choosing Datapol
in the first part of item 4 providing another example. These low risk choices increased
expected costs. Such decisions were low risk from their current personal perspective because
they anticipated hassle if they took the ‘high risk’ alternative of not adding extra memory
or choosing a subcontractor with unknown costs, and they implicitly assumed that Astro
preferred they avoided such decisions because of the hassle. But this low risk approach
lowered the chances of Astro winning bids, and it lowered the expected margin if they did
win. In terms of winning bids and making a profit from an overall Astro marketing perspec-
tive, decisions like adding extra memory were high risk choices, not low risk choices, in the
sense that they led to a high risk of losing bids and less profit on average when bids were
won. This lowered average profitability, increasing the risk that Astro as a whole would fail.
Across all bids such decisions increased the chances of Astro ‘going broke’ (being ‘broken
up’ in practice) because Astro did not have enough business, and what it had was not profit-
able enough, now a very serious threat. There was ‘no such thing as a no- risk business’, just
as there was ‘no such thing as a free lunch’. Astro could balance the risk of losing money
on some contracts and the risk of going out of business as currently structured because it
did not get enough contracts or make enough on those it got. But Astro was no longer in
a position to avoid taking significant risk on individual contracts. Technology and market
changes had transformed Astro’s competitive position, and they all needed to adapt or face
the implications of Astro going out of business.
Martha then made it clear that she saw the need for significant changes in operating policy
linked to significant culture changes. One component key issue was taking more risk on each
bid to maximise expected return for Astro Wales over all contracts over time. She did not
want to labour this complex set of issues now. But she was grateful Sian had triggered the
need to flag them – she would organise a separate session later to make sure everyone was
comfortable with all her new policy directions, including the way enlightened governance
would ensure that staff who took enlightened gambles which proved unlucky did not have
problems as a result.
Sian was still not entirely convinced, but she realised Martha had anticipated this chal-
lenge. She didn’t fully understand Martha’s position but felt it was not sensible to prolong
the discussion at this stage. You may also have reservations about Martha’s position, sym-
pathising with Sian and Sam. What is involved is important as is its contentious nature,
and there was a case for Sian’s position if Astro could not be trusted to reward people for taking enlightened gambles whether or not they were lucky. It is crucial for you to under-
stand Martha’s thinking on this issue, but prolonging discussion at this point would not be
helpful.
Ben then eased Sian’s discomfort slightly by indicating that the textbook decision tree
format of Figure 6.7 did make discrete events (like a need to post- install additional memory
at Astro’s expense) seem risky, but the cumulative distribution decision diagram format of
Figures 6.5 and 6.6 could be used for discrete as well as continuous variable outcomes, as
indicated in Figure 6.8.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
252 Employing planning tools in practice
Ben suggested that in the format of Figure 6.8 the pre- install or post- install choice was
clearly analogous to the Sysdoc 2 versus Datapol choice in Figure 6.6, and it might seem
less threatening in this format. The expected value of £0.1m plotted on the ‘post- install
if necessary’ curve of Figure 6.8 was on the horizontal part of the curve because £0m or
£1m discrete values were assumed to be the only feasible outcomes and point estimates of
probabilities were used (P10 = P90 in minimum-clarity terms, implying both equal the
P0 and P100 values). This makes Figure 6.8 seem a bit odd. However, Figure 3.8 and the
associated discussion in Chapter 3 in an IBM UK and BP context makes the implications
of the assumed situation very clear, and it highlights the rather special nature of either- or
outcomes involving both assumed certain outcome values and exactly known probabilities
as conventionally used with decision trees.
These conventional textbook decision tree assumptions could be relaxed, and sometimes
should be relaxed to aid understanding. For example, if the 0.9 probability Astro would not
have to pay defined by 1 – 0.5 × 0.2 was assumed to be an expected value with an appropriate
asymmetric and skewed probability distribution P0 to P100 uncertainty range from 0.5 to 1.0,
and the £1 million post- installation cost was assumed to be an expected value with another
appropriate asymmetric and skewed probability distribution P0 to P100 uncertainty range
from £0.8 million to £5 million, then the ‘post- install if necessary curve’ will leave the y- axis
at a probability of 0.5 and approach a probability of 1.0 at a direct cost value of £5 million. If
Astro used a Monte Carlo simulation package to accommodate these asymmetric and skewed
smooth curves for both component distributions, the resulting curve would be smoother.
Some decision analysis textbooks refer to using simulation in this way as ‘risk analysis’, but in
an EP framework it is more usefully seen as simply uncertainty analysis which goes beyond the
basic textbook first order uncertainty modelling of Figures 6.7 and 6.8.
Item 4 – the ethical issue and a red face test amber light example
Martha then intervened, suggesting that Sian and Trevor might both have somewhat dif-
ferent reservations about assuming a 0.8 probability Transcon would have to pay if extra
indicates expected cost
pre-install
post-install if necessary
0.5
0.2
0.4
0.6
0.8
1
0 1.0
cumulative
probability
direct cost in £m
Figure 6.8 Decision diagram: possible additional memory example.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 253
memory was needed without telling Transcon – was this ethical? In her view it was a red face test amber light issue – a source of concern on an ethical issue which caused her to pause
and think about the implications. The ITT stated that current documentation was ‘believed
to be reasonable – 95% accurate’. This implied that the customer’s consultants who pre-
pared the ITT were well aware that one or more errors and omissions making Transcon
responsible for extra memory could be expected, and Transcon should be prepared for most
tenders to assume this, using some variant of Figure 6.7 implicitly if not explicitly. How-
ever, in her view there was an important potential ethical problem if the customer did not
understand this. On reflection Martha thought there was a very good business policy case
for being honest and helpful, with due concern for the prospective customer’s best interests
as well as protecting Astro’s commercial position, assuming the customer might or might
not understand the issues involved. As an illustrative example of how they might proceed,
she suggested a three step plan for addressing a ‘bonus opportunity option’. The bonus
opportunity in question here was ‘bonus additional memory’, and she would explore further
bonus options shortly.
The first step was exploring what proposition they might reveal to the customer if they got the job after they had developed a strategic implementation plan with the customer but before they made a final commitment to memory size assuming they would prepare the ground and perhaps avoid the issue arising with a suitable bonus memory option at the bid
stage.
The second step was exploring how they would condition the customer for this proposi-
tion by offering a ‘bonus additional memory option’ as an add- on to the basic bid.
The third step was developing a strategy linking all related bonus opportunity options
plus later post- bidding issues. She would outline her current thinking on steps 1 and 2 now
and deal with step 3 later.
If they won the bid, then in the process of planning the job with the customer, before they were committed to the memory size the bid was based on but after they had under-
taken a review of the current position and developed implementation plans at a strategic
level, they should raise the issue directly and explicitly using an updated and generalised
variant of Figure 6.7 showing expected cost for Transcon as well as for Astro – at this stage
assuming no mark- up charges or potential legal costs involved. She then showed them a ver-
sion of Figure 6.7 with the text in brackets included and suggested they consider it assuming
for the moment that the Figure 6.7 probabilities were still their best estimates. Assuming
they showed it to their customer and outlined its implications in terms of the basic bid
assuming the customer had not taken their offered bonus additional memory option, their
proposition now could be as follows.
Astro would re- offer the bigger memory for pre- instalment as a bonus additional memory
option at a special cost of £0.4 million (the £0.5 million it would cost Astro less a £0.1 mil-
lion discount with no mark- up, implying an Astro cost of £0.1 million). This would involve
the same expected cost for both options for both parties, but it would eliminate the risk for
both, it would eliminate the need for a possible argument, and it would give Transcon a 0.5
probability of significantly exceeding their minimum spare memory capacity requirement.
From the perspective of Transcon and Astro as a collaborating partnership, it was a risk
efficient choice which was best for both, with benefits and costs shared in a manner which
was arguably fair to both.
To prepare the ground for this revelation, at the bid stage they would offer this same
bonus additional memory option, described as ‘pre- installation of additional memory for
£0.4m, a very deep discount price (expressed as a discount of about 50% on the normal price
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
254 Employing planning tools in practice
defined by £0.5m plus overhead and profit), because it might be useful for Astro to meet compliance guarantees as well as for Transcon’. They would emphasise this extra memory
would probably not be needed by Astro in the short run, probably significantly exceed-
ing Transcon’s requirements, but it might be needed in the short run by Astro, and in the
long run it probably would prove useful to Transcon, an added- value feature of the Astro
approach. They need not reveal their assumed probabilities or any other working assump-
tions, but they could if pushed admit a mix of ‘looking after their customer’s best interests’
plus ‘enlightened self- interest’. If they did so they should make it clear that other bidders
could not be assumed to take a similar line of thinking – many other bidders would not have
the integrity Astro prided itself on, an important non- price advantage of Astro which ought
to be discreetly but firmly pushed in advance. Indeed, if they wanted to emphasise this issue,
they might be explicit and completely open in the bid. The associated commercial risk was
competitors copying their lead.
To put it a bit differently, Astro could do a bit of planning ahead, to allow them to offer
a draft version of their anticipated proposition as an option, without revealing their prob-
ability estimates or upsetting or confusing a customer who did not understand the situation
if that was the preferred route. They might revise their option if it was not taken to reflect
new information. If their option was taken the risk associated with possibly needing more
memory was eliminated for them and for Transcon and a useful early step towards a part-
nership approach to contracting taken, perhaps finding a subtle way of selling the benefits
of this approach to their potential customers in advance, with or without being completely
open. Numbers which were not as conveniently simple as this example would obviously
complicate matters, but the basic ideas would generalise in reasonably clear ways.
Both Sian and Trevor looked very happy at this stage, so Martha indicated that Ben
should now carry on with his presentation, moving on to item 5.
Item 5 – the distributed scheduling system
Ben suggested they began addressing item 5 as assessed on Table 6.3 by looking at the line
dealing with estimates for option (b) Zoro software. The 20% to 40% uplift from preliminary
quotes to the agreed prices for Zoro noted in the item 3 information implied the £1.2 mil-
lion to £1.4 million range associated with a £1.0 million base value for Zoro, computed as in
earlier examples. Then looking at the line dealing with option (a) Astro software, the Astro
potential variability was unclear from Sam’s note, but a base cost of £1.10 million and a 20%
contingency for a total of £1.32 million suggested a £1.2 million to £1.4 million range and
an expected value of £1.3 million, directly comparable to option (b) to the nearest £0.1 mil-
lion. The line dealing with hardware cost took an approach directly comparable to that used
for item 1, with £3.0 million for all estimates. As shown in the fourth line, the implication
of these three preceding lines was the same expected cost and associated uncertainty ranges
for both options. Diagrams like Figure 6.5 and 6.6 could be used, but they would probably
not suggest a significant difference in terms of the quantified costs, and it would not be clar-
ity efficient to attempt to discriminate between them for present purposes. Ben suggested
Sam’s preference for the all- Astro option seemed reasonable given the Zoro hostile takeover
risk, and the presumed desirability of keeping the work in- house given the recent spate of
lost contracts. However, he indicated Martha had three serious problems with the all- Astro
option choice which she wanted him to explain.
First, Martha had discovered the hostile takeover bid for Zoro emanated from one
of Astro’s key competitors. If it went ahead, Zoro would probably become an effective
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 255
instrument of the competition, no longer a considerable asset available to Astro. Astro’s
position would become seriously restricted in relation to work like that on item 5, competi-
tion on cost would become very difficult for this kind of work, and work like that associated
with item 2 could become a further problem. This implied the need for a friendly takeover
as discussed earlier, plus the need to add an opportunity cost to the all- Astro approach which
should cause them to make a definitive switch to option (b) to reassure Zoro that their new
strategic partnership with Astro would work in Zoro’s interests.
Second, in this case they had another opportunity cost issue which was the flip side of
the earlier item 2 adjustment. Some of Sian’s staff who would have to execute the all- Astro
option were currently very busy, with very important major contracts from Astro Inc to
develop software for global markets. Delay on this work would be extremely serious. This
implied a further opportunity cost associated with the all- Astro approach which should
cause them to switch to option (b). Sian’s preference for a £19 million bid and £13 million
or £14 million cost estimate in the context of PD2.5 no doubt implicitly reflected this con-
cern, and it was very important to provide Sian with a simple tool for explicitly feeding this kind of concern into bidding decisions to avoid ineffective choices from an overall corporate
perspective. Sian and Sam might have used a red or amber light driven by a frowns test to
flag the issue earlier, and in the future, that is what Martha wanted Systems to do. It would
help communication and teamwork if Trevor understood this concern directly and explicitly,
Trevor and Sian jointly adjusting the bid on an agreed- on basis to reflect it without Sian
indirectly biasing cost estimates with unidentified opportunity costs.
Third, the application of a 30% overhead cost uplift on an efficient, well- managed subcon-
tractor like Zoro and the same 30% overhead charge on an in- house development using staff
already stretched with important work was biased against the use of efficient subcontractors
and towards overloading internal staff already fully engaged with profitable work. Astro
overheads on their own staff time were actually higher than the 30% nominally assumed,
perhaps closer to 50%, and in this particular case, a much bigger opportunity cost rate was
relevant. Astro overheads on a contractor like Zoro were actually lower than 30%, perhaps
only 10%. If this kind of bias forced ineffective choices it was a major strategic problem for Astro. Even if it didn’t force ineffective choices, it would adversely affect the level of
bid price chosen, itself a major problem. This implied that they should adjust the costs to reflect actual overhead cost differences whenever doing so mattered, and being aware of this issue ALWAYS mattered.
In the goal programming framework outlined in Part 1, seeing the protection of the time
of internal Systems staff involved in other very valuable contracts was a very important issue
for Martha. So was the protection of valued subcontractors. Both implied objectives requir-
ing opportunity costs driven by underlying shadow costs as an essential aspect of achieving
corporate optimality. Normal overhead rates of 30% were too crude to address this aspect of
the relevant big picture. Martha did not need to understand goal programming to see this,
but she did need an intuitive grasp of opportunity costs of this kind, and you may find it
helpful in the context of this book as a whole to take a goal programming perspective. Doing
so involves an intermediate level link between the very simple and conventional opportunity
cost issues addressed in Chapter 5 and the much more complex and controversial use of the
same basic foundation level goal programming framework to come in Chapters 7, 8 and 9.
Martha intervened in Ben’s presentation again at this point. She outlined her under-
standing of these opportunity cost issues and then indicated that she did not know what
these opportunity cost effects together amounted to in terms of a total opportunity cost
adjustment to option (a). However, £1 million seemed to her a very modest and robust
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
256 Employing planning tools in practice
‘best current guess’, and the exact number did not matter. In summary, her concerns
were protecting already busy Systems staff, ensuring trust was maintained with Zoro, and
unequivocally overriding the 30% overhead mark- up norm as a clear precedent for future
opportunity cost assessments and associated departures from standard mark- ups, the basis
of the Table 6.3 adjustment line. This meant that the £1 million adjustment was usefully
interpreted as a nominal opportunity cost adjustment associated with mandating a switch to
option (b) to protect critical Systems staff and Astro’s relationship with Zoro without wor-
rying about the exact amount, clearly understanding that there was no actual change to the
item cost for bid purposes or for monitoring estimated and actual outturn costs later if they
won the contract. In future it would be helpful for everyone involved if this kind of frowns
test opportunity cost adjustment issue and the Zoro friendly takeover opportunity (turning
a frown into a smile) were flagged earlier, as red light and blue light issues in stage 1.
Martha emphasised that from her perspective the need for a friendly takeover of Zoro
was a key strategic insight which came from her involvement in the details of the Transcon
bid. She explained that she wanted to make this clear to Trevor and Sian as part of a general
effort to ensure that they were comfortable with her ongoing involvement in the PD3, and
would not see it as just ‘looking over their shoulder all the time’, especially as she had earlier
suggested they ought to take responsibility for the details of specific stage 3 analysis choices
plus leading on process development.
At this point Martha also made it clear that all the PD3s would be assessed by her on a
case- by- case basis initially, but just a sample audit basis later, providing Trevor and Sian kept
her fully informed of crucial new insights. The reason was the learning processes involved
for her, as well as wishing to make ongoing contributions to the evolution of the SP for bid-
ding. Even the PD1s and PD2s involved learning processes, like identifying strong synergies
between a contract and Astro’s objectives in the PD1. However, the insights provided by the
PD3 process were much deeper and more general, and there was a much stronger and wider
need to share the learning involved. She thought this issue would be crucial and ongoing,
but particularly intense early on. It could be viewed as gaining strategic clarity, building on
key aspects of tactical clarity.
Ben then picked up his presentation again by noting it was now very clear that the item
5 Zoro choice was preferred, because of the opportunity cost adjustment to protect busy
Systems staff and the trust central to Astro’s relationship with Zoro following on from the
friendly takeover. However, had the two expected cost estimates been identical and these
other concerns not been involved, as a process issue it was very important everybody under-
stood that Astro did not usually need to commit to decisions where there was no direct cost
benefit either way at the bid stage. All that mattered for bidding purposes in such cases was
an unbiased estimate of the expected cost. That is, had the takeover and appropriate average
overhead and other opportunity cost issues not been present, it might have been useful to
keep their options open, assuming an expected direct cost of £4.3 million without making a
commitment to option (a) or (b).
Item 5 – the additional performance issue
Ben then suggested they move on to the decision tree of Figure 6.9.
The format of Figure 6.9 was much like that of Figure 6.7, used to consider the possible
additional memory issue. However, three differences were noteworthy. First, in this case
Martha had suggested that if additional performance was needed the probability Astro would
have to pay for it was 0.4, twice as likely. Second, he had not included the expected cost
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 257
0.5
cost to Astro
£1.0m
£2.0m
£0m
£0m
pre-install
expected Astro cost = 1.0 × £1m = £1.0m
post-install if necessary
expected Astro cost = 0.5 × 0.4 × £2.0m = £0.4m
extra performance needed
0.5
Astro pay
0.4
Transcon pay
0.6
no extra performance needed
£1.5m
£2.5m
key:
decision node choices available indicated above ‘choice branches’ expected values indicated below ‘choice branches’
chance node alternative outcomes indicated above ‘chance branches’ probabilities indicated below ‘chance branches’
spread of outcomes – assume a midpoint expected-value minimum-clarity estimate
Figure 6.9 Decision tree: possible additional performance example.
estimate for Transcon or an ‘additional performance bonus’ calculation to keep their analysis
simple in terms of this complication. And third, Sam had suggested a spread of costs, from
£1.5 million to £2.5 million, which he had treated as a nominal range between P10 nominal
minimum and P90 nominal maximum values as earlier, with a midpoint expected value of
£2.0 million. The latter would slightly complicate the additional performance equivalent
of Figure 6.8, suggesting the sort of Monte Carlo simulation approach he had hinted at
earlier might be useful, but the only complication for Figure 6.9 was replacing the nominal
range of £1.5 million to £2.5 million with its £2.0 million expected value prior to comput-
ing the expected value of the post- install if necessary option. Table 6.3 showed that Martha
preferred the post- install if necessary option because of the £0.6 million (£1.0 million –
£0.4 million) advantage relative to ‘pre- install’ given the assumed probabilities, although
Sam had indicated a preference for ‘pre- install’.
The nominal maximum of £2.5 million shown on Table 6.3 was not a simple P90 in the
same sense as the other Table 6.3 nominal maximums, but the difference was not worth
explaining immediately unless it was raised by Sian or Trevor. A similar comment applied to
the nominal maximum entry for additional memory.
Martha then interjected to pointed out that the probability Astro would have to pay given
extra performance was needed was higher (0.4) than in the additional memory case (0.2),
by a factor of two. She wanted them to understand why. In the additional memory case, the
invitation to tender statement about current documentation ‘believed to be reasonable – 95%
accurate’ was ‘an invitation to make claims’. Accepting this invitation could be based wholly
on insisting that the customer pay for the additional memory in full if it became clear it would
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
258 Employing planning tools in practice
be needed post- installation because of inaccuracies in customer documentation. However, she
suggested that the 0.4 probability that Astro had to pay if additional performance was required
would be higher still if Astro took on the whole of item 5, because Astro Wales could not pass
the cost of extra performance on to its own staff, but Astro Wales could probably pass at least
some of it on to Zoro. Even if Astro UK owned Zoro, this structural difference mattered.
This comment, coupled to Martha’s rationale for the £1 million adjustment just dis-
cussed, caused Sian to twitch noticeably. However, she did not yet fully understand the
oblique hint, which you might like to understand now. The key is Martha could pass some
of the additional performance risk on to Zoro without damaging the Astro– Zoro relation-
ship. Because all Zoro staff had incomes geared to Zoro profit levels, all Zoro staff were strongly motivated to perform in an effective profit- driven and loss sharing way which Astro had never approached for Sales staff or even contemplated for Systems staff. In the recent
past Astro Sales staff had incomes highly geared to sales commissions, but the trend had
been to reduce the gearing for many years, and Sales staff did not suffer immediate, directly
geared financial penalties if the contracts they won lost money. The Astro reward structures
for performance were reasonably clear in terms of promotions for fast- track staff like Mar-
tha, but comparatively woolly and ineffective for most other staff. An inability to deal with
this was in part what was driving the ‘redefining core business’ approach to outsourcing cur-
rently sweeping through Astro Inc. This insight was part of the key strategic clarity spin- offs
of Martha’s involvement in the details of the Transcon bid, and a key part of the motivation
for the friendly takeover of Zoro which Martha proposed. Had some of Sian’s staff not been
very productively occupied with other kinds of work, they might have found their transfer
to Zoro part of the takeover deal. Given the positive opportunity cost in the Transcon con-
text it would have been inappropriate to even raise this possibility in the Transcon contract
context, but it might need addressing soon.
Sian was aware that Martha was sending an oblique message, despite her inability to
decipher this message. This took away her appetite for discussing the particular probabilities
associated with Figures or 6.7 and 6.9. She would revisit them later.
A set of bonus performance options as a new feature of the Astro bidding strategy
Martha then indicated that a ‘bonus performance option’ offer as part of their basic bid
should be developed, as discussed in the possible additional memory context. Furthermore,
having done so they would then need to consider the third step of her three step plan for
dealing with all bonus opportunity options – developing a strategy linking all related bonus
options at a pre- bidding stage, and thinking about related post- bidding issues. She explained
that what she had in mind was making customers ‘strategic partners’ to the maximum extent
that this was feasible. They should pursue this ‘strategic partnership’ goal at all stages in the
overall process of bidding and then executing the work, using whatever approaches seemed
most appropriate in the particular circumstances.
Pre- bidding strategy development linking all bonus options identified earlier might be
explained to the potential customer in terms of a further overall discount of say £0.5 mil-
lion if they took the full set of three bonus opportunity options: a Datapol approach plus
extra memory plus extra performance. This would provide assurance for both parties that
their interests would be protected in a collaborative manner, sharing the benefits, and laying
sound foundations for an ongoing relationship built on trust.
Briefly exploring the post- bid benefits and the robustness of the simple pre- bid models
used so far, Martha began by reminding everyone that to be compliant with the Transcon
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 259
invitation to tender, a ‘fixed price’ bid was required. However, there was a significant chance
Transcon would have to pay the successful contractor’s claims for additional costs in rela-
tion to additional memory and performance provisions, whether or not Transcon could
see this coming and whether or not they selected Astro. In addition to the red face test,
her uncle Paul had suggested that ‘the moment the contracts come out of the drawers,
everyone was in trouble’, an adage which she thought worth remembering at this point.
She strongly believed that as a matter of good business practice it was important to try to
avoid all ‘contracts out of the drawer’ situations and to try to avoid arguments about fine
print clauses and fault. More generally, Astro had to comply with contractual terms speci-
fied by potential customers, but sometimes it might be in a customer’s best interests to vary
these terms as well as in a contractor’s best interests, and if this was the case an enlightened
contractor would try to use this possibility as a competitive advantage so long as it did not
weaken their bargaining position. More generally still, building a strong partnership- driven
customer– contractor relationship had significant short- term benefits as well as crucial long-
term benefits in some cases, with knock- on repeat business and reputational implications.
Developing and maintaining trust was central.
When it came to negotiating post- contract, Astro and Transcon would obviously have
different views about probabilities, claims could in practice arise and include very large
contributions to profit, claims could be structured to cover Astro failures indirectly, and
legal costs could be crucial. ‘Claims engineering’ was a highly developed skill set for some
contractors, and Astro needed some of these skills. But a very simple trade- off deal like ‘you
pay for the extra performance and we will pay the extra for Datapol’ might be the anticipated
outcome of a complex negotiation, and avoiding the need for a situation requiring complex
negotiation can be a crucial opportunity to focus on as part of the bidding process – a non-
price advantage for the customer as well as the contractor, usefully marketed in these terms.
As soon as Martha was sure Sian was reasonably comfortable with this approach she sug-
gested that Ben start to bring their PD3.2 discussion to a close.
Drawing the PD3.2 process discussion to a close
Ben now drew attention to the highlighted total direct cost of £11.8 million in the final por-
tion of Table 6.3, rounded to £12 million to reflect the ‘nominal precision’ of this estimate as
a whole. He indicated that it was important to use a nominal precision level like working to the
nearest £0.1 million earlier for individual items to clarify choices, and some of these choices
might benefit from slightly more precision, like the Sysdoc 1 to 3 choices. However, Astro
staff would be misleading themselves if they believed the Transcon PD3.2 of Table 6.3 pro-
vided an accuracy level greater than the nearest £1 million for expected total cost, and it was
particularly useful to formally acknowledge this now to simplify the example analysis to follow.
The associated nominal range of £10.4 million to £15.7 million for the anticipated actual
outcome was rounded to £10 million to £16 million to reflect a comparable level of nominal
precision for the overall project- level direct cost estimate.
The additional stage 3 pass 1 estimating effort had significantly refined the earlier infor-
mation basis. The PP3.2 expected overall cost was lower than the PP2.2 expected overall
cost (as anticipated by Trevor). The associated range was also reduced. However, both the expected overall cost and the range might have increased. A reduced range should be expected if the process is functioning properly, but the expected overall cost moving up or down with
equal likelihood suggests a process functioning properly.
Ben then indicated they could associate the £10 million to £16 million range with a 10– 90
percentile range consistent with the use of other ‘nominal ranges’ throughout the analysis.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
260 Employing planning tools in practice
He passed over the additional memory and performance minor complications. However, he
did point out that internal consistency implied perfect positive correlation when defining
the nominal range for total cost in terms of the nominal ranges for each of the five items, a
coefficient of correlation of 1.0 assuming well- behaved dependence. Well- behaved depend-
ence and less than perfect positive correlation implied that five 10– 90 percentile ranges
added in this way yield a percentile range like 5 to 95, with less chance of being outside the
plausible minimum and maximum P values defining the range. An assumption of independ-
ence (a coefficient of correlation of zero) was the only other simple assumption, this was
much further from the truth, and it was inappropriately wildly optimistic. Ben suggested
that in this context it was important to use a simple assumption which should be conserva- tive (pessimistic) most of the time to preserve some robustness for those occasions when
dependence was not well behaved and not understood, and even perfect positive correlation
proved seriously optimistic.
Despite the probable conservative nature of this correlation assumption, the nominal
range of £10 million to £16 million revealed no significant threat to Astro Wales from cost
variations as estimated, although an outcome of £16 million or more which could not be
explained might be a threat to the careers of those involved in the Transcon bid and the
execution of the project. Ben indicated the key reasons for assuming perfect positive cor-
relation were keeping it simple and ensuring that Astro did not ignore the uncertainty range
unless it could be tolerated, but another useful reason from Sian’s perspective was it made
it easier to explain departures from expectations. The last comment was to some extent
designed to cheer up Sian, although Ben was making what he knew to be very important
practical points. Sian managed a smile, although her unease remained. She resolved to see
Martha after, which Martha anticipated.
Ben then drew attention to the 30% overhead and 10% profit margin calculations which
finished off Table 6.3, indicating a highlighted nominal price of £17 million with a nominal
margin of £5 million. He made the point that this nominal price was a cost- based price as a guide to what Astro Inc ‘normally expected’ in terms of anticipated average mark- ups on
cost. It was not a market- based price, reflecting what the market would bear in terms of their competitors’ behaviour.
Martha now intervened again, thanked Ben for his presentation to this point, and indi-
cated that she hoped it was clear to Sian how her Systems input to the PD3.2 via the PPD
was transformed in an efficient and effective way. She indicated that she needed to further
pursue some important issues which the process highlighted, part of the purpose of the
process. But she suggested that they all had a good basis to proceed with discussing the rest
of pass 1 of the stage 3 Transcon bid process for that day’s purposes.
Martha then indicated that six things should come out of the PD3.2 process, and their
pursuit should be the clear and conscious goals of this phase:
1 An unbiased estimate of the expected overall direct cost, without being concerned
about choices associated with project execution at this stage except insofar as such
choices affected expected cost or contractual commitments or other concerns like their
Zoro relationship and the system’s commitment to Astro Inc work.
2 Contractual options offered to the customer when mutual benefit could provide
non- price advantages as a competitive strategy in appropriate cases, with a range of
approaches perhaps proving worth developing.
3 Insights about negative and positive adjustments to direct costs as normally assessed
which were critical to reflect lower marginal costs or opportunity costs associated with
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 261
other uses of resources, distinguishing between transfer costs which actually change
(like the Service Bureau costs) and nominal opportunity cost adjustments to protect
overstretched resources without changing actual costs for accounting purposes (as in
the item 4 distributed scheduling system case to protect critical Systems staff).
4 Insights about qualitative benefits of the options chosen which could be played up as
non- price advantages in the bid, like those associated with starting to train Transcon
staff earlier on a nearby Astro site, were also critical.
5 Insights about more general issues which go beyond the case in question, like the need
to mount a friendly takeover for Zoro, and the need to consider differential overhead
rates, were of strategic importance.
6 Comfort with the fairly wide nominal range for anticipated outcomes was important,
in this case, £10 million to £16 million, in particular, recognising that some contracts
are going to lose money, and this should not be a problem provided it is not a crippling
unexplainable surprise because of incompetence.
In summary, the pass 1 stage 3 cost uncertainty analysis of phase 3.2 was not just a matter
of refining estimates of the expected cost and anticipated range of actual outcomes. It was
about an enhanced understanding of all the issues which drive such cost estimates, including
associated non- price advantage and other qualitative benefit issues, along with other work-
ing assumptions which need explicit understanding. A core concern was using this under-
standing to add value to the proposition offered to their potential customer.
At this point Martha suggested they might all enjoy a tea break, to catch their breath
before moving on to Trevor’s role and phase 3.3. She said she appreciated this was going to
be a long day, with a lot of new ideas to absorb.
You may feel you need the equivalent of a tea break at this point, and if you have a limited
interest in tactical clarity concerns in a bidding context, you may be tempted to just skim-
read the next few sections. However, while the next few sections may look specialised and
technical, with no immediately obvious role in more general terms, they are important in
terms of providing a foundation for some of the key strategic clarity insights to follow.
Phase 3.3 – enhanced probability of winning analysis on pass 1
After a tea break, while Sian was still trying to take in the full implications of the process
underlying the PD3.2, Martha opened the session and turned her attention to Trevor’s role.
She moved on to what she had earlier designated the PD3.3, obtaining an enhanced prob-
ability of winning curve using the phase 3.3 process on pass 1.
Martha began by reminding Trevor and Sian that the preliminary probability of winning
curve of Figure 6.3 was new to Astro as an explicit tool, although Trevor and other sales staff
had obviously used such a curve intuitively. This was not unusual. Very few organisations
made direct use of an explicit Figure 6.3 equivalent. Astro was going to obtain significant
competitive advantage by using Figure 6.3 in phase 2.3. But Astro was going to obtain much more additional competitive advantage by using a refined version in phase 3.3. Ben and
Trevor had not yet addressed this issue, but she had worked with Ben to develop and then
illustrate the basis of what she wanted.
She reminded them that one of the key advantages of Figure 6.3 was that it enabled the
use of Table 6.2, which explicitly brought together costing and pricing issues in a manner
which clarified what was involved, including providing useful sensitivity analysis. For exam-
ple, it clearly indicated the bid which maximised expected margin, and it also indicated the
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
262 Employing planning tools in practice
sensitivity of expected margin and the probability of winning to alternative bids, to allow
explicit consideration of key trade- offs. Crucially, this made trade- offs very clear as the bid
was changed, between the expected margin, the actual (conditional) margin if they won,
and the probability of winning, letting them consider these trade- offs in relation to other
key trade- offs involving qualitative issues identified in stage 1. The more refined version
of Table 6.2 which pass 1 of phase 3.3 would enable was going to be a major competi-
tive weapon for Astro because it would extend this kind of insight to specific competitors
who mattered, including qualitative issues like non- price advantages specific to particular
competitors which might be built into their bid, encouraging and facilitating a process of
systematically seeking these advantages using their understanding of key competitors.
Among the advantages of Figure 6.3 or refined versions discussed earlier was the ability to
record the estimated P(B) for whatever bids were chosen over time to also record how many
bids were won over time and to compare these predictions and outcomes to provide feed-
back on the accuracy of the P(B) estimates. That is, only if P(B) values associated with bids
were explicitly recorded could data be collected which would help to improve the estimation
of P(B). Stage 3 would refine this capability in ways not considered earlier, to assess Astro’s
ability to estimate probabilities of beating the specific competitors who really mattered.
Martha stressed that as useful as Figure 6.3 was, it was very crude relative to the refined
estimate for the PD3.2 provided by Table 6.3. It should be clear that significant benefit was
attached to using the PD3.2 process just discussed relative to Sian’s direct cost estimate in
the PD2.2. A comparable increase in benefit from the PD3.3 process relative to the PD2.3
process was the goal.
One of the keys to the benefits provided by the PD3.2 was the decomposition of the con-
tract cost into five component items, and different option choices within some items. This
decomposition allowed a clearer view of what was involved, the use of different people with
different expertise to work on different items, and the use of past experience with similar
items and option choices to facilitate more effective use of data associated with past experi-
ence. The effective pursuit of non- price advantages was also triggered by this decomposi-
tion. A comparable decomposition was required for the third stage enhanced probability of
winning curve for different but broadly comparable reasons.
Martha then asked Ben to illustrate the basis of how they might approach the Transcon
case if they explicitly associated the Figure 6.3 preliminary probability of winning curve
with a ‘composite competitor’ – and asked the question, ‘How should we decompose this
composite competitor?’
Ben took the floor and began by explaining that the theoretical basis for the decompo-
sition Martha wanted had been published many years ago, and his critical review of the
extensive literature had included a classic 1950s’ paper (Friedman, 1956) which provided
a useful basic generic bidding model starting point. If there were n competitors, 1, . . . , n,
if they assumed that the lowest bid wins, and if they knew the probability each competitor
would bid at or below any given bid level, they could use well- known statistical procedures
to combine these probability distributions to derive a probability of winning (being below
all competitors) for any given bid level. Several important practical difficulties were stopping
most people using this basic theoretical model to build practical operational models for
widespread routine use. But he and Martha now had the basis of a practical clarity efficient
way of overcoming these difficulties, and with Trevor’s help they believed Astro had an
effective way forward.
The first difficulty he would comment on was in practice a significant number of com-
petitors might be involved, not all these competitors might be worth formal analysis on an
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 263
individual basis, and some might be unknown. To begin to deal with this issue he suggested
ordering the competitors 1, . . . , n by their importance, and assuming that only the first one
or two would receive individual attention most of the time, perhaps as many as half a dozen some of the time but probably never the complete set if n was larger than about half a dozen. Those not considered individually would need collective treatment as a residual ‘composite
competitor’ which explicitly embraced ‘all unknown competitors as well as all known com-
petitors not worth decomposition’ – a closure with completeness component.
The second difficulty was ‘non- price advantages’ – the extent to which Astro might com-
mand a premium relative to different competitors (possibly negative) because of Astro’s
‘quality’ (non- price advantages) relative to competitors as perceived by Transcon.
To address both of these difficulties, Ben suggested they limited their discussion that day
to working with clearly identified working assumptions which could be generalised later that
were currently very strong in order to keep their initial illustrative analysis simple:
1 Separate assessment was worthwhile for only one competitor – designated ‘competitor
1’. All other competitors could be treated as a closure with completeness residual not
worth decomposition – designated ‘competitor 2’.
2 Astro had a known constant £2m non- price advantage relative to competitor 1 and a
known constant £1m non- price advantage relative to competitor 2, and both of these
non- price advantages were constants in the sense that they were independent of what
competitors 1 or 2 bid.
3 Competitors 1 and 2 could both be assumed to understand that Astro was the market
leader, to be capable of using their expected cost estimate to estimate that Astro’s cost-
based nominal price had an expected value of about £17 million, and to plan to beat
Astro by about £1 million in adjusted price terms, understanding Astro’s non- price
advantage relative to their own position.
To interpret the implications of these assumptions, if for example competitor 1 bid £13 mil-
lion, then Astro would beat competitor 1 if Astro bid less than £15 million, and if com-
petitor 2 bid £14 million, then Astro would also beat competitor 2 if Astro bid less than
£15 million, with the same £2m million and £1 million non- price advantages applying what-
ever competitors 1 or 2 bid.
Ben indicated that they could use more sophisticated working assumptions in various ways,
including associating non- price advantages with probability distributions and making these
distributions dependent on what competitors bid. These were complications which Martha
and Ben correctly anticipated would prove crucial later, deliberately avoided at this stage.
To start to make use of his assumptions and further interpret the implications, Ben sug-
gested that Astro Sales staff might find it useful to consider direct estimation of the prob-
ability that competitor 1 would bid at a given level and the associated adjusted bid using
the three scenario tabular format of Table 6.4. This table could then be used to generate
Figure 6.10. He would explain the generation of Table 6.4 first, then the construction and
meaning of the density and reverse cumulative portrayals of Figure 6.10.
You might find it helpful to observe that Table 6.4 involves a three scenario version of the
HAT approach discussed in Chapter 3, with a corresponding rectangular density function
portrayal illustrated by the first part of Figure 6.10. The Table 6.4 reverse cumulative prob-
ability distribution provided by the right- hand column was the basis of the second part of
Figure 6.10, a curve defining the probability of Astro beating competitor 1 for the relevant
range of Astro bids. Ben’s explanation incorporated relevant Chapter 3 discussion.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
264 Employing planning tools in practice
Table 6.4 Tabular format for estimating the probability that competitor 1 will bid at three sce- nario values.
Actual bids (£m) Scenario labels Probability Adjusted bids (£m) Reverse Cumulative Probability
13 low bid 0.2 15 1.0 14 medium bid 0.6 16 0.8 15 high bid 0.2 17 0.2
Expected bid = £14 million, and adjusted expected bid = £ 16 million
probability density for
competitor 1 adjusted
bid
adjusted bid for competitor 1 in £m
area = 0.6
0
0.6
14.5 16.5 17.515.0 15.5 17.016.0
area = 0.2 area = 0.2
probability of beating
competitor 1 if Astro bid as
indicated
1.0
Astro bid in £m
0.2
0 0 14.5 16.5 17.515.0 17.015.5 16.0
0.8 piece-wise linear
reverse cumulative
format
rectangular
density format
Figure 6.10 Three scenario representation of Astro’s probability of beating competitor 1.
Ben suggested that preparation of Table 6.4 might start by estimating a ‘plausible mini-
mum’ (P10) ‘low bid’ of £13 million, next estimating a ‘plausible maximum’ (P90) ‘high
bid’ of £15 million, and then interpolating a single ‘medium bid’ of £14 million. These three
scenarios implied corresponding adjusted bids of £15 million, £17 million, and £16 million.
The £1 million constant interval basis of Table 6.4 could be interpreted to imply a range of
+/−£0.5m associated with each central value in the rectangular density function intervals.
This meant that a low bid scenario adjusted value of exactly £15 million interpreted as a
P10 was consistent with a probability of 0.2 for the range 14.5 to 15.5, a similar logic made
the high bid scenario probability 0.2 as well, with 1 – 0.2 – 0.2 = 0.6 being the probability
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 265
residual for the medium bid scenario. In practice more than three scenarios could be used
and asymmetry might be involved.
Ben emphasised that he had kept Table 6.4 and Figure 6.10 as simple as he could for
illustrative convenience, but the approach easily generalised; he would provide seminars and
coaching on the whole family of available associated approaches as required, and he would
develop simple computer software which would help them to produce Table 6.4 and use the
fourth and fifth columns to produce Figure 6.10.
Using the fifth column of Table 6.4 to produce the second part of Figure 6.10 involved
recognising that the adjusted bid by competitor 1 probability density function in a reverse
cumulative form defined the probability that Astro would win against competitor 1, clearly
the case in this simple example. With a non- price advantage of £2 million, if Astro bid less
than £14.5 million, the probability of Astro beating competitor 1 was one, because there
was no chance that competitor 1 would bid less than £12.5 million, an adjusted bid less than
£14.5 million. If Astro bid £15.5 million, there was a 0.8 probability of beating competitor
1, because there was a 0.2 probability competitor 1 would bid in the range £12.5 million to
£13.5 million with an adjusted bid equivalent of £14.5 million to £15.5 million, and so on.
This Table 6.4 approach could also be applied to competitor 2, a composite of all other
competitors, as illustrated by Table 6.5.
As indicated in Table 6.5, for illustrative purposes he had assumed competitor 2 would
involve a smaller non- price advantage associated with a broader distribution of actual bids
and probabilities. His rationale was based on its composite nature. Table 6.5 could be por-
trayed as a five scenario equivalent of Figure 6.10 if this was useful.
Ben had made the competitor 1 and 2 estimates of Tables 6.4 and 6.5 different in part to
make the illustration easier to follow, but the differences also made the example more cred-
ible. Both were simple and symmetric, with a common expected value of £16 million for
the adjusted bids, for convenience not credibility. This meant that he could provide a par-
ticularly simple but useful illustration and interpretation of the adjusted bid for composite
competitor 1 + 2, the composite derived when competitors 1 and 2 were formally combined
using the information in Tables 6.4 and 6.5. The symmetry and the same expected value for
both competitors’ adjusted bids implied that both competitors were equally likely to win.
He could extend the computer software handling the individual competitor Tables 6.4
and 6.5 to combine the associated two probability distributions using the HAT probability
tree discrete probability computation scheme of Table 6.6.
A comparable Monte Carlo simulation approach was an option. However, he had used
the Table 6.6 approach because this was the easiest framework for understanding how any
two distributions could be combined, allowing for later generalisation as needed.
Table 6.5 Tabular format for estimating the probability that competitor 2 will bid at five scenario values.
Actual bids (£m) Scenario labels Probability Adjusted bids (£m)
13 very low bid 0.1 14 14 low bid 0.2 15 15 medium bid 0.4 16 16 high bid 0.2 17 17 very high bid 0.1 18
Expected bid = £15m and adjusted expected bid = £16m
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
266 Employing planning tools in practice
Table 6.6 The probability of adjusted bids from competitor 1 + 2, a composite of competitors 1 and 2.
Competitor 1 Competitor 2 Competitor 1 + 2 adjusted bid probability contributions
Adjusted bid Probability Adjusted bid Probability 14 15 16 17
15 0.2 14 15 16 17 18
0.1 0.2 0.4 0.2 0.1
0.02 0.04 0.08 0.04 0.02
16 0.6 14 15 16 17 18
0.1 0.2 0.4 0.2 0.1
0.06 0.12
0.24 0.12 0.06
17 0.2 14 15 16 17 18
0.1 0.2 0.4 0.2 0.1
0.02 0.04
0.08 0.04 0.02
Composite competitor 1 + 2 probabilities (column sums) 0.10 0.34 0.50 0.06
Reverse cumulative probabilities for plotting Figure 6.11 at Astro bids of
1.0 13.5
0.90 14.5
0.56 15.5
0.06 16.5
0.00 17.5
A tabular form of probability tree was defined by the body of Table 6.6, with probability
contributions to common result values associated with different probability tree branch
end points collected on the right- hand side. Understanding the probability tree involved
was facilitated by this sort of tabular form, especially important when dependence was a
concern.
To clarify line 1 in the body of Table 6.6: a bid by competitor 1 of £13 million was associ-
ated with an adjusted bid of £15 million and a probability of 0.2 as indicated in Table 6.4, a
bid of £13 million by competitor 2 was associated with an adjusted bid of £14 million and
a probability of 0.1 as indicated in Table 6.5, and the Table 6.6 joint scenario with a prob-
ability of 0.2 × 0.1 = 0.02 was associated with the composite competitor 1 + 2 adjusted bid
of £14 million because competitor 2 would beat competitor 1.
All the other lines did the same, for each of the possible combinations of competitor 1 and
2 bids, giving 3 × 5 = 15 lines in all, associated with 15 probability tree branch end points.
If you are interested in the detail, the pattern of adjusted bid probabilities should make
sense if you work down row by row. You do not need to follow all the how to do it detail if you are not mathematically inclined or you are not interested in the arithmetic details, but
you do need to understand what needs to be done in terms of an overview appreciation of
what is involved. This comment will apply to all future HAT portrayal tables of this kind,
throughout the rest of this book, but it will not be repeated.
The competitor 1 + 2 column sums collected these probability calculations together, to
define a probability distribution for the adjusted bid for composite competitor 1 + 2.
These probability contribution products were then used in the form provided by the last
two lines to produce the basis for plotting a probability of winning curve for Astro (beating
competitor 1 + 2), interpreting discrete bid scenarios as £1 million ranges.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 267
For example, in the ‘composite competitor 1 + 2’ (column sums) row the probability of
0.06 of a bid of 17 (bold entry on Table 6.6 to assist identification) was associated with a 0
probability of Astro winning with a bid of 17.5, rising to 0.06 by 16.5. In conjunction with
a probability of 0.50 of a competitor adjusted bid of 16 (next to the bold entry), this implied
a chance of Astro winning with a bid of 15.5 was 0.56. Similarly, Astro’s chance of winning
with a bid of 14.5 was 0.9, and Astro was certain to win with a bid of 13.5.
As soon as Trevor and Sian were comfortable with this aspect of Table 6.6, Ben showed
them Figure 6.11.
The solid curve for composite competitor 1 + 2 in Figure 6.11 simply plotted the final row
values of Table 6.6, a rectangular probability histogram assumption being used in its equiva-
lent piecewise linear reverse cumulative probability distribution form to yield the probability
of Astro winning.
Figure 6.11 also used a dashed curve to display the comparable distribution for competi-
tor 1 on its own, based on Table 6.4 as portrayed in Figure 6.10, to indicate the contribu-
tion to composite competitor 1 + 2 provided by competitor 1. The gap between the two
curves indicated the impact of competitor 2. There was no need to plot competitor 2 as well
as 1 and 1 + 2, but competitor 2 could have been plotted instead of competitor 1 for a similar
gap. Both competitors made a comparable contribution to the 1 + 2 total in this particular
example to illustrate what equal contributions implied.
Adding a cumulative probability curve for one of the two components involved in this way
made Figure 6.11 a form of ‘sensitivity diagram’ which is directly comparable to Figure 4.1
as discussed in Chapter 4. The same HAT computational process framework underlies both,
albeit with a different mathematical operation. These features of a HAT approach, including
the probability arithmetic option used for the tables, the sensitivity diagram options, related
use of decision diagrams to judge risk efficiency, and the related framework to seek overall
opportunity efficiency, are central to using the EP toolset to understand the way compo-
nents of uncertainty from different sources accumulate and the implications for alternative
composite competitor competitor 1
1+2 curve curve
the space between the curves for competitors 1 and 1+2 indicates the impact of competitor 2
10 15 20
0.2
0.4
0.6
0.8
1
Astro bid in £m
probability of Astro beating
competitor 1+2
(winning overall) if Astro bid as
indicated
0
1+2 1
Figure 6.11 Sensitivity diagram format for a probability of winning curve for Astro using Table 6.6 plus Table 6.4 (Figure 6.10) for the competitor 1 curve.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
268 Employing planning tools in practice
planning options, as discussed in the somewhat different context of BP offshore projects in
Chapters 3 and 4. In this Astro bidding context a sensitivity diagram provides a useful basis
for building up an understanding of the impact of successive competitors, with the lowest
adjusted bid of a complete set of n competitors’ bids defining winning or losing for Astro.
You may find it useful to recall the Chapters 3 and 4 discussion of BP’s separately identified
and analysed sources of uncertainty associated with an activity like offshore pipe- laying in a
SCERT framework, instead of using a PERT model treating each activity as a single source
of uncertainty, which has directly comparable features. The way all these features of the HAT approach can be employed in a flexible toolset which adapts to very different contexts is an impor- tant overall characteristic of the EP toolset, also illustrated and exploited in Chapters 8 and 9.
Ben explained these separately decomposed components plus a ‘closure with complete-
ness’ component with minimal reference to the ideas illustrated by the earlier BP illustra-
tions, but he did emphasise understanding the way the build- up of uncertainty occurred was
of crucial importance in any context. In this bidding context he explained that the curve
for any composite competitor 1 + 2 would always be to the left of the curve for one of its
components, regardless of the ordering. But three different variants need distinction and
anticipation in practice:
1 If competitors 1 and 2 were a fairly even match, approaching the equality assumed by
Tables 6.4 and 6.5, a shift like that illustrated in Figure 6.11 would be observed.
2 If competitor 1 completely dominated competitor 2, the competitor 1 and competitor
1 + 2 curves would coincide, approaching this position as partial dominance increases.
3 If competitor 1 was dominated by competitor 2, a more significant shift to the left
would be observed, and plotting the contribution of competitor 2 instead of 1 might
be more instructive to show the difference made by the less important of the two com-
petitors, perhaps also formally reordering competitors to amend the earlier assumption
about which mattered most and clarify the assumption required revision.
Ben made the point that Table 6.6 might look a bit complicated, but they needed to under-
stand its structure to follow a later discussion of dependence concerns, and he could write
simple computer software to produce Figure 6.11 from the basic inputs. Sales staff, Martha,
and everyone else involved needed to be confident that they understood what was going
on, including how dependence issues could be managed. However, they did not have to
‘crunch the numbers’.
Even if you have no interest in the mathematics of the modelling issues, you do need to
understand what can be done in dependence terms building on Table 6.6 and Figure 6.11
plus associated what needs to be done concerns to follow where this book is going. In many
different contexts our ability to combine the estimates for two probability distributions and
understand graphically how important the contribution of each is to the overall result is
important. So is our ability to deal with more than two components in a generalised version
of the same approach. Particularly crucial in the current context is tactical clarity about how
to generalise this basic framework to cope with non- price advantages which are not known
constant parameters as assumed for the examples discussed so far, because this kind of tacti-
cal clarity in a flexible and powerful framework is central to strategic clarity. Also, particularly
crucial in the current context is our ability to decompose to whatever level is clarity efficient
on each occasion, using the ‘closure with completeness’ concept with a suitably robust
understanding of the implications. Astro could usefully see the stage 2 approach of their
new bidding process as a particularly simple special case of their stage 3 approach when a
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 269
single composite competitor was addressed – no individually decomposed competitors. In
a stage 3 context it would be crucial to develop the ability to make the decomposition level
judgement skilfully as experience is gained, and it would be important to err on the side
of too much decomposition initially to develop and test their understanding of how much
decomposition provided extra insight that more than paid for the effort involved.
These were all issues that Trevor needed to understand as well as Ben and Martha. The
reason you need to understand them at an overview level, whatever your perspective and
management level and areas of interest, is comparable concerns may need to be addressed
whenever multiple sources of uncertainty and their relationships need clarity in the possibly
very different contexts of interest to you, a central strategic clarity concern.
Ben now suggested that they could relax all their earlier working assumptions, but the only
one he wanted to explore that day was treating more than one competitor separately, and
he would do so very briefly at an overview outline level. He would use a high clarity n = 6
illustrative example, involving five separately identified competitors plus a sixth closure with
completeness composite competitor. Composite competitor 6 was, by definition, a residual of
all those competitors not worth further decomposition. They could use an even bigger n, but
in practice this was very unlikely to pay, even for early very high clarity learning process studies.
He and Martha believed that usually one or two separate competitors would suffice after they
had gone down the requisite learning curve and become used to the approach.
He then showed them Figure 6.12.
Ben’s version of Figure 6.12 had named companies beside competitors 1 to 5, and com-
posite competitor n = 6 involved an incomplete list of possible competitors with illustrative
examples. The competitor 6 information explicitly recognised that some competitors might
be unknown, and which of the known competitors would bid might be uncertain.
Ben explained that a pairwise approach to this grouping like that portrayed by Figure 6.12
was not essential, but it could be useful for reasons associated with ‘common features or
characteristics’ of members of each pair plus dependence assumptions which they would
explore later if and when they became relevant. One reason worth mentioning immediately
was a pairwise build- up using a Figure 6.11 format sensitivity diagram was facilitated if a
competitors 1 to n
composite competitors n + 1 to 2n – 1
1
2
3
4
5
6
11
7
8
9
10
UK major in Wales
UK major in Wales
UK major not in Wales
UK major not in Wales
French major not in Wales
composite competitor n – all other competitors
three majors outside Wales
two UK majors in Wales
all major competitors all
competitorstwo UK majors not in Wales
Figure 6.12 Competitor grouping structure example, n = 6.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
270 Employing planning tools in practice
structure like Figure 6.12 was used. One diagram could be used for competitors 1 plus 2
yielding composite competitor 7, followed by another for 3 + 4 = 8, followed by a third for
8 + 5 = 9, followed by a fourth for 7 + 9 = 10, followed by a fifth for 10 + 6 = 11. Separate
sensitivity diagrams for each pair could be summarised using a single diagram (like Fig-
ure 4.1), and a simpler sequential pairing would facilitate a direct route to a single diagram
generalisation of Figure 6.11 (like Figure 4.1).
Using any pairwise structuring meant that n competitors required n − 1 pairs, 6 − 1 = 5
pairs in this example. The pairing structure of this particular example assumed that competi-
tors 1 and 2 were the most important at least in part because they were both UK majors
with offices located in Wales, it was convenient to think about 3 and 4 as the next strongest
pair with common features, and it was convenient for the trio 3 to 5 being paired with 1
and 2 before pairing 1 to 5 with the residual composite competitor n = 6. Any conveni-
ent structure could be used, but there were underlying separability working assumptions
involved which he would explain later (a purely intuitive grasp of what pairwise composi-
tion approach involves will suffice at present for all target readers, and Ben did not want to
explore these issues at this stage). Ben did suggest that if any variant of this kind of pairwise
nested build- up structure was used, a recommended approach to keeping the notation as
simple as possible was always dealing with competitor i and competitor j to form the com-
posite competitor k, with k = n + 1 to 2n − 1 for n − 1 pairs. For any variant of Figure 6.12,
he would ensure the computer software equivalent of Table 6.6 would allow the definition
of Figure 6.11 equivalents for each pair of competitors so that it would be very clear which
competitors mattered most within each pair.
At this point Ben showed them a version of Figure 4.1 to explain how BP used sensitivity
diagrams as explored in Chapter 4, and how they might look for similar insights in terms of
which competitors mattered most, and then direct more time and effort to exploring what
seemed to matter most as more insight was gained.
Ben indicated that if the Figure 6.12 pairing structure was appropriate, they could start
by estimating bid probabilities for competitors 1 and 2, their two closest and most obvious
competitors, both international organisations with bases in Wales who could be expected
to bid. This would define the probability of winning curves for i = 1 and j = 2, and allow a
composite probability of winning curve associated with both competitors to be calculated
and associated with k = 7, the composite of i = 1 and j = 2.
Similarly, bid probability distributions and probability of winning curves could be estimated
for other groups of competitors. The pair designated i = 3 and j = 4 to define k = 8 could
address two UK majors with no base in Wales, and j = 5 could be a French major with no base
in Wales, all three being known competitors who were expected to bid. Then a probability
of winning curve for k = 9 could compose known major competitors with no base in Wales,
and a probability of winning curves for k = 10 would compose all known major competitors
who were expected to bid. Ben indicated that composite competitor n = 6 would be a residual
concept to include all unknown competitors, and k = 11 could then define a probability of
winning curve for all competitors, a pass 1 stage 3 equivalent to Figure 6.3.
Ben said Martha wanted Trevor and other sales staff, with help from Ben as needed,
to provide adjusted bid distributions for competitors 1 to 6 in whatever structure Trevor
thought was most appropriate, initially using about five separately identified competitors. Ben would then provide Figure 6.11 format probability of winning curves showing how
the components build up to the overall composite competitor k = 11 in whatever way best
explained the relationships between the individual competitors and the residual composite
competitor, with an overall summary in the Figure 4.1 format.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 271
The simplest version of this approach assumed that the probability distributions for
adjusted bid values for competitors i and j are independently distributed in terms of prob-
ability distribution shapes which can be estimated approximately by a rectangular histogram
probability density function and combined using a discrete probability tree approximation.
Rectangular histogram forms could approximate any distribution to whatever level of preci-
sion was required so that was not a restrictive assumption, simply a useful working assump-
tion. Discrete probability arithmetic was approximate when continuous variables were
involved, but the error was modest and it could be controlled or eliminated. Alternatively,
standard Monte Carlo simulation software could be used to achieve the same ends. Statisti-
cal independence in this case seemed a reasonable assumption if collusion between competi-
tors was not suspected, but statistical dependence could be modelled if appropriate to reflect
possible collusion or other sources of dependence.
The key assumptions which this approach made included the ability of Astro Sales staff to
estimate all the required input distributions, and the levels of decomposition used for each
bid was clarity efficient. They could employ any variant of Figure 6.12 they were comfort-
able with. Once they had used about five separately assessed competitors a few times, they
would probably find that just two or three separately identified key competitors was more
than enough most of the time, and just one separately identified key competitor might even-
tually be enough much of the time.
Martha interjected here on cue. She suggested that sets of versions of Figure 6.11 with
a summary like BP’s Figure 4.1 based on variants of Figure 6.12 had to be estimated for as
many separately identified competitors as experience with about five suggested was worth-
while. This was because she was determined to have a refined version of Table 6.2 to refine
the bid price and the treatment of non- price advantages with a clear understanding of what
a clarity efficient level of decomposition of their competitors looked like.
If Astro sales staff were uncomfortable estimating the non- price advantages and competi-
tor bid distributions probabilities this required, they were just going to have to get over it.
In her view, an inability to estimate the equivalent of Figure 6.11 for each pair of the overall
composite of all competitors using an appropriately decomposed competitor structure was
tantamount to not understanding the marketplace in a way which was directly useful to her
and Astro more generally.
However, she could assure them that what she was asking for would not be as difficult
to deliver as it might seem, and successive use of figures of the form of Figure 6.11 or ver-
sions with more than two components (like Figure 4.1) would help them to understand
the implications of their assumptions. Furthermore, this kind of decomposition would yield
benefits which were broadly analogous to those observed from the five item breakdown of
costs. For example, Astro could start to acquire data on winning bids by individual major
competitors relative to Astro’s nominal price, to develop a database relevant to their prob-
ability distribution estimates. As this data grew, Astro could start to look for formal analysis
explanations for changes in these distributions over time – systematic variations due to the
state of the market, the state of the order book of particular competitors, and so on. These
models would provide a framework to test hypotheses about the implications of strategy
changes by competitors as well as helping to identify changes.
The curve for the final composite of all competitors, k = 11 in Figure 6.12, would be
the bottom line, what Sales staff had to produce before any bid was considered. If they
started with n = 6 they might soon find that n = 2 or 3 or 4 was actually adequate, but
a plausible case for aggregate treatment of competitors who did not matter much would
be needed.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
272 Employing planning tools in practice
How they produced it would be up to the sales team leaders like Trevor, with extensive
support from Ben and Martha’s help whenever she could be useful. How they proceeded
would be audited initially by Martha on a case- by- case basis, to see if the approach adopted
made sense at all levels, with constructive feedback.
Obviously, there was no guarantee that any Figure 6.11 format curve or n greater than
two equivalents (like Figure 4.1) would be correct. However, Martha could apply two tests.
First, did all the underlying assumptions seem to make sense? Second, as data accumulated,
did it confirm the order of magnitude of the estimates used at the final ‘all competitors’ level
and at the level of individual competitors they understood?
Martha stressed that it would be some time before significant data were available, but the
first of these tests would be used with immediate effect, with some intensity, until she was
convinced that everyone had a joint understanding of what they were doing. The technical
details of this approach could clearly take a while to get used to, but Martha did not want
to spend more time on technical details now. Ben would arrange for special seminars to take
everyone through the basics, and workshops to develop the details, which she would support
whenever this was useful. She wanted to finish off the current meeting with a formal presen-
tation by Ben on the enhanced bid setting summary table and discussion based on that table.
Martha knew that Trevor would be struggling to visualise how to deal with her require-
ments at this stage, and if you are not struggling to see were this chapter is going at this
point you are unusually perceptive. However, when we begin to discuss the way Trevor
started to respond to his challenges a few sections on, you should begin to see the point of
the framework established in this section, and its wider relevance.
To finish on an encouraging note, Martha reemphasised the message that once they got
used to using about five individual competitors employing a structure like Figure 6.12, they
should find competitor 1 defined as ‘the key competitor’ plus a composite competitor 2
embodying the residual was good enough much of the time, and Ben would assume that
this was the case for the rest of their discussion that day.
Phase 3.4 – enhanced uncertainty syntheses on pass 1
Ben began his discussion of the first pass through the phase 3.4 process to produce an
enhanced margin evaluation table by indicating that these enhanced tables could have the
same basic structure as Table 6.2, but the illustrative version he would use that day was
slightly complicated because it used both Figure 6.3 and Figure 6.11 information – two P(B)
curve information sets to facilitate useful comparisons. He then showed them Table 6.7.
Ben started to explain and explore the implications of Table 6.7 by noting that he wanted
to use this table to clarify how crucial an unbiased and enlightened estimate of the P(B)
relationship was. That was why Table 6.7 made use of P(B) curves from both Figure 6.3 and Figure 6.11, showing the different results for these two alternative P(B) curve information
sets. He would focus on the differences, assuming Figure 6.11 was ‘correct’ for illustrative
purposes because these differences illustrated key points he wanted to make. In practice they
could use a simplified variant of Table 6.7 with only one set of P(B) values.
The Figure 6.3 curve through c1 had been used to extend the bottom end of the range to
match that of the Figure 6.11 curve range. He appreciated that this was not necessarily what
Trevor had in mind. Its choice here was purely illustrative.
Furthermore, the Figure 6.11 estimate as a whole was not intended as a suggested revi-
sion to Trevor’s Figure 6.3 estimate – its role in Table 6.7 was purely illustrative to show the
extent to which differences of the scale illustrated by these two P(B) curves matter.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 273
Table 6.7 Enhanced margin evaluation table (the PD3.4 produced in stage 3 phase 4) with two P(B) curves.
Expected direct cost E(C) = £12m working to the nearest £m with an outcome in the range of £10m– £16m anticipated
Bid values B (£m)
P(B) Figure 6.3*
P(B) Figure 6.11
Margins (conditional) M = B – E(C)
Expected margins E(M) = P(B) × M
Comments
Figure 6.3 Figure 6.11
13 0.97 1.00 1 0.97 1.00 New minimum B 14 0.94 0.95 2 1.88 1.90 15 0.80 0.72 3 2.40 2.16 New maximum E(M) 16 0.66 0.30 4 2.64 1.20 Figure 6.3 probability
curve maximum E(M) 17 0.52 0.03 5 2.60 0.15 18 0.38 0 6 2.28 0 No chance of winning
at a bid level of £18m or more assuming the Figure 6.11 probability of winning curve is correct
19 0.24 0 7 1.68 0 20 0.10 0 8 0.80 0
* using the dotted line through point c1 in Figure 6.3 for illustrative purposes.
As an initial comparison, for a bid B of £15 million as indicated in the first column of
Table 6.7, the second column indicated a P(B) of 0.8 taken from Figure 6.3, while the third
column indicated a P(B) = 0.72 taken from Figure 6.11, a difference of about 10%, which
was within the range of acceptable variations they might expect between stage 2 and stage 3
estimates. There was no cause for concern here.
As the bid B was reduced from £15 million to £14 million and then £13 million, the
probability of winning increased at a faster rate using the Figure 6.11 column than even the
dotted line through point c1 in Figure 6.3, but the two curves crossed and then remained
close. That is, over the bid B range from £13 million to £15 million, the two curves roughly
coincided if the extension through point c1 of Figure 6.3 was assumed. There was no cause
for concern here either, in terms of different P(B) curves implying significantly different
implications.
But the difference started to show and its implications became significant as B values
increased above £15 million. When the bid B increased to £16m, the Figure 6.3 P(B)
dropped to 0.66 while the Figure 6.11 P(B) dropped to 0.30, a difference of more than 50%
using the larger number for percentage comparisons, highly significant. When B increased
to £17 million, the Figure 6.3 P(B) dropped to 0.52 while the Figure 6.11 P(B) dropped to
0.03, a difference of more than 90% using the larger number for percentage comparisons, a
huge difference. At a B of £18 million or more the P(B) for Figure 6.11 is zero, massively
different from the Figure 6.3 situation although not meaningfully associated with a percent-
age difference.
The E(M) column of Table 6.7 associated with the Figure 6.3 curve suggested a maxi-
mum expected margin of £2.64 million, with a bid of £16 million, and a 0.66 probability of
winning. Ben used bold type for these P(B) and E(M) values to highlight the implications of
a B = £16 million decision assuming Figure 6.3 was the basis for this decision. This looked
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
274 Employing planning tools in practice
quite attractive. It was an even more attractive scenario than the Table 3.2 E(M) maximising
bid of £17 million with a probability of winning of 0.52 and a maximum expected margin
of £2.08 million because of the lower E(C).
However, if the Figure 6.11 columns of Table 6.7 was the reality, the chance of winning
with this bid of £16 million was only 0.30, and the expected margin was only £1.20 million,
about half the P(B) and E(M) values assumed on Table 6.2.
Furthermore, if Figure 6.11 was the reality, a maximum margin was achieved with a bid
of £15 million, and a lower bid like £14 million was suggested because a probability of win-
ning rapidly approaching 1.0 can be obtained with very little loss of expected margin in the
£14 million to £15 million range.
A key message Ben wanted to make using Table 6.7 was it clearly mattered that they knew which probability of winning curve they were dealing with when differences of this order were involved. More generally, being capable of estimating the P(B) curve without bias and with a clarity efficient level of uncertainty about its location really mattered.
Figures 6.3 and 6.11 might be roughly the same in the bid B = £13 million to £15 mil-
lion range, but the differences outside this region were of crucial importance. Astro need to
confirm (or revise) and refine Trevor’s probability of winning estimates.
To get started, Ben recommended working with best estimates associated with close com-
petitors in a structure comparable to Figure 6.12 to define a single best estimate P(B) set
of Figure 6.11 equivalents for use in a simplified version of Table 6.7. They might always
estimate the bounds first and then interpolate the best estimate P(B) curve to be used as a
way of minimising bias and making themselves aware of their uncertainty about the location
of a best estimate curve.
As a form of sensitivity analysis while they were in the early stages of acquiring experience
with this new framework, they might find it useful to explore the implications of nominal
optimistic and pessimistic bounds for each best estimate P(B) curve by showing the best
estimate plus its bounds on a three P(B) curve variant of Table 6.7. However, he envisaged
this kind of sensitivity analysis as a possible learning tool for the team preparing the estimates
in the early stages, probably not necessary once they got used to the approach.
Phase 3.5 – bid evaluation and development on pass 1
Ben finished his final presentation of the workshop by indicating that the sort of bid
‘pricing’ level and underlying cost estimation discussion that Sian and Trevor started in
the context of Table 6.2 and the PD2.5 would clearly be much better informed by a vari-
ant of Table 6.7 using one best estimate probability of winning curve they trusted as an
unbiased estimate. The discussion would not just be working with more precise numeri-
cal estimates; it would be working with a quite different level of understanding of all the
uncertainties involved.
He reminded Sian and Trevor that Martha believed stage 1 and 2 of their bidding process
should jointly target the use of less than 10% of the effort required for the full bid prepara-
tion process (about 2% followed by about 5%). They should initially anticipate the elimina-
tion of about 50% of the potential bids in the first two stages (about 30% followed by 30% of
the remaining 70%, midpoints in the 10%– 50% ranges). This would leave a residual of about
50% of the bids for stage 3 consideration. Martha wanted them to try to do better than these
crude expectations, with more aggressive stretch targets to maximise the time available for
more effective analysis of ‘the right contracts to win’. The aim was to spend about 80% of
their time on the 20% of the bids that really mattered – a variant of the traditional 20:80
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 275
rule. However, Martha’s initial estimates were just intuitively based initial guidelines to be
refined over time as they gained experience.
Another variant of this rule would apply to their approach to successive passes through
the stage 3 process of Figure 6.4. Ben indicated that Martha thought pass 1 through stage
3 should target using 10% to 20% of the effort required for the full bid preparation process
once they became familiar with the process. However, initially they would be in learning mode and might take 10 to 20 times what they could expect after six months or a year of experi-
ence. Occasionally the first pass through stage 3 would prove enough to clarify a ‘minimum
effort’ bid, or a late no bid decision.
Most of the time, an optimal bid preparation process would involve further effort build-
ing on the insights provided by pass 1, in subsequent passes and parallel bid management
processes which exploited the insight gained earlier, to make bids that Astro wanted to win
more attractive to the potential customers. It would not just be a matter of refining earlier
estimates; it would be a matter of addressing different kinds of uncertainty and complexity
in different ways as appropriate. Even the pass 1 form of the PD3.5 should address these
kinds of issues as a basis for possible further stage 3 passes and other forms of analysis.
Sometimes a very late no bid decision would be an unplanned and unwelcomed outcome
of later stage 3 passes – but it was important to keep an open mind.
Martha then interjected to remind Trevor and Sian that Ben would support the PD3 pro-
cesses and make the provision of necessary information as easy as possible for Systems staff,
Sales staff, and any other Astro staff involved. They did not have to understand the technical
computational issues in detail, or the separability structure nuanced implications, just the
basic principles. She knew that Trevor and all other Sales staff would find the new approach
to explicit use of P(B) curves particularly challenging to get to grips with in the early stages,
but they would find it very rewarding as well, and she was convinced that Trevor had cru-
cially important expertise which he could share with other Sales staff, her included, via the
new process, playing a leading role in the shaping of their stage 3 approach. She also knew
that Sian and all other Systems staff would find knock- on implications of the new bidding
process challenging, but she hoped she could help Sian deal with the underlying changes
impacting them both which had triggered the need for a new approach to bidding.
Martha closed their workshop by emphasising that she was determined to make the Sales
contribution to Astro Wales a success, and she would support Sian’s wider role as well as
Sian’s relationship with Sales in any way she could. She would now let them to recover from
the information overload of this very long and intense workshop.
Trevor’s transformative insights
Martha needed time to reflect on how the workshop had gone, update and extend her
plans for further development of the new bidding process, brief Rhys on her plans and get
his feedback, and begin some conversations with Sian. However, she attended to these
tasks in parallel with a top priority series of meetings with Trevor and Ben, beginning the
next day.
Martha started the first of these meetings by indicating to Trevor that she thought the very
basic known constant non- price advantage bid adjustment assumption which Ben had used
to produce Figure 6.11 was a useful simplification for their workshop discussion. However,
it lacked sophistication relative to the feel for non- price advantage issues which she believed
Trevor had and Astro needed to capture. She knew the framework of the approach Ben
had provided would easily accommodate the more insightful working assumption structure
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
276 Employing planning tools in practice
which she thought Trevor could help them to develop. She then outlined what she had in
mind based on extensive prior discussions with Ben.
Trevor quickly responded to this opportunity, with a view to making his mark on the new
bidding models and processes. He ‘thought out loud’ with a whiteboard and marker pen
for several hours while Martha prompted him and Ben made notes and asked clarification
questions. Trevor, Ben and Martha then started to put together an initial presentation for
Trevor to make to his Transcon bid sales team. This was used to trigger further ideas, which
eventually evolved into what Martha referred to as ‘Trevor’s plan’.
Before considering the detail necessary for strategic clarity about Trevor’s plan at this
stage in its development, you may find it useful to consider a brief outline of the transforma-
tive insights which created the basis for and then shaped Trevor’s plan, ordered to provide
a simple initial overview.
First, Trevor and his Transcon bid sales team needed to start by focusing considerable
effort on the competitor which mattered the most – ‘competitor 1’ using Ben’s labels. Tre-
vor identified competitor 1 as Tauris. As Trevor would put it later, to compete effectively
the Astro bid team ‘needed to get to know the opposition’, with a focus on what really mat-
tered, initially in the context of their relationship with the competitor which mattered most.
Second, they needed to formally acknowledge that Tauris was ‘the market leader’, not
Astro. This was a mindset change for Trevor and for Astro more generally. As Trevor would
put it later, ‘getting to know Tauris was a significant reality check result for him, and Tauris
was by no means the only competitor they were behind in ways needing explicit and focused
attention’.
Third, they needed to assume that Tauris used an approach to bidding which never led to
a no bid decision, but always varied the effort expended in proportion to how much Tauris
wanted to win any particular bid and how tough Tauris thought the competition was going
to be. That is, Tauris would always bid for sure, but with significant and crucially important
variability in the level of effort expended on their bidding effort.
Fourth, Astro adopting the working assumption that Tauris would never no bid but could
produce a ‘no bid equivalent’ involving a very low effort but very high price bid made the
Astro analysis framework much simpler. The primary purpose of this working assumption
was making Astro analysis simpler, but it was a reasonable working assumption for a sophis-
ticated competitor like Tauris which Astro could learn from for direct use.
Fifth, a no bid equivalent approach by Astro should be built into Astro’s SP for bidding
as a ‘diplomatic option’, for possible use whenever Astro wanted to maintain a market pres-
ence or not offend an organisation which they wanted to maintain good relationships with
despite not wanting this particular contract.
Sixth, in stage 3 a no bid equivalent approach by Astro should be viewed as the extreme
end of a spectrum of ‘provisions for cost ambiguity uncertainty for a possible “nice to have”
contracts which they would not mind getting at a high enough mark- up but did not believe
were worth further analysis effort’. In effect, stage 3 iteration decisions have to consider a
number of trade- offs, including a judgement about given other possible uses for the time
of those involved in the bidding process, is it worth spending more time trying to better
understand how to increase the expected margin and reduce the risk associated with not
winning this particular potential contract, or would it be better to just add a bigger provi-
sion for what they still do not know and stop investing any more time on this bid now,
proceeding with a fairly high bid and not being too worried if they lose it.
Seventh, they should assume that the systematic relationship between how much Tauris
wanted to win and the price Tauris would bid could not be used to predict the Tauris bid
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 277
at this stage by Astro. An attempt to penetrate the ambiguity associated with predicting
what Tauris would bid might be contemplated later, if a particularly important contract
Astro really wanted to win was involved. They might base this on approaches employed
by oil majors when bidding for oilfield exploration sites which they really wanted. It was
known that some oil majors sometimes assessed likely bids from other oil companies, but
Trevor, Martha, and Ben were agreed that this level of sophistication was not a current Astro
concern.
Eighth, the systematic relationship between the price which Tauris decided to bid for
the Transcon contract and the non- price disadvantage which Astro would suffer could be
assessed by Astro fairly simply using crucial working assumptions which were central to his
transformative insights. Given Tauris made a plausible minimum bid, a plausible working assumption was that Tauris really wanted to win this contract and had put a maximum level of effort into the bid; given Tauris made a plausible maximum bid, a plausible working assump- tion was Tauris were not too concerned about winning this contract and had put a minimum level of effort into the bid; given Tauris make a mid- range bid, a mid- range scenario in terms of Tauris effort could be inferred. Astro’s immediate priority was developing clarity about what this view of the relationship between Tauris bids and associated effort implied in terms of
Astro’s non- price disadvantage, and how this understanding could be used by Astro.
After Trevor had acquired strategic clarity about the what needs to be done implications
of these eight components of his transformative insights, he became comfortable with tacti-
cal planning for their next steps, briefing his Transcon sales team to bring them up to date
with progress so far, and beginning to seek input from his Transcon team for developing the
new bidding process for Transcon.
The next section addresses the resulting first meeting with the Transcon sales team. But
before moving on you might find it useful to understand that we might perceive Trevor’s
transformative insights as roughly comparable in terms of process and some aspects of the effects to Nicola’s pass 3 plus Bob’s plan 4 evolution in Chapter 5. The key difference is Trevor’s transformative insights did not lead to putting aside the equivalent to Nicola’s
formal EOQ model – they led to tactical clarity which enhanced Ben’s basic models within
the bidding framework envisaged by Martha all along, making the overall evolving bidding
process much easier to use as well as much richer and much more robust. Full understanding
of the implications is central to the enhancement of your strategic clarity which this chapter
provides.
An alternative analogy is the intended evolution of your thinking in Chapter 3 when
moving from William’s minimum clarity approach to unbiased estimation towards Peter’s
eventual understanding that unbiased estimates were just one useful spin- off from an oppor-
tunity efficient approach to all relevant aspects of contingency planning.
Trevor takes up the leadership role which Martha planned
Trevor began his first presentation to the Transcon bid sales team by outlining all their pro-
gress to date, using all the tables and diagrams discussed earlier in this chapter to ensure that
his sales team understood the tools developed thus far, and the lessons learned. He finished
his summary by confirming the importance of using a structure like Figure 6.12 to order
competitors in terms of importance, with competitor i = 1 being ‘the key competitor’ by
definition, competitor 2 the next most important, and so on.
Trevor then suggested they spend some time ‘getting to know competitor i = 1’, and they
start exploring their relationship with this key competitor by explicitly asking the question
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
278 Employing planning tools in practice
‘Was the key competitor the market leader?’ If the answer was ‘no’, then Astro was the mar-
ket leader by definition. For the Transcon bid he was now assuming that Tauris was the key
competitor, and Tauris was the market leader.
The core issue they needed to explore immediately was generalising Ben’s non- price
advantage adjustment approach. Ben’s Table 6.4 for competitor i = 1 assumed a known
constant £2 million adjustment between all actual and adjusted bids, with competitor j = 2
involving a known constant £1 million adjustment on Table 6.5. This working assumption
was far too restrictive and unrealistic. It oversimplified in a counterproductive way, to such
an extent that until he, Martha, and Ben had agreed on an effective practical way to accom-
modate a more sophisticated approach to non- competitive disadvantage assessment which
was both probabilistic and conditional upon the bid made by Tauris, he had not been con-
vinced there was any feasible way to get beyond his very basic Figure 6.3 direct estimation
of a preliminary probability of winning curve.
Trevor’s generalised variant of Ben’s approach involved associating each competitor
with formal modelling of systemic uncertainty about non- price advantage or disadvan-
tage using what Trevor (with Ben’s prompting) had labelled ‘conditional probabilistic
bid adjustment tables’. These tables would capture crucial uncertainty and associated
dependence issues which Trevor had understood intuitively for some time. He was now
convinced that he plus all his Sales team needed to develop significant additional clar-
ity about these issues via formal analysis. Crucially, these ‘conditional probabilistic bid
adjustment tables’ would allow Trevor to develop more clarity and then explain his
understanding to others so that they could all develop and exploit comparable under-
standing. Trevor believed that all of Martha’s sales teams ought to understand these rela-
tionships, although previously he had no tools to explain his understanding to anyone.
Using the new tools which he, Ben, and Martha had evolved had greatly clarified his own
understanding as a direct consequence of formalising what had previously been intuitive.
Trevor now saw developing collective Sales team clarity about key aspects of systemic
uncertainty as very important, and well worth thoughtful direct formal treatment. Clarity efficient formal treatment of key aspects of uncertainty was going to prove the key reason why the new Astro approach to bidding was going to revolutionise Astro’s competitive
capability.
Trevor indicated that they would initially concentrate on a ‘conditional probabilistic bid
adjustment table’ for Taurus, their most important competitor for the Transcon bid. But all
other important competitors plus a composite of the residual could also be assessed via the
same tabular format, then combined via Table 6.6 format computations. In practice, they
would develop and use computer software to keep the input requirements minimal and the
computations as easy as possible, and they would later have to learn how to use these tools.
The focus that day would be the principles behind what had to be done and why. Formally
modelling the systemic uncertainty associated with the adjustment process would help to
give them confidence that the expected values used for their analysis were unbiased, because
they could express their lack of certainty about the adjustment process visibly and explicitly,
and later use actual outcomes to monitor and control estimation error.
Trevor then showed his team Table 6.8.
Trevor began by indicating that Table 6.8 had been developed using a highly iterative
process with considerable detailed help from Ben and crucial oversight help from Mar-
tha. While it had not been straightforward evolving Table 6.8, he was now confident that
with suitable training and background understanding all Astro sales teams could construct
similar tables for any key Astro competitor which was judged to be worth careful thought.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 279
Table 6.8 Conditional probabilistic bid adjustment table: Taurus, assuming a £12m Astro cost estimate.
Probability tree estimates Probability contributions for an adjusted Taurus bid of
Actual bid
Probability Adjusted bid
Probability (0) 15 16 17 18 19 20
19 0.2 20 19 18 17
0.1 0.2 0.5 0.2 0.04
0.10 0.04
0.02
18 0.6 18 17 16 15
0.1 0.2 0.4 0.3 0.18
0.24 0.12
0.06
17 0.2 16 15 (0)
0.1 0.3 0.6 0.12
0.06 0.02
Adjusted bid probabilities (column sums) 0.12 0.24 0.26 0.16 0.16 0.04 0.02
Reverse cumulative probabilities for plotting Figure 6.13 at Astro bids of
1.0 (0)
0.88 14.5
0.64 15.5
0.38 16.5
0.22 17.5
0.06 18.5
0.02 19.5
0 20.5
He thought they needed a simplified acronym label for this table, but he had yet to take the
time to think of one.
Trevor started to explore the thinking behind Table 6.8 by indicating that the first four
columns defined a probability tree which Sales had to estimate. Constructing Table 6.8 had
to be grounded on understanding this probability tree. They might begin by noting that
the second level branches were conditional on the first level, a defining characteristic of the
model they would be using.
The ‘actual bid’ probability distribution in the first two columns of Table 6.8 defined
three bid scenarios which reflected the range of actual bids by Taurus anticipated by Trevor,
with illustrative initial probabilities estimated by Trevor. The ‘adjusted bid’ distributions in
the third and fourth columns were conditional on the actual bids, with illustrative uncer-
tainty ranges and probabilities also estimated by Trevor. They could debate and perhaps
modify these estimates later, and several approaches to more refined portrayals were feasible
if they thought the effort was worthwhile.
To explain the rationale for the actual bid distribution in the first two columns, Trevor
started by observing that Taurus had low costs relative to Astro’s. That was part of the reason Tauris were the market leaders. Astro’s current cost estimate of about £12 million
would be high relative to a comparable cost estimate by Taurus for its direct cost, and Tauris
would recognise that its approximate cost estimate equivalent was less than £12 million. But
Tauris also had technology quality advantages in some areas, which were further sources of
non- price advantage. In addition, Tauris was very capable of generating non- price advan-
tages comparable to the set of ‘opportunity options’ which Martha had identified during the
workshop he had outlined.
Trevor had assumed that even if Tauris had decided to put minimal effort into a bid for the
Transcon contract, because they were very busy or not very interested in winning this particular
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
280 Employing planning tools in practice
contract for other reasons, and they used a ‘high bid’ of £19 million, they still had a low but
plausible chance of winning. If Taurus wanted to win this contract very badly, for a range of
possible reasons which Trevor assumed Astro could only guess at, Tauris might put in a ‘low
bid’ of £17 million. Trevor assumed that the most likely Tauris bid was £18 million, given their
cost base plus their non- price advantages and no obvious reasons for high or low bids currently known to Trevor. Trevor had used a ‘plausible maximum’ P(90) ‘high bid’ (and low Tauris effort) estimate followed by a ‘plausible minimum’ P10 ‘low bid’ (and high Tauris effort) esti-
mate with an ‘intermediate bid’ scenario on an integer scale to keep his example simple, and
his current working assumption about Astro’s inability to forecast the extent to which Tauris
wanted to win a contract made assuming a simple symmetric probability distribution plausible.
The adjusted bid probability distributions in the third and fourth columns completed
the probability tree which defined appropriate adjusted bids given the actual bids assumed.
The probabilities in this case captured Trevor’s best estimates of bid adjustment uncertainty
associated with non- price advantage and disadvantage issues. For example, if Taurus bid
£19 million, Trevor thought that appropriate adjusted bids might be £20 million, £19 mil-
lion, £18 million or £17 million, with the probabilities indicated. His example stuck to the
same integer value £ million scale and only four probability tree branch end point scenarios
for simplicity, but demonstrated clearly that he believed an asymmetric conditional distribu-
tion should be expected.
Trevor suggested that they think of adjusted bids as ‘the maximum bid Astro could
make to beat the associated actual Taurus bid’. The adjustment for non- price advantage
and disadvantage concept was usefully conceived and numerically estimated in these terms.
For example, looking at the first line in the body of Table 6.8, if the Taurus actual bid was
£19 million (assumed probability 0.2), and given this the Taurus adjusted bid was £20 mil-
lion (assumed probability 0.1), this probability tree branch end point scenario implied that
Astro should win with a bid of less than £20 million, lose with a bid of more than £20 mil-
lion, £20 million being the flip point. This involved a non- price advantage for Astro of
+£1 million, which he would refer to as a non- price disadvantage of −£1, because this would
reduce the number of negative numbers they had to work with and clarify the new mindset
they all needed. The associated unconditional probability of this probability tree branch end
point scenario was 0.2 × 0.1 = 0.02.
They should see the right- hand columns of Table 6.8 as a way of recording the probability
products associated with each actual bid and adjustment combination to facilitate later com-
putation. For example, in the body of the table the probability product 0.2 × 0.1 = 0.02 in
line 1 was associated with an adjusted bid of £20 million.
The rationale provided by Trevor for the bid adjustment range plus the probabilities for
each value in the range for each actual bid value was then developed row by row.
Line 1 was a convenient starting point for exploring the assumptions underlying the
estimates. It involved combining the highest plausible Taurus bid and the lowest plausible
non- price disadvantage for Astro. This defined the most optimistic plausible scenario from
an Astro perspective. It was worth further brief elaboration for several reasons.
Trevor had assumed that if Taurus bid at £19 million, Transcon might not recognise
that Taurus was not particularly keen to get the job, but the minimal effort which Tauris
expended on developing a package of non- price advantages would mean that Astro might be able to bid as high as £20 million and still win, implying a non- price disadvantage of
−£1 million for Astro relative to Taurus as perceived by Transcon, an adjusted Taurus bid
of £20 million. Trevor associated this scenario with Astro bidding in its ‘premium end of
the market’ manner with the package of opportunity options and other non- price advantage
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 281
features developed earlier in their bidding process, and Transcon being prepared to pay
for Astro quality given the consequences of a lack of enthusiasm for developing non- price
advantages on the part of Taurus. Trevor thought that the probability of this very optimistic
outcome was only 0.1 given a Taurus bid of £19 million, implying an unconditional prob- ability of only 0.2 × 0.1 = 0.02. Trevor emphasised that this was the only negative non- price disadvantage in Table 6.8. An obvious part of the rationale for his use in Table 6.8 of non-
price ‘disadvantage’ instead of ‘advantage’ was minimising the negative numbers they had to
deal with, but the implied mindset change was actually more important in his view.
Lines 2, 3 and 4 continued to assume a Taurus bid of £19 million, considering succes-
sively lower adjusted bid values together with the probabilities of associated probability
tree scenarios until the successive conditional probability estimates summed to one. These
scenarios explored a £19 million adjusted bid scenario followed by £18 million and £17 mil-
lion adjusted bid possibilities. Together with line 1 this defined a probability distribution
of adjusted bids (and associated non- price disadvantages) conditioned on Taurus bidding
£19 million. With Taurus bidding at £19 million, Trevor was confident that Astro could win
for sure with a bid of £16 million, assuming that a non- price disadvantage of £3 million or
more being ascribed to Astro as perceived by Transcon was not possible using probabilities
defined to the level of precision employed in his example. Working to the nearest £1 million,
non- price disadvantages of −£1 million, £0 million, £1 million, and £2 million were pos-
sible but no values outside this set. Trevor thought that given a Taurus bid of £19 million,
the most likely adjusted bid was £18 million, involving a non- price disadvantage to Astro of
£1 million, with a 0.5 probability. He associated a 0.2 probability with the adjusted bid pos-
sibilities of £19 million and £17 million, both twice as likely as the £20 million possibility.
Line 5 of Table 3.8 associated a Taurus bid of £18 million with a maximum adjusted bid
value of £18 million, because Trevor assumed that the most optimistic possible scenario was
Transcon would see an £18 million bid from Taurus and the associated package of non- price
advantages for Tauris as directly comparable to an Astro bid of £18 million, with no chance
of Astro doing any better than parity in non- price advantage or disadvantage terms. Indeed,
even a parity situation had only a 0.1 conditional probability in Trevor’s view.
Lines 6, 7 and 8 continued assuming a Taurus bid of £18 million and considered decreas-
ing adjusted bid values (increasing non- price disadvantages to Astro) until the conditional
probabilities summed to one and a highest plausible non- price disadvantage was addressed.
If Taurus bid £18 million, Trevor was confident that Astro could win for sure with a bid
of £14 million, implying a non- price disadvantage for Astro of £4 million was not possible.
Trevor believed that with Taurus bidding at £18 million, the most likely equivalent Astro
bid as perceived by Transcon was £16 million, implying a non- price disadvantage of £2 mil-
lion with a probability of 0.4.
Trevor observed that the conditional probability distribution shapes were different for
the actual Taurus bids of £19 million and £18 million, in addition to the range shift, to
emphasise the flexible approach to systemic dependence provided by Table 6.8 without
trying to overstate the implications in this example. If the Table 6.8 structure was used to
model exactly the same adjustment probability distribution for each actual bid value, this
would imply a probabilistic approach which assumed that the non- price disadvantages were
unconditional – independent of the actual bid. If in addition there was only one common
possible outcome for each plausible Taurus actual bid, this would be a very simple special
case directly comparable to Ben’s earlier model.
Line 9 was associated with a Taurus bid of £17 million and a maximum adjusted bid of
£16 million, a further conditional distribution shift. This was because Trevor assumed that
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
282 Employing planning tools in practice
Transcon would see a £17 million bid and the associated package of non- price advantages
for Taurus as very competitive, with no chance of Astro doing any better than a £1 million
non- price disadvantage, and only a 0.1 probability of this occurring.
Line 10 continued the definition of a conditional probability distribution given a Taurus
actual bid of £17 million, with a £15 million adjusted bid associated with a 0.3 probability.
With a big but noticeably sad smile on his face, Trevor indicated that when he initially
addressed line 11, he had associated an actual Taurus bid of £17 million with an adjusted bid
of £14 million, a non- price disadvantage for Astro of £3 million. He was sure many of them
were puzzled about why Table 6.8 used (0). But in discussion with Ben, they had both come
to the conclusion that in practice if Taurus bid £17 million and put the maximum effort into
non- price advantages he was assuming, there was actually a zero probability Astro would
win with a bid of £14 million and further Astro price reductions would not help – there was
a 0.6 probability Astro could not win at any price. This could be interpreted as implying an
adjusted bid of zero because the market leader Taurus wanted this job so much they had put
together a package which was beyond Astro’s reach.
A useful way to interpret this was Taurus had made Transcon an offer Transcon could not
refuse, with a probability of winning of unity. Trevor suggested that this scenario should be
called ‘a zero bid equivalent’ to signify that there was a probability of about 0.6 that ‘Astro
couldn’t give it away in the face of a £17 million Taurus bid’.
The 0.6 adjusted bid probability entry defined by (1 – (0.1 + 0.3)) was associated with a
zero- adjusted bid in brackets so that the adjusted bid probabilities associated with a Taurus
bid of £17 million sum to one. The (0) notation signified that this was the zero- bid equiva-
lent for Astro bids in the range 0 to 14, a simple way of dealing with the complete set of
Astro bids which they had no chance of winning.
The columns collecting together the adjusted bid unconditional probability contributions
incorporated the ‘zero bid equivalent’ under the (0) heading with a 0.12 entry (0.2 × 0.6).
The ‘adjusted bids probability’ row near the bottom of the table just collected the column
sums, and the last two rows just converted the discrete outcome distribution of adjusted
bids into a continuous distribution approximation for later use in Figure 6.13.
13 15 20
0.2
0.4
0.6
0.8
1
Astro bid in £m
probability
of Astro beating Tauris
0
The dashed curve was reproduced from Figure 6.3 without the extrapolations for bids below £15m. The solid curve taken from Table 6.8 terminates at an Astro bid of £14.5 because Trevor was assuming that Astro could not win with a bid below this price.
Figure 6.13 Probability of Astro beating Tauris curve from Table 6.8.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 283
As soon as Trevor thought his Transcon Sales team understood how Table 6.8 worked,
he showed them Figure 6.13.
Trevor then began his explanation of the relationship between Table 6.8 and Figure 6.13.
He pointed out that the last two lines of Table 6.8 allowed the preparation of Figure 6.13
by transforming the discrete probability tree end point probabilities collected by the right-
hand columns of Table 6.8 into a piecewise linear continuous variable distribution for the
probability that Astro would beat Tauris. Working from the right- hand end: if Astro bid
£20.5 million or more, the probability of beating Tauris was zero; if Astro bid £19.5 mil-
lion, the probability of beating Tauris was 0.02; if Astro bid £18.5 million, the probabil-
ity of beating Tauris was 0.06; and so on until the curve simply ends at an Astro bid of
£14.5 million – if Astro bid at £14.5 million the probability of winning was 0.88. The prob-
ability of winning curve just stops at that point, in effect, dropping to zero for bids below
£14.5 million.
Trevor then explained that as illustrated by Figure 6.11, any competitor i curve plus com-
petitor j curve would yield a composite competitor k curve result which was to the left of i or
j. Assuming that i = 1 and j = 2, if i dominated j the shift to the left would be negligible, but
if j dominated, i then the shift to the left would be substantial. The curve for k would never
shift to the right. This meant that as they consider other competitors, things could only get
worse relative to the solid curve associated with Tauris in Figure 6.13. As they could see, this
solid curve was already well below the dashed curve associated with his preliminary curve
addressing all competitors in Figure 6.3, which he had overlaid for comparison.
The solid curve in Figure 6.13 showed only a 0.22 probability of winning with a bid
of £17.5 million, only a 0.38 probability of winning with a bid of £16.5 million, rising
to 0.64 for £15.5 million and 0.88 for £14.5 million. Trevor suggested that if this Fig-
ure 6.13 assessment based on Table 6.8 was reasonably robust, they need to work out how
to attack the massive non- price and cost advantages Taurus had before worrying about other
competitors.
Trevor indicated that this result had surprised him greatly. Relative to his preliminary
assessment it was very pessimistic. But he believed addressing the specifics of their key com- petitor relationship in the way he had illustrated provided a much more realistic basis for
understanding the competition as a whole than anything they had undertaken in the past,
and they could build on it. Furthermore, when they built upon it they could use their time
much more effectively than they had in the past. They would be able to better focus on
what really mattered, which they had earlier lacked the tools to address systematically and
effectively.
Trevor then suggested that they could use everything they had learned since Martha and
Ben began the Transcon bidding exercise to stimulate what he called a ‘further search for
value added’, building on the value added opportunities they had identified earlier. Using
the Astro site for training illustrated what he was after. So did the three bonus opportunity
options identified via Martha’s and Ben’s revisions to system input: the option of Datapol at
a low extra price, the option of additional memory at a minimal extra price, and the option
of a bonus level of performance at a very low extra price. More value added was needed to be
competitive with the market leader Taurus, and other strong competitors, targeting poten-
tial Astro strengths and weaknesses in relation to identified specific strengths and weaknesses
of their key competitors.
Trevor now showed his Transcon bid team Figure 6.14.
Trevor indicated that Ben could produce software to provide this rather different rep-
resentation of Table 6.8 in addition to Figure 6.13, all three portraying the same basic
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
284 Employing planning tools in practice
non-price
disadvantage for Astro
as perceived by Transcon as estimated
by Trevor in £m
Taurus bid in £m
17 18 19
-1
0
2
4
5
-2
0.1
3
1
discontinuity region
expected value 1.9
probability
probability
expected value 0.8
probability
0.3
0.3
0.4
0.2
0.1
0.2
0.5
0.2
0.1
Figure 6.14 Non- price disadvantage probability distributions from Table 6.8.
information in different ways. The Figure 6.14 format could be used as a basis for estimat-
ing the Table 6.8 equivalents as well as a way of portraying those estimates, and he found
the Figure 6.14 format a particularly useful basis for thinking about the information they
needed to generate and fully understand for each of the important competitors.
To understand the relationship between Table 6.8 and Figure 6.14, given a Tauris bid
of £19 million, the fourth column of Table 6.8 defines the adjusted bid conditional prob-
abilities shown for the right- hand distribution in Figure 6.14. Similarly, given a Tauris bid
of £17 million, the fourth column of Table 6.8 defines the adjusted bid conditional prob-
abilities shown for the left- hand distribution in Figure 6.14. The middle case is directly
comparable.
They could use the Figure 6.14 format to explore the implications of the less important
competitors too, including the residual composite n = 6 in a structure like Figure 6.12. He
thought some of them might prefer thinking about what they were doing by starting with
the Figure 6.14 format and using it to generate all the Table 6.8 input estimates, and he was
quite interested the possibility of developing the use of this option.
Trevor also indicated that once they were comfortable using Table 6.8 and Figure 6.14
formats for estimating curves in the Figure 6.13 format, they could revisit their estimates
post bidding whenever they were able to acquire relevant information using the Figure 6.15
format.
For example, if they believed that Table 6.8 and Figures 6.13 and 6.14 were appropriate,
made a bid of £19.3 million and won, and then discovered Tauris had bid £18.3 million, the
Figure 6.15 data point a implied they had been plausible in their assessment of what Tauris
would bid but excessively pessimistic about Astro non- price disadvantage. Their Figure 6.15
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 285
distributions for Tauris bids of £18 million and £19 million interpolated to define point b
for a Tauris bid of £18.3 million suggested winning with an Astro bid of £19.3 million was
not plausible. Their win was an unanticipated outcome implying that they needed to shift
the lower tail of these two conditional distributions downwards, or the whole of these dis-
tributions downwards, to take a more optimistic view of non- price disadvantage relative to
Tauris in the future.
Winning given a key competitor’s bid when this event had been assumed non- feasible
would obviously be good news in part, but a bad news sting in the tail might be relevant
if they had bid significantly lower than they needed to, a form of lost opportunity. Losing
when this event had been assumed non- feasible given a key competitor’s bid would be
unambiguously unwelcome, implying points comparable to points a and b above the Fig-
ure 6.15 conditional distributions. Most of the time data points should fall within ranges
assumed to be feasible, and assuming this was the case it would clearly take time to start to
confirm the reasonableness of expectations and ranges and refine them appropriately. How-
ever, the conditional framework of Figures 6.14 and 6.15 and Table 6.8 clearly gave them
the tools to model their subjective understanding of the issues and then start to acquire data
to further develop their understanding.
When closing this meeting, Trevor reaffirmed that after the initial full- day workshop with
Sian, Martha, and Ben, he had not understood how Martha’s and Ben’s stage 3 enhanced
process for Sales was going to work, and he had significant concerns about it, as indicated at
the beginning of this meeting. However, he now understood clearly how the Tauris exam-
ple approach worked in the conditional probabilistic framework he had just outlined, and
non-price
disadvantage for Astro
as perceived by Transcon as estimated
by Trevor In £m
Taurus bid in £m
17 18 19
-1
0
2
4
5
-2
0.1
3
1
discontinuity region
expected value 1.9
probability
probability
expected value 0.8
interpolated point b data point a
probability
0.3
0.3
0.4
0.2
0.1
0.2
0.5
0.2
0.1 expected values
key:
Figure 6.15 An example data point on the Figure 6.14 non- price disadvantage probability distributions.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
286 Employing planning tools in practice
he had no doubt that they could collectively develop it to deal with the Transcon bid. He
expected they all still had some of the misgivings he had earlier, but he was convinced they
would be resolved as experience was gained. They now needed a series of working meetings
to use these new tools to do the task they had always done well even better, with a much
better focus on what mattered most, to shape the SP for bidding.
This opening meeting for the Transcon bid team was followed by a series of further work-
ing meetings, some involving Sam and Sian, some involving Martha, almost all involving Ben.
With Ben’s support, plus Martha’s strategic overview guidance, Trevor and the other Sales
staff worked with Sian’s staff to find more ‘value added’ for the Transcon bid. They put a lot
of effort into the Transcon bid because it was viewed as a learning project. It was what Trevor
later referred to as ‘a game- changing investment in corporate learning, not just a routine bid’.
Martha then asked Trevor to take a central role in the joint leadership of two more initial
learning project test case bids, nominally led by a lead salesperson from other Astro sales
industry areas, to start spreading the joint learning process. All three industry leads then
took a central role in presentations and training programme development for all other Astro
Sales and Systems staff.
This bidding process development project was an important part of the start of the turna-
round in the fortunes of Astro Wales. On average, once the process was established and in
routine use, only about 20% of the effort put into a full bid was associated with the stages 1
to 3 bidding process to the end of pass 1 of stage 3 as described in this chapter. The other
80% was associated with later pass refinements of stage 3 which incorporated value added
searches to develop the bid in terms of non- price advantages for Astro, better- quality deliv-
erables for their customers.
What emerged from the Transcon exercise as a general consensus and basis for building
upon was summarised as follows:
1 Starting with an extensive analysis of Astro’s key competitor in stage 3 was crucial.
Furthermore, they had learned so much from this Tauris exercise that it was beginning
to look like a simple approach to a key competitor curve with a following simple adjust-
ment for a residual of all other competitors might be the best basis for the preliminary
probability of winning curve used for the PD2.3 as well.
2 It was important to ‘anchor’ the development of Table 6.8 to an estimate of Astro’s
direct costs on the assumption that key competitors had a correlated comparable anchor
reflecting their relative cost advantages or disadvantages. But it was also crucial to assume
that each key competitor would vary their bid according to how much they wanted to
win this particular contract, a low bid being assumed to reflect extensive effort put-
ting together packages of non- price advantages, a systematic uncertainty assumption
associated with key competitors. Relatively unimportant competitors incorporated in
composite competitor n would not justify this assumption, part of the rationale for not
expending effort decomposing composite competitor n.
3 The format of the conditional probabilistic bid adjustment portrayal of Table 6.8 pro-
vided a simple tabular framework for capturing the key issues which needed to be quan-
tified, in particular the range and relative likelihood of actual bids by the key competitor,
and the associated possible non- price advantage scenarios which defined adjusted bids
conditional on the key competitor bid.
4 It was important to explicitly consider systemic uncertainty associated with non- price
advantages and to assume this uncertainty was related to the bid levels of key competitors,
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 287
because this dependence and the associated conditional estimates of uncertainty were
very important, and they required explicit and focused discussion by those with relevant
knowledge.
5 Figures 6.13, 6.14 and 6.15 all provided useful frameworks for interpreting Table 6.8,
and having a single key competitor underlie Figure 6.13 made the graph very credible.
Everyone could accept that it approximated reality. The story it told was clear. The mes-
sages it sent were clear. Its basis was subjective but sound.
6 Once a way of improving the picture portrayed by Figure 6.13 had been identified
and developed, it was relatively straightforward to consider an equivalent of Table 6.8
and Figures 6.13 through 6.15 for the next most important competitor. The approach
employed in Table 6.6 and Figure 6.11 could then be applied to combine these two
competitors. Again, the results were accepted as subjective but sound. A second wave
of value added searches, perhaps in slightly different directions, could follow if needed.
This process could continue as necessary, as the basis of a revised later pass version of
the emerging PD3. However, the dominance of the first few competitors, because of
the ordering approach, made everyone reasonably confident of a final composite com-
petitor n approach to considering ‘all other competitors’, once the process and the
model results became reasonably familiar.
7 There was a very clear need to use this basic approach to the PD3 even when no data
was available to confirm or deny the validity of subjective estimates agreed on by rel-
evant staff, but there was also a clear need to start thinking about data acquisition from
the outset, making use of the Figure 6.15 starting point.
8 On the first pass through the stage 3 process early non- price advantage creation and
analysis by Systems were separated from later comparable non- price advantage creation
and analysis by Sales. But the iterative framework of the process meant that Sales and
Systems could jointly understand and manage the interconnections, plus the way these
interconnections were linked to particularly important individual competitors as well as
the composite of all other competitors.
In summary, Martha initiated and provided strategic shape and direction for an overall set
of changes within the Astro Wales Sales team approach with significant implications. With
Martha’s support, Ben facilitated and provided the basis for appropriate operational mod-
els and processes used to facilitate these changes. However, Trevor was also a clearly vis-
ible leader for Sales staff, using a generalisation of Ben’s very basic non- price disadvantage
model, and Martha was very shrewd to have engineered this, as well as lucky to have some-
one with Trevor’s skill set and mindset. The models chosen and the processes evolved were
simple but effective. The new ideas were an unqualified success.
Trevor and Ben, as well as Martha, established reputations which soon went well beyond
Astro Wales. Sian shared some of the glory, and Rhys felt some of the benefits too, in both
cases because of their direct contributions to the changes led by Martha, as well as further
spin- offs not yet discussed. All their careers took a strong turn for the better because of these
bidding process developments.
Key steps in Martha’s approach to change management
Martha had synthesised six key steps from her analysis of the ‘change management’ litera-
ture which she believed were relevant to the approach she needed to adopt for changing
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
288 Employing planning tools in practice
the way her Astro Wales Sales team addressed bidding processes. They are a useful general
checklist which you might build on, summarised as follows:
1 Make sure everyone involved understands that ‘no change’ is not an option.
2 Make sure that the strategic direction of the change and the key deliverables from the
change are clear.
3 ‘If you cannot change the people, then change the people’ – redeploy (transfer, retire,
or fire) anyone who, given reasonable support, is not willing or able to make the change
as soon as feasible to minimise the negative effects of unconstructive dissent on overall
team morale, as well as the operational burden of people who are not pulling their
weight. Crucially, make it very clear from the outset that this is your intention, with a view to motivating the willing parties to focus on meeting the challenges involved and
encouraging the difficult parties to leave voluntarily as soon as possible with minimal
aggravation before they go.
4 Manage the whole change process as a very high risk programme (portfolio of high
risk projects), with an initial focus on flexibility and effectiveness rather than efficiency
(initially doing the right things is usually much more important than doing things right, although later continuous improvement also matters).
5 Facilitate joint development of the new models and processes that the change requires
by everyone involved so that they all have a strong sense of ownership.
6 Facilitate learning processes so that everyone involved can develop the skills necessary
to develop and implement the new models and processes.
By the end of the developments discussed so far Martha had done what she could to mini-
mise uncertainty with respect to the first three steps, and she had initiated the last three.
Martha had learned a lot from being involved directly in the bid process development
project, as well as what she learned directly from Trevor, Ben, and the others involved. She
knew that this would be the case. More generally, from the outset Martha knew roughly
what she wanted all key Astro staff to learn and what she wanted the Transcon bid process
development project to deliver. Any project requires as much of this kind of strategic clarity
about all intended deliverables as the context and the capabilities available allow.
Wider impacts on Astro Wales
Martha could think top- down within her Astro Wales Sales responsibilities, a key capability.
She could also think sideways to Sian’s concerns, and upwards to the issues that mattered
for Astro Wales, Astro UK, and Astro Inc.
Martha had been clear from the outset that she would have to proactively manage side-
ways and upwards as well as downwards within her Sales team. Her highest priority outside
Sales was changes within Systems. She knew there were changes in Systems which Sian was
going to have to deal with if they were going to jointly resolve the core problems for Astro
Wales, and a coordinated joint effort was essential. Towards the end of her focus on the
enhanced cost estimate process with Sian, Martha had deliberately provided Sian with some
clear hints that they needed to get together on these joint concerns. She made it clear that
not all Sian’s staff would attract favourable cost adjustments, and all groups of staff within
Systems had to be competitive to survive beyond the accommodation of short- run dif-
ficulties. Martha would try to ensure that Sian’s underutilised groups had work generated
for them in the short run if they could compete on price and quality in the medium-term.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 289
But nobody in any group could be carried for long if Astro Wales as a whole was going to
succeed.
The degree of difficulty inherent in adopting the new models which Sian and her staff
had to make use of in Table 6.3 was trivial in comparison to those Trevor had to develop.
Sian did not need Ben’s support to a significant extent with the costing of bids. Table 6.3
and the supporting figures could be facilitated with software tools, and Ben helped with
the specification and testing of these tools, but Ben’s support for Sian was low key com-
pared to his support for Trevor and other Sales staff in terms of developing formal models
and processes to implement these models. However, the new internal bidding process
which Table 6.3 cost adjustments made explicit involved a major culture change within
Systems. Sian needed Martha’s support to a significant extent to cope with this, and she
got it. She also needed and got support from Rhys, with Martha playing the facilitator
role here. More generally, Rhys, Martha, and Sian needed to work as a team, with flexible
role changes when needed. How they did this is briefly elaborated now, starting with a
focus on Sian.
Sian’s Astro career had always been in Systems, and she had always seen the role of a
head of Systems like herself in terms of a conventional matrix- based ‘service’ organisa-
tion: Sales people were responsible for getting the work and making a profit on the
sales; Systems people were responsible for doing the work and ensuring that the quality
generated a good reputation which would lead to future work. This built in a quality–
profit tension management process with clear ‘champions’ for each. Sian had recognised
the ambiguity about responsibility for Astro losses on contracts was a long- term Astro
problem, and the growing use of subcontractors like Zoro involved a complication of
the simpler historical situation, but she had never seen marketing her Systems people in
competition with external suppliers of the same or similar skills as a key part of her job.
Nor had she fully appreciated the conflicts of interests inherent in her current role as
judge in a process which entailed her Systems staff bidding against external subcontrac-
tors. She now recognised these complications were crucial. Her job, as she now saw it,
involved serious conflicts of interest, and it was not clear to her what ought to be done.
Massive uncertainties and complexities she had not previously seen suddenly revealed
themselves. Martha anticipated this because she had thought about and understood to
some extent the coming changes. They had been forced on her thinking when she con-
fronted the fact that as a salesperson in Astro Scotland she had done very well selling on
commission, but she now had no commission, and she was responsible for profitability
within her Sales silo with staff reporting to her paid commissions on sales (not profit) in
addition to those responsible for delivery like Systems motivated very differently to those
working for organisations like Zoro. In terms of the capability- culture concept discussed
in Chapter 2, Martha could see some important liabilities needing effective immediate
accommodations and important longer- term resolution, although she might use different
language to explain them.
Rhys, the manager in charge of Astro Wales, had a systems background like Sian. He was
a native of Wales, but most of his modestly successful career had been spent in various other
locations in the UK and US. Martha and Sian were both keen to involve him in Sian’s dif-
ficulties from the outset, and he was very keen to be involved. He had held his current post
longer than Martha or Sian, but he had never been very comfortable telling either of them
or their predecessors what to do. He kept himself busy acting as a front man for Astro Wales
in ways which did not always seem as useful as he wanted to be. For some time he had been
as uncertain about his proper role as Sian had suddenly become.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
290 Employing planning tools in practice
Martha, Sian and Rhys began a series of discussions about these issues, initially usually just
two at a time in various combinations, but increasingly as a three person team working in a
coordinated manner. Five significant changes emerged from these discussions:
1 With crucial support from Rhys, Sian should divide Systems into a number of groups,
and ensure that each had a strong ‘group head’ who would take full responsibility for
the quality and cost of work by that group, with requisite authority to hire, promote,
and recommend redeployment in order to maintain a group size agreed on with Sian.
Sian would be responsible for the performance of her group heads, but not the quality
or cost of the work by groups so long as a group head was not in obvious trouble as a
manager. In effect, important roles Sian currently played would be devolved to group
heads, drawing lines around what was devolved and what Sian kept with a view to
effective management of uncertainty and complexity without direct conflicts of interest
for any of the players. Sian was ‘subcontracting’ part of her job to her group heads in
order to enhance her performance with the residual responsibilities, analogous to Astro
Wales subcontracting parts of systems integration projects in order to enhance overall
performance. Martha was the one who guided their thinking in this direction and who
suggested the subcontracting analogy, but the solution was obvious to all of them as
soon as they understood what was involved and the rationale for the changes.
2 Each Systems group would be regarded as a profit centre, not just a cost centre. Each
group head was responsible for cost and quality, but Sian would accept responsibility for
the profit aspect of each group, in conjunction with group sizing decisions. The size of
each group would be a focus of discussions between Sian and her group managers, with
oversight by Rhys. Groups which made a sustainable profit would grow; groups which
could not make a sustainable profit would shrink or ‘relocate’.
3 Sian would see one key aspect of her job as an internal marketing manager for her Sys-
tems groups, to Martha’s Astro Wales projects, to Astro UK, and to Astro Inc more
generally. She would take a reasonably reactive role to the first, but a highly proactive
role to the second and third. That is, she would trust and rely on Martha to take a
proactive interest in generating local (Welsh) work for her people, given support and
guidance, but she would proactively seek systems contracts with Astro UK and Astro
Inc because Martha’s remit did not include this kind of sales, and in the past such con-
tracts did not really have a ‘champion’. Sian had to take on some of this ‘champion’
role, working explicitly towards this end with Rhys.
4 Rhys would serve as Sian’s ‘front man’ and ‘chief salesman’ for internal ‘sales’ to Astro
UK and Astro Inc, making effective use of his extensive network of connections and
his wide experience. Martha would also support Rhys and Sian’s internal sales in the
Astro UK and Astro Inc areas in terms of the marketing experience which Sian and
Rhys lacked. Ben might provide backup analytical support as well. Sian would be ‘in
charge’ of marketing in this internal markets for Systems sense, but she would be fully
supported by all Astro Wales staff whenever they could contribute, whether they were
technically a level above (Rhys), on the same level (Martha), or somewhere below in
someone else’s reporting structure (Ben). Martha’s responsibility for Sales would have
a local (Welsh) focus, explicitly restricted to non- Astro markets in terms of proactive
leadership.
5 Rhys would take personal responsibility for developing and maintaining a suitable list of
‘partnership external subcontractors’, and for looking after their interests in any com-
petition with Astro Wales Systems groups on Martha’s contracts unless and until this
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 291
task warranted a separate manager. Rhys, Martha and Sian all agreed that an internal
bidding process involving sealed bids, replicating what Martha faced with customers
like Transcon, would not be helpful. However, an open process with a ‘level playing
field’ was necessary. A ‘strategic partnership’ relationship with subcontractors was a pro-
ductive way of levelling the playing field, working towards putting them on a par with
internal Systems groups, rather than the other way around. It also had direct benefits,
which business fashion was beginning to recognise. The concept of non- price advan-
tage was central to these considerations. One of the key non- price advantages was trust
that subcontractors and internal groups could deliver what they promised when they
promised, without arguments and without close management. Another was confidence
that ‘partners’ would be open and honest about uncertainties which could become
sources of risk. Partners who could not be trusted should be rejected as partners, and
all partners needed to understand this clearly from the outset.
Rhys, Sian and Martha did not attract the immediate attention with these developments
that Trevor, Ben and Martha generated with their bidding models and processes. To some
extent they were competing with other Astro UK and Astro Inc groups, so they did not
want to shout about these changes and give away competitive advantage. However, within
a few years Astro Wales became prominent for attracting many large systems development
contracts from other parts of Astro, as well as for working effectively with a wide range of
partner subcontractors in a growing volume of local business, and it was clear where the
credit was due. Zoro was one of their partners, as were several Astro groups which were
floated off as separate businesses, usually to everyone’s benefit, although sometimes these
float- offs failed, and sometimes very few people were surprised by such failures.
The uncertainties and complexities to be faced, the uncertainty and complexity associ-
ated with the best way to handle them, and the uncertainty generated by associated changes
which Sian had to address, were all quite different from those Trevor and the other sales staff
had to address. The support Sian needed was very different as a consequence. The ‘softer’,
less clear- cut nature of the issues made formal models less useful, but the need for flexible
formal processes and attention to organisational structure correspondingly more important.
There was no need to redefine Sian’s job and the organisational structure around her in
formal terms for all to see, but there was a crucial need to redefine Sian’s roles in agreement
with those she worked with. Equally important was the flexibility implied in both Martha
and Rhys ‘working for Sian’ in some contexts when this was useful. The senior staff of Astro
Wales worked as a close team, with role changes according to context, to effectively man-
age the opportunities inherent in the complex uncertainty associated with redefining Sian’s
role in relation to their own roles. The experience was very positive for Martha and Rhys, as
well as for Sian, and enlightening for everyone involved. It was an example of enlightened
teamwork fully integrated with enlightened governance. Ben regarded his limited involve-
ment in it as an important part of his apprenticeship, learning what management was all
about beyond his immediate role as facilitator for the development and use of Astro’s SP
for bidding.
The example set by Rhys, Sian and Martha in terms of their approach to teamwork and
governance was infectious. It was supported by Trevor and other Sales staff turning to Sys-
tems staff to help in the search for ‘value added’ in order to develop winning bids. It was
reinforced by heads of Systems groups seeking a degree of marketing expertise from Tre-
vor and other Sales staff as well as Martha. There was a visible flow of ideas and help from
Sales to Systems and from Systems to Sales, which expanded to include the Service Bureau
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
292 Employing planning tools in practice
operation, Accounting and other support groups. To a lesser but significant extent it also
embraced the partner external subcontractors. Apart from the direct benefits of the ideas
and help, as silo boundaries softened and barriers went down, trust grew, and new synergy
was identified and exploited.
‘Managing upwards’ and ‘managing across’ became operational aspects of the Astro office
jargon – natural complements to ‘managing downwards’. This collaborative style of man-
agement was recognised as the glue which closely bonded the Welch office as a team. Man-
aging the uncertainty and complexity facing Astro Wales as a whole was very much a ‘team
sport’. Teamwork became the key driver of Astro capability- culture evolution.
Uncertainty and complexity management models and processes helped to ‘shape the key
and lubricate the lock’. Astro Wales as a whole prospered. Astro Wales was seen to be in the
vanguard of the turnaround in Astro UK, and a very useful contributor to the turnaround
in Astro Inc.
Effective and efficient external bidding processes were often held up as illustrative of the
contribution Astro Wales made to Astro Inc. Those directly involved knew these bidding
processes were a catalyst or a driver, in some ways only a very small part of the whole Astro
recovery story, but arguably this kind of catalyst was crucial. All major corporate changes
usually need a portfolio of catalysts and drivers.
The credibility of this tale and linked issues
The basis of this chapter’s tale is a case study which I initially developed for an IBM UK
culture change programme based on ideas and information provided by ‘Edward’, an execu-
tive at a very senior level in IBM UK. Edward provided most of the context information
of Table 6.1 and its background. Bidding was the topic Edward chose because the market-
driven nature of IBM meant that all IBM UK staff could relate to what was involved. This
‘Transcon case study’ was used about 40 times in its initial form with about 25 IBM staff
each time from a wide range of roles. During this period IBM staff initiated a number of the
ideas incorporated in this chapter, including the friendly takeover of Zoro. Subsequently,
Stephen Ward and I have used this case study hundreds of times, on in- company short
courses and on open short courses in the UK, the rest of Europe, North America and Africa,
with participants from many companies and countries. We have also used it for university
courses.
The three stage structure aspects of this chapter’s approach were developed earlier when
Stephen and I were working with Southern Gas and one of our MSc students on tendering
for large service contracts with local authorities and housing trusts. It involves a particular
form of sophistication not used for the IBM- related work but tested elsewhere.
Some of the contracting ideas derive from a Science and Engineering Research Council
(SERC) research contract with Stephen Ward and Bernard Curtis plus follow- on work with
Stephen and others, and some of the internal contracting and outsourcing ideas were trig-
gered by my work with BP.
All the initial full- day workshop ideas were part of the framework IBM employees were
exposed to, but the Table 6.8 conditional probabilistic bid adjustment approach plus the
linked clarification of ‘Trevor’s transformative insights’ were not developed in this particular
form until this book was in its fourth draft. Some much more sophisticated ideas of a differ-
ent kind were developed with IBM in the 1990s on a follow- on consultancy basis.
One aspect of this follow- on work with IBM considered a potential customer with a
very large project (say, £200 million plus) who invites up to a dozen bids from different
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 293
companies, perhaps paying them a fee to tender, with a view to eliminating all but three
or four who then get a chance to bid in a second round. The key purpose of the multiple
rounds approach from the potential customer’s perspective is the second round specifica-
tion is based on a synthesis of the best ideas on project strategy from the first round tenders.
This means first round tendering has to include a portfolio of ideas which are good enough
to get into the second round, but all tendering organisations have to preserve competitive
advantage to win the second round by holding back some key ideas not to be shared with
the potential customer during the first round. In the first round the potential customer is
in effect buying relatively low cost or free consultancy on project strategy, with a view to
a fairly crude form of partnering approach to contracting the project. An aspect arguably
needing refinement from the customer’s perspective is the gaming strategies of the first
round players decrease the value to the customer of what could be learned. Having won the
second round stage, third round partnership negotiations are anticipated, with some further
negative implications of potential gaming. What this multiple rounds study illustrated to
me, incorporated into Astro’s approach when a single round is involved as far as possible,
is the importance of the ‘bonus opportunity’ concept and other ‘added value’ ideas as a
contractor’s way into an efficient and effective form of ‘strategic partnership’ relationship
with the client. Customer organisations need to understand these ideas as well as contrac-
tor organisations, as part of seeking ‘strategic partner’ relationships with their contractors.
Furthermore, contractor organisations which use subcontractors need to understand they
are simultaneously involved in both.
Using the Transcon case study for many years with people from a wide range of organi-
sations it has become increasingly clear that customer organisations almost always need
much more clarity about the significant difference in the nature, quality and performance of
‘deliverables’ provided by different contractors. They also need a very clear understanding
of the need to avoid the significant foolishness of assuming that the lowest price is the best
deal for the customer, unless they are JUSTIFIABLY very confident that they know exactly what they want and how to unambiguously specify what they want, all competing contractors are equally competent and honest, and all competing contractors are equally keen to win the
contract. The probability of all these working assumptions holding at the same time is zero
for most practical purposes. More generally, there are a number of reasons why the tale of
this chapter is directly relevant to customer organisations, whatever they have to purchase,
understanding how to seek best ‘value for money’ from purchases being one aspect. To
some extent all organisations are both ‘customers’ and ‘contractors’.
Astro is comparable to IBM, but Astro is not just a disguised version of IBM. Astro is
a fictitious basis for a tale. However, all the Astro capability- culture assets assumed for the
tale are consistent with those observed in IBM UK groups when I worked with them as a
consultant in the 1990s and in IBM Canada colleagues when I worked for IBM in various
roles in Toronto in the 1960s. IBM has a well- deserved long- standing international reputa-
tion for periodically reinventing itself with teams like Martha, Sian, Trevor, Rhys and Ben
working together in the way described in this tale.
Martha’s capabilities may pose a credibility issue for some people, but in my experience,
what is needed to make Martha’s success plausible includes a deep understanding of what
is interconnected in the industry involved, a deep understanding of how processes can be
designed around models to make decision making more effective, what is needed to lead
teams making significant changes, and the personality to carry it off. Martha did not have
a formal degree- based background in MS/OR, information technology or management.
Suitable in- house or open course experience to fill in gaps can clearly be crucial, but a very
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
294 Employing planning tools in practice
general initial background like philosophy or history or physics or simple learning by doing
on the job from an early age is not an issue, nor is youth, nor is a limited depth of knowl-
edge of the industry. Martha’s characterisation deliberately clarifies the need for requisite
in- house or open course experience to fill in the gaps, technical input like that provided by
Ben plus experience based input like that provided by Trevor, but their inputs plus those
provided by Sian and Rhys need the synthesis plus the understanding of process design,
strategic vision, and change management capabilities which the tale assumed Martha could
provide. A technical background in an area like MS/OR would be a useful asset if it did not
detract from Martha’s abilities to think and plan top- down as well as synthesise sideways and
upwards, and it might make her less dependent on Ben’s capabilities. However, when draft-
ing the tale it seemed worth making the point that the capability set provided collectively
by Edward and myself plus the IBM sales teams we interacted with via the case study and
follow- on consulting exercises could be provided by the Martha, Trevor, and Ben trio plus
Sian and Rhys or some other combinations with the same overall capability- culture asset set.
The implications if you associate yourself with Martha, Trevor, Ben, Sian, Rhys or higher
levels of management could be particularly important.
Another credibility issue which you might reflect on is the way an absence of any of
these key inputs or a capability- culture asset set which did not facilitate synthesis in the way
achieved by Astro Wales would damage and perhaps destroy the ability of the team to suc-
ceed on this scale. The way Trevor, Sian, Rhys and Ben all supported Martha was crucial, as
was the way she supported them. This teamwork issue could be particularly important if you
associate yourself with comparable teams or have them working on your behalf.
The only really debatable credibility issue is whether someone just out of university like
Ben would have the experience necessary to help Trevor be as creative in technical terms
as the tale suggests, unless Trevor himself had skills he had been hiding for some reason,
or Martha had chosen her professional development short courses very effectively, or some lucky combination of all these factors. Arguably someone with much more experience than
Ben might be a safer proposition which was very worthwhile, and this is clearly an option
you might prefer in practice. However, this aspect of the tail is arguably just a good way to
challenge those who associate themselves with Ben, or managing the teaching of people like
Ben at universities, and those who associate themselves with Trevor or Martha. Hopefully
your credibility has been usefully stretched in a good cause and not permanently damaged.
As in Chapter 5, the UP and the EP perspective as a whole have been used throughout
to bring the interpretation up to date. It should be clear that although the context of the
tale was the 1990s, the substantive ideas are all relevant to the 2020s and beyond. Further-
more, a periodic need to reinvent successful organisations as well as dealing directly with
comparable organisational difficulties in organisations yet to achieve significant success are
perennial concerns.
A clear and explicit modelling framework for competitive bidding has been available for
some time. The literature I drew on post Friedman (1956) includes Ruthkopf (1983), Sam-
uelson (1986), King and Mercer (1985, 1991), and Tweedley (1995), to illustrate some of
the relevant sources. However, in contexts directly comparable to Astro’s, very few organisa-
tions I have experience of have even approached the effective implementation of these ideas
in the way achieved by Astro in this tale, with notable exceptions like IBM UK. Use of the
case study underlying this chapter on professional short courses for more than two decades
provides considerable evidence of this. The perception of organisations which deliberately
avoid such formalisations must be that such formalisations will not be helpful or that they
do not want to change their ways and the competition will not force them to change. These
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 295
perceptions are very risky, as is a failure to even consider the issues. The power of formalisa-
tions which can effectively separate the quantitative and the qualitative aspects of complex
decisions are increasingly recognised as such formalisations become better developed. If one
significant player in any given marketplace adopts the effective use of such approaches, their
competitors will have to follow suit or pay a heavy price.
In somewhat different contexts some very sophisticated bidding processes and models
are now common practice because others have had to follow where one successful pioneer
led. For example, purchasing an airline ticket now usually involves a price set by a ‘yield
management’ system, a bidding process and model allowing airlines to maximise revenue
via separate price bids for each flight for each prospective passenger. My understanding is
this approach was originally pioneered by an in- house American Airlines MS/OR group.
Booking hotel rooms and many other services now involves widespread use of the same
approach. The growth of Internet booking and other Internet- based online purchasing
decisions can be expected to accelerate the use of simpler variants of the kind of marketing
approach explored in this chapter, and more sophisticated variants could also find accelerat-
ing interest, both increasing the direct relevance of Astro’s approach in post- 2020 contexts.
This chapter’s stage 2 approach to the probability of winning curves and the stage 3
approach to ‘enhanced cost estimation’ have been tested in practice as well as being pub-
lished earlier (Chapman, Ward, and Bennell, 2000). Some aspects of the stage 3 formali-
sations of the probability of winning curve developed by Trevor and Ben were key new
material provided by Chapman and Ward (2002, chapter 3), further evolved in this book,
which has not yet been tested in practice. However, one of the reasons many organisations
may have failed to adopt formal bidding processes might be resolved by the further new
material associated with Table 6.8 and ‘Trevor’s transformative insights’. The lack of simple
explicit ways of considering important individual competitors in the conditional non- price
advantage framework of this chapter’s tale, plus simple ways of accommodating the rest,
may have been a key barrier in the past, a ‘missing link’ developed into an easier form for
practical application during the drafting of this chapter.
This ‘enhanced probability of winning’ material may help to spread the development of
approaches to bidding like that explored in the tale of this chapter, breaking new ground.
But its real purpose in this book is illustrating ideas with wider implications. In particu-
lar, you might find it useful to understand that the Table 6.8 and Figures 6.14 and 6.15
approaches coupled to Trevor’s transformative insights did not ‘pop out’ until I spent some
time thinking about how to make this missing link as simple to use in practice as possible.
This implies a key message from this tale is when any aspect of an overall approach does not
seem to be working very well, taking the time to explore what alternative working assump-
tions might work better can be fruitful if you are both persistent and lucky.
A general point worth emphasis in this section is bidding processes are simply a special
case of selling a product or service or combination of the two by a ‘contractor’ (vendor) to
a ‘client’ (purchaser or customer), one way of approaching marketing or purchasing. From
both vendor and purchaser perspectives value for money is or should be the crucial issue,
not just price. And crucially, a strategic partner approach by both parties can be in the best
interests of both parties. These two issues are central to this chapter’s tale, and they are key
linkages to all other Part 2 tales.
A related final general point worth emphasis in this section is ‘marketing’ as it is approached
by the Astro team in this tale is not separable from ‘product development’ and ‘corporate
innovation’ more generally, and breaking down the ‘Sales’ and ‘Systems’ silos is an obvious
early step in dealing with crucial corporate interdependencies which also have to embrace
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
296 Employing planning tools in practice
outsourcing with residual strategic partnership linkages as well as forms of insourcing like
the friendly takeover of Zoro.
Synthesis and reflections at an overview level
This final section summarises synthesis and reflections on a few aspects of the tale of this
chapter at an overview level which you may find helpful for stimulating your clarification of
the emerging messages about EP.
Formal analysis of uncertainty to take more risk on some decisions, not less risk
Astro’s strategy of developing formal probabilistic based models to take enlightened gam-
bles involving more cost risk on each bid, not less cost risk, was central to the tale of this chapter and a core component concern in the risk efficiency discussion of Chapter 3. It was
also central to the IBM culture change programme that both draw upon. This strategy has
general applicability in a very broad sense. Risk efficiency and its generalisation as a central
part of opportunity efficiency always imply that it is important to take all risk which is bear-
able and worth taking, to increase expected performance, and this may be crucial to reduce the risk of higher order concerns, like avoiding bankruptcy.
However, taking enlightened gambles in terms of more cost risk in all bids does not imply anything about reputation risk, and this chapter’s tale introduced a red face test to avoid any reputation risk the board did not explicitly condone. Martha explicitly assumed that the
board would and should be risk averse in terms of reputation risk, and the board should
take responsibility for leading on this issue. Furthermore, there was an unstated assumption
that Astro needed to manage higher levels of decision making with due concern for relevant
cost and revenue risk comparable to those to be addressed in Chapter 8 but not even men-
tioned so far in this chapter. For example, at board level Astro Inc needed to address what
portfolio of products and services they should be targeting now to characterise where the
organisation wanted to be at five or ten year strategic planning horizons. They would have
to do this bearing in mind the need for enlightened caution in the context of diversifying
efficiently because the information technology business is inherently unpredictable, ideas
explored in Chapter 8 needing attention. They would also have to bear in mind the need for
each development team to focus on what they believed was the best bet, taking enlightened
gambles when appropriate.
A focus on using unbiased expected values when associated variability is ‘noise’
If variability associated with expected values is ‘noise’, not significant ‘risk’ which needs
to be reduced, then we can focus on maximising opportunities in terms of expected out-
comes and forget about risk in terms of downside variability relative to unbiased estimates
of expected outcomes. Biased estimates of expected values remain an important risk, be it
biased estimates of expected outcomes, biased estimates of the probability of key events like
winning a contract at a given bid price, or any other key parameter in any context. However,
being able to focus on decision making using expected values which can be assumed unbi-
ased is useful because it improves clarity efficiency, letting us focus on what really matters
using simpler approaches.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 297
Part 1 introduced this idea, the tale of Chapter 5 made use of it, and this chapter makes
a central point of further demonstrating the rationale. Further illustrations are part of all
following chapters. A key message in this book as a whole is do not worry about risk that
does not matter. Focus your attention on being more effective and efficient at capturing the
opportunities that really matter and avoiding the risks that really matter. Opportunities that
really matter should be your top priority or first order concern, almost always. Risks that really matter should be your second order concern, whenever relevant. Persistent bias is usu- ally a risk that matters. A follow- on issue is the almost inevitable need for explicit initial plans
to minimise bias plus subsequent data gathering and analysis to monitor and control bias.
Simple strong assumptions which are robust but potentially counter- intuitive
Sometimes it is very useful to employ strong assumptions that are robust but require an
appreciation of the underlying complexity to fully understand them because they go beyond
natural intuitions, and they may also go beyond common practice. One example is using a
perfect positive correlation assumption when adding up nominal minimum and maximum
values of cost items to define the overall cost range, as in Table 6.3.
Over many years numerous seminar participants who clearly believed they had a good
understanding of correlation have challenged this working assumption, arguing that it is
far too pessimistic. They are almost always implicitly arguing that the common practice of
assuming independence is preferable. Without making too much of it, my response always
begins with the observation that their intuitive preference for assuming independence makes
the analysis computations much more complex – not simpler. However, the more serious concern is an assumption of independence is almost always seriously and inappropriately
optimistic because variability largely cancels out if independence is assumed, even for as few
as five items as for this chapter’s example. Empirical work with BP on correlation between
cost items for refinery projects suggested 0.6 to 0.9 was a common range for the coefficient
of correlation, and 0.7 or 0.8 are reasonable best guesses in other contexts if more careful
consideration is not worthwhile. This is much closer to 1.0 (perfect positive correlation)
than zero (statistical independence). If knock- on, cascade effect and other feedback loop
effects are involved, but not fully appreciated and built into the ‘unconditional’ estimates
of component distributions, a perfect positive correlation assumption can actually prove far
too optimistic. If the simple procedure of just adding nominal maximum costs suggests a
risk level which is tolerable, then a reasonable inference is ‘what is probably a modestly pes- simistic estimate of the range is acceptable and reasonably robust, but it could actually prove
optimistic, and any well- founded suspicions about knock- on, cascade effect and other feed-
back loops concerns should be pursued’. If this estimate of the range suggests associated risk
is not tolerable, then much more sophisticated estimation of the range is potentially impor- tant and well worth undertaking. A common practice default assumption of independence
generally involves a failure to even consider these concerns. The seminar participants raising
the question are almost always immediately convinced because the underlying complexity
is appreciated in a way which adjusts their intuition. My surmise is those who are not con-
vinced simply do not yet understand dependence at an acceptable level.
Managers at senior and intermediate levels in organisations are often dependent on advice
being reliable. To be confident that their organisations are effective and efficient, board
level managers need to ensure that all those designing and implementing all the decision-
making processes used by their organisations fully understand the complexities underlying
important simplifying assumptions. Those responsible for selecting or designing processes
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
298 Employing planning tools in practice
and training staff to use them need to meet the challenges this raises. Especially for specific
processes which are core to an organisation’s success, keeping it simple is very important,
but doing so appropriately is crucial, a very general concern.
A key risk is organisations using stealth assumptions which matter – assumptions which
simplify in the ‘wrong way’ to a significant extent.
More complex general models which are easier to use than ‘simpler’ models
The way Ben explained adjusting each competitor’s bid for non- price advantage using
a constant value which was assumed to be independent of the bid prices was a sensibly
simple way to start, keeping it simple to build an initial understanding of the basic frame-
work. That was why this approach was taken in the section dealing with the enhanced
probability of winning aspect of the full- day workshop. However, when the generalisa-
tions which led to using the ‘conditional probabilistic bid adjustment table’ portrayed
by Table 6.8 and the follow- on use of Figures 6.13– 6.15 were explored, it soon became
clear that this more complex conditional probability approach was actually much easier to use. It provided a much clearer understanding of what was involved than the ‘simpler’ and more restricted portrayals Ben had employed initially, the basis of ‘Trevor’s trans-
formative insights’.
I was surprised, as I often am, by how a more complex general conceptual frame-
work actually makes it simpler to see what is going on because the additional complexity
captures important features of the reality. Assuming away important relevant features
of reality is part of the conceptual difficulty people sometimes have with the ‘simpler’
model, because the gap between the simple model and a more complex reality becomes
dysfunctional – the model is not a credible proposition. An ability to keep an open mind
on when further generality is likely to prove useful with a nuanced understanding of
the implications is a valuable capability- culture asset which all organisations need. It
is closely coupled to making understanding easier for those who need new and deeper
understanding in order to communicate effectively what they already understand at an
intuitive level. Both are closely coupled to understanding when simplicity of particular
kinds pays or does not pay. An overview understanding of all three is part of the strategic
clarity you need to acquire which all Part 2 chapters build in a layered manner, whether
or not you are also interested in associated further tactical clarity. They are also part of
the generalisation of constructive simplicity to the aspirational enlightened simplicity
target which EP needs to seek.
To be confident that their organisations are effective and efficient, board level manag-
ers need to ensure that all those designing or selecting or contracting for their corporate
processes fully understand what is involved to a meet the challenges this raises, especially
crucial when specific processes which are core to an organisation’s success are involved. All
those using corporate processes can also contribute to corporate progress in this area if they
understand the issues.
Explicitly addressing ‘traffic light’ issues in qualitative terms
Martha’s ‘traffic light’ approach to stage 1 addresses important concerns needing qualitative
treatment via red and blue lights plus an intermediate spectrum. This facilitates giving quali- tative concerns appropriate direct and explicit consideration when addressing related quan- titative issues, avoiding inappropriate bias of cost estimates or judgements of appropriate
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 299
prices as part of the associated process. Even ethical issues can be addressed using this
framework. Biased cost and other parameter estimates leading to inappropriate decisions
are endemic in many organisations. Part of the problem may be no formal provision for the separate qualitative consideration of important objectives and constraints which should not be addressed by the metrics being used but do need systematic and effective consideration
by all relevant parties.
The general issue is making sure that all important concerns are properly addressed in a
way everybody understands which is appropriate, effective, and efficient. If important con-
cerns cannot be effectively addressed directly by the cost and margin metrics being used,
then addressing them effectively by some variant of Martha’s traffic light concept may be
crucial. A spinoff will be the removal of bias in estimates of what does need measuring.
These issues need attention in a very wide variety of contexts.
The red face test aspect was introduced to me fairly recently by Paul Thornton. But I have
found the smiles test for blue light concerns conceptually and operationally useful for many
years. Both the frowns test and the halo test concepts have clear complementary roles, com-
pleting the ‘traffic light’ concept expounded by Martha in a form you could build on and
adapt to suit the needs of other kinds of organisations.
It is worth reiterating the crucial importance of avoiding a ‘white light’ averaging of
multiple coloured lights – there is no numeric content or inference involved. The com-
mon practice use of probability- impact graphs (PIGs) with red to green colour coding
addressed in Chapter 7 not only fails to deal with an equivalent ‘white light’ averaging
concern; it fails also to deal with very different basic meanings for ‘amber’. In a PIGs
context these issues present in a slightly different form. ‘Low impact’, ‘high probability’
scenarios and ‘high impact’, ‘low probability’ scenarios which may both have comparable expected outcomes are both rated ‘amber light’ concerns in a PIG framework when they
obviously have very different implications in an EP framework. Comparable confusion
exists for other combinations given the same colour, explored in Chapter 7 in terms of
shifting corporate cultures which favour this kind of dysfunctional simplicity towards a
more enlightened approach.
A formal corporate approach to not doing ‘bad’ as well as doing ‘good’
The notion that ‘doing good is great for business’ and associated underlying trends identi-
fied by Overman (2014) was discussed in Chapter 5, but some of the underlying issues and
trends have different implications in this chapter, directly linked to the way Martha’s red
face test with board level approvals was managed upwards by Rhys for Astro Wales, as well
as opening channels for downward communication by the board. Arguably all organisa-
tions need comparable qualitative communication channels primarily for positive opportunity efficiency reasons which are an internal concern. That was the way Martha dealt with the red face test in her generalised traffic light framework, with ‘smiles’ and ‘frowns’ playing the
primary roles, ‘halos’ and ‘red faces’ treated as special cases of smiles and frowns. However,
from an external perspective with a view to regulation and law enforcement, it is important
to protect employees who are not happy with corporate ethics who feel a moral obligation to
do more than just vote with their feet (leaving if they are not comfortable), and it is impor-
tant to make boards accountable for effective processes to ensure that employees who do
not uphold corporate standards are required to leave, also facing prosecution if appropriate.
This is not an issue in any of this book’s tales, but it is a general concern which needs to be
addressed effectively.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
300 Employing planning tools in practice
Process stages, by design, to filter potential opportunities plus other process objectives
Martha developed a formal three stage process to clarify why Astro should not expect to bid
in response to every invitation to tender, to begin to use a formal process to discriminate
between the opportunities presented at an early stage before too much effort was wasted,
and to begin to order the acquisition of knowledge appropriate to rejection or not in an
efficient sequence.
An important further feature or benefit of this stage structure was the orderly acquisition
of qualitative information which might get overlooked once quantification of a bid began,
exploiting the focus on one kind of task at a time to do a better job of each task.
Arguably all potential corporate opportunities need this kind of multiple stage clarity effi-
cient filtering process to exploit systematic acquisition of all relevant information, qualitative
and quantitative, in a wide variety of contexts and forms. Seeing they get it should be part of
an enlightened approach to all aspects of corporate decision making, at all levels.
The benefits of doing a better job with less effort if carefully planned separability is used
are of general relevance – even if filtering is not an issue.
Also of general relevance is our ability to design processes which deliver complex sets of
multiple objectives if we are clear about what we want from the processes and how oversim-
plification in the ‘wrong way’ can be avoided.
Process iterations within stages by design, to seek more clarity as needed
Astro developed a formal iterative process within stage 3 to mobilise efficient searches for
value added and better understand competitors. It was crucial that Astro staff understood
why they should not expect to finish pass 1 of stage 3 ready to bid.
A target of expending 10% to 20% of the effort available to achieve completion of the first
pass of stage 3 to produce the first version of the PD3 would be reasonable for any variant
of this tale, with 80% to 90% left to enhance the bid further using further passes or parallel
complementary processes like Trevor’s ‘search for value added’. This was a simple applica-
tion of the traditional 80:20 rule – a small proportion of the effort available should be used
to identify where a large proportion of the effort should be expended. It is widely applicable
and usually usefully approached by a multiple phase approach embedded in a multiple stage
approach.
A broader implication is that iterative processes are not just about efficient knowledge
acquisition. Iterative processes can be focused on shaping any ‘project’ being developed,
and most sets of corporate opportunities and threats can be viewed as ‘projects’ which need
explicit shaping. Making a bid more attractive in Trevor’s competition focused ‘value added’
terms or Sian’s more customer needs driven ‘quality’ terms are interdependent variants of
developing a better business proposition. Arguably all potential corporate opportunities
need this kind of clarity efficient shaping process, to add value to opportunities in what-
ever way this is most usefully understood by all relevant players. Seeing opportunities get
this kind of development should be a central part of an enlightened approach to corporate
decision- making processes which board level managers understand and insist upon, with
corporate teams capable of delivering what is required.
Arguably all potential corporate threats can and should be approached in comparable
terms, usually in a single coordinated process which addresses uncertainty and underlying
complexity with respect to both opportunities and risks.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 301
Spending more time on what is ill understood, less time on what is understood
Most people involved in bidding processes have rather different expectations to those devel-
oped by this chapter’s tale. They usually expect to spend much more time on costing options using well- understood approaches. They usually expect to spend much less time on under- standing what advantages competitors can offer, and the implications for pricing decisions if
more added value cannot be developed. Astro staff should not have been expecting to spend
80% of the effort available to prepare a bid developing a specification and costing it, fol-
lowed by 20% considering the competition and pricing the bid, the usual starting position.
These ratios can and should be reversed.
A closely related issue is the crucial need for formal systems to help communication
between those who currently understand the key complexities in intuitive terms and all
other relevant players who also need this expertise in a formalised structure.
A very simple variant of these two related issues is often encountered in terms of a common
preference for people in meetings spending a lot of time discussing what is well- understood
and easily communicated, much less time on what is unfamiliar or ill understood in a struc-
tured sense. It is sometimes referred to as ‘the car park syndrome’ – because everybody
understands and wants to talk about the difficulty involved in parking their car. It is related
to a common preference for polishing well- worn tools instead of developing new tools and
the new skills needed to use them when it is not clear what these new tools might do or
how they might be used. A general implication worth emphasis is the need to focus on what
really matters most, even if doing so is uncomfortable because it is difficult. We need to con- sciously avoid just spending our time on what we are most comfortable addressing. We need
to be prepared to routinely ‘go beyond our comfort zone’, as part of aspiring to do better.
Spending more time (not less) on what is ill understood is not common practice for a
complex set of behavioural reasons. The need to do so may include coming to terms with
why the key to past organisational success is no longer working. Most organisations which
have a long- term reputation for success have had to reinvent themselves several times as
their history evolved. Astro is portrayed as an organisation needing reinvention, in this
sense directly emulating IBM’s many very different re- inventions. But addressing a deeper
understanding of current success, to leverage it to achieve greater success, is relevant to all
organisations which do not take a complacent approach to current success.
This chapter’s tale is a story of corporate reinvention, but that is just an extreme variant
of the general need to improve organisations on a scale from ‘reinvention’ to ‘continuous
improvement’. At any point on this scale what is crucial is the right mix of people, who are
collectively able to identify what needs to change and capable of responding to an enlightened
appreciation of what needs to be done. Central to corporate reinvention or more modest
ongoing incremental change for continuous improvement is getting to grips with what is cur-
rently ill understood. Also central is strong corporate leadership which is not necessarily top-
down. Nor is it likely to be bottom- up in a simple sense. Corporate change of this kind is often
most successful when it is led by intermediate levels of management with crucial teamwork
characteristics involving all levels of management, with receptive and supportive top levels of
management being particularly crucial. Ensuring that an organisation has the ‘right stuff ’ to
deal with these issues is a broad general concern the tale of this chapter tries to address.
Understanding these ideas at board level is important. Even if the board does not lead
on these issues, it needs to accept responsibility for encouraging and facilitating those who
do, with a clear understanding of what needs to be done and whether it is being achieved.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
302 Employing planning tools in practice
Top- down use of the UP concept to develop the SP for bidding
Martha had always followed global Astro Inc concerns as well as national Astro UK issues
as far as this was feasible given her current Astro roles so that she could understand ‘the big
picture’. She began her new sales manager job by deliberately avoiding becoming immersed
in the details too soon so she could develop an understanding of the strategic level issues
which were new to her new role. Her top- down UP approach to the development of an SP
for bidding began with a ‘capture the context’ phase which reflected this perspective.
For example, changing the Astro Wales culture to put Astro Wales firmly on the trajectory
she believed the rest of Astro Inc would have to follow sooner or later was part of her initial
planning concerns. She did not have any colleagues to discuss this perspective and sense of
direction with directly. But she was confident her Astro Wales colleagues would help her
to shape her ideas as well as play their role in their development and implementation. The
key was convincing them that the basis of her plans made sense from an Astro perspective
and from their own perspectives. She was also confident that both Astro UK and Astro Inc
would welcome successful bottom- up strategy development with a wider potential applica-
tion if they could see the merit involved, even if they did not arrive at the same or similar
conclusions before this became an issue.
Letting Trevor and Sian develop the PPD as a single stage basis for the three stages of her
new approach was an effective starting point for developing the details of the SP for bidding.
It captured a current best practice basis for part of her proposed stages 2 and 3, and it pro-
vided a clear basis for exposing the imbalance between Sales and Systems inputs in addition
to highlighting what was missing in relation to stages 1 to 3.
These are example aspects of Martha’s use of the capture the context phase of the UP
in a top- down manner, but they also demonstrate some flavour of the select and focus the
process, create and enhance plans, and shape the plans using models of key issues phases, as
do some further aspects of Martha’s approach. For example, encouraging Ben to develop
and present the simple, constant non- price advantage version of the stage 3 approach, and
then facilitating Trevor to lead development of the more complex but easier to use com-
petitor bid price dependent variant, might have been just good luck of the kind that tends
to follow starting with what looks like the simplest approach cast in a general framework
with everyone encouraged to test the simplifying assumptions. However, the tale implies
that Martha suspected Trevor would be more comfortable using his knowledge in the more
general framework he was credited with but deliberately avoided suggesting this framework
was a requirement or letting Ben push it at their initial full- day workshop. This was because
everybody benefitted from the increase in motivation which cascaded through Sales when
Trevor took on important aspects of the shared leadership roles, soon further shared with
other industry leads. Encouraging everybody involved to contribute to the shaping of the
new bidding process as far as possible was an explicit part of Martha’s top- down strategy, a
key aspect of her strategic clarity.
The three stage structure of the SP for bidding is itself usefully interpreted as a direct
result of the top- down UP select and focus the process approach that Martha used. It facili-
tated a carefully sequenced series of related but different component processes addressing
different sets of issues.
Stage 1 was a question and answer- based qualitative approach which was very simple and
quick if only green lights were encountered. It was only moderately complicated by col-
oured lights. But Martha’s red face test was an example of explicitly addressing potentially
complex ethical issues which matter a great deal in practice. It used a simple approach which
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 303
should lead to consistent decisions the board can take responsibility for, as they should. The
frowns test, halo test, and smiles test generalisations are also illustrative ways to empha-
sise simple but effective processes for capturing and continuing to reflect important non-
quantitative issues, avoiding some people embedding them in bias but nobody addressing
them systematically and effectively. Finding a simple way to capture important corporate
concerns in danger of being completely overlooked or ineffectively dealt with is an impor-
tant opportunity, part of the opportunity efficiency of the approach as a whole.
Stage 2 was a separate minimum clarity quantitative approach, designed with consider-
able care to achieve clarity efficiency in conjunction with stage 1 in terms of terminating the
SP for bidding as quickly as possible if the time saved was better spent in stage 3 for other
potential contracts.
Stage 3 used the same phase structure pattern as the stage 2 process, but the stage 3 pro-
cess added significant complexity in a clarity efficient manner. Understanding broadly what
the stage 3 process involved was essential to designing the stage 2 process as a minimum
clarity special case and stage 1 as a separate and somewhat different effective and efficient
front- end qualitative approach. The three stages were separable by design for sequential use
but designed as an integrated interdependent whole.
The create and enhance plans approach in top- down UP terms can also be interpreted as
part of Martha’s joint approach with Sian and Rhys. Martha was clear from the outset that
‘her top- down plans’ had to address concerns like dealing with competition between Astro
internal groups, subcontractors that were strategic partners and other subcontractors, in
addition to outsourcing some current in- house group functions when appropriate to reduce
costs, and this had to be incorporated in Astro’s SP for bidding to deal with it directly for
each specific bid.
The shape the plans using models of some key issues phase was initially focused on models
for the stage 2 and 3 processes rather than the wider plans so far as Sales and Systems staff
were concerned. These models were important, but they were not a complete answer to the
questions being addressed. Crucially, they facilitated both Sales and Systems staff working
together to add value and improve Astro competitiveness in terms of all aspects of their
bidding plans. The SP for bidding was not just about models. It was about creating and
enhancing plans involving both qualitative and quantitative approaches in order to produce
a winning bid. Shaping the quality and competitiveness of what Astro offered a customer
was much more complex than just assessing the cost of a particular assumed approach plus an associated bid price, and formally planning to do so in a clarity efficient manner for all
bids was a significant challenge.
The test the plans phase in top- down UP terms was crucial throughout, as was ‘inter-
pret the plans to exploit creativity’. ‘Implement appropriate aspects of the plans’ followed
naturally, with Martha taking considerable care to avoid rushing into limiting the flexibility
of any key aspects of the approach before testing was based on experience as well as critical
thinking and reflection.
Aspects of a bottom- up UP embedded in the SP for bidding
In Chapter 5 the ‘models of some key issues’ addressed in the shape the plans phase of the
UP were generic EOQ model variants in the first three passes. In the tale of this chapter, the
models of some key issues from a UP perspective was the three stage SP for bidding – itself
a process. Some but not all the features of a bottom- up UP approach are built into this SP
for bidding. For example, phase 3.1 (initiation of a pass in stage 3) is directly based on the
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
304 Employing planning tools in practice
‘capture the context’ and ‘select and focus the process’ phases of the UP, phase 3.5 (bid
evaluation and development) is directly based on the ‘test’, ‘interpret’, and ‘implement’
phases of the UP, and there are some clear relationships in the middle phases of these two
processes. However, the ‘select and focus the process’ aspects of the UP are very restricted
within the SP for bidding because the SP for bidding has been carefully tailored in advance
to a very specific context to gain efficiency and effectiveness within that context.
In the tale of Chapter 7 an SP for projects is explored which is much closer to an unre-
stricted UP variant. This chapter’s SP for bidding illustrates a midway position.
Comparing top- down and bottom- up UP approaches
In UP terms, ‘bottom- up’ planning means starting at a fairly low level in terms of a tactical
to strategic level of planning spectrum, decomposing in a ‘reductionist’ mode, and then
using a broad synthesis process to reassemble what was decomposed and perhaps take crea-
tive leaps further upward. The ‘analysis’ part of the process is focused at a low level initially,
with tentative movement upwards a significant possibility later.
For example, in Chapter 5 Nicola began with a focus on the two key components of the
average cost per unit of time for the inventory of controls given the use of an optimal order
quantity Q0*. Her pass 2 approach began by extending the EOQ model to address planned
stock- outs and then shifted the starting point upwards (to a broader set of concerns) by
considering Pete’s safety stock for dealing with unplanned stock- outs as well as the order
quantity, but it also decomposed the implications of safety stock using a different kind of
model, the need for different model forms to address different issues being part of the
insight necessary to deal with this higher level perspective. Her pass 3 further decomposed
holding costs, recognising that whose money was tied up was crucial, as well as introducing
price changes linked to inflation. Her creative leap at the end, further developed by Bob’s
pass 4 to the My- Place- Your- Timing approach, released TLC from the key limitations asso-
ciated with the optimal order quantity assumption.
This quite subtle moving up and down was associated with both ‘looking wider’ and
‘looking deeper’, key alternative basic UP options as discussed in Chapter 5. This looking
wider/up and looking deeper/down feature needed building into the SP for bidding, with
the looking wider aspect including going beyond quantitative analysis to address quality and
value for money competitiveness concerns, plus a level playing field for internal groups and
key subcontractors.
In UP terms, ‘top- down’ planning means starting as close to the top as is feasible, pre-
dominantly working down, although sometimes moving back up beyond the starting point
may be very important. The same looking wider, as well as looking deeper, concepts are
involved in the UP test and evaluate phases, but the predominant direction of travel is down
rather than up.
Ajit was the key top- down thinker in Chapter 5, Martha playing this role in this chapter’s
tale, as well as guiding the development of the SP for bidding in terms of bottom- up features.
In general, if it is not clear that we need to start low down as in Nicola’s case, or as near
to the top as we can as in Martha’s case, we need to make a judgement and start at whatever
level seems suitable for the introduction of UP- based EP approaches. However, working
downwards in analysis decomposition terms and the follow- on synthesis putting compo-
nents together again has to consider a broad approach to synthesis which questions starting
point working assumptions to take the process into emergent strategy issues whenever this
is appropriate.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building ‘specific processes’ 305
Using a top- down UP approach to develop an SP with strategic concerns
When using any variant of the UP concept, the core issues to be modelled when shaping
plans are often perceived as particular decisions. However, the core issues modelled can
involve the design of appropriate processes for addressing a common set of decisions. In
this chapter, the focus was the design of an appropriate SP for bidding to help Astro make
systems integration bidding decisions. Figure 6.1 illustrates the stage structure of this SP for
bidding, elaborated by Figures 6.2 and 6.4.
One useful way to interpret this SP for bidding is as a form of UP concept which has been
made specific to a particular assumed context.
The stage 3 process of Figure 6.4 is itself a specialised variant of the UP concept, designed
to develop a successful bid for potential customers when a bid is deemed worth spending
time on. Martha’s team needed to spend most of their time on the development and then
the use of this stage 3 process model, part of the overview process model of Figure 6.1.
The stage 2 process of Figure 6.2 is a special case of the stage 3 process, simplified to a
single pass using minimum clarity models to filter out potential bids which are unlikely to
provide sufficient margin.
The stage 1 process is a further filter and context capture process designed to clarify
whether any further bid development effort makes sense in terms of qualitative issues and,
if so, help to clarify what level of margin might be appropriate.
The process model as a whole, as portrayed by Figure 6.1, draws on general purpose UP
tools, and the stage 3 component process of Figure 6.4 involves a phase structure which can
be directly compared to the UP portrayal of Figure 2.1. However, this stage process and the
SP of Figure 6.1 as a whole are specific to bidding for systems integration projects by Astro;
they are not ‘universal processes’.
The ‘specific’ nature of this SP for bidding means that its loss of generality carries a price.
Each time the Astro SP for bidding is used it may need some further shaping in the context
of background ongoing use of the higher level UP. This further shaping may be of a minor
kind, to reflect modestly different features of a specific bid. But it might be of a major kind
to reflect very different features or unfamiliar characteristics of a novel kind of bidding con-
text or significant lessons learned from bids just lost or won. A radically different type of bid
or a significant change in the corporate context may require a major rethink. This price is
significant but worth paying because it buys an increase in efficiency and effectiveness in the
context the process was designed for.
If the SP for bidding of Figure 6.1 is viewed as a ‘process model’ created by using the UP
of Figure 2.1 applied in top- down and bottom- up terms, it becomes clear that the team of
people using the UP need all the associated requisite capabilities, but routine users of the SP
for bidding do not require these skills. Martha, Ben, and Trevor working as team were the
central contributors to the management of the development of this specific process, Mar-
tha largely on her own managing the higher order top- down use of a UP to define the SP
for bidding, Ben leading some of the bottom- up UP based content of the SP for bidding.
These capabilities in terms of effective use of the UP concept are in the background of this
chapter’s development because the chapter structure had to focus on the nature of the SP
for bidding, but this does not mean these issues are not important. Most of the Sales, Sys-
tems, and other staff involved did not need the same UP concept user capability skills as the
core team developing the basis of the SP for bidding, but it was very important everybody
involved had a reasonable degree of understanding of what was going on, and Martha’s
leadership with a top- down UP perspective was crucial. Martha and Ben characterise those
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
306 Employing planning tools in practice
who need this kind of UP based understanding in depth, but Trevor also needed to develop
some directly relevant capabilities, and it will help all those involved in using the SP for bid-
ding if they have strategic clarity about what these capabilities involve at a level of sophistica-
tion appropriate to their role.
When the tale of this chapter began, Martha had strategic clarity in terms of a broadly
defined game plan which evolved, shaped by her top- down UP approach. She knew that she
had a change programme to manage from the outset, and bidding for systems integration
projects was recognised as a useful vehicle for this broader ambition early in the tale. But
Martha did not know how the story would evolve at the outset. The conceptual framework
she used is usefully interpreted as a UP applied to building an SP for bidding, but mapping
higher level formal processes onto successful practice is rarely simple or straightforward.
All these ideas can be applied in any area of an organisation’s operations which needs
change. If you want to apply them yourself, or have others apply them for you, the way
forward may not be obvious or straightforward. However, developing new tools to deal
with what your organisation does not understand very well can be much more fruitful than
polishing existing tools.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Planning for ‘projects’ has received a lot of attention in the management literature. How-
ever, when drawing on this literature, it is often worth remembering that some common
practices receiving a lot of attention adopt MUCH narrower interpretations of both ‘pro- jects’ and associated ‘project risk management’ than the perspective adopted in this book.
In this book ‘projects’ are VERY broadly defined, and it is useful to understand the nature of this breadth in three separate senses:
1 embracing not only the creation of physical assets like buildings but also the manage-
ment of change which may not involve a physical asset, as in a culture change pro-
gramme or a decision- making process;
2 accommodating distinctions among projects, programmes, and portfolios of projects
and programmes, when this is important, but for simplicity sticking to ‘projects’ termi-
nology to cover all three most of the time;
3 addressing the whole project lifecycle, defined in terms of the lifecycle of the asset pro-
duced by the project, from concept initiation to termination of the asset’s relevance.
Developing a ‘specific uncertainty and complexity management process for projects’, an ‘SP for projects’ concept, is the central mission for this chapter. The SP for projects is an opera-
tional tool which also serves very broad background conceptual roles – it is a very high level
gateway concept as well as an operational tool for direct application.
Chapter 1 introduced project planning at an overview level and developed a ‘four Fs’
concept in general framing assumption terms, a high level gateway concept involving four
component frameworks which are each lower level gateway concepts: a project lifecycle
framework, the seven Ws framework, a goals– plans relationships framework, and a process
framework that builds on the other three frameworks. This process for project planning is
given an operational form in this chapter, as the SP for projects, and it makes use of opera-
tional examples of the other three components of the four F, using working assumptions
suitable for the example context of the tale. The resulting operational form for all four Fs
is relevant to all the organisation’s projects in the tale. It is specific to the context of the
tale, but it is an illustrative example with a very general nature, and each of the four Fs
operational frameworks discussed can have their working assumptions tested and adjusted
as necessary for all contexts of interest to you and your organisation.
Most of the issues addressed in this chapter draw on a wide range of both successful and
unsuccessful engagements with organisations in a wide range of industries. In the successful
category relevant examples which stand out include BP International, National Power, Sir
William Halcrow and Partners, Severn Trent Water Authority, British Maritime Technology,
7 Adapting ‘generic processes’ – a project planning example
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
308 Employing planning tools in practice
Gulf Canada, Petro- Canada, Fluor Engineers and Contractors Inc, Statoil, NatWest Bank,
Consorzio Venezia Nuova, IBM UK and the UK Highways Agency. In the mixture of suc-
cess and failure category examples which stand out include the Alaska Power Authority,
UK Nirex and the UK Ministry of Defence (MoD). Each of these engagements taught me
important lessons and helped to shape the approaches discussed. The reasons will be indi-
cated when directly relevant and generally useful.
Even if you have a limited interest in project planning at present, the basic ideas explored
in this chapter are a core component of the overall EP framework, and there are several
useful implications for all aspects of decision making in all organisations. For example, the
SP for bidding developed in Chapter 6 can be usefully viewed as a special case of the SP for
projects developed in this chapter. In addition, the approach which Martha took to develop-
ing the SP for bidding can be usefully viewed as a further higher order application of the SP
for projects approach developed in this chapter. That is, Chapter 6 uses the SP for projects
developed in this chapter in two different ways at two levels as part of its use of the UP
concept at various levels. Deferring formal consideration of project management until now
allows us to build on concepts explored in Chapters 5 and 6 as well as Part 1 at a convenient
point in the development of EP as a whole.
Generic processes used for project planning
Project management as a separate management profession with its own professional socie-
ties has been growing at a significant rate since the 1950s. There has been a related growth
in the associated professional knowledge base, including a wide range of models, ‘tech-
niques’ and ‘generic processes’. However, for what might be perceived as territorial rea-
sons in part, some professional project management societies tend to promote a relatively
narrow view of projects, with what might be seen as a natural tendency to overemphasise
the current roles of the majority of the practicing project managers who belong to the
societies, to some extent inhibiting evolution to broader roles for ‘project management’ as
a central aspect of all management. Everyone interested in management decision- making
needs to understand the techniques and generic processes involved in what needs to be
done terms.
‘Techniques’ for project planning can be interpreted as simple SPs: generic models embed-
ded in a simple associated process which have been generalised to accommodate comparable
contexts in all organisations. For example, CPM (Critical Path Method) is often referred
to as a ‘technique’, sometimes using the alternative label Critical Path Analysis (CPA). In
its initial basic form CPM involves a finish- to- start activity- on- arrow precedent relationship
network diagram model. This model is embedded in a simple iterative process for planning
project activities. It was initially developed by a team of consultants for Westinghouse in the
US in the 1950s, primarily for factory construction and refurbishment projects. It was very
quickly adopted as a basic tool by most project planners and used across a very wide range of
organisations and projects, along with PERT (Program Evaluation and Review Technique).
PERT is a simple stochastic model variant of CPM, developed by another team of consult-
ants for the US Navy and the Polaris Missile project. For an early but still useful introduc-
tion to both and subsequent generalisations see Project Management with CPM and PERT (Moder and Philips, 1970). CPM and PERT both exploited newly available computer sup-
port approaches and built on earlier models. For example, Henry Gantt’s bar chart, widely
used for project planning in the early 1900s, became the basis of the ‘linked bar chart’ con-
cept, illustrated earlier in this book in Figure 4.4 and employed again later in this chapter.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 309
Generic processes are a more recent development. A broad interpretation of generic pro-
cesses is that they are SPs which have been generalised beyond the scope of specific techniques
like CPM or PERT to accommodate much broader families of relevant models and processes,
and they should help users to choose appropriately from the available families of options.
If this broad interpretation of generic processes is used, an important feature is some vari-
ant of the UP ‘select and focus the process’ phase being incorporation in the associated SP
to explicitly facilitate the use of different variants of the generic process for different kinds of models for different projects in different contexts. For example, the Association for Project
Management (APM) PRAM Project Risk Analysis and Management Guide (APM, 1997) involved a working party of about 20 people synthesising what they saw as best practice
processes across all UK users of project risk management models, techniques, and processes.
From my perspective, the inclusion of a ‘focus the process’ phase variant of the UP select
and focus the process phase was a defining characteristic of the PRAM approach, making it
unequivocally a generic process. I drafted the process chapter of the 1997 PRAM guide, and
co- authored a 2004 version, having earlier served as the founding chair of the APM Specific
Interest Group for Project Risk Management which produced the PRAM Guides as part of
its evolving role (Hillson, 2012).
The initial trigger for including an explicit focus the process phase in PRAM, the first
time I used this concept explicitly, was addressing the need to accommodate a wide range
of views about what models and associated processes work best in different contexts, based
on an implicit assumption that members of the working party producing the PRAM Guide
held different views because we had worked in different contexts, and agreement on ‘best’
practice would be reached this way. Alternative views were strongly held and firmly advo-
cated, sometimes with more heat than light, although we all remained civil, empathetic, and
friendly. It is now clear that achieving a clear consensus amongst all those involved in PRAM
and other guides about ‘best’, ‘good’, and ‘bad’ practice was a crucial goal at the time
which we failed to achieve. It is still an ongoing issue in a PRAM context, with comparable
differences of opinion across the project management profession, only partially reflected by
differences between guides. The way PRAM as published uses the focus the process phase
was agreed and remains an important evolutionary step in generic processes, but the PRAM
focus the process phase did not resolve all the differences of opinion which triggered mak-
ing it explicit.
The aspect of project management which I have been most closely associated with since
the 1980s is commonly referred to as ‘project risk management’. However, my interpreta-
tion of what needs to be done under this heading is more accurately referred to as ‘project
uncertainty and complexity management which addresses opportunity first, risk second, in a
manner fully integrated with project management as a whole’. For example, How to Manage Project Opportunity and Risk: Why Uncertainty Management can be a Much Better Approach than Risk Management (Chapman and Ward, 2011) provides a generic process referred to as a ‘performance uncertainty management process’, contracted to ‘PUMP’. This 2011 book
is the third edition of Project Risk Management: Processes, Techniques and Insights (Chapman and Ward, 1997, 2003), retitled to make it clear that our interpretation of project risk man-
agement is not the conventional interpretation. All three editions were based on an evolving
synthesis of SPs for projects and other closely coupled frameworks developed for a range
of client organisations, plus my experience and Stephen Ward’s experience contributing to
the working parties producing several ‘generic’ project risk management guides, including
PRAM (APM, 1997, 2004) and Risk Analysis and Management for Projects – RAMP (ICE and IFoA, 1998, 2005).
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
310 Employing planning tools in practice
Project risk management generic processes are very widely deployed. The most widely
used is probably the Project Management Institute (PMI) approach described in the PMBoK
Guides (PMI, 2008, 2009, 2013, 2017a), but there are many others.
Both relatively simple techniques like CPM and PERT and relatively sophisticated generic
processes like PRAM and PUMP are usefully seen as SPs which are relevant to project man-
agement. An important common feature of all SPs is they require initial development using
a UP concept (or comparable general process – implicitly if not explicitly). Furthermore,
their ongoing use in different contexts requires adaptation using a UP concept (or com-
parable general process) unless some variant of the UP concept is effectively built into the
SP to make it ‘universal’ in the full set of contexts where it might be used. Making an SP
for projects ‘universal’ in all project management contexts is a strategy explored in an initial
manner towards the end of this chapter, but the focus is transforming a very broadly defined
generic process into an even broader SP for projects within a particular organisation.
This chapter uses the UP concept plus other relevant concepts to generate an SP for
projects which is suitable for all aspects of project planning within an illustrative specific
organisation. In doing so, the SP for projects developed in this chapter makes extensive use
of the performance uncertainty management process, or PUMP, developed more fully by
Chapman and Ward (2011) – PUMP is a central component of the overall approach taken.
However, there are two very fundamental and crucial differences in perspective.
First, PUMP uses what is referred to as a ‘performance lens’. The purpose of this per-
formance lens concept is emphasising a focus on project performance when PUMP is used
to shape project plans. What is crucial in the context of this chapter is appreciating that the PUMP approach assumes that the plans were probably initially created and developed prior to ‘project risk management’ shaping of project plans using separate processes involving separate people if PUMP is replacing widespread common practice. The rationale was accommodating the common perspective that project risk management should contribute to project man-
agement but maintain a separable role executed by separate people. PUMP as developed
by Chapman and Ward (2011) introduces, but does not fully develop or exploit, a linked
‘knowledge lens’ associated with ‘What else do we need to know to get to the next stage
of the overall project planning process?’ Although Chapman and Ward (2011) argue that
full integration with the rest of project management is essential, using both lenses, the case
for doing so is made from a perspective that those familiar with common practice project
risk management approaches can identify with and feel comfortable with. This chapter uses
an explicit ‘planning lens’ which operates as a performance lens plus a knowledge lens.
The intention is fully embedding the PUMP concepts in an SP for projects concept which
addresses all relevant aspects of project planning in a direct manner from an EP perspective.
The starting position is a fully integrated approach which addresses creating and enhancing
all relevant plans as well as shaping them using PUMP concepts. The embedding of PUMPs
in a holistic approach to project planning argued for by Chapman and Ward (2011) is imple-
mented directly in the tale of this chapter.
Second, the ‘basic’ PUMP discussion that Chapman and Ward (2011) start with is
focused on the execution and delivery strategy stage of the project lifecycle. In part this
was because the origins of PUMP in terms of my early work with BP were focused on this
stage of the project lifecycle, and I had a much wider range of experience with this stage
than other stages. A more important second reason was most members of the PRAM work-
ing party (which significantly influenced PUMP in terms of mid- 1990s developments) were
more concerned with this stage or the later detailed planning, execution, and delivery stages,
as were most APM members. An underlying third reason was the project management
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 311
profession as a whole tends to have minimal involvement or interest in the early portion of
project lifecycles, with notable exceptions like RAMP (ICE and IFoA, 1998, 2005). This
chapter explicitly addresses the implications of the third reason. It starts with a focus on the
concept strategy stage of the project lifecycle. The use of PUMPs from the outset of the
project lifecycle, which is argued for by both Chapman and Ward (2011) and the RAMP
guides, is implemented directly in the tale of this chapter.
The SP for projects approach developed in this chapter might be interpreted by some
advocates of project risk management as a takeover of project planning which is likely to
offend project managers. On this basis they might judge it as ‘a bridge too far’ (a World War
II analogy involving a strategic failure because a stretch target proved too ambitious). Some
people concerned with project management as a whole who have a limited interest in project
risk management may share this view.
Furthermore, the SP for projects approach developed in this chapter might be interpreted
by some advocates of project management as a significant unwarranted extension of project
planning and management, likely to offend many people project managers report to or work
alongside, a second bridge too far. However, the intended and much more fruitful inter-
pretation is a long overdue takeover of project risk management, including PUMP and all
other generic processes like PRAM and RAMP, by a more enlightened approach to planning
projects which incorporates coordinated involvement of operations and corporate manage-
ment with a much broader and more effective form of project management.
This book promotes the perspective that enlightened planning of projects should start
with project concept creation, involving all relevant players, addressing all relevant uncer-
tainty in a systematic manner. It then has to build on this foundation throughout the whole
lifecycle of the asset created by the project. Sometimes the conventional ‘from the cradle
to the grave’ expression may be apt, but usually a much more appropriate phrase is ‘from
conception to “legacy” – for however long the legacy implications need considering’. Net
Present Value (NPV) issues become particularly crucial when long time horizons need effec-
tive treatment, and legacy issues can involve relevant time horizon issues which are infinite
for all practical purposes.
This EP starting position conflicts with the traditional position of some project manage-
ment professionals. An example is those who take a contractors’ perspective and see their
role in terms of the project execution and delivery aspects of project lifecycles, assuming
that project owners know what they want and that after delivery of the asset is achieved the
project is complete – the rest of the asset lifecycle is somebody else’s problem.
This EP starting position also conflicts with the idea that project risk management can
be treated as a separable silo within the project management profession, and the project
management profession can be treated as a silo separable from operations and corporate
management.
Key implications of eliminating this nested silo approach need to be seen as important
by everyone who is interested in any areas and aspects of management in all organisations.
The concerns addressed in this chapter are not just project management concerns – they are
central to all planning.
The need to clarify ambiguity
One useful aspect of using multiple stages in a project lifecycle framework is a framework
for explicitly cultivating an enlightened view of ambiguity which involves seeking ‘clarified
ambiguity’. This involves recognising that some ambiguity needs early resolution, but leaving
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
312 Employing planning tools in practice
‘the right kind of ambiguity’ in plans may provide very important benefits, and we have no
choice about other kinds of ambiguity. Put slightly differently – not all ambiguity is equally
bad. Some ambiguity may not matter, some may be irritating but relatively harmless, some
may be serious and resolvable but only at considerable cost, and some is inevitable even if it
is a problem. This implies a sound case for formally planning a roughly optimised sequence
and timing for addressing all the different kinds of ambiguity, with a view to clarifying it in
an opportunity efficient manner as part of our clarity efficiency concept.
Five basic reasons for developing this clarified ambiguity perspective are the following:
1 Some aspects of both strategy and tactics are usefully given priority because of inherent
precedence relationships – it is simply not possible to do everything at one time, and
some things are better done earlier, leaving other things until later.
2 Telling people what you want done, and why, is usually much more effective and effi-
cient than telling them what to do in detail, PROVIDED they are competent and imagi- native, and you can trust them to pursue aligned interests. The basis of this idea can be linked to a useful quotation attributed to General George Patton (Ambrose, 1994):
‘Tell people what you want done, not how to do it, and you will be amazed by their
ingenuity’. But it also involves very important provisos. 3 It is always cost- effective to clarify strategic choices before spending a lot of time on
detailed planning for implementation purposes, although we do need to be aware that sometimes ‘the devil is in the detail’. This idea can be usefully generalised to include ‘angels in the detail’. Strategic clarity includes clarity about how to capture the oppor-
tunities which different kinds of ‘angels’ offer, as well as how to deal with the ‘devils’,
and this aspect of strategic clarity is a very important corporate capability.
4 At the outset of a substantial project all plans are necessarily very ambiguous. The
people who are being asked, ‘What will it cost?’ and ‘How long will it take?’ to deliver
objectives which those formulating the project have yet to fully define face ambiguity
which is so large, and so uncomfortable, that most organisations have some degree of
‘a culture of denial’ about this kind and level of uncertainty. This is understandable,
but it can be seriously dysfunctional. The first step to avoid the problems which usually
flow from this ‘uncertainty denial’ is recognising the size and shape of the ambigu-
ity involved, its implications, how best to confront it explicitly and directly, and how
best to exploit potential positive aspects as well as avoiding potential negative aspects.
Using interval (range- based) estimates, with any key ‘conditions’ (associated assump-
tions) clearly stated and understood, are essential aspects of this. So are the ABCs of
targets, with clear distinctions drawn between all three target types and expected values
whenever this is appropriate – most of the time in practice.
5 Usually, as soon as any single- value ‘point estimates’ of a project’s profit, cost, dura-
tion or quality measures are agreed on, opportunity efficient decision making starts to
become increasingly handicapped until it becomes obvious that the point estimates
adopted have become inappropriate and need to be changed. This is because the focus
soon becomes staying within bounds on key performance measures when these bounds
have become inappropriate. Frequently, cycles of successive changes which are inher-
ently opportunity inefficient to an obvious extent become ‘the norm’. Along the way,
opportunities are lost and threats are dealt with ineffectively. Furthermore, usually the
process of successively changing point estimates results in an inevitable loss of trust and
permanently damaged relationships. All these effects are dysfunctional for some key
parties, although other parties may prosper from dysfunctional relationships, and may
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 313
be motivated to encourage them. The biggest cost may be the opportunity inefficiency
associated with failing to pursue aggressive stretch targets to manage good luck, failing
to respond effectively and efficiently to bad luck, taking inappropriate amounts of risk
(too much or too little), and generally failing to even understand the opportunities or
risks involved.
To respond to reasons four and five an enlightened approach to ambiguity always avoids point estimates unless appropriate aspirational targets and commitment targets are identical
(zero provision plus zero contingency is appropriate). This is because point estimates fail to
address any kind of uncertainty, never mind facilitating monitoring and managing ambigu-
ity with clarity efficiency. If significant ranges are a feature of early estimates, this clarifies
the need to systematically reduce associated uncertainty, making VERY clear distinctions between uncertainty and risk. Reasons one, two and three underlie reasons four and five, and collectively all five reasons may help to clarify why the very general notion
we always need to be clear about the nature, size and shape of all the ambiguity involved
is not the paradox it may sound. It is the basis of a useful ‘clarified ambiguity’ concept. You
may even find a version of this notion a useful mantra.
Some aspects of clarified ambiguity underlie the discussion of Chapters 3 and 4. In
addition, some features were developed in the three stage SP for bidding of Chapter 6.
This chapter makes the clarified ambiguity concept explicit as a useful way to visualise the
management of knowledge as a project evolves in a more traditional project context than
‘forerunner to a traditional project’ operations management ‘bidding project’ context of
Chapter 6. There are closely coupled implications for all operations and strategic manage-
ment processes.
The context and characters of this chapter’s tale
The context used for the tale of this chapter draws upon my experience as a non- executive
director of Southern Water, for six years beginning in 1997. This was a period of significant
changes in the UK water and sewage utility sector. It also draws on important positive fea-
tures of Southern Water approaches, including capability- culture assets of their executive
directors throughout my involvement, but a disguised version of Southern Water is not
involved.
The water and sewage utility context of this chapter is useful because the key features of
project planning ‘best’, ‘good’ and ‘bad’ practice which this particular example context helps
to clarify have general implications for most decision making in all organisations, including
the importance of dealing appropriately with the cost of capital and other relevant factors
when determining the discount rates used to shape what projects deliver and how their
execution is approached as well as the decisions about which projects should be undertaken.
The main characters in the tale of this chapter are seven of the directors on the board of
Water and Sewage Limited (WSL):
Paul, Projects Director, recently appointed;
Ollie, Operations Director;
Frank, Finance Director;
Richard, Regulation Director;
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
314 Employing planning tools in practice
Curt, Customer Relations Director;
Michael, Managing Director;
Charles, Chairman of the Board.
When he was appointed Projects Director of WSL in 2012, Paul had 30 years of project
experience, mainly with design and construction contractors. His initial training was civil
engineering, and if asked what he did in casual conversation he replied, ‘civil engineering’.
But he had always been a serious student of all aspects of management relevant to his job.
WSL had been privatised in the early 1990s, along with all other UK water and sewage
providers. Like many of the other UK water and sewage companies, WSL used bond fund-
ing for about 80% of its capital value to gear up the return on shareholder equity capital in
the face of multiple regulator pressures on profits from two directions: constraining prices
to look after water and sewage customer interests in terms of cost and constraining water
supply and wastewater quality to look after health and environmental interests.
Paul’s most recent job was programme manager for a major sewage scheme constructed
by a civil engineering contractor for WSL. This project was completed on time and under
budget, part of the reason Paul was head- hunted for his new post, his first in a ‘client- side’
organisation. Paul had limited earlier experience in the water and sewage industry but sig-
nificant relevant earlier experience in the UK oil and gas industry and some road and rail
construction industry experience in other European countries and in the US. Crucially, he
had introduced significant changes in the practices and processes of his previous employer
which were highly relevant to WSL needs.
Paul had attended a 2011 IPMA (International Project Management Association)
advanced training programme in Milan based on Chapman and Ward (2011) and under-
stood its concepts. He had attended an earlier version of the same course in Copenhagen
a decade earlier, and he became an experienced user of early variants of most EP concepts
soon after.
After taking up his WSL appointment, one of Paul’s first actions was to hold an initial
workshop for his ‘Projects Group’ staff to explain:
1 the role of ‘project management’ within WSL as he saw it,
2 the role of the Projects Group within WSL and his role as he saw it,
3 a new Projects Group structure they would target as of that day,
4 some very basic problems the new Projects Group structure would address, and
5 four frameworks which all WSL projects under his control would use, the four Fs.
The role of project management within WSL
Paul began his initial workshop by suggesting that from the perspective of all those seeking
overall corporate opportunity efficiency, project management had to be seen as an inte-
gral part of three related perspectives on management: corporate management, operations
management, and project management. He used his variants of the EP language devel-
oped earlier in this book to a significant extent, always explaining it to his colleagues when
appropriate in language they were comfortable with. The concepts involved in terms like
‘opportunity efficiency’ will not be considered again in this chapter unless specific aspects
need emphasis, but the terms ‘corporate management’, ‘operations management’, and ‘pro-
ject management’ are worth brief reconsideration now, with a focus on the concerns which
Paul emphasised.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 315
Operations management involved a creative top- down approach to business as usual, inte-
grated downwards with all appropriate bottom- up planning of business as usual, integrated
sideways with all projects concerned with new relevant assets and other related corporate
changes, and integrated sideways and upwards with corporate management.
Corporate management involved a creative top- down approach to corporate strategy
formation, fully integrated with all relevant bottom- up approaches. It also involved over-
all corporate responsibility for governance and all corporate resources and competencies.
Not all corporate management was at a ‘higher’ level – a simple example at a ‘lower’ level
was human resource management involving operations management roles like water metre
reading.
One key role of project management was the change management component which
interfaced corporate and operations management. However, many of the ongoing opera-
tions and corporate management activities within an organisation could also be given a
project management perspective, including decision making and change management. ‘Pro-
jects’ in this very broad sense included programmes and portfolios of projects.
The role of the Projects Group within WSL
All WSL staff in the Projects Group reported to Paul. The remit of the WSL Projects Group
was projects involving the construction of physical plant and facilities – engineering projects
like new water mains and sewage collection pipes, new water treatment plants and sewage
treatment works. The Projects Group’s remit did not extend to projects undertaken under
the auspices of other directors. For example, information systems projects were the respon-
sibility of the information technology director plus the directors whose systems were being
addressed, like Finance. Nor did the Projects Group have any responsibility for project
management in the very broad sense which included decision making and change manage-
ment in operations or corporate planning contexts. In the longer term, Paul wanted to
influence all WSL project management practice, by examples of Projects Group best practice and by helping other groups to embed compatible project management practices in their
groups to facilitate joint working. His long- term overall ambitions included full integration
of best practice project management with corporate management and operations manage-
ment. However, Paul had no wish to redefine the current boundaries of responsibilities for
the Projects Group, and he saw it as very important that all Projects Group staff understood
both the wider possible implications of the changes Paul would be making and the consider-
able sensitivities associated with group and departmental boundaries that might be triggered
by some of the changes he would stimulate. All members of his group needed to understand
and support his concerted efforts to ensure that nobody outside the Projects Group felt
inappropriately threatened by the significant changes Paul was planning to make.
Paul explained that he saw his WSL board membership role in terms of contributing to
overall WSL governance and some aspects of overall WSL strategy shaping in addition to
reporting on his responsibilities as head of the Projects Group. He indicated that Ollie, the
operations director and head of the Operations Group, had a directly comparable role for
operations in the physical plant and facilities sense. However, Frank, the finance director;
Richard, the regulation director; Curt, the customer relations director; Ian, the information
technology director; and Hannah, the human resources director, all had responsibilities at
board and group or department levels which were somewhat different in terms of interac-
tions. All these interfaces had to be managed by Charles, the chairman of the board (a
non- executive chair appointed by the company which owned WSL), Michael, the managing
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
316 Employing planning tools in practice
director (the chief executive officer or CEO), plus all other board members working as an
integrated group.
The new structure of the WSL Projects Group
Paul then indicated that his view of the role of the Projects Group meant that as of that day
they needed to start working together as a single ‘Projects Group planning team’ – the ‘PG
planners’ or the ‘PG team’ when these contractions were useful.
By PG ‘planners’ he meant PG ‘EP planners’ using an ‘enhanced planning’ label for the ‘enlightened planning’ concept developed earlier in this book they would all adopt as their basic planning framework, which he would outline that day and continue to clarify in a
project planning context. He made it clear that there was a need to distinguish current WSL
‘planning’ from a very different EP approach they would all adopt starting that day, but it
was important to avoid any hint of pretention and possible associated empire building which
might threaten other groups and reduce their willingness to collaborate, his rationale for
interpreting EP as ‘enhanced planning’ for WSL purposes.
A key feature of this new PG team was eliminating the existing departmental boundaries
which currently placed everybody into one of three departments: the estimating depart-
ment, the planning department, or the risk management department. Currently staff in
these three departments initially became involved in projects separately, in a sequence which
began with estimating, put risk management in the middle, and put planning supported by
further estimating and risk management at the end.
From the outset the PG team was going to provide cost estimates as a part of a revised
fully integrated planning remit. The new remit involved uncertainty and complexity man-
agement which included risk management plus opportunity management built into all plan-
ning and costing involving the PG team from the concept stage initiation of projects. All
existing skills and experience would need integration with new concepts and capabilities
in a new framework. In the short run, the existing management support functions would
remain in place, but as of that day they would start working together in teams which cut
across departmental boundaries, developing new processes which helped to integrate their
toolsets, skill sets and mindsets.
The new PG team structure would involve the current Projects Group heads of estimat-
ing, planning and risk management being promoted to senior PG planning managers, with significantly broader and more flexible remits than at present. He had spoken to all three
departmental heads earlier and had their support. All other PG staff – previously in one of
these three groups – would be given comparable PG team positions. He had not spoken to
any of them individually yet, because he thought it best to let them all know at the same
time as soon as possible, but he would have individual meetings with everyone as soon as
this could be arranged.
The basic idea in terms of staff structure was breaking down the current departmental
silos, moving to a much more flexible multi- skilled team approach, initially with larger teams
than would be the case later, to facilitate learning from each other and joint ‘learning by
doing’ while they were transforming the way the Projects Group operated. There would
be scope for ‘promoting’ a number of people because most people would need a wider
and more flexible skill set in relation to new roles, but some people with specialist expertise
would be less affected than others. Meeting mutually agreed- on personal capability devel-
opment goals, with sustained WSL support from him and all senior PG managers, was the
general idea.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 317
Paul made it clear that he would work with everyone involved to ensure that the new
more flexible structure worked as smoothly as possible from the outset. He would always
welcome one- to- one confidential discussions about problems of any kind, and he was look-
ing forward to ensuring that tensions created by the changes were constructively and crea-
tively resolved. He emphasised that in private conversations open and candid constructive
criticism of his approaches would always be welcomed, but the direction and degree of
change being pursued would have to be embraced positively to remain a member of the
PG team.
Paul indicated that more workshops and some training exercises would be an early prior-
ity, but to some extent they would all have to learn by doing, initially working with larger
teams than usual to ensure cross- fertilisation of ideas from the currently different perspec-
tives of planning, costing, and risk specialists. Their first few exercises should be seen as
development projects initially and then demonstration projects. These early projects would
be investments in the development of their new working practices, which could then be used
to demonstrate what was involved. The initial focus would be on effectiveness, with later
refinement also concerned with achieving efficiency.
Paul explained that he had outlined his proposed approach to restructuring the Pro-
jects Group and the way it would operate to Michael, their managing director, prior to
accepting his appointment. He now had Michael’s approval for a reasonably detailed change
management programme, which included changing the current PG departmental structure
and that day’s workshop. Charles, chairman of their board, was fully informed and support-
ive. But Paul had not yet discussed his approach with other directors.
Basic problems the new structure would address
Paul indicated that the new PG team structure was just part of the move to a new PG
approach to projects involving a wide range of changes with advantages and implications
he would gradually explain in detail. He would start by indicating some basic aspects of
the underlying ‘mess’ which the new PG team structure would help the new PG approach
address. By ‘mess’ he meant a set of important basic ‘problems’ which were fundamental
and not easily separated or decomposed because of complex interconnections, but they
could also use a plain English interpretation of ‘mess’.
Currently, when a new project was being conceived, early PG team involvement was lim-
ited to the estimating department providing a point (single- value) estimate of construction
cost. Later the risk management department used event- based common-practice project
risk management techniques to provide a cost estimate ‘risk adjustment’, also a point esti-
mate, derived via the expected value of a probability distribution defined by ‘risk’ probabili-
ties and their impact probability distributions. Later still, following on from considerable
expenditure on project development in terms of operating and design features, the Plan-
ning Department became involved, with further follow- on involvement from the Estimating
Department and the Risk Management Department. About this time serious cost escalation
usually set in, building on some earlier cost escalation while design issues were addressed.
By the time a contract was signed for execution and delivery, the original ‘risk adjusted esti-
mate’ had almost always been increased, by an average of about 50% with a range of about
20% to 150%. Even then most projects still came in over budget and late, sometimes badly
over budget and seriously late.
The recent WSL sewage plant project which he had been involved in as the contractor’s
programme manager came in under budget and on time, but this was an exception to the
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
318 Employing planning tools in practice
norm, and the contracted budget was 60% more than the original board- approved ‘risk
adjusted cost estimate’. Some of this 60% increase arose because the head of planning in the
Projects Group had become involved after the initial budget was approved, and Planning
Department input plus follow- on input from the other two Projects Group departments
had consolidated a cost escalation process triggered by Planning Department input which
should have preceded setting the initial budget. But Paul had been forced to make a con-
vincing case for a significant further increase, his ability to justify doing so, and then deliver
the project within budget and on time, being part of the reason for his appointment to his
current post.
These large cost increases after initial project approval based on a ‘risk adjusted estimate’
were not just a minor common problem to be accommodated. They were significant symp-
toms of a large set of serious underlying concerns. Paul understood why (in terms which
build on your understanding of all previous chapters), but he needed to transfer this under-
standing to his team members gradually, layer by layer, building on their current base point.
One of the concerns this systematic cost growth revealed was a lack of WSL planning
which fully integrated the management of opportunity, risk, uncertainty and complexity
from the outset. This, in turn, required fully integrating not just the skill sets and toolsets of
the PG team but also the skill sets and toolsets of other WSL groups and departments and
WSL contractors. What was involved was not a single ‘problem’ with a simple solution – it
was a very large mess of interconnected sets of problems which needed holistic resolution.
The PG team would have to undertake a portfolio of programmes to resolve this mess in
conjunction with other WSL groups. Paul had clear ideas about the resolution of the mess
in terms of the overall shape of his strategy and targets, but the details were unclear, the
strategy was still fluid, and some of the WSL staff they were going to have to work with
closely were not yet involved.
Paul indicated that this mess was the result of some common corporate practices which he saw as problems to be resolved or dissolved from a systems analysis perspective, not solved
individually. He made it very clear that some common practices were very good practice, but some common practices were not good practices, and some of the common practices adopted by WSL were so bad they had to be changed as soon as possible.
In his view, endorsed by the WSL board, ‘the current WSL overall approach to Planning
Group projects was not fit for purpose, and its transformation needed immediate attention,
beginning with the WSL approach to the concept strategy stage of the project lifecycle for
all projects within the remit of the Projects Group’. Paul made it clear that Projects Group
were not fully responsible for this situation, but he had been hired to sort it out, he was
directly accountable for sorting it out, and he was determined to see to it that all the needed
changes were made within a reasonable timeframe.
The first step to resolve the aspects of this mess which he had direct control over was
starting to move to a new PG team structure. The next step was everybody in WSL starting
to understand four crucial frameworks which he would refer to as the ‘four Fs’. The four
Fs were the basis of the new PG team approach to project planning they would be working
with. Each of the four component frameworks was usefully seen as a ‘tool’ in the conceptual
model sense, with some crucial operational toolset implications. Everybody within WSL – including all board members – would need to make direct use of the first two of these new
frameworks. The PG team would need to make direct use of all four Fs in a fully integrated
manner.
Each of the four Fs would now be considered, in turn, in some detail: a project lifecycle
framework, the seven Ws framework, a goals– plans relationships framework, and a specific
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 319
planning process framework for projects (SP for projects) with an initial focus on the front-
end strategy development stages of the project lifecycle.
A project lifecycle framework
Paul indicated that a project lifecycle was the first framework which WSL as a whole needed
to understand in terms of a common portrayal for all WSL projects involving the Projects
Group because the rest of the four Fs built upon this foundation. He would refer to a ‘pro- ject’ lifecycle to follow convention, but it was the ‘asset’ lifecycle he was talking about – the lifecycle of the asset created by the project from its conception until its disposal or in some
cases until the end of its ‘legacy’. In their case, physical assets like water mains and sewage
works were the immediate focus. However, corporate change management projects also
produced an ‘asset’ which warranted design and execution and utilisation planning within
the same project lifecycle concept.
The basic role a project lifecycle concept serves is providing a time- based framework for
staged evolution of ‘the project’ so that different parties can play different roles in an effec-
tive and efficient sequence as inherent ambiguity in a project is systematically clarified.
A traditional four stage project lifecycle structure
To clarify what was involved, Paul explained that he would initially consider a traditional
four stage view of a project’s lifecycle from a client organisation’s perspective, incorporating
stated assumptions about the dominant management aspects of each stage. Because they
would have to consider several project lifecycle structure definitions, he would refer to his
initial example as ‘stage structure one’, contracted to ‘SS1’. This SS1 version of a traditional
four- stage project lifecycle from a client perspective used an asset lifecycle basis, because from a client perspective the utilisation stage was central to the rationale for the project. He
then showed his team the SS1 lifecycle portrayal of Table 7.1.
His first observation was that a common practice variant often used by project manage-
ment professionals working for contractors involved dropping the ‘utilisation’ stage and
separating ‘execution’ and ‘delivery’, with delivery of the asset ending the project lifecycle.
From a contractor perspective this could make a lot of sense, but the utilisation stage was
crucial for a client organisation like WSL, as part of a full asset lifecycle view of the pro-
ject lifecycle. One inference of this observation of general importance was that contractors
often had different concerns driving different conceptual frameworks. Enlightened clients
should ensure that their contractors used the same project lifecycle and employ incentives
Table 7.1 SS1 – the dominant management aspects defining a traditional client’s four stage pro- ject lifecycle.
Basic lifecycle stages Dominant management aspect
Conceptualisation Operations or corporate management initially, then corporate management
Planning Corporate management initially, then operations and project management
Execution and delivery Project management initially, then project and operations management Utilisation Operations management and corporate management
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
320 Employing planning tools in practice
to motivate their contractors to concern themselves with the project’s performance over the
whole lifecycle as effectively as possible. For example, a client might use explicit rewards for
providing quality which ensured low operating costs.
In his view SS1 was a reasonable portrayal of the way most WSL staff currently thought
about project lifecycles at this very basic level of decomposition, but various more detailed
decomposition structures were also employed within WSL, for various purposes.
This kind of four stage structure plus versions with five or more stages were very widely
used beyond WSL, with many different variants of the stage labels, the assumed dominant
management aspects, and the assumed approach to those management aspects. A feature of
the SS1 portrayal was emphasising the relationship among the four generic project lifecycle
stages employed and the corporate, operations and project perspectives on management
discussed at the outset of this workshop. The role of these three perspectives and the best
way to manage their interactions were crucial to the elaborations of SS1 which WSL needed.
The ‘conceptualisation’ stage included initial concept development and the development of
an early version of the business case for investing in the asset produced by the project – be it a
physical asset like a water main or a less tangible asset like a new process or corporate culture.
This stage might be initiated from the bottom up to meet operations needs or from the top
down to meet corporate level strategic needs, but corporate management considerations usu-
ally dominated by the end of the conceptualisation stage, with a focus on business case issues.
The ‘planning’ stage involved a complex and potentially lengthy process that began at a
strategic level and progressively refined the design of the asset, an understanding of intended
benefits from the asset, how it would be used, how it would be created, what resources
would be needed, and when and how it would be delivered. The focus by the end was
detailed tactical plans for execution and delivery, but the starting point was an initial version
of the concept and business case plans which needed refining from a corporate perspective.
There was a great deal of ground in between, with a central role for planning associated with
design development based on operations plans for the utilisation stage.
The ‘execution and delivery’ stage involved implementation of the plans for creation and
then delivery of the asset to its users. The ‘planning’ focus became a ‘doing’ focus, although
important ongoing detailed planning and control incorporating possible replanning of stra-
tegic issues were also involved.
The ‘utilisation’ stage involved the operation of the asset throughout its operating life
to eventual termination of use. The doing and planning focus was very different, and while
operations management now dominated both tactical and strategic planning, corporate
planners had to maintain an interest for their strategic planning purposes.
The way the dominant management aspect pattern portrayed in Table 7.1 changed over
time, the lack of real separability between these management aspects, the very broad scope
of the planning stage, and the complexities introduced by divisions of responsibilities, all
encouraged a wide range of different more detailed project lifecycle structures in different
project contexts to ensure clear definition of who does what, when and how in an orderly
manner. WSL needed one common way to deal with these issues.
Looking at Table 7.1 from an EP perspective, one important issue was which group led
the effort carried out in each stage in the lifecycle. A second was which other groups sup-
ported their efforts in what ways. A third was ensuring that all uncertainty and complexity associated with different stages of the lifecycle received appropriate attention by the right
people at the most appropriate time, bearing in mind the clarified ambiguity concept men-
tioned earlier. Maximising the opportunities presented by the creation of proposed assets
warranted careful attention to all stages of the project lifecycle, individually and in terms of
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 321
integrated teamwork by all the appropriate people. Furthermore, maximisation of opportu-
nities required careful attention to what role the asset created by the project would play in
the context of WSL’s other investments and operations.
Characterisation of the project lifecycle as four sequential stages in Table 7.1 started to
indicate the scope of the tasks involved from corporate, operations and project management
perspectives, and the scope of the uncertainty and complexity that warranted attention.
However, a significantly more detailed decomposition of these four basic stages could pro-
vide much deeper insight into the scope of the decisions involved in different parts of the
lifecycle, the goals being addressed, who were the main players, and the extent and nature
of the uncertainty involved. To deal with these issues effectively WSL badly needed this
additional insight in a framework common to all associated WSL decision making.
Distinguishing progress stages and gateway stages
Paul indicated that one aspect of this additional detail was a clear distinction between what
he would call ‘progress stages’ and ‘gateway stages’. In broad terms:
1 ‘progress stages’ would progress the planning process by getting the planning work
done and achieving the execution of the plans,
2 ‘gateway stages’ would provide independent scrutiny of the quality of progress stage
efforts with a view to corporate governance considerations before moving on to the
next progress stage, plus ‘lessons learned’ reviews to address corporate learning con-
cerns as and when appropriate.
Each progress stage should be followed by a closely coupled gateway stage. Most of their
discussion today would focus on progress stages at a strategic level. He would start by
explaining his recommended use of nine progress stages which cover the whole lifecycle. He
would briefly outline the nature of associated gateway stages later.
The nine progress stages Paul was proposing for WSL
In Paul’s view WSL needed to change immediately to ‘stage structure 2’ (SS2), and he
began his explanation of SS2 by showing his group Table 7.2, which relates the nine SS2
progress stages to the four SS1 stages.
Table 7.2 SS2 – Nine nominal progress stages for the WSL project lifecycle.
Basic lifecycle stages Nine nominal progress stages
Conceptualisation Concept strategy Planning Design, operation and termination strategy (DOT strategy)
Execution and delivery strategy (E&D strategy) Concept, DOT and E&D ‘devils and angels’ in the detail tactics testing
to confirm the overall strategy (D&A strategy) Concept, DOT and E&D implementation tactics (Implementation
tactics) Execution and delivery Execution
Delivery Utilisation Operation
Termination
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
322 Employing planning tools in practice
Table 7.2 used terms like devils and angels, which he would clarify shortly, and useful contractions (like D&A for ‘devils and angels’) serving as labels in bold. It portrayed stage
alignments with the four basic lifecycle stages of Table 7.1 which were a convenient starting
point for understanding SS2 stage roles and their relationships with SS1 but approximate for
reasons worth understanding which he would explain in a few moments.
SS2 was a convenient contraction, following SS1 and anticipating an SS3 possibility which
other directors might require. He knew that some of his fellow directors might want to
argue for modifications, and because the whole of WSL needed a common and agreed- on
project lifecycle definition, he clearly had to be flexible about capturing complexities which
others thought important. He would use SS3 as a contraction of stage structure three to
represent what his fellow directors might later agree. The nine stages of SS2 were ‘nominal’
in the sense that variations could and should be adopted as appropriate by WSL and any
other organisation using an SS2 variant.
In some organisations simplifications of SS2 might be sensible. However, in his view
simplification was not a viable option for WSL, so SS3 would need to be more detailed than
SS2 if changes were made. He would assume SS2 was accepted by the board as a working
assumption for present purposes.
Paul began a more detailed explanation of the SS1 transformation into the SS2 structure
by indicating that the central change driving much of the other changes was the single ‘plan-
ning’ stage of the Table 7.1 SS1 structure being decomposed into four separate components
for SS2:
1 ‘DOT strategy’ (DOT was a useful contraction for design, operation and termination),
2 ‘E&D strategy’ (E&D was a useful contraction for execution and delivery),
3 ‘D&A strategy’ (D&A strategy was a useful contraction for finding and dealing with all
‘devils in the detail’ of tactics plus any ‘angels in the detail’ in the context of addressing
strategic planning), 4 ‘implementation tactics’ (to deal with tactical planning for implementation purposes).
This four components SS2 decomposition of SS1 planning was directly linked to decompos-
ing the single SS1 ‘conceptualisation’ stage of Table 7.1 into five SS2 components, embed-
ding all but the first in the four components of SS2 planning listed earlier:
1 an opening ‘concept strategy’ progress stage – the first of four SS2 strategy stages,
2 a concept strategy development component in the DOT strategy progress stage,
3 a concept strategy development component in the E&D strategy progress stage,
4 a concept strategy development component focused on crucial tactical concerns in the
‘D&A strategy’ progress stage, and
5 a residual ‘concept tactics’ component in the ‘implementation tactics’ progress stage.
The resulting first five SS2 stages in Table 7.2 were followed by the SS1 ‘execution and
delivery’ stage decomposed into the SS2 ‘execution’ and ‘delivery’ stages, then the SS1 ‘uti-
lisation’ stage decomposed into SS2 ‘operation’ and ‘termination’ stages, with some aspects
of the strategic planning for utilisation and termination being embedded in earlier stages.
There were several reasons for treating the first five progress stages of SS2 in this way
which required understanding by all relevant WSL staff.
First, clarity efficiency suggested that all strategic planning should precede all tactical
planning apart from a form of planning intended to test tactics with respect to the ‘devils
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 323
in the detail’, plus equivalent ‘angel in the details’, in terms of the robustness of strategic
choices. This was because planning costs escalate when detailed tactical planning begins,
the effort involved in detailed tactical planning may be largely wasted if strategy changes,
and late changes in strategy can have hugely expensive knock- on cost implications because
of interdependent commitments. The ‘D&A strategy’ stage facilitated a strategy- oriented
focus on tactical planning to find and test key tactical aspects of earlier strategic planning for
‘the devils or the angels in the detail’. The ‘implementation tactics’ stage could make very
good use of D&A strategy stage tactical plans, but it addressed tactical planning for imple-
mentation purposes, a very different task in terms of objectives, concerns, the time required
and the effort involved. Usually it was not worth starting detailed planning for implementa-
tion purposes until a rigorously tested appropriate strategy had been agreed on. An inability
to distinguish which areas of the plans needed devil or angel in the detail testing and which
did not, plus time pressures which made it worthwhile risking wasted effort by overlapping
lifecycle stages, were both examples of potential complications which he would consider
later but wanted to avoid now.
Second, it was important to distinguish between planning for concept, design, operations
and project execution purposes, because although they were closely related, they involved
different purposes and different people in different WSL groups. Design and operations
issues plus key termination concerns like the design life of an asset were particularly closely
coupled concerns and by tradition reasonably well integrated within WSL. However, much
better integration was needed for some key aspects of concept strategy currently given early
consideration by an inappropriately limited set of people, with directly related aspects of
execution strategy currently given later separate consideration by people who should have
been involved earlier.
Third, it was important to distinguish between strategic planning and tactical planning for
all purposes because they involved different goals as well as different people who might be in
different organisations. For all WSL purposes, strategic planning would be in- house with sup-
port from consultants. Some of these consultants might later become contractors. All detailed
planning for asset design, execution and delivery would be a contractor responsibility with
WSL support. The WSL Projects Group did not have the capability to undertake detailed
planning for major projects. Part of the historical reason was his predecessor had taken the
policy decision ‘it is best to let contractors do their own detailed planning within a WSL
approved strategy’ – and he supported this policy with the modest revision of more emphasis
on WSL support and early dialogue when late strategy changes looked advisable. For related
reasons, the WSL Operations Group did not have in- house detailed design capability.
The rationale for decomposing the last four SS2 progress stages as portrayed by Table 7.2
was relatively straightforward. The execution and delivery stage in Table 7.1 needed decom-
position because the Projects Group would be the lead WSL group interfacing with the con-
tractors while they were undertaking execution, but the Operations Group would become
the lead WSL group during delivery, by tradition for very good reasons. The utilisation stage
in Table 7.1 needed decomposition to separate termination because the Operations Group
was the responsible WSL group during operation, but although the Operations Group usu-
ally took the lead on design life issues, the initially planned timing and later implementation
of termination usually gave rise to wider WSL involvement.
The same decomposition structure would be needed for all nine of the gateway stages
corresponding to each of the progress stages. Each of these gateway stages was very differ-
ent from the associated progress stages, requiring very different processes, because different
people and different objectives with different kinds of outcomes were involved.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
324 Employing planning tools in practice
Paul explained that the four strategic level gateway stages after the first four Table 7.2
strategic level progress stages were primarily board led governance exercises to control pro-
jects proceeding to the next stage while WSL was addressing strategy directly. Each should
build on all earlier stages, testing strategic decisions made in earlier stages, making full use of
all information arising from interim planning progress and other updating where appropriate.
The fourth of these strategic level gateway stages was the last gateway stage addressing
strategic planning before detailed tactical planning for implementation began. This meant
the D&A strategy gateway stage was a particularly crucial stage – an important ‘watershed’ –
the last chance to amend the strategy without serious cost implications. Part of this stage
was usually setting a budget which was linked to granting permission to let contracts which
would start detailed tactical planning for implementation purposes. ‘Overall strategy gate-
way stage’ and ‘contracting strategy gateway stage’ were probably more useful alternative
labels for the D&A strategy gateway stage, a suggestion he would be making to the board.
The fifth gateway stage, the implementation tactics gateway stage, would also be a crucial
lifecycle stage if all the tactical planning had to be completed before board level approval for
detailed plans and the beginning of implementation. However, he was assuming this was not
the case, and he would explain why in detail later. In outline, he was assuming that implemen-
tation tactics gateways would be partial as tactical planning rolled forward during execution,
and these gateways plus the gateway stages after the next four progress stages were primarily
ongoing project control exercises plus lessons learned exercises for WSL as a whole, with gov-
ernance oversight to ensure that no important lessons were overlooked and no inappropriate
actions were condoned. Sometimes agreeing to major plan revisions could become crucial,
abandoning a project after detailed tactical planning or during execution or delivery remained
an option, and in all gateway stages the transfer of responsibility involved important issues.
Paul finished his introduction to this SS2 lifecycle structure with four observations that he
particularly wanted his team to bear in mind.
First, as they all knew, currently the Projects Group initially became involved in projects
when someone in other WSL groups asked them for an initial cost estimate for the execution
and delivery of a project which was still early in its concept strategy stage, often defined in
terms best described as a ‘preliminary statement of intent by project initiators’. The Projects
Group were not invited to assist with ‘planning’ within the current equivalent of the SS2
concept strategy stage – they were asked to base their estimate on someone else’s ‘plan’,
This initial estimate was currently referred to as a ‘best estimate’, but in his view this was a
euphemism for what in practice was a ‘best guess’ based on an unacceptably informal and
ambiguous understanding of the DOT strategy to be created and shaped later, plus the
E&D strategy to be created and shaped later still. D&A strategy was not part of the cur-
rent agenda, which was an important oversight. Despite this unacceptably ambiguous best
guess nature, point estimates were used, with a failure to secure clarity about the range of
uncertainty or the nature of the point estimate. Point estimates might be a very ambitious
aspirational stretch target, a commitment, an expected value, or something else that might
mean different things to different people. This point estimate approach had to be scrapped
immediately, and effective and efficient formal linkages between the four strategic planning
stages had to be established as rapidly as possible.
Second, as they all knew, currently the WSL Projects Group was asked for a ‘risk adjust-
ment’ estimate towards the end of the current equivalent of the SS2 concept strategy pro-
gress stage. This practice also had to be scrapped immediately, with effective opportunity
and risk management input to the concept stage provided by the Projects Group from the
outset as part of the required formal linkages between strategic planning stages.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 325
Third, the Projects Group contribution to the new concept strategy progress stage
needed to be a balanced and integrated package providing opportunity, risk, uncertainty
and underlying complexity management input. This input needed to include estimating
expertise driven by and empowered by holistic planning input of a kind which was currently
completely missing without WSL even understanding the nature of the implications.
Fourth, if any WSL project did not pass the concept strategy gateway test when it should
pass, or if it passed when it should fail, or if it passed in a form which is not appropriate with
an initial budget which is not sensible, serious corporate inefficiencies would be involved.
As the first gateway following the first progress stage, the concept strategy gateway had to
address all key objectives for all relevant parties, allowing for all the inherent uncertainty and complexity underlying all the ‘best guesses’. Following gateways had to further test the validity of the earlier gateway decisions in the light of further plan development. A lack
of well- defined plans and the lack of objective data in all concept strategy stage estimates
involved inherent difficulties which required effective treatment. They needed systematically
derived subjective range estimates using the best available judgements in the most appropri-
ate structures. A core issue in the concept strategy stages (both progress and governance)
was appropriate WSL understanding of the size and shape of the ambiguity involved, and
the implications of clarifying it in an enlightened manner. As head of Projects Group a very
early concern was addressing the most appropriate form of Project Group input into WSL
concept strategy progress stage estimates, a topic he would return to shortly.
However, he would first look at the other three frameworks WSL needed as part of their
four Fs concept. The ‘seven Ws’ framework, which builds on the project lifecycle, was next,
followed by a goals– plans relationships framework, building on the first two. Fourth was a
‘specific process for projects’ framework (sometimes usefully contracted to SP for projects
framework or SPP framework), building on and embedding the first three.
The seven Ws framework
Paul started discussing the seven Ws framework by observing that the motivation for intro-
ducing and developing formal risk management in a project planning context since PERT
approaches were first introduced in the 1950s had varied widely. In part this variability was
driven by a range of very different understandings of how opportunity, risk, uncertainty and
complexity were related. In part this variability was also driven by particular concerns which
needed systematic consideration for each project in each organisation, like:
1 sources of risk in design and logistics issues which involved the large- scale use of new
and untried technology;
2 sources of risk associated with shortages of key resources, including finance;
3 contractual issues;
4 communication and trust issues; and
5 conservation, pollution potential or political imperatives.
However, it was now very clear to Paul that WSL and most other organisations had to
embrace the view that whatever the areas of concern:
1 opportunities in a general EP sense should always be a key issue,
2 clarity about the relationship between opportunity and risk in terms of underlying
uncertainty and complexity was always crucial, and
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
326 Employing planning tools in practice
3 uncertainty about the full set of performance objectives and associated relationships between the full set of stakeholders and other key interested project parties was always central.
For example, a common and ongoing issue was ‘Did everyone involved in a project know
what they were trying to achieve in agreed and clearly defined terms that linked the corpo-
rate objectives of the project’s owners and those of other key parties to project plans?’ WSL
as a whole needed a clear corporate understanding of why such concerns arise, and WSL
had to respond effectively in a coherent manner in any project context at any stage in the lifecycle, starting with the concept strategy stage. This was not just a PG team issue.
Paul then showed his group Table 7.3.
Paul’s first observation was this seven Ws structure had evolved from the notion of ‘six
honest serving men’ used by Rudyard Kipling for a comparable purpose at a personal plan-
ning level, and Kipling did not claim it was a new idea. Table 7.3 had provenance, a basic
practical idea which had been tried and tested in well- known earlier variants.
His next observation was the key questions in the middle column provided a starting
point for thinking about issues which were crucial for all WSL staff, not just the PG team.
For convenience, WSL staff could refer to key questions in the middle column of Table 7.3
as ‘the seven Ws’, using the seven ‘W labels’ in bold in the left- hand column as a short form
when appropriate. But in practice the ‘further clarification labels’ in italics in the right- hand
column would usually be more useful than the ‘W labels’.
Paul acknowledged that there was clearly an artificial flavour about some of the W labels,
especially ‘wherewithal’ and ‘whichway’ – he had smiled when he observed several of his
group wincing noticeably when the ‘seven Ws’ designation was first revealed. He suggested
that the seven Ws terminology for this conceptual model might irritate some sensibilities,
but it was just a handy reminder that an operational form of this framework would help
them to consider all seven of these aspects of a project, their multiple components in some
Table 7.3 The seven Ws – key questions needing answers in the basic project definition process.
W labels Associated key questions Further clarification labels
1. who Who are the parties involved? parties 2. why What do the parties want to achieve? motives 3. what What is the deliverable product (asset)
that the parties are interested in? design (plans for design purposes)
4. whichway How will all relevant plans in each life-cycle stage deliver what is needed?
plans for relationships and contracts, business case purposes, operations processes, project execution activities
5. wherewithal What resources need prior planning to achieve the execution of these plans?
resource plans for operations, project activities, other plans?
6. when When do all relevant activities and events have to take place?
integration of all plan- based timetables
7. where Where is the project in its lifecycle? Where will the project take place physically, and where will its deliverables be located if different? Any other relevant context issues?
context where
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 327
cases, their inherent interdependence, the comprehensive set of ‘plans’ involved as clarified
in the third column, and the basis for the link between these plans and the life-cycle stages
in Table 7.2.
He then introduced Figure 7.1, an influence diagram which builds on Table 7.3 to begin
to clarify the operational implications of the seven Ws framework plus the project lifecycle
framework. Table 7.3 and Figure 7.1 were taken directly from Chapman and Ward (2011).
Figure 7.1 employs Table 7.3 ‘further clarification label’ designations as well as W labels
to show a set of relationships which elaborate how Table 7.3 purposes can be linked to plans
as the project lifecycle evolves. Paul used Figure 7.1 to explore the nature of these relation-
ships with his group.
He began by observing that in the concept strategy stage the plans for business case pur-
poses usually became a central concern fairly quickly. An NPV cash flow model at the heart
of the business case might act as the axle, the plans for business case purposes as the hub of
a wheel. However, from the outset it was crucial to avoid too much inward focus on the cash flow model aspects of the business case. As soon as possible, the ‘plans for business case pur-
poses’ should reflect all the key objectives of all the key players. Some people were inclined
to limit the business case concept to cost and revenue concerns, but it was misleading to
what
the design of the product
of the project
who
all project parties ultimately involved
plans for relationships and contracts
plans for operations
why
project motives: e.g. profit (revenue cost) and other motives
plans for business case
purposes
wherewithal
resource plans for activities
when
integration of all plan-based timetables
whichway
plans for activities
resource plans for
operations
where – the project life-cycle position, the physical location of the project or its deliverables, and other relevant ‘context where’ issues
Figure 7.1 The basic project definition process – the clarified seven Ws structure.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
328 Employing planning tools in practice
define a business case in such narrow terms. Who ‘owned’ or controlled a business case
could be crucial to the scope of that business case, and the breadth of vision that a project
planning team as a whole brought to the interpretation of a business case was critical.
Paul observed that the ‘hub’ status of ‘plans for business case purposes’ was modestly
emphasised by slightly bolder lines for its box. The comparably bold lines of the overall box
signified the wheel rim. The other boxes were analogous to spokes. He suggested those
worried about square wheels and hubs (and square pegs in round holes) might prefer using
circles instead of boxes – he had stuck to boxes, in part, because they were a more conveni-
ent shape for most purposes of interest to him, in part to remind him that all analogies of
this kind can produce problems if taken too far. The ‘spokes’ and ‘wheel’ notions centred on
the plans for business case purposes were useful but by no means perfect analogies. An alter-
native hub, associated with an alternative axis, was often a useful focus for different people
for different purposes. For example, the Operations Group might view plans for operations
as their focus sometimes, but the design of the product of the project their focus at other
times. The Projects Group might see plans for activities as the focus sometimes, but plans
for relationships and contracts as the focus at other times.
The ‘who’ was a good place to begin considering the spokes of the wheel which circled
the business case hub and the way the clarified seven Ws structure of Figure 7.1 related to
the project lifecycle of Table 7.2.
The who included ‘project initiators’ – the key players initially, who kicked the whole pro-
ject off in the concept strategy progress stage of the lifecycle. The who also included a much
wider set of ‘project parties ultimately involved’ – some of whom might become dominant
players later in the project lifecycle. Usually projects of concern to the Projects Group were
initiated by the Operations Group, but the groups headed by the Regulation Director or the
Customer Relations Director might be the initiators.
One or more project initiators usually identified the basic purpose of the project at the
outset, or the intended benefit from it, the ‘why’ or motives for the project at conception.
The objectives of the project initiator might dominate the why at first, but identifying and
aligning all WSL corporate concerns needed early attention, anticipating issues which if
overlooked could become serious problems later, including shareholder, contractor, cus-
tomer, and regulator concerns. When working at the board level as a director, Paul would
be trying to take a balanced view and trying to ensure that the board took a balanced view.
Relevant motives would usually include WSL profit, involving revenue and cost. But ‘other
motives’ would usually be important too, like meeting growing public concern about water
main leakage rates and linked regulator actions. Initially, the nature of all these motives
might not be defined very clearly, and even when motives were clearly defined they might
not be quantified objectives. In terms of the four part mission– goals– attributes– criteria hier-
archy of objectives often used to move from an overall corporate mission or vision state-
ment to quantified objectives at a criteria level like expected cost, the initial focus of the
why might be on mission and broadly defined goals rather than more detailed, specific
performance objectives. Significant ambiguity initially might be coupled to changes as other
players become involved.
The initial ‘what’, an outline design, was usually driven by the initial why, the initial
conception of the project’s purpose. But as soon as possible the design should be driven
by competing agendas of all who parties which have been aligned in the corporate interests
of WSL as a whole, including appropriate concessions to other parties like customers and
regulators. This implied initial attention to all the parties to be ultimately involved, and asso-
ciated relationship plans and contracts, before getting too deeply into design. WSL needed
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 329
to give effective early attention to the who– why– what trio, plus the ‘plans for relationships and contracts’, plus the ‘plans for design purposes’, as an integrated set. For example, if
WSL wanted to put a new sewage works in a sensitive location for the residents of that area,
it might be important to minimise intrusive characteristics of all the new facilities, as well
as providing recreational facilities for local residents if some of the land involved could be
used as a public park. All these features might be designed into the facility and the project
management process from the start to plan to manage resistance by local residents in a
constructive and transparent manner. Some aspects might be addressed before the local
residents were even aware of the possibility of such a project, although very early consulta-
tion with residents might be crucial in establishing trust and generating ideas. Some of these
features might be anticipated concessions to demonstrate a willingness to listen to concerns
and respond, if needed, later in the process. Some might be seen as worth clear presentation
from the outset. And some might be tentative suggestions which local residents could use-
fully shape as part of a joint approach.
Useful ways of addressing clarified ambiguity might take a variety of forms, evolving as the
project lifecycle progressed, with some clear views of where and how to look for opportuni-
ties from the outset.
The initial and evolving ‘design’ concept was part of an interdependent set of plans of
interest. What was realised during delivery might not be what was intended earlier, and
managing the implications of design changes was often an important part of the wider plan-
ning process.
Especially during the concept strategy progress stage, it could be very useful to see the
initial outline ‘plans for design purposes’ as central in a different way to ‘plans for business
case purposes’ – a separate and different hub on a different axis for everyone involved to
focus on and think creatively about. ‘Plans for design purposes’ were arguably central to
the creativity which ought the drive the overall creation and shaping of the complete set
of plans – in this sense usurping the role of ‘plans for business case purposes’. At the very
least, there was often value in some constructive tension between creative and imaginative
approaches to meeting the ambitions of other interested parties and the purely financial
aspects of the business case plans when producing an outline design in the concept strategy
progress stage.
This outline design, be it of a physical facility like those of interest to the Projects Group, a
less tangible asset like a service, or a relatively intangible organisational change, should drive
and be driven by the initial outline ‘plans for operation’. These outline plans for operation
needed consideration in conjunction with the outline plans for design from the start, as did
initial outline ‘resource plans for operation’, both of which might involve important feed-
back into design. The timing of all these aspects also needed consideration.
The who– why– what trio plus the ‘plans for activities’ and the ‘resource plans for activi-
ties’ were also interdependent, with linked timing implications. But in the concept strat-
egy progress stage the plans for activities and associated resources might be driven by
business case considerations from both ends – capital cost and delivery time for a very high
level design concept might be all that was required by the owners of the plans for business
case purposes. If the business case was the focus, how the asset produced by the project
would be delivered by the Projects Group was often seen as relevant only insofar as cost,
duration and ‘quality’ assumptions made for the business case might prove non- feasible.
Paul indicated that the current minor and subordinate role played by the Projects Group
in the concept strategy stage of projects was causing difficulties for WSL projects which
needed clarity of understanding in the framework portrayed by Figure 7.1, and effective
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
330 Employing planning tools in practice
resolution at the board level. Furthermore, the board also needed to understand the
need for corporate process changes linked to problems arising because of the minor and
subordinate role often played by the Operations Group at the concept stage of projects.
With support from other WSL directors, he needed to achieve this understanding by the
board as a whole as quickly as possible. Charles and Michael had already approved the
basic ideas, but he now needed to involve other directors like Ollie and Frank as a matter
of urgency.
The initial development of the who– why– what plus ‘plans for relationships and contracts’
might be driven later by the ‘when’ and the business case as a whole. The initial pass around
the hub might be largely clockwise, feeding the business case through the spokes, with feed-
back shaping in reverse directions.
If producing or implementing ‘plans for relationships and contracts’ had critical resource
implications, they might need picking up as ‘other plans?’ in Table 7.3, omitted for simplic-
ity in Figure 7.1. A similar comment applied to all ‘other plans’ needing direct attention
within the Figure 7.1 ‘wherewithal’ box. For example, it might be important to recognise
that WSL did not currently have all the in- house expertise to ensure that early concept plan-
ning involved local residents as effectively as possible.
In the DOT strategy progress stage the focus shifted to the way the plans for opera-
tions and associated resources drove the design and the business case instead of the other
way around, with feedback as appropriate. The purpose of the exercise and the team of
people involved usually changed significantly. Involving operations staff or other ultimate
users and linked stakeholders as well as design staff at this point in the project definition
process usually had significant benefits, particularly in terms of building in ‘operability’ and
‘user- friendly’ opportunities via feed- forward and feedback between the ‘plans for design
purposes’ and the ‘plans for operation’. This might lead to feedback to modify the why or
who. This stage usually provided a refined quantification of operating cost, possibly linked
to refined revenue feedback. It often also provided a more refined view of why in terms of a
more developed understanding of performance objectives.
One of the first allies Paul would seek at the board level was the Operations Director
Ollie, by helping to ensure that the Operations Group was not just the lead player in a clearly
defined DOT strategy stage, but also the contributor of a prototype version of the DOT
strategy in the concept stage – even if some other group initiated the project. He would like
the Projects Group to start seeing the Operations Group as very important members of a
joint team from the outset.
In the E&D strategy progress stage the purpose of the exercise and the team of peo-
ple involved usually changed significantly. In the WSL context the Projects Group now
became the lead group. The focus shifted to the way the who– why– what trio drives execu-
tion and delivery strategy, shaping the project from an execution perspective, with feedback
as appropriate. When execution stage strategic plans were fully developed, delivery stage
strategic plans could receive similar consideration. The design developed earlier drove the
execution plans, associated resource use plans followed on in a natural manner, and the
delivery timetable might just follow on again. But as execution and delivery strategic plans
were developed from a Projects Group perspective there might be significant feed- forward
and feedback between which- way plans for activities, wherewithal resource plans for activi-
ties and when. Some ‘build- ability’ opportunity for feedback to the plans for design and
operations may also prove useful, and the Projects Group needed to involve the Operations
Group directly if this was the case. It was important the Projects Group understood all these
internal loops and interactions. It was also important some prototype variant of all these
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 331
concerns was built into the concept strategy stage, at the very least in terms of suitable provi-
sions for initial business case purposes.
In the D&A strategy progress stage all the parties involved earlier, plus some who might not have been involved so far, needed to address the potential opportunities, problems and
risks which later detailed tactical planning and subsequent implementation and use might
reveal. The focus was key areas of tactical planning for project execution and asset use which
might give rise to this kind of issue. It was the potential for involving new players who per-
haps should have been involved earlier that made the overall strategy testing aspect of this
stage complete, within both the progress and the governance components. For example, if
local authority planning application issues including actions by potential objectors to plans
were not considered earlier in terms of talking to the people directly involved, it might be
crucial to rectify this omission.
From a corporate strategy perspective, as the four strategic planning progress stages
unfolded, it might be appropriate and important to bring in key stakeholders not initially
involved in a carefully pre- planned manner. For example, when the business case had reached
a suitable stage of development the who might be extended to banks or bond funders via
the Finance Group for financial resource reasons. As another example, enlarging the who
to regulators via the Regulation Director for relationship development purposes at a point
when useful advice might be sought could be invaluable. Whether or not this pre- planning
takes place, the D&A strategy progress stage needed to ensure that any residual ‘devils and
angels in the detail’ were identified and managed effectively.
Paul picked this point to indicate that for some purposes it might be useful to see ‘angels’
as ‘agents of opportunities’. Such angels might include local residents living near a new
sewage works site who could make WSL aware of important opportunities if they were
appropriately approached – or local planning authority officers who might be helpful if
approached with this possibility in mind. Both were examples of parties usually associated
with problems and threats, potential ‘devils’, and the way they were approached might con-
dition their response.
In the ‘implementation tactics’ progress stage detailed planning for implementation
began, and for WSL projects involving the PG team this meant contractors became the
lead party, a major transfer of responsibility with big implications for the PG team role. The
letting of contracts might introduce a significant lag between the overall strategy gateway
stage and the actual start of detailed planning, which he would embed in the plans for the
following stage.
It would not be useful to consider the seven Ws in terms of other lifecycle stages at
this point in their discussion, but later the role of the seven Ws would need exploring in all
lifecycle stages through to termination.
The ‘context where’ involved potentially important location issues that might be over-
looked without this focus. For example, the physical location of the project execution activi-
ties and delivered assets might be important if ‘ground conditions’ proved troublesome
during construction and should have been anticipated much earlier.
Paul concluded his discussion of the seven Ws by suggesting that the interactions among
Ws were key aspects of any project which needed to be explicitly recognised and appropri-
ately managed. The clarified seven Ws framework portrayed by Figure 7.1 which he had
just explained in outline would help the PG team to see their role within WSL projects
as a key part of the bigger picture, helping them to integrate their work with other WSL
staff and other players. The seven Ws were central to each stage of the project lifecycle,
although the emphasis and focus would vary as the lifecycle unfolded. All seven Ws should
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
332 Employing planning tools in practice
be addressed by WSL in an orchestrated manner, initially in the context of project defini-
tion during the concept strategy progress stage, with further development as appropriate
in all subsequent stages.
A goals– plans relationships framework
Paul emphasised that every project needed an explicit and carefully defined goals– plans
relationships framework built on the project lifecycle and seven Ws frameworks to develop
clarity about the relationships between the very broad goals being pursued by WSL at a
corporate strategy level and the project plans being used to contribute to the achievement
of some aspects of those goals. These relationships had to address goals directly concerned
with project execution and delivery plans which the Projects Group were directly responsi-
ble for. They also had to address all other plans within the seven Ws structure over the whole
of the lifecycle – in some cases an indirect responsibility of Projects Group which other WSL
groups might feel was their territory. At its simplest this meant the WSL project team as a
whole had to address concerns like through- life costs for the assets and associated revenue
issues. But a complex related concern he thought important was return on equity issues for
shareholders which had to reflect the cost of borrowed capital to gear up returns on equity
capital.
For most purposes the Projects Group would be focused on criteria like expected project
execution cost and associated cost risk driven by possible significant delays or materials cost
overruns. However, they needed to be very aware of the broader goals of interest to WSL,
and formally consider associated intermediate level objectives whenever they were relevant,
involving other WSL groups when appropriate.
Furthermore, WSL had corporate goals which included building well- founded trust with
their customers in a reliable and safe source of water at low cost plus reliable and efficient
wastewater removal at low cost, without smells, local flooding, pollution of streams, rivers,
or beaches. WSL also had to maintain good relationships with their regulators and with local
councils in relation to planning permissions and issues like approvals when they needed to
dig up roads to install new water mains or sewers.
A goals– plans relationships framework was not easily characterised in a seven Ws frame-
work like Figure 7.1 or a lifecycle framework like Table 7.2. Its nature was very specific to
each context. However, it had to become a WSL corporate tool, used along with the project
lifecycle framework and the seven Ws framework that it built on, plus the specific process for
projects framework to be considered next.
A specific process for the strategy progress stages
Paul indicated that all WSL staff also needed to understand a fourth framework in broad overview terms, some getting directly involved in part of it. This fourth framework was an
integrated set of specific process framework components which accommodated each of the
progress stages of the project lifecycle plus each of the gateway stages. The PG team had
to make the whole of this fourth framework a central part of their mindset as well as their
toolset.
He would concentrate his attention on the specific process framework components for the
progress stages that day, in particular the four strategy progress stages.
The specific process the PG team was going to use in the four strategy progress stages
was the ‘basic specific planning uncertainty and complexity management process for projects’,
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 333
contracted to the ‘basic specific process for projects’ or ‘basic SPP’. In the context of discuss- ing the ‘basic SPP’ the further contraction of ‘SP for projects’ kept the terminology simple
and easy to interpret.
It was useful to see this basic SPP as a composite of three contributing components mak-
ing interdependent contributions:
1 one ‘framing process’,
2 one ‘named contributing specific process’,
3 plus an open set of ‘all further contributing processes’ which might prove useful.
The framing process was the ‘universal planning uncertainty and complexity management process’, contracted to ‘UP’. The named contributing specific process was the ‘basic perfor- mance uncertainty management process’, contracted to the ‘basic PUMP’. The basic PUMP was associated with the strategy progress stages in Chapman and Ward (2011). The fur-
ther contributing processes included all other processes relevant to WSL project planning, another EP usage of the closure with completeness concept.
Paul explained the nature of the UP concept using a few examples and a summary of some
of the key relevant features explored in Part 1 and Chapter 5, with special attention to the
generality of the framing assumptions and the value of clarity about working assumptions.
He next provided a brief background introduction to the PUMP concept. His explanation
included earlier comments in this chapter. He explained that the basic PUMP was developed
for the E&D strategy progress stage, but used in its basic form for all earlier project lifecycle
progress stages and adapted for all later progress stages. The basic SPP was developed for
the concept strategy progress stage, to be used in its basic form for all later strategy progress
stages and adapted for all following progress stages. He indicated that the focus of his dis-
cussion that day would be using the basic SPP in the concept strategy progress stage, but he
would also consider using the same basic SPP structure for all three of the following strategy
progress stages. Using the same strategy gateway stage PUMP after all four SPP strategy
progress stages would also be addressed briefly. Adaptations to the remaining five progress
stages and their gateway processes would be deferred until later meetings.
The nature of other relevant ‘specific planning uncertainty and complexity management processes’, contracted to SPs, was then briefly explored using a summary of Chapter 6 ideas plus an overview of the broad scope of all aspects of project planning beyond PUMP – build-
ing on a brief overview of all CPM and PERT based approaches as discussed in Chapter 3.
Paul then showed his group Table 7.4.
At an overview level the basic SPP phase structure largely follows the UP phase structure,
with some features drawing from related PUMP phases. The ‘create and enhance plans for
all relevant concerns phase was not part of the basic PUMP, nor was the following shape
base plans phase. These two phases of the SPP have to draw on a wide range of other project
planning model and process concepts, working within a UP perspective. The joint role of
these two phases is a central aspect of the SPP not provided by the PUMP because it was
assumed that other people had provided this aspect of planning separately. The integration
of the UP shape, test, interpret, and implement phases followed by a different decomposi-
tion structure involved relatively complex SPP/PUMP relationships, which he would out-
line shortly.
This overview portrayal of UP/SPP/PUMP relationships was a useful starting point for
the more detailed understanding which was needed. Paul indicated that he would use the
second column to structure his discussion, outlining the basic SPP one phase at a time, but
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
334 Employing planning tools in practice
he would link his comments to the other two columns. Table 7.4 showed the key phase
alignments, but all alignments would need clarifying.
Paul then showed his group Figure 7.2. He used this flow chart portrayal to help every-
one to focus on the basic SPP phases one at a time within a UP- based framing assumption
perspective, as outlined in the following subsections.
The basic SPP ‘capture’ phase – capture the context with appropriate clarity
Paul indicated that the first pass through the first phase in the first stage of a project was
about using the four Fs to capture an initial corporate understanding of the context which
gave rise to the project – all the relevant immediately obvious aspects of the ‘project con-
text’, portrayed in a systematic structure which was designed for clarity efficient use as it
evolved throughout the project lifecycle.
The basic SPP ‘capture’ phase involved seeing the context of a project in terms of the
project lifecycle, the seven Ws, and the goals– plans relationships as currently understood,
with a view to the requirements of all following process phases. This initial use of the four Fs
concept as an integrated toolset provided usefully structured focus from the outset, increas-
ing the clarity efficiency of the process relative to the completely general capture the context
phase of a UP. However, the capture phase of the basic SPP and the capture phase of the
unrestricted UP as discussed in Part 1 and Chapter 5 were otherwise directly comparable.
Comparing the capture phase of the basic SPP to the define phase used to initiate the
basic PUMP, the fundamental difference was the basic PUMP assumed that responsibility
for creating base plans lay with others who had already executed this responsibility, but the
basic SPP assumed that responsibility for creating the base plan lay with those using the
basic SPP. This meant that the capture the context phase of the basic SPP had further scope.
Making use of the four Fs in this way was not difficult in principle for well- trained and
experienced planners, and they would soon get used to capturing initial project information
Table 7.4 Key phase alignments for the basic SPP with the UP and the basic PUMP.
UP phases Basic SPP phases Basic PUMP phases
Capture the context with appropriate clarity
Capture the context with appropriate clarity
Define the project
Select and focus the process for appropriate clarity
Select and focus the process for appropriate clarity
Focus the process
Create and enhance plans for all relevant concerns
Create and enhance plans for all relevant concerns
Shape the plans using models of some key issues, test the plans to ensure robustness, interpret the plans to exploit creativity, and implement appropriate aspects of the plans
Shape base plans using models of some key base plan issues
Identify all relevant sources, responses and conditions
Structure all uncertainty
Clarify ownership
Quantify some uncertainty
Evaluate all the implications
Identify all relevant sources, responses and conditions
Structure all uncertainty
Clarify ownership
Quantify some uncertainty
Evaluate all the implications
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 335
identify all relevant sources, responses, and conditions
structure
all uncertainty
clarify ownership
quantify
some uncertainty
evaluate
all the relevant implications
to the appropriate gateway
stage
capability-culture
assets
capability-culture
liabilities
create and enhance plans for all relevant concerns
from project
initiation or a gateway
stage
select and focus the process for appropriate clarity
capture the context with appropriate clarity
shape base plans using models of some key issues
Figure 7.2 The basic SPP (specific process for projects).
and updating that information as the project evolved in this structure. However, working
in the concept stage would be a new experience for the PG team, with some important new
toolset, skill set, and mindset implications.
Two projects of current interest would serve as illustrative initial test and demonstration
case studies, to be added to as further proposals arose. One was a new sewage works –
with a concept strategy progress stage study initiated by the Operations Group still in early
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
336 Employing planning tools in practice
development. The other was a programme of water mains replacements – initiated by the
board after prompting by the Regulation Director because of adverse publicity about water
leakage rates, with a concept strategy progress stage study undertaken by the Finance Group
still in early development.
In both cases the Projects Group had been asked for early cost estimation input but no
other planning input. Paul would approach Ollie, the Operations Director, and Frank, the
Finance Director, about both studies. An approach to Ollie about both these new projects
would be his first step. His objective would be Projects Group ‘planners’ working alongside
Operations Group ‘planners’ and Finance Group ‘planners’, to jointly capture the context
as early as possible in a new manner which was much more appropriate for WSL than cur-
rent practice.
The basic SPP ‘select and focus’ phase – select and focus the process for appropriate clarity
Paul began explaining the ‘select and focus’ phase by indicating that this phase was use-
fully seen as planning the planning, what might be viewed as a higher order form of project
planning. The basic SPP process needed to be adjusted to whatever the capture the context
phase had uncovered so far. The way the initially selected and focused version of the SPP
process would be implemented needed initial planning, with further adjustments and associ-
ated planning control as the process continued. Any broadly defined planning process which
does not have an explicit phase embodying a planning the planning component involves a stealth assumption of the form ‘one process suits all relevant circumstances’.
You may recall that in terms of the evolution of the UP concept, the UP ‘select and focus’
phase concept was a generalisation of the basic PUMP ‘focus’ phase. However, for present
purposes the basic SPP select and focus phase is best seen as a special case of the UP select
and focus phase. In its basic SPP form the select and focus phase is richer in nature than
in its focus phase PUMP form, because the basic SPP is a broadly defined project planning
process explicitly addressing all aspects of project planning within one overall project plan-
ning framework. However, it is less general than the UP select and focus phase because an
acknowledged focus on project planning provides very useful efficiency gains.
In summary, the role of the basic SPP select and focus phase is a more focused version of
the UP select and focus phase, which is much more general than the basic PUMP focus the
process phase. You should find an overview understanding of this relationship useful now,
but the full implications will have to emerge as this chapter progresses.
Paul suggested that understanding this planning the planning concept in any detail
required a comprehensive understanding of the options available, which he would not dis-
cuss until later, so it would not be helpful to dwell on the select and focus phase at present.
But it would be useful to briefly consider Figure 7.3 now.
In the bar chart (Gantt chart) format of Figure 7.3, the capture (the context) phase is
the first phase, followed by the select and focus (the process) phase, and then the create and
enhance (plans) phase – an obvious basic precedence ordering. However, these three phases
are portrayed as proceeding in parallel during the first complete pass, with linked start- to-
start precedence relationships plus common finish- to- start precedence links to the shape
base plans phase. Paul explained that this portrayal of the formal plan by Figure 7.3 allowed
for extensive unplanned iterations within the first pass throughout these three phases – it
made an explicit provision for informal use of multiple iterations which it would be fruitless
to try to plan in advance. All other phases could then follow using finish- to- start precedence
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 337
relationships to complete the first full pass, iterating back to the capture phase after the eval-
uate phase. Sometimes partial passes through some of the later phases might prove useful,
provided for in Figure 7.2 by the multiple loop- back options shown using pairs of arrows in
both directions, but Figure 7.3 ‘kept it simple systematically’ in a useful manner.
You may find it helpful to note that the Figure 4.4 portrayal of the UP concept is simi-
lar, with comparable implications, although Paul did not explore this relationship with his
group.
Paul did emphasise that a highly iterative process should be planned for, completing the
first complete pass in 10% to 20% of the time available.
He also explained there was no point in detailed planning in the Figure 7.3 framework
beyond the first pass until the evaluate phase was completed at the end of the first pass, but
an 80 to 90% time provision for later iterations needed planning from the outset. This was
comparable to the idea that there was no point in attempting to even portray the iterations
needed within the first three phases because in capable hands these three phases seem to
proceed in parallel, nor was there a need to portray the occasional use of incomplete itera-
tions later. Figure 7.3 was a guide, not a prescription, and it provided strategic clarity, not
detailed specific tactical advice.
The ‘ongoing effort’ aspect of the select and focus phase (as distinct from the initial
‘intense effort’ aspect and the intermittent intense effort nature of all the other phases)
portrayed by Figure 7.3 involved adjusting the initial planning the planning plans as they
were implemented in two senses: dealing in real- time terms with unexpected outcomes and
learning from experience about what to do differently later.
This overview pushed the ‘formal’ planning aspects of planning the planning about as far
as it was useful for Paul to go in his presentation. In practice he would be relying on informal
ownership
quantify
evaluate
intense effort ongoing effort no effort or intermittent effortkey:
structure
phase
create & enhance
select & focus
start of the process
end of the first complete cycle
capture
identify
shape base plans
Figure 7.3 An example basic SPP bar chart (Gantt chart) for the first pass.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
338 Employing planning tools in practice
planning by skilled planners leading the overall effort to build on this overview using a new
skill set, assuming those wanting and needing the requisite skills could build them later,
starting with a comprehensive understanding of the SP for projects concept. Paul observed
that some members of the PG team would need to develop their skills to lead the undertak-
ing of this phase. It was not a straightforward skill set to acquire, but he would ensure they
had all the training and support needed.
The basic SPP ‘create and enhance’ phase – plans addressing all relevant concerns
Paul indicated that the first pass through this third phase in the first stage of a project was
about creating and then enhancing an initial corporate understanding of ‘the plans’ which
described what the project involved, with a focus on the base plans, what he would call the
‘initial plans’. He believed it was important for these ‘initial plans’ to be seen by all WSL
groups involved as ‘the initial creation of the project initiators’ – generally not members of
the PG team.
In future he was going to ensure that all WSL project initiators had support from PG
team planners from the outset so that everyone involved was using the same concepts and
language. When he was hired, he had insisted on a mandate to do this. However, it would
be important to ensure that ownership of the initial plans created by project initiators was
not challenged by the PG group or any other groups, and part of doing this involved PG
planners being seen by project initiators as ‘working for the project initiators’, helping them
with ‘their’ ideas, not ‘taking over their creation’. ‘Ownership’ of plans (being in control)
was usually an extremely sensitive issue, which the PG team would have to manage with care
and diplomacy.
The basic SPP ‘shape base plans’ phase – using models of some key base plan issues
Paul wanted to couple this ownership of ‘initial plans’ by the project initiators, with support
by the PG team which reinforced their ownership, to moving as soon as possible towards a
broader ownership of what he would call ‘grounded plans’. He would associate the initia-
tion of ‘grounding the plans’ with beginning the first pass through the ‘shape base plans’
phase using models of some key base plan issues. This, too, might be led by the project
initiators, but in some areas joint ownership involving a broader project team might start to
develop, as part of a gradual evolution of ownership to the comprehensive integrated team
basis which he wanted to encourage.
To begin with a minimum clarity position for the aspects of initial plans based on activities,
Paul suggested project initiators might model the construction portion of project duration
using just one activity. A directly related minimum clarity position for estimating construc-
tion cost might use two construction cost components associated with the cost of the con-
struction contract: a fixed direct cost component linked to contractor costs like materials
and a variable direct plus indirect cost component linked to the joint effects of resources
committed and project duration. For example, he knew Finance Group planners were cur-
rently assuming one contract undertaken by one national contractor for all the new water
mains, so in principle a one activity structure could be used for a PG team– supported initial
estimate of project construction duration. Overall project duration would involve additional
pre- construction activities like getting permissions, plus post- construction activities like the
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 339
testing of completed facilities. Furthermore, partial delivery of the new water mains network
would need attention in practice. But maximum simplicity could be maintained for project
initiators if that is what they wanted for their initial plans.
Paul made it clear that he wanted to encourage project initiators to start with a minimum
clarity position on the project construction period duration and associated construction cost
estimates if that was what the project initiators really seemed to want. However, as soon as
possible he wanted the PG team to diplomatically move them towards a ‘grounded version’
of their initial plans which ensured that the uncertainty associated with duration and cost
estimates was clearly and explicitly articulated and linked. All the components would have
to be interval (range- based) estimates portraying very significant associated uncertainty, and all underlying uncertainty would need identification. He would insist that the PG team sup- porting initial plans should refuse to reduce their view of uncertainty if they did not have
a clear basis for doing so. He indicated that project initiators creating ‘initial plans’ might
very reasonably concentrate on their areas of expertise, but they should be encouraged to
collaborate with others in order to address all other relevant concerns even if some concerns
were not modelled.
One of Paul’s basic premises, backed by a mandate when he was hired, was that by the end
of the concept strategy progress stage, the business case focus of ‘the plans’ had to be sup-
ported by plans for all other aspects of the seven Ws which were in effect relatively low clarity
prototype variants of the plans the Operations Group would produce in the DOT strategy
stage, the Projects Group would produce in the E&D strategy stage, and the Operations
Group plus the Projects Group plus all other relevant WSL groups would produce in the
D&A- strategy stage. Furthermore, these concept strategy progress stage plans had to reflect
effective treatment in a coordinated manner of generally understood and tested goals– plans
relationships.
A terminology which he thought WSL as a whole might find useful was the ‘initial plans’
had to be ‘grounded’ – developed and documented by all relevant WSL planners. PG plan- ners would take the lead in their traditional activity- based planning roles, but these efforts
would be coordinated with cost estimating and uncertainty management roles, and PG
planners would also work as facilitators alongside all other WSL planners to ensure that all of
the seven Ws plus the goals– plans relationships received balanced and integrated treatment
within the lifecycle framework before the concept strategy stage was completed. Helping
project initiators to develop their initial plans to achieve a grounded form would be the
focus of the PG planners’ role as facilitators. Operations Group planners, Finance Group
planners, and others with planning input roles, would have to join this ‘grounding the initial
plans’ effort, working as an integrated team, with some of the coordination provided by PG
planners in this shape base plans phase.
This shape base plans phase (using models of some key base planning areas) and the pre-
ceding create and enhance plans (for all relevant concerns) phase were both central parts
of the UP concept, but they were not part of the PUMP structure, and the way they were
approached would be new to all WSL groups.
Paul wanted the project initiators to lead the create and enhance plans phase with PG
team support initially, recognising that many important aspects of plans are clear to project
initiators but not expressed explicitly in models of any kind. He then wanted the PG team
to unobtrusively lead the coordination of the grounding aspect in a facilitating role with an
EP agenda and the seven Ws structure plus the goals– plans relationships as an operational
framework. They would begin with a model- based approach to shaping base plans drawing
upon all relevant project planning models and processes in the shape base plans phase, like
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
340 Employing planning tools in practice
CPM. They would then move on to a PUMP based approach to the concept strategy pro-
gress stage as a whole. The concept strategy gateway stage was often highly focused on the
plans for business case purposes, and some people expected this business case focus for the
whole of the preceding concept strategy progress stage, with very little background plan-
ning in a full seven Ws plus a complete goals– plans relationships structure sense. Paul would
focus on the development of grounded plans underlying the concept stage business case
strategic plans today, with an immediate focus on the activity- based plans.
Most projects would need more than one main activity plus some secondary activities to
provide a credible estimate of project construction period duration and construction cost for
concept strategy gateway purposes. But a low level of detail relative to later planning would
be a defining feature of appropriate concept strategy stage planning. Paul wanted a clarity
efficient approach to this PG team responsibility which would serve as a demonstrator for
all other WSL groups.
He started by observing that most of the current Planning Department activity modelling
he had seen used finish- to- start activity- on- arrow diagram CPM modelling. This was simple,
it was widespread common practice, and it worked well for detailed planning. However,
sometimes it did not work well for E&D strategic planning stages because the low level of
detail involved made precedence relationships ambiguous, and it was certainly not suitable
for the prototype preliminary forms of planning needed in the concept strategy stage. Plan-
ning Department staff had not been involved in concept strategy planning before, which
was particularly demanding if opportunity efficiency was the goal.
A generalised and much more flexible approach than activity- on- arrow CPM models was
activity- on- node models (precedence diagrams), sometimes synthesised with Gantt charts
(making activity nodes time- scaled) to form ‘linked bar charts’, as illustrated by the basic
SPP diagram of Figure 7.3. Figure 7.3 illustrated start- to- start precedence relationships
(one activity cannot start until another activity starts), plus finish- to- finish precedence rela-
tionships (one activity cannot finish until another activity finishes), as well as the finish- to-
start precedence relationships assumed by activity- on- arrow CPM models. Also feasible if
the activity- on- node format was adopted were either– or precedence relationships (one activ-
ity must follow the other but with the order unrestricted), and other flexible choices like
overlaps of an uncertain size. The activity- on- arrow CPM format was popular because it was
very simple if finish- to- start precedence relationships were the only precedence relationships
involved. But this kind and level of simplicity carried a serious cost in terms of generality,
and PG planners needed more generality to cope with their new roles.
Paul then observed that even the basic CPM model was always employed using an iterative
‘method’, this method included steps which addressed timetable constraints and resource
constraints, sometimes generalised to deal with timetable and resource optimisation, and these
issues were all relevant to all WSL project planning in the shape base plans phase. Books like
Project Planning with CPM and PERT (Moder and Philips, 1970) provided a useful guide to the models and associated processes available and used by most professional planners in a large
number of countries for many years. He believed WSL planners needed to be fully capable of drawing on all these models at a conceptual level, operationally employing particular practical examples when relevant, to make effective contributions to the initial plan create and enhance
phase and the following shape base plans phase. He wanted to start with very simple variants.
He would then illustrate the range of complications which could be addressed.
He began by observing that there was a significant project planning literature on models
which looked at optimising the balance between direct and indirect costs as project expected
duration is shortened or lengthened.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 341
In their simplest form these models assumed that any project had ‘direct costs’ associ-
ated with executing each activity plus ‘indirect costs’ (overhead costs) which were a simple
linear function of overall project duration. These models assumed that each activity had a
‘normal’ duration which involved a ‘normal’ direct cost, and most ‘normal’ activity dura-
tions could be shortened to a minimum (‘crash’) duration if the direct cost of execution was
increased. Furthermore, they assumed that the activity direct cost increase from the normal
cost to the crash cost involved a linear intermediate relationship. They used an iterative
search procedure to achieve ‘crash cost efficiency’ – identifying an efficient frontier defined
by the minimum direct cost for all project activities for a feasible range of project duration
compressions, from a ‘normal project duration’ (minimum direct cost) to a ‘crash project
duration’ (all activities on the critical path at crash duration). The results could then be
portrayed using Figure 7.4.
As shown in Figure 7.4, the minimum project cost level which defined the optimal project
duration point was the point on the joint cost curve when it departed from a zero slope.
This was determined by the point where the rising direct cost curve had the same rate of
increase as the falling indirect cost curve as project duration declined.
This simple model was a useful conceptual model as a starting place to consider trade- offs
between project duration and cost, but in practice it was far too simplistic to use for opera-
tional purposes in most contexts. One simple generalisation of this model recognised that
indirect cost was never just a simple linear function. If the cost of capital invested in a pro-
ject (not yet productive) was considered, then project direct cost rises at an increasing rate.
Opportunity costs (forgone pay- offs from a completed project) can also be very important.
However, an obvious and reasonably robust implication of this basic model was most critical
path activities would need some compression from their normal duration, most paths would
become critical as compression took place, and the optimal project duration was often closer
to crash project duration than normal project duration. More sophisticated inferences could
be drawn from more complex generalisations of this basic model.
These issues were of direct relevance for all projects in all stages of their lifecycle. They
were especially important during the first and third strategy progress stages – when a key
project
cost in £
joint cost
minimum
cost
project duration in months
crash duration optimum duration normal duration
0
indirect cost
direct cost (efficient frontier)
0
Figure 7.4 Joint cost relationship graph for the basic project duration- cost trade- off model.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
342 Employing planning tools in practice
strategic concern was project duration– cost trade- offs. The inferences involved needed
sophisticated consideration by skilled planners within the iterative process framework even if
formal models were not used directly. This meant PG planners needed to understand what
all these models could and could not do, even if they did not use them directly in formal
planning terms to a significant extent.
The underlying more sophisticated models he was talking about included project plan-
ning models addressing resource allocation and resource usage smoothing, which in their
most sophisticated forms looked at optimal resource use, generalising the simple linear crash
cost and duration models.
One of the most general deterministic project planning models of this kind in some
respects, referred to as ‘Decision CPM’, recognised that activity definitions could be
changed as well as activity durations, at a cost, by changing construction technology choices
and design choices, and activity sequence choices involved important working assumptions
which could also be changed. This implied project planners had to enhance the determin-
istic modelling basis of their plans by considering all these degrees of freedom in a creative
fashion, using iterative processes to test initial working assumptions in a flexible manner.
This involved a form of ambiguity uncertainty inherent in base plan creation which needed
systematic attention. It needed consideration over and above uncertainty in the event, vari-
ability, systemic, and other ambiguity uncertainty senses. Managing uncertainty driven by
ambiguity in this particular sense, plus all associated systemic uncertainty, was a demanding
craft which all PG planners needed to acquire and share with other WSL planners because
it was central to an EP view of uncertainty and complexity. In practice Paul recommended
largely informal comparisons of options addressing the full range of generalisations avail-
able. Informal planning of this kind would require understanding the nature of the options,
facilitated by an understanding of models that could be used but using them directly if
at all only on selected occasions when this looked likely to pay in opportunity efficiency
terms. He observed this was not project risk management as commonly conceived, nor
was it addressed in a direct way by the PUMP approach – it was basic project management
addressing an aspect of base plan option choice uncertainty and underlying complexity with
a view to achieving overall opportunity efficiency. In this phase they were starting with
a focus on shaping base plans prior to addressing contingency plans. Later phases would
consider contingency plans and joint assessment of all aspects of the plans. He observed
that the implications for useful ‘searches for opportunities’ beginning with base plans were
important. He would later organise in- house courses developing all these ideas for everyone
in the Projects Group.
In the meantime, to demonstrate what he meant, he suggested that the Finance Group’s
current assumption of ‘one contract with one large national contractor’ for all the new
water mains was probably deeply flawed. It was certainly questionable. He suspected that
half a dozen local contractors plus several national contractors working on a programmed
sequence of water main section contracts would meet WSL needs very much better and sig- nificantly improve overall WSL opportunity efficiency. Some projects benefit from a single
contractor ‘fast track’ option. But in his view this project needed a multiple contractor ‘slow
track’ option ‘base plan A’, with ‘faster track’ options B, C, D and so on which might be
used as construction costs varied over economic cycles and regulatory pressures varied, in
conjunction with ongoing allocation of new contracts limited to contractors demonstrating
good performance. He wanted PG planners to enrich WSL concept strategy progress stage
planning so that this kind of alternative ‘concept strategy stage’ base plan assumption could
be tested. This particular base plan option set required a project work package structure
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 343
which accommodated a multiple contractor programme view of the project. They needed
to convince Finance Group planners this was a good idea fairly soon, but a more urgent
need was persuading Operations Group planners to explore 100 year design life technology,
because the Finance Group’s current assumption of a 50 year design life was also deeply
flawed in his view, for reasons driven by discount rate assumptions in the business case con-
cept strategy plans.
To further demonstrate what was implied by the scope of this approach, Paul indicated
that Operations Group planners addressing the new sewage plant were currently assuming
that WSL would replicate the last sewage plant built, and they were currently considering
urban site A as the preferred choice, with urban site B as an option. He suspected that a
very different layout would improve overall WSL opportunity efficiency. In his view, this
preferred layout would involve a bigger footprint but a lower profile, landscaped into a rural
setting at locations like site C, with surrounding landscaped grounds made available to the
public. He thought this would be better than squeezing a new sewage plant into an urban
location like A or B for a number of reasons. He wanted to persuade Operations Group
planners to explore this approach to a site like C.
In the near future, he wanted all PG planners to get used to, and become skilled at, devel-
oping these kinds of opportunity, and stimulating other WSL planners to do the same, at a
‘grounded base plan’ level, later building on this with ‘grounded contingency plans’.
If the PG team had been familiar with Chapter 5 of this book, Figure 7.3 might have been
compared to Figure 5.2, discussing the comparable need for subsequent generalisation of
associated analysis. Paul might have pointed out that WSL had to get beyond the equivalent
of a basic EOQ model to explore Nicola’s equivalent of a My- Place- Your- Timing approach
in the basic SPP create and enhance plans phase. You may find these relationships useful at a
conceptual level. The direct cost (efficient frontier) curve of Figure 7.3 is a piecewise linear
curve obtained iteratively in this case, a more complex basis than Nicola had to deal with,
but otherwise comparable trade- off issues are involved in similarly simple base plan models
of complex associated realities.
The ‘identify’ phase – identify all relevant sources, responses, and conditions
Paul indicated that the ‘identify’ phase of the basic SPP could be viewed as the next step
of the UP shape phase. It addressed a further aspect of ‘shaping the plans using models
of some key issues’, moving on from those aspects of ambiguity uncertainty addressed by
traditional planning approaches for base plans in the previous phase to the full treatment of
all aspects of uncertainty using a PUMP framework. Any serious potential misconceptions in the base plans were sources of uncertainty, along with many other sources of uncertainty
not addressed by current WSL project risk management.
The base plan issues which he had just explored were so complex PG planners had to not
only understand underlying models which can in principle be applied, but also understand
that in practice they had to depend to a significant extent upon informal planning craft
skills and relatively simple formal models. They had to use the potential models as a way of
understanding what was involved – usually employing them as conceptual tools as opposed
to operational tools. The contingency planning issues were also very complex, but well-
chosen simple operational models plus a well- developed understanding of the more complex
models underlying them would be invaluable.
The ‘sources, responses and conditions’ modelling initiated in the identify phase embodied
the PUMP equivalent of the phase of common practice project risk management processes
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
344 Employing planning tools in practice
which involved producing a ‘risk list’ or ‘risk log’ plus a later phase which addressed responses
to the risks. The ‘source’ concept was a ‘source of uncertainty’ concept which generalised
the traditional project risk management ‘risks’ concept. The rationale for closely coupled
joint treatment of sources and responses included dealing with the inevitable ambiguity
uncertainty associated with a source of uncertainty which as yet had no assumed proactive
or reactive treatment, but encouraging a much richer overall approach, including facilitating
later specific and general response distinctions, was actually more important. There were
other very important departures from common practice involved, but for present purposes
the key thing the PG group needed to understand was the identify phase was about begin-
ning a clarity efficient approach to structuring the grounded initial plans in a form suitable
for managing contingency plans including base plan adjustments to deal with uncertainty
associated with key plan components, with a view to shaping the project concept to achieve
opportunity efficiency as outlined in Chapters 3 and 4. This involved decomposing key
sources of uncertainty within each of the key concept strategy stage plan components, asso-
ciating that uncertainty with responses to the extent this was currently relevant, and identi-
fying important underlying assumptions to be treated as conditions.
As a very simple example, if a single cost component approach was taken to construc-
tion cost for the new water mains or sewage plant projects, a single source of uncertainty
prior to decomposition would have to include everything that Estimating Department staff
traditionally thought about when they produced an initial point estimate, everything that
Risk Management Department staff traditionally thought about when they produced a
‘risk adjustment’, everything that the Planning Department then thought about with fur-
ther Project Group support, plus everything else that was relevant but traditionally got left out, the basis of current duration and cost overruns plus less visible but equally important lost opportunities.
He would later hold a series of workshops, which would address how best to decompose
a complex single source of uncertainty like this, and particularly for the benefit of PG plan-
ners with mainstream project risk management expertise, he would explain the differences
between SPP approaches based on the PUMP approach which they would be using and
common practice project risk management equivalents.
In practice, WSL would have to address a programme- based decomposition of the water
main construction period and all the additional activities involving setting up contracts,
getting planning permission, and so on, well before the four strategy progress stages were
completed, and well before a reliable project cost estimate could be developed. This would
raise important practical issues to address in the concept strategy progress stage. For exam-
ple, if there was uncertainty about how long it might take to get planning permission for the
new sewage works currently being considered by the Operations Group, this might drive
related uncertainty about where to locate it and how to design it, a systemically related set
of concerns associated with his suggested site C option needing attention in the concept
strategy stage.
The basic SPP ‘structure’ phase – structure all uncertainty
Paul indicated that the ‘structure’ phase of the basic SPP was another elaboration of the UP
shape phase which also incorporated some aspects of the UP test phase. The basic SPP struc-
ture phase embodied the PUMP structure phase. The model structuring concerns included
identifying any important dependencies between components of the plans across all seven
Ws. One particularly crucial aspect of the structuring was identifying ‘general responses’ to
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 345
sets of identified sources – a key basis for building robustness to deal with significant sets of
both identified sources and unidentified sources (the unknown unknowns). The SPP struc- ture phase was also concerned with testing all the earlier working assumptions involving the structuring of plans.
An illustrative example was a failure to get planning permission for a new sewage works
at a stage when knock- on delays would be very expensive. This might be avoided by much
earlier exploration of alternative sites where the local population might be quite receptive if
significant recreational facilities as part of the site development plans were on offer, and con-
cerns about current local flooding risks or other linked issues were also addressed. As part
of the planning of activities following on from obtaining planning permission, they would
not only have to address what responses might be used to respond to delays, but they would
also have to deal with what might be done to reduce the risk of this problem arising – both
proactive and reactive responses.
A closing aspect of the structure phase was addressing a preliminary ordering of WSL
preferences for all relevant options available, within the specific and general response sets
which clarified which responses would only deal with specific reasons for delay and which
responses would deal with sets of reasons for delay.
The basic SPP ‘ownership’ phase – clarify ownership
Paul indicated that the ‘ownership’ phase of the basic SPP was yet another elaboration of
the UP shape phase plus some aspects of the UP test phase. The basic SPP ownership phase
embodied the PUMP ownership phase. Basic questions addressed included the most appro-
priate forms of contract for all aspects of the project, plus being clear about which party was
responsible for managing which aspects of all opportunity and risk, and who benefited or
suffered if opportunities and risks were realised.
To illustrate the role of this phase, applying the thinking which it involved was in
part what lay behind Paul’s suggestion that a single contractor for the water mains con-
tract which the Finance Group had recommended should be avoided. Assuming a mul-
tiple contractor approach, if any of the set of contractors Paul envisioned working on the water mains contracts failed in any significant respect, they could immediately be excluded from further work, perhaps after terminating their current contract, and they
would understand this was the position from the outset. In conjunction with trusting all
contractors to use their initiative and sharing the benefits this provided when they per-
formed well, this form of ‘partnership contracting’ approach would give WSL a robust,
low- cost approach to a wide range of sources of uncertainty not even worth identification
given this kind of contracting concept. If a single contract with a single contractor was
the approach adopted, this kind of simple robustness was not available. The ownership
phase should be looking for this kind of built- in robustness if any useful options were
not identified earlier, and the sooner this concern was addressed in the concept strategy
stage of the project the better. You might interpret this advice coming from Paul as the
words of ‘a poacher turned gamekeeper’, but all enlightened clients should understand
the issues involved.
Paul emphasised building and sustaining trust was a central aspect of EP approaches,
usually a complex matter. The ownership phase facilitated a careful context specific focus on
these issues. All aspects of ownership concerns would need attention in the ownership phase,
and the robustness of all associated thinking would need rigorous testing. Some aspects
might not be immediately obvious. For example, could WSL use the same contractor for
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
346 Employing planning tools in practice
consulting advice in the strategic planning stages and in the later implementation tactics and
execution and delivery stages, and if so, how would WSL deal with the associated potential
conflict- of- interest problems? Paul thought the multiple contractor approach might make
this issue reasonably easy to deal with, but there were some obvious difficulties needing
resolution.
This kind of testing of the contracting strategy and more general relationships strategy
included some particularly important aspects of the UP test phase in PUMP form within the
SPP structure. This phase provided a separate point of focus for ownership, contracting and
other relationship issues because associated opportunity and risk concerns were somewhat
different in nature and particularly difficult to deal with in some circumstances. Most com-
mon practice project risk management processes, including those currently used by WSL,
did not have a similar process component, nor was it provided as a formal aspect of any other
contributor to the concept strategy stage of project planning.
The basic SPP ‘quantify’ phase – quantify some of the uncertainty
Paul indicated that the ‘quantify’ phase of the basic SPP was a further elaboration of the UP
shape phase, and it embodied the PUMP quantify phase. Like the PUMP quantify phase,
this SPP phase started to address ‘quantitative’ modelling concerns, normally incorporat-
ing an explicit probabilistic (stochastic) component, building on the earlier ‘qualitative’
(non- numeric) modelling framework. He emphasised the importance of understanding that
probabilistic quantification was not just a matter of addressing risk in the event uncertainty
sense associated with common practice risk management approaches – it was about address-
ing all relevant sources of uncertainty. A minimum clarity approach to all sources of uncertainty deemed worth quantification
on a first pass might be appropriate. Paul showed them a simple variant of Figures 3.1 and
3.2 and explained briefly in outline using Chapter 3 material what minimum clarity involved
and how modest clarity efficient enrichment of this kind of model could be used. He also
indicated much more sophisticated approaches might be relevant, and he provided a very
brief overview of immediately relevant aspects of the rest of Chapter 3. He assured them the
planned workshops and demonstration projects would allow them to quickly acquire the
capability to develop a clarity efficient approach tailored to the context of each project they
had to deal with which would build on all the insights revealed by the process thus far. He
also emphasised that he had experience to base this assurance on.
At this point Paul emphatically emphasised a key issue – the process was not mechani-
cal. For example, at this point in the basic SPP each PG team would have to choose
what to quantify and what to treat as unquantified uncertainty associated with working
assumptions which implied ‘conditions’ for the quantified analysis. The choices made
should depend upon whether or not the PG team members making the choices thought
quantification would be useful. This need for careful and enlightened judgement was
true of earlier process decisions too, and many of the earlier choices were even more
important and much subtler. For example, back in the identify phase they needed to
decide whether to further decompose each source of uncertainty of significant interest,
depending on whether doing so looked useful. He linked this need for insightful thinking
to the creative nature of the process. The need for both thoughtful analysis and creative
synthesis was the reason most EP planners were stimulated by the work, which, in turn,
was why the quality of their work usually rose to meet the challenges which an ambitious
EP process involved.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 347
The basic SPP ‘evaluate’ phase – evaluate all the relevant implications
Paul began discussing the ‘evaluate’ phase by recommending that they avoid putting any
faith in simple analogies with other ‘evaluate’ concepts they might be familiar with. It was
crucial to see the basic SPP evaluate phase as a pivotal phase with a very rich set of roles.
The evaluate phase of the basic SPP began as yet another elaboration of the UP shape
phase, by finishing off the process of ‘shaping the plans by modelling the core issues’
concerns, like combining quantified sources of uncertainty and modelling all associated
dependence. It then moved on to residual aspects of ‘test the plans to ensure resilience and
robustness’. After that it moved on to the whole of ‘interpret the plans to exploit creativity’
plus ‘implement appropriate aspects of the plans’. The basic SPP evaluate phase was based
upon the PUMP evaluate phase.
In the context of combining sources of uncertainty, it was important to understand that
some common practice project risk management approaches involved using Monte Carlo simulation packages to combine all probabilistically quantified risks at the same time, with
a statistical independence default assumption and limited departures from this default posi-
tion. The PG team would proceed in a very different manner. Key differences would include:
1 starting with the two or three most important sources of uncertainty,
2 gradually adding further sources as a growing set of sources was understood,
3 avoiding statistical independence assumptions unless they were clearly robust,
4 using whatever approach to dependence best suited each particular context,
5 modelling all important dependencies with increasing clarity as they progressed,
6 testing all working assumptions as they went,
7 gradually building a layered understanding of overall quantified uncertainty using sen-
sitivity diagrams when appropriate,
8 making provisional decisions as they went using decision diagrams when appropriate to
seek opportunity efficiency for the project concept as a whole, and
9 always being clear about the role and implications of important assumptions which
condition the quantitative analysis and underlying qualitative analysis.
Paul emphasised the workshops he would organise would develop all the necessary skills for
those facilitating this aspect of the quantitative and underlying qualitative modelling. They
would build a significant skill set on the Chapters 3 and 4 discussions you have already been
exposed to plus Chapman and Ward (2011). However, what all PG group staff needed to
understand immediately was everybody involved in WSL projects would need to become
familiar with the use of ‘sensitivity diagrams’. He then showed his group the Figure 7.5
example, identical to Figure 4.1 example used in Chapter 4.
Paul began by explaining the Figure 7.5 example in its original BP offshore North Sea
context, as discussed in Chapter 4. Repeating some of this discussion here as a reminder for
clarity, Figure 7.5 portrayed six sources of uncertainty addressed in quantitative terms in
relation to a jacket fabrication activity during the E&D strategy stage of an offshore North
Sea project. The jacket was the steel structure fabricated in a dry dock referred to as a ‘yard’,
floated out to the oil production site, upended and pinned to the ocean floor, and then fitted
with modules for producing the oil and sending it via an oil pipeline to the shore. The six
sources of uncertainty were ordered during the structure phase to facilitate a planned senior
management discussion, noting that causal dependency relationships could also influence
the ordering of sources. The notes addressed sources of uncertainty treated as conditions –
the quantification was conditional on no major fires for example.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
348 Employing planning tools in practice
Outlining what was involved usually started with the curve for source 1 – in this case start-
ing with a composite of two mutually exclusive reasons for a delayed start to the activity (the
yard was not available because another jacket in the yard was behind schedule and had to
be finished before work on BP’s jacket could start, or mobilisation delays because the yard
was free but had not been used for some time, so a labour force with all the requisite skills
had to be built up again). The curve for source 2 portrayed the joint effect of 1 + 2, adding
to source 1 the impact of construction and weather problems relevant for the whole activity
duration. The closeness of curve 2 to curve 1 indicated a very limited contribution by source
2 to the total 1 + 2. The curves for sources 3 to 6 indicated successive further additions to
the total uncertainty. The significant size of the contributions made by sources 5 and 6
reduced a 0.9 probability of completing by the base plan end of March date to 0.15. The
ordering of sources 5 and 6 facilitated a discussion focused on the elimination of sources 5
and 6, starting with source 6.
Source 6 could be eliminated if commitment to a fabrication contract for the jacket
involved was made as soon as possible – a key role for the diagram was stimulating early
award of the fabrication contract so that uncertainty associated with source 6 was eliminated
and overall total uncertainty was now portrayed by curve 5. This was followed by discussion
about the possibility of also eliminating source 5.
Mar NovDec Jan Feb Apr May Jun Jul Aug Sep Oct
probability jacket fabrication will be complete by the dates indicated
0.2
0.4
0.6
0.8
1
0
0.3
0.5
0.7
0.9
0.1
base plan completion date
1 2
3 4
5
6
probability curves show the cumulative effect of the following sources of uncertainty:
1. yard not available or mobilisation delays 2. construction problems/adverse weather 3. subcontracted nodes delivery delays
notes: 1. the curves assume a minimum fabrication period of 20 months 2. no work is transferred offsite to improve progress 3. no major fire, explosion, or other damage
6. delayed award of fabrication contract
4. material delivery delays 5. industrial disputes
dates with relevant years indicated
Figure 7.5 Sensitivity diagram: North Sea oil project platform jacket fabrication example.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 349
Having introduced Figure 7.5 in the original BP context, Paul then linked this discussion
to the way the PG team might use this approach in the concept strategy stage to address
the new sewage works project. The WSL new sewage works equivalent to source 6 might
be possible local planning permission delays leading to additional costs which could be
avoided by committing to a preliminary design strategy contract on a consultancy basis with
one of the potential DOT strategy phase contractors for the project before the end of the
concept strategy progress stage with instructions to liaise with the local authority planners
to ensure that all their concerns were addressed to speed all associated regulatory processes.
The WSL sewage works equivalent of source 5 might be further possible local planning
approval delay driven by local residents’ objections, leading to additional costs which could
also be avoided by somewhat different responses. In this case, the best way to avoid delays
might be incorporating significant consultation with and carefully listening to local residents during the DOT strategy stage, prior to letting the preliminary design strategy contract. The WSL equivalent of source 4 might involve possible delays and additional costs driven
by ground condition and associated design uncertainty. This might be eliminated if ground
conditions for several possible sites were considered and perhaps tested if this seemed advis-
able before engaging local populations and then letting preliminary strategic design stage
contracts. This might all be part of the Projects Group and Operations Group saying to the
WSL board when the board were asked to approve the concept strategy stage plans ‘we do
not yet have a clear recommendation for site A, B or C, and they all involve very different
design and construction scenarios, but we believe it would be worth approving the concept
strategy plans and associated cost estimates now, and getting on with ground condition
surveys followed by preliminary consultations with local residents followed by a design
strategy contract’.
If a diagram like Figure 7.5 was used to develop this kind of dialogue with the board, the
board would want to revisit concept strategy stage approval at the gateway stage following
the DOT strategy progress stage, and again after the E&D and D&A strategy progress
stages. But it would not be helpful to overlook important ambiguity at the concept strategy
stage which could be resolved in various ways depending upon how WSL chose to shape
all relevant plans. This kind of thinking involved the PG team providing duration planning with built- in contingency planning. This went well beyond the construction cost concerns
usually addressed by PG group estimators, because explicitly considering design uncertainty
and linked planning permission uncertainty had significant knock- on implications which
Project Group planners could address with Operations Group planners.
Paul explained that everybody involved with WSL projects should have a shared under-
standing of the implications of all quantified sources of uncertainty using ‘sensitivity dia-
grams’ in the Figure 7.5 format or simpler linear versions beginning in the concept strategy stage. Sensitivity diagrams of this kind were a key EP tool. More than half a dozen sources on any one diagram usually made reading them too complex, but sensitivity diagrams could
be used in a nested form, employing from two to ten sources per diagram to cope with
any required number of sources in any framework structure which suited the context. For
offshore North Sea projects in the E&D strategy stage, BP targeted 20 activities (with a
nominal upper limit of 50), and the number of quantified sources per activity was usually
in the range 5 to 15, each associated with one to five responses specific to that source, plus
general responses. WSL activity and source- response structures would be much simpler in all lifecycle stages, at their simplest in the concept strategy stage, and they would seek sim- plicity in a systematic manner. He would later organise workshops which would help them
to develop an initial feel for judging what level and kind of complexity would work best in
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
350 Employing planning tools in practice
WSL contexts. They could build on this basis employing a learn- by- doing approach like all
other organisations which had adopted EP approaches.
These sensitivity diagrams would be used in conjunction with ‘decision diagrams’, which
Paul now explained using Figures 3.5 through 3.8 to develop a summary of the Chapter 3
discussion of clarity efficiency, risk efficiency and opportunity efficiency, followed by WSL
concept strategy progress stage reinterpretations.
The purpose of the evaluate phase as a whole was evaluating all the relevant implica-
tions of a creative planning process. This had to be done in a way which allowed all the
important insights to be communicated effectively to everyone who needed to understand
them after the PG team had used these insights to shape, test, and creatively interpret the plans. Combining the uncertainty sources which were quantified in the previous phase using
standard commercially available Monte Carlo– based simulation software would be a part
of the evaluate phase effort, essential to beginning the evaluation required. But it was a
relatively minor part of the evaluation phase as a whole. Of crucial importance was clearly
communicated collective understanding of the implications of assumptions about statistical
and causal dependence when combining sources, assumptions about what combinations of
responses were the preferred options, and assumptions about the significance of sources not
quantified – treated as conditions. Furthermore, early passes through the evaluate phase had
to attempt to clarify for the analysts which areas need further work to let them plan further passes in a clarity efficient manner. More generally, PG planners and Operations Group plan-
ners, plus the relevant project initiators, needed to develop a joint understanding of why
they needed to iterate the process and revisit areas identified by the evaluate phase until they
were satisfied that their plans were fit for the purpose before they recommended moving on to gaining approval for their plans.
Paul reminded his team that Figure 7.3 suggested that the first pass might take 20% of
the time available, leaving 80% to focus on the now identified 20% of the issues needing 80%
of the effort. This was an application of the widely cited 80:20 rule, often called the Pareto
rule, although a more ambitious 90:10 rule might be a stretch target aspiration. Sensitivity
diagrams like Figure 7.5 were a key tool for deciding whether or not further iterations were
likely to be useful, and if so, which sources of uncertainty looked worth further decomposi-
tion, before as well as after the kind of management discussion just considered. They cur- rently did not use sensitivity diagrams of this kind, or decision diagrams, clear symptoms of
a lack of understanding of what an effective and efficient evaluate phase ought to involve.
Paul then completed his phase- by- phase discussion of Figure 7.2 by linking each phase of
the basic SPP back to their alignment with the UP phases and PUMP phases in Table 7.4.
He first noted that the UP interpret the results- creatively phase was central to the basic
SPP evaluate phase. That is, the ‘combining probability distributions’ aspect of the evaluate
phase was a necessary component part of the modelling, but doing so incrementally using
sensitivity diagrams to interpret the results plus decision diagrams to build up an integrated
set of provisional decisions (contingency plans) was the central concern.
Paul pointed out that the UP test the process rigorously phase was embedded in earlier
phases to a significant extent, but a final summary ‘test’ in the evaluate phase had to address
both ‘the project’ plans and ‘the process’ plans. That is, iteration control as envisaged in the
basic PUMP concept meant considering the possibility of revising both the project plans and
the process plans on each successive potential loop back until moving on to the following
gateway stage could be judged appropriate. Planning the project and a higher order plan-
ning the planning were both involved, and they should not be confused. The final aspect of
the UP phase ‘implement the results’ was moving on to the next stage when iteration was
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 351
completed, in this case moving on to the concept strategy gateway stage when the concept
strategy progress stage conclusions and recommendations were judged fit for the purpose.
The flexible iteration structure of the Figure 7.2 format and the underlying Figure 2.1
format generalised the simpler iterative structure of the Figure 7.3 linked bar chart por-
trayal, with implications which Paul would explore at a later date.
The role of the capability- culture assets and liability concepts which Figure 7.2 included
within the outer box would also need exploring at a later date (building on all the ideas
developed in previous chapters).
An interim overview for the PG team
At this point Paul used the PG team’s initial understanding of the project lifecycle, seven
Ws, goals– plans relationships and basic SPP frameworks as an integrated toolset to give an
overview of changes in the WSL approach which would impact the PG team.
One key set of changes was working together in a fully integrated manner, dissolving
the current departmental boundaries within the PG group, and softening boundaries with
other groups, especially their boundary with the Operations Group. Other closely cou-
pled changes included beginning intensive PG input in the concept strategy stage, working
directly on all of the activity- based plans, but also helping to coordinate development of all
the other plans associated with the seven Ws, including helping to develop common corpo-
rate understanding of key goals– plans relationships concerns.
The linked set of all changes driven by their new EP approach had a number of very broad
implications which were important. For example:
1 The risks which current Risk Management Department staff looked for in earlier concept
stage estimates would be more broadly framed as ‘sources’, a convenient contraction
of ‘sources of uncertainty’. These sources could include issues like ‘Should a different
design for a sewage plant in a different location be addressed?’ The use of the word risks in the current Risk Management Department sense would be dropped. It was essential
to do so because the ‘baggage’ which a common practice view of risks carried had impli-
cations which needed to be left behind by everyone involved in WSL planning.
2 The ‘responses’ to sources interpreted in this general way might have to include con-
cept strategy progress stage planning associated with consulting customers, organising
planning permissions processes, and relating both these activities to the timing of prop-
erty purchases and ground- condition tests. This planning activity might not be under-
taken with a view to resolving detailed planning concerns. Its aim might be tentative
planning with a view to developing a better understanding all the ambiguity involved,
and sizing all relevant uncertainty in approximate but unbiased terms, before deciding
whether several possible variants of the project concept were the basis of the best way
forward. It required the use of prototype variants of the DOT, E&D and D&A strategy
progress stage approaches to be employed later in more developed forms.
3 The generalisation of traditional project risk management this involved meant that cur-
rent Planning Department, Estimating Department and Risk Management Department
staff needed direct and coordinated joint involvement from the outset.
4 Operations Group staff should be involved to a significant extent in a coordinated
manner as soon as possible, and facilitating closer integration of effort across all groups
involved in projects within and outside the WSL corporate structure would be an
important feature of the new PG team approach.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
352 Employing planning tools in practice
5 The basic SPP framework provided a practical overall process framework to facilitate
this integration of both process ideas and input to particular projects.
6 Thinking about all plans from the outset in terms of shaping ambiguity in an enlight-
ened manner, and ensuring WSL avoided their current misleading rush towards
inappropriate certainty, would be a fundamental change in mindset for everyone
involved.
An extended overview not provided by Paul for his first workshop
Paul did not want to overload his first workshop more than he had to, and you probably feel
more than ready to move on too, but you may find a brief treatment of further observations
on the Figure 7.2 links to earlier discussion useful now.
Everyone involved in the use of a process like the basic SPP needs to remain aware of the
important difference between useful models and much more complex realities. The basic
SPP is a concept model to be used operationally as a map or plan – like the UP as outlined in
Chapter 2 and demonstrated in Chapter 5 and the SP for bidding of Chapter 6. It is always
important to avoid confusing a model with the more complex reality the model represents.
Anyone familiar with going for a walk or driving a car or sailing a boat with a plan based on
a map will understand the possibility of important differences between reasonable plans and
what actually happens if best use is made of the plans, with effective departures as intruding
realities are encountered along the way. However, process models used for management
decision making have a tendency to lead those using well- worn processes to lose sight of
the limitations of familiar process portrayals. This is especially true if the processes being
used are presented in simple and highly prescriptive terms as mandated corporate processes,
perhaps based on professional guidelines.
This limitation of process models is crucial, but equally crucial is the need to employ the
best maps we can get based on empirical observation, simplified to a level which is conveni-
ent for current usage. The basic SPP and its embedded four Fs are based on a synthesis of
what works in practice in three integrated base and contingency planning spheres: general
unspecified planning contexts, project planning contexts, and project risk management con-
texts when a broad uncertainty and complexity management interpretation is used. If the
basic SPP is not used, then something demonstrably better needs to replace it. If this is not
done explicitly, something which is probably inferior will replace it implicitly. Like Figures 2.1 and 4.4, Figures 7.2 and 7.3 are process models which help to shape
creative thinking, to be used to guide what people actually do – not to constrain their think-
ing. They indicate ‘what needs to be done’ to make effective and efficient decisions at an
overview level, not ‘how to do it’. They provide guidance on the most appropriate sequence
for doing things, but their iterative nature means a very flexible approach to sequencing
issues is crucially important in practice. The capability- culture assets and liabilities of those
involved in implementing this approach will determine ‘how it is done’ and the overall effec-
tiveness and efficiency of the planning.
The role of the basic PUMP in the basic SPP is the provision of a generic process model
for addressing opportunity and risk in terms of underlying uncertainty and complexity in
a project planning context which is the most general I am aware of. It has been used suc-
cessfully in practice by a range of organisations in various prototype forms for more than
30 years, and the prototypes themselves rest on earlier successful toolsets. Your organisation
can add other approaches which prove useful in the context of ‘further contributing pro-
cesses’. But keep testing what you use for clarity efficiency.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 353
The primary role for ‘further contributing processes’ is up- front integration of use-
ful processes for creating and enhancing initial plans and then developing shaped base
plans, drawn from all literature and practice relevant to projects, including everything
relevant which you and your colleagues currently know how to use, and all relevant
future learning.
The role of the UP is expanding the scope of PUMP and all the other embedded project
management process approaches so the basic SPP looks at all relevant aspects of creating, enhancing, shaping, and implementing project plans in an integrated manner.
The capability- culture assets and liability concepts can make important contributions to
this generalisation role for the UP. For example, ‘effective and efficient monitoring and
review within the PG team’ might be defined as a key capability- culture asset. This asset
might address issues like ensuring all estimates are unbiased and everyone involved learns
from experience (in the manner outlined in the Astro tale of Chapter 6) under the over-
all supervision of a PG management team which understand what this means and what it
requires in the way of aptitudes, skills, training, and on-the-job support. A facilitating role
for PG staff when other WSL planners were the central players was a crucial component of
Paul’s need, as a board member, to deliver unbiased estimates to the board for opportunity
efficient plans, and to demonstrate that he had done so. A failure in any of the plans covered
by any of the seven Ws undertaken by other WSL groups would mean a project failure which
Paul would bear some responsibility for, even if the plans involved were not within his direct
remit. Persistent bias was a guaranteed source of failure. Any guaranteed source of failure
implies a form of ‘premeditated failure’.
A specific process for the strategy gateway stages
Paul reminded his group that the basic SPP portrayed by Figure 7.2 was designed to facili-
tate WSL planners making progress until all the project plans embraced by the seven Ws
and associated goals– plans relationships were judged ‘fit for the purpose’ at the end of the
concept strategy progress stage. At this point independent governance was needed by the
WSL board, treated as a separate concept strategy gateway stage. Different people were
involved, the gateway and its preceding progress stage process purposes were very different, and an appropriate gateway process was not iterative by design.
He then showed them the gateway stage process of Figure 7.6, which he would refer to
as the ‘gateway SPP’, and began his explanation by indicating that it was simply the PUMP
gateway process discussed by Chapman and Ward (2011) embedded in the UP framework
of Figure 2.1. It was a gateway SPP which would serve all the WSL strategy gateway stages
in conjunction with the basic SPP of Figure 7.2 serving all four of the WSL strategy progress
stages. Later progress and gateway stages would require modest modifications, discussed by
Chapman and Ward (2011), which would not be considered that day.
The ‘consolidate and explain earlier progress stage deliverables’ part of the gateway SPP
involved producing an executive summary for the board of all relevant base plans, contin-
gency plans, and commitment plans, with backup detail as needed. What was presented to
the board should be judged ‘fit for purpose’ if it allowed the board to decide whether:
1 the project should be passed to the next progress stage as recommended,
2 the project should be terminated now for identified reasons as recommended, or
3 the project should be sent back to an earlier progress stage for identified reasons as
recommended.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
354 Employing planning tools in practice
If these deliverables were judged ‘not fully fit for purpose’, the project plans might be sent back for modest revision or clarification or both – a form of ‘gateway maybe’ outcome which
required the resolution of unacceptable aspects of the plans or unacceptable ambiguity in
the basis for current recommendations. This outcome might imply a degree of incompe-
tence during the progress stage without making too much of an issue about it. This was
quite different from a form of gateway maybe which accepted the validity of the progress
stage but acknowledged that further ambiguity resolution or potential new opportunities
identified by the board during the gateway stage was worth the investment. Crucially, it was
also very different from a form of gateway maybe which rejected the validity of the progress
stage explicitly, explicitly calling into question the competence of those making the recom-
mendations. Important complexity in a maybe might be left ambiguous or spelt out very
to the next progress
stage
capability-culture
assets
capability-culture
liabilities
from an earlier progress
stage
deliverables fit for
purpose?
gateway maybe gateway yes
gateway no
consolidate and explain earlier progress
stage deliverables
board level support and convince
process
commitment plans
contingency plans
base plans
summary report
stop the
project
back to an earlier progress
stage
Figure 7.6 A gateway SPP to follow the basic SPP of Figure 7.2.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 355
clearly, at the board’s discretion, and spelling it out very clearly might be a precursor to
somebody being subjected to explicit personal criticism plus proportionate action.
No planned iterations were involved in this gateway SPP. Usually recommended options for further ambiguity resolution would only be offered because current timetable, scope, or
financial constraints did not allow a clear recommendation to proceed to the next stage by
the planning team. Usually new opportunities or risks identified at the board level would be
surprises which might have been anticipated and explored earlier.
What the PG team needed to understand that day was when the PG planners and all the
other WSL planners involved thought they were ready to move from a strategy progress
stage using Figure 7.2 to a strategy gateway stage using Figure 7.6, a very different process
began, usefully associated with an entirely different project lifecycle stage. Unlike preceding
strategy progress stage SPPs, a following strategy gateway SPP was not iterative by design,
and the decision makers were the board, not the PG planners and other WSL planners.
The PG planners would be directly involved in the first part of this gateway SPP, working
with other WSL planners to produce a summary report. This report would be based on doc-
umentation developed during the progress stage basic SPP. It would incorporate expected
outcomes and associated ranges for all key parameters. It would include associated base
plans (based on balanced targets or on aspirational targets equivalent to P10 plans in dura-
tion, cost, and other performance measurement terms). It would include basic contingency
plans (indicating fall- back options to cope with targets not being achieved plus ‘fall- forward’
options to capitalise on good luck not incorporated in the target values used for base plans).
It would include commitment plans (indicating the equivalent of P90 or P80 plans in dura-
tion, cost, and other performance measurement terms). It might also include further contin-
gency plans to deal with failing to meet commitments which need prior planning.
Paul would take a direct interest in helping Project Group and Operations Group team
members and other WSL planners meet the board’s expectations, coordinating all their
input to the board, and presenting their plans to the board. But the decision making based
on that input in the gateway SPP of Figure 7.6 would be a boardroom exercise.
Paul would later explain to the board that embedding the PUMP gateway in the UP con-
cept to define the gateway SPP of Figure 7.6 emphasised several key features:
1 The full UP process in its basic SPP adaptation had been used by PG planners and
other WSL planners to prepare the basis for a board decision, but the board members
could now use all relevant UP concepts to add their own judgement to any aspect, they
should do so wherever and whenever they had knowledge not available earlier, and they
should expect to play an important role in these terms.
2 The board members should also add their own creativity when appropriate, asking the
question ‘Do the plans as a whole, given all the assumptions underlying them as just
considered, offer any further scope for creative improvements in opportunity efficiency,
perhaps involving a redefined wider scope or more financial flexibility?’
3 Even if the board members did not add knowledge or creativity which might lead to
revised plan suggestions, they needed to ask the question ‘Do the plans as a whole, as understood by the board, make sense, making full use of all board level expertise?’
4 Robustness testing in very broad terms within a UP framework was a key board respon-
sibility which could not be delegated.
All the capability- culture asset concepts discussed in Chapter 2 were relevant, not just the governance role capabilities. So were the associated capability- culture liabilities and
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
356 Employing planning tools in practice
accommodations. Paul understood these issues should not be laboured with the board, but
as a responsible board member he had to confront them when and where appropriate.
The board might, of course, reject a concept strategy stage recommendation that a pro-
ject should proceed to the next stage or be stopped now and send it back for reconsidera-
tion with a very different line of thinking. The board might even reject a recommendation
to return to an earlier stage, because the grounds were held to be inappropriate, and insist
that different considerations were addressed in a different manner. However, this kind of
iteration suggests earlier errors of judgement. It explicitly finds the recommendations are not fit for purpose. This ought to be avoided if possible, and it should be unusual, possibly
career- limiting.
The planned iteration which Paul argued must be built into the progress stage basic SPP evaluate phase involved using insight generated by the evaluate phase to decide whether a
further pass of the whole SPP or part of the SPP would be useful. In this context, sensitivity
diagrams like Figure 7.5 played a central role which was very different from their commu-
nication role for the board. For example, if a large contributor to overall uncertainty like
source 5 in Figure 7.5 was not well understood, and it looked as if this mattered, further
effort to understand what was involved and how it might be managed needed careful con-
sideration as a process management decision before going to the board. However, given the effective communication provided for the presumed one pass gateway
SPP, board members might well draw different inferences based on wider experience or
different perspectives, a central aspect of governance interpreted in positive terms, and this
might lead to further unplanned iterations.
Wrapping up his first workshop
Paul brought his first workshop to a close with a brief review session. He made sure that all
the main messages had been absorbed, and there was a positive perspective on the changes
ahead. He knew that he had just started a change management project of considerable
complexity which would raise many challenges, for many practical purposes usefully seen
as a programme or portfolio of change management programmes and projects. As soon as
possible he wanted to start to build an appetite for change, confronting the associated chal-
lenges working with individuals and small groups. He made this very clear and emphasised
that this first workshop was about a very high level summary overview of where they were
going which they could all share at the outset to help them get started. Paul wanted to move
each member of his group forward at a pace they could cope with, and he wanted to ensure
that capability development and culture changes kept pace with the tasks they addressed.
Paul’s overall change management strategy – initial goals
Paul had a much more developed overall strategic change management plan than Martha
in Chapter 6. In part this was because he had extensive project planning experience using
PUMP- based approaches which provided a sound foundation for his current understanding
of the basic SPP, the associated gateway SPP, and the ‘SPP pack’ of all progress and gateway
SPP concepts for all lifecycle stages. In part, it was also because he could use a variant of all
these SPP concepts to address the management of change within WSL – a higher level use
of the same basic conceptual framework. But he was open to changes in his strategy and the
operational form of his underlying concepts as he learned during the initial implementation
and ongoing development of his strategy, and he took a very flexible approach to his rate of
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 357
progress. He wanted to avoid rushing key developments that had to bed down to provide
firm foundations for later developments.
While his group was adapting to their new approach via a programme of technical work-
shops to be discussed shortly, trying out the ideas using some basic aspects of current pro-
jects in progress as simple test beds, and having one- to- one conversations about their own
roles, Paul started parallel detailed discussions with Ollie. As soon as he could, Paul devel-
oped a close working relationship with Ollie at a personal level, working towards a strong
partnership between the Operations Group and the Projects Group.
When Paul was confident that he and Ollie had a rich set of planning options which
would integrate Operations Group and Projects Group input to the business case for
the proposed new sewage plant project, Paul did some joint strategic change manage-
ment planning with Ollie. They then jointly briefed Michael. As part of briefing Michael,
they outlined the purpose of a proposed meeting with Frank and ensured that they had
Michael’s support for this next stage. They then approached Frank with a view to a joint
presentation about their progress so far on the sewage plant and some related issues
involving the water mains project. When Frank agreed, and a date was set, Paul had sev-
eral rehearsals with Ollie, before a meeting designed to lead Frank to the conclusion that
he had to significantly redefine his role in WSL business plan modelling. For all projects
within the PG remit, near the outset of the concept strategy stage WSL needed much more
focus on plans for design purposes created by the Operations Group to identify options
which would drive early versions of all the other plans, including dependent business case
plans. Crucially, the business case plans were not the primary driver for the sewage works,
nor should they be viewed as the primary driver for the new water mains programme. Fur-
thermore, the discount rate applied to future cash flow was not a parameter WSL could
treat as a technical judgement within a straightforward NPV process approach best left to
Frank without board level scrutiny. Board level responsibility for testing the robustness of
all underlying assumptions needed a very different approach. At their heart was the impli- cations of assumptions underlying the discount rate and the decision- making process it
was embedded in for business case purposes which had important interdependencies with
central concerns for Ollie and Paul.
The discount rate used to shape and select projects
Paul started the meeting with Frank with an outline of the proposed agenda. Paul did not
hide the importance of a discussion about the discount rate which he would soon come to.
However, in addition to the usual opening pleasantries, Paul wanted to begin by indirectly
selling Frank the benefits of an EP approach in broad terms. He particularly wanted to sell
the proposition that an EP approach for all aspects of WSL decision making would give
Frank’s role as custodian of WSL finances more opportunities. He also wanted to ensure
that Frank understood why uncertainty and complexity associated with their sewage plant
case – like no choice of site as yet with all the follow- on uncertainties – was an opportunity
to be embraced by the board, not risk to be avoided.
Paul began his discussion of the appropriate discount rate to use for NPV calculations by
acknowledging that this was usually seen as the Finance Director’s responsibility in com-
panies like WSL. As an immediately relevant comparison, it was usually seen as the respon-
sibility of bodies like HM Treasury for UK government decisions. Furthermore, those
responsible for choosing the discount rate to be used usually drew on financial economics as
outlined by standard texts which were based on well- established schools of thought.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
358 Employing planning tools in practice
However, HM Treasury had recently acknowledged earlier misconceptions of a very seri-
ous and fundamental nature, which had important, direct implications for the private sector
as well as the public sector, especially private sector companies like WSL which used bond
funding to leverage return on shares. A fundamental underlying flaw in most (not quite all) NPV modelling was a failure to recognise the need for more than one hurdle rate test if more than one objective or any constraints on investment were involved, keeping most of these tests OUTSIDE the discount rate. This was essential to avoid bias associated with a discount rate set too high in an attempt to get away with a single- hurdle rate test of the form ‘Is the NPV
positive?’ Paul emphasised that HM Treasury (2003a) recognised this fundamental problem
and mandated a multiple test approach for all UK public sector decision making, explicitly removing the ‘risk premium’ and ‘the opportunity cost of other forgone investments’ from
the discount rate mandated by HM Treasury prior to 2003, acknowledging a long history
of previously getting this important issue wrong.
Paul said he had undertaken considerable background reading – enough to verify that the
question of choosing an appropriate discount rate had a very long and confusing history,
with no stable generally accepted ‘received wisdom’ as yet, although many people would
disagree and argue that their perspective was ‘the received wisdom which was correct’. This was the case 50 years ago, and central underlying issues were still unresolved. His position
would be seen as highly contentious by many people, and he acknowledged he was not a
financial economist. But anyone disagreeing with him would have to acknowledge a particu-
lar school of thought and demonstrate a clear understanding of the nature of disagreements
with other schools of thought if they had any real clarity about the issues. There was a very
old ‘joke’ that could be stated in the form ‘If you ask 10 economists a serious question you
should be prepared for at least 11 answers’. It was not very funny if the serious question
urgently needed an answer, but it was relevant to understanding what was going on here.
So was the EP position on this controversy, and the concept of stealth assumptions from an
EP perspective.
The starting point for understanding an EP position, which Paul made direct use of, was
the paper ‘An Optimized Multiple Test Framework for Project Selection in the Public Sector
with a Nuclear Waste Disposal Case- based Example’ (Chapman, Ward, and Klein, 2006).
This paper argued that in public sector projects the government’s expected future bond
rate was the only consideration which should be embedded in the discount rate – all other
considerations needed separate tests, building on the HM Treasury (2003a) multiple- tests
approach. A positive NPV using the expected future bond rate for discounting purposes
involves passing a ‘bond test’. Passing this bond test confirms that if expectations about the
bond rate and the cash flows were correct, then the cost of borrowing could be covered. No other objectives or constraints were addressed by this test, and its validity depended on unbiased estimates of all expected outcomes being used – not just the discount rate.
A separate ‘return test’ was needed to address the question, ‘Does the surplus return
after borrowing costs compare favourably with other available options given constraints
on public sector investment in this particular sector, like education, defence or transport?’ This was a generalised variant of the opportunity cost considerations often embedded in the
discount rate for all possible government investments, but it was specific to expenditure cat- egories as well as being outside the discount rate. A high priority expenditure category implied a relatively low return requirement and vice versa – priorities were reflected by the shadow
prices associated with constraints defining how much investment to make in each expendi-
ture area. As government priorities changed for political reasons, different priorities could
be reflected by different return tests in different expenditure areas, with significant negative
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 359
returns allowed in very high priority cases. As an extreme example, military expenditure
would not be tightly constrained in the middle of a war of survival, but it would be in a
time of peace with no obvious prospects of war on the horizon. From an EP perspective,
embedding these return tests in opportunity cost contributions to the discount rate was an
extremely serious conceptual mistake, with crucial operational bias implications, because it
muddled different return requirements for different sectors in a discounting process which
should not address considerations other than the expected cost of capital raised via govern-
ment bonds.
Further necessary tests included ‘risk’ tests – to address all relevant sources of uncertainty giving rise to important risks. Many of the most important risks were not financial or finan-
cial markets related – like the possibility of contaminating groundwater with nuclear waste
if disposal of nuclear waste using deep mines was allowed to proceed or the risk of failing
to preserve a UK capability to develop nuclear power if disposal was deferred and this had
an impact on political pressure to stop producing more nuclear waste and the demise of
national nuclear engineering capability. From an EP perspective, ‘risk premiums’ in a dis-
count rate were another extremely serious conceptual mistake, with crucial operational error
implications. Even if the risks involved were purely financial in nature, which often they
clearly were not, the market- based assumptions, plus any other assumptions needed, require
effective robustness testing. Any aspects of projects which may change fundamental market
perceptions of an organisation needed very special care and attention, even if the organisa- tion involved is a national government.
In a national government context, important further tests were needed to address inter-
generational transfer issues – was this generation having economic advantages paid for by
future generations, or vice versa, to an extent that either generation would find unaccepta-
ble? This ‘heritage’ test was a good example of a test which had to be addressed at a port-
folio level (not a specific project level or a particular sector level) because it was the overall
balance across all new projects which mattered, one of several ongoing flaws in the 2003
HM Treasury position from an EP perspective. Furthermore, this heritage test was also a
good example of a test with ethical content – simply observing past revealed preferences (as
the 2003 HM Treasury approach does) involved working assumptions which needed to be
questioned and tested effectively.
The Chapman, Ward, and Klein (2006) paper showed why a 1990s decision by the UK
Department of the Environment (DoE) to postpone nuclear waste disposal in order to save
£100 million using the HM Treasury mandated real discount rate of 6% real (DoE, 1994) at
that time was seriously ill informed. It was actually a £2 billion mistake, ‘other things being
equal’, in terms of the less than 3% real discount rate mandated by HM Treasury a decade
later (HM Treasury, 2003a). On its own figures, given its own assumptions, what might be
reasonably interpreted as a £2 billion error was involved because of mandated HM Treasury
advice. It was very obvious that ‘other things were by no means equal’, but it was also very
obvious these ‘other things’ had been ignored in terms of the single test NPV approach
mandated by the Treasury before 2003, and not effectively considered within the rest of
the decision- making framework. In effect, the 1994 decision using a discount rate which
was far too high gave excess weight to the money saved by putting off doing anything new
and too little weight to the future implications, bearing in mind sooner or later something
different had to be done, with no direct consideration of the risks and opportunities which
really mattered.
Paul emphasised that it was the implications of immediate relevance to WSL which mat-
tered to WSL, not the nuclear waste issue or the more general debate about ‘best’ practice
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
360 Employing planning tools in practice
decision making in a discounted cash flow framework. As explained by Chapman and Ward
(2011) when using a private sector water mains investment example directly comparable to
the WSL context, in the simplest case this boiled down to two WSL discount rates for two
portfolios, one associated with bond funding and the other associated with corporate work-
ing capital funding, and separate tests for all relevant objectives other than covering the cost
of capital for associated planning.
For example, Paul, Ollie, and Michael were all agreed that the new sewage works project
should be part of WSL’s ‘bond funded portfolio’, and he was sure Frank would also agree.
This meant Frank ought to make the best estimate he could of the expected cost of bond
funding over the life of the asset and use this bond rate in his NPV calculation. He should
also take a view on the best timing for this funding, which might or might not coincide with
the currently planned timing for the execution stage of the project. The board should use
Frank’s views on bond market rates and funding availability when considering trade- offs
between alternatives like the proposed new sewage works and the water mains replacement
programme and the timing for both because bond funding could be used for both and
relative returns after the cost of bond funded capital investment as well as other benefits
and risks had to be addressed explicitly and directly by the board. Issues beyond return on
investment, including regulator and public pressure, might be crucial. Most important, in
the case of both these projects the low bond rate relative to the much higher rate of return
on shares expected by shareholders meant that WSL would avoid the well- understood prob-
lems of bias associated with using discount rates which were too high – a bias towards ‘quick
buck’ high risk projects, away from long payback period low risk projects. Current WSL
practice meant that the discount rates used were too high because they incorporated a risk
premium as well as other factors which should be outside the discounting process.
At this point Ollie led the conversation for a while, as prearranged. He began by empha-
sising that it was not just a question of which projects were selected when choosing between
projects. The discount rate also had a dramatic effect on the way projects were shaped prior
to the project selection process. He used current planning for the new sewage works to
explain that WSL traditionally used a high discount rate which was driven by shareholder
rate of return expectations. This led to what might be termed ‘low quality’ choices, in the
sense that they involved relatively low capital cost estimates coupled to relatively high oper-
ating cost estimates. He was now exploring the use of a relatively low discount rate which
was largely driven by bond funding rates. This was leading to what might be termed ‘high
quality’ choices, involving relatively high capital costs coupled to relatively low operating
costs. It was also leading to a very rich set of new kinds of options because of new levels of
creativity triggered by the new perspective. He emphasised the obvious direct importance
of the lower discount rate in justifying much higher quality plant with much lower main-
tenance cost expectations as well as a much longer life expectancy. He also emphasised the
importance of the new insight he now had, which he hoped would become a mindset shift
associated with EP thinking in WSL: quality now matters much more to WSL than it did previously because a lower discount rate means the future now matters much more – chasing a ‘quick buck’ is now much less attractive, despite a very aggressive sustained focus on a high expected rate of return for WSL shareholders.
This can be seen as simply recognising the implications of the well- established benefits of
using other people’s money (extending a key aspect of the Chapter 5 tale).
Paul then suggested that they briefly discuss the ‘working capital funded portfolio’. This
second portfolio would involve NPV calculations using a discount rate based on Frank’s
best estimate of WSL shareholder rate of return expectations. This was the working capital
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 361
funded portfolio for investments which clearly should not be viewed as bond funded. A rel-
evant current example was a project to encourage customers to use less water initiated and
managed by Curt, the Customer Relations Director. It involved significant investment in
advertising, advertising having an impact half- life measured in months, as well as subsi-
dised sales of water butts for storing rainwater by customers and other devices which would
impact consumption for some time. A shareholder expected rate of return was clearly more
appropriate than a bond rate as the basis of the discount rate for some aspects of this kind
of investment.
There might be some projects which involved the middle ground, especially if funding
issues linked to project lifecycle concerns were highly uncertain. In such cases generalising a
two portfolio approach to three or more portfolios might be useful.
Paul then observed that the arguments used to support received wisdom approaches
which always used the shareholder expectation rate or a weighted average of this rate and
the bond rate were using stealth assumptions which did not withstand scrutiny from an
EP perspective. If WSL discounting used an expected rate of return on shares or an 80/20
weighting of the bond and shareholder expectation rates for all business cases as their central
test, they were implicitly assuming there was no distinction between investing in new water
mains designed to last 100 years exclusively funded at bond rates and advertising to reduce
water consumption with a half- life impact measured in months funded via working capital.
Frank quickly understood in intuitive terms the stealth assumption implications of the
‘at least two separate portfolios and a multiple tests approach to discounting’ EP perspec-
tive. He had always been slightly uneasy about the received wisdom used in WSL’s current
approach, but unwilling to say so and depart from his textbook basis without a clear ration-
ale for an alternative. Paul anticipated this understanding and pressed on, but in case it was
needed he had prepared additional explanation, explored with Michael earlier. You might
find a brief exploration of Paul’s contingency plan explanation useful now.
The EP perspective on discount rates underlying Paul’s two or more portfolios, and a
multiple tests discount rate approach, can be explored in terms of four steps.
The first step involves understanding that the most general (unrestrictive) basis for adding
monetary values in time period t and time period t + 1 for any relevant t is a decision maker
preference (or revealed preference) function which effectively asks, ‘How much money in
time period t + 1 is equivalent to £1 in period t, as a basis for “adding apples to apple equiva-
lents” to get a total in terms of apple equivalents?’ If the answer is £1.04 the implication is
a 4% discount rate equates to the implied indifference needed in order to work with ‘apple
equivalents’ instead of ‘apples, pears, bananas and so on’.
The second step involves recognising that a 4% discount rate is a reasonable working
assumption if the options being considered can be funded at 4% real (net of inflation) on
average over the horizon being considered, and all cash flows are in today’s (un- inflated)
money values. However, a positive NPV using this discount rate simply indicates a ‘borrow-
ing rate test’ has been passed, which means the decision maker can expect to get their money
back after borrowing costs.
The third step involves appreciating that putting any other considerations into the dis-
count rate to facilitate a single test approach produces bias, distorting the role of discount-
ing. Future cash flows (positive or negative) will be excessively discounted. For example,
adding an opportunity cost premium to impose capital rationing through a higher hur-
dle rate means the decision maker will reject low risk projects involving a long payback
period which should be accepted, replacing them with high risk quick- buck projects which
should be rejected. Opportunity cost considerations should be considered separately, using
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
362 Employing planning tools in practice
appropriate shadow prices outside the discount rate, even if this is the only additional con-
sideration and its nature is very simple. Many economists made this very clear in the 1960s
and 1970s, in the different context of discussing IRR (Internal Rate of Return, the discount
rate that produces a zero NPV). The key concern then was arguing the case for never
choosing between projects based on relative IRR, although using a cash flow defined by the
difference in alternative projects was a very useful parametric analysis approach in an IRR
framework if decision makers were unsure what interest rate should be used, and all relevant working assumptions were understood.
The final step, as outlined in Chapman, Ward and Klein (2006) and Chapman and Ward
(2011), is recognising that a multiple criteria mathematical programming framework is the
most general conceptual framework to deal with all relevant objectives and constraints other
than the cost of borrowing embedded in the discount rate, but simple and visible working
assumptions can be used to provide simple practical approaches which do not require direct
use of mathematical programming. For example, capital rationing in particular bond funded
portfolio areas (water mains versus sewage plants for example) could be imposed by guess-
ing at the shadow price associated with an annual capital spending budget constraint in each
spending area, testing each new possible investment with this shadow price in terms of rate
of return over and above borrowing cost, passing clear winners immediately and rejecting
clear losers immediately, but ordering all the ‘maybe’ outcome projects towards the end of
the financial period used for planning to spend whatever remains of the available funding.
WSL operation of some variant of a two- portfolio approach means Frank has to ensure
WSL does not invest more at any given time than the bond market can take (with bank
bridging funding if appropriate). Frank also needs to ensure that this bond funding is not
used in ways the bond market would not endorse. But Frank is not a gatekeeper for pro-
jects operating a single hurdle rate test. The board as a whole will have to take a view on
the monetary expected return and associated risk after borrowing costs plus all the other expected benefits and risks of options like new sewage works, water main replacement and
conservation measures. Other benefits and risks include customer perceptions of WSL and
regulator relationship concerns.
Furthermore, Frank, Ollie and Paul needed to work together to manage other constraints
and factors, like the limitations of local contractor availability, and the implications of traf-
fic disruption caused by WSL construction activities or WSL operations disruption issues.
Many of these issues need qualitative judgements at the board level, and some implied
ongoing planning to deal with complexity needing attention outside the basic business case
metrics embedded in common hurdle rate tests.
Having omitted this aside, which Paul had ready to go in the form outlined earlier, Paul
went on to clarify his other concerns by suggesting that he had two lines of input he would
like to pursue later with Frank. One was the use of a bond rate discount factor made higher
capital cost but longer life and lower maintenance water mains a preferred option – as for
all bond funded WSL capital assets. The other was replacing the one national contrac- tor approach currently proposed with a portfolio of contracts for tranches of new water
mains installation. They could use several local contractors as well as several national major
players. This would yield significantly lower costs. They could ensure a degree of ongoing
competition between multiple contractors to keep delivered quality high and costs achieved
low. Frank, Ollie and he could work together with other directors to speed up or slow
down the programme depending on both bond rates and construction cost rates, as well as
regulator and customer views on leakage rates from old water mains, and local resident and
council (local government) views on traffic disruption due to mains replacement projects.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 363
Low interest rates coupled to low economic activity driving low construction cost expecta-
tions was the time to press ahead with the maximum rate of progress they could manage.
The issues they would have to address in this maximum rate of progress scenario would
include avoiding traffic and WSL operations disruption that was unacceptable, involving
close coordination with Ollie and local government road traffic planners. He indicated that
Michael was very enthusiastic about this possibility, as was their board chairman Charles. If Frank were to present a business case to the board based on this approach, Paul was sure the support for Frank would be unanimous.
Paul did everything he could to ensure that Frank felt he was a key player in the new
thinking that was going to implement an EP perspective across the whole of WSL’s plan-
ning, and in no way remiss in not being aware of the flaws in conventional wisdom about
NPV discount rates from an EP perspective. Paul had been clear from the start that he
needed to treat Frank as an equal but senior partner, and Paul could not afford to offend
Frank, unless his very best efforts failed, and both Michael and Charles were on his side of
an argument that was best avoided.
The basis of the critique of conventional NPV analysis using a single hurdle rate test just
outlined is conceptually sound and the suggested alternative approach using multiple tests
in a goal programming conceptual framework is operationally feasible. But it is important
to acknowledge that accepted wisdom does not yet reflect these views, and all views in this context are highly contentious. This book is not the place for a technical debate about this
controversy. However, consider the following paragraph, and reflect on its relevance to your
organisation when you review this book’s implications as a whole.
The view ANY organisation takes on the correct discount rates to use in all investment appraisals matters greatly, and distinguishing between different kinds of investments may be an important factor. These issues may matter more than avoiding inappropriate time scale and cost estimates. They do not just affect all project selection choices. They also affect the way projects are shaped, and they may affect underlying corporate culture. Any organisation which risks continuing with discount rate policies which may be inappropriate is taking a risk which really matters. Taking this ‘discount rate risk’ is both wholly inappropriate and totally unnecessary – a particularly dysfunctional form of opportunity inefficiency.
Regulation risk – privatisation and nationalisation risk concerns
Frank had no difficulty understanding the rationale or the very significant implications of
the two or more portfolios and multiple test approach as explained by Paul and Ollie. He
also saw some possible further implications of importance. He reciprocated Paul’s coopera-
tive spirit by suggesting they both needed a conversation as soon as possible with Richard,
WSL’s Regulation Director, to explore these further implications.
‘Regulation risk’ was a WSL category of ‘enterprise risk’ which Richard had to lead on.
Ofwat, the regulator concerned with customer prices, was currently advocating privatisation
of primary water supply sources like rivers and reservoirs on a basis which would be separate
from earlier privatisation of water and sewage companies – purportedly to lower prices for
consumers, although some read these privatisation proposals as playing silly political games
motivated by seriously unsound economic beliefs. The water industry as a whole was resist-
ing as strongly as it could, with some success, arguing that this form of privatisation would
not lower customer costs – it would increase customer costs.
Frank thought that Paul’s multiple tests and multiple portfolios approach might help to
clarify the way bond price destabilisation driving share price destabilisation would work to
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
364 Employing planning tools in practice
increase the cost of capital underlying water and sewage costs for consumers if this form of
privatisation was pursued – but it also suggested a further possible opportunity.
WSL had major reservoirs used for leisure and recreation purposes. WSL might counter
Ofwat by more explicitly linking WSL bond funding to assets like reservoirs, as well as water
mains built to last 100 years, and sewage works built to very high standards for long life,
low operating costs, and reliability. They might use this link to encourage their water and
sewage customers to buy bonds to preserve local control of their reservoirs and keep their
water rates low and encourage their customers to attack the Ofwat plans to devalue its bonds
and drive up water rates, interfere with their local amenities, and ‘steal their local natural
resources’, perhaps providing financial support for a lobby which also addressed government
views in a direct manner.
When he met with Frank and Paul, Richard was very interested in Frank’s line of think-
ing. Richard warmly endorsed Paul’s multiple portfolios and tests approach as well as the
broader EP approach, but he suggested they also involve Curt, the Customer Relations
Director. His reasoning was Frank’s strategy might decrease the risk of privatisation of this
kind, but increase the risk of nationalisation of all primary water sources, plus all the cur-
rently privatised closely coupled assets like reservoirs, plus perhaps the water and sewage
industry as a whole, a risk which was very much in the background but a concern Curt
had raised.
Curt was very pleased that Richard had suggested a meeting, enthusiastic about Frank’s
involvement, and delighted that Paul had triggered consideration of this aspect of his EP
approach. As Curt put it when he elaborated on Richard’s introduction, most customer
groups had raised some significant objections when all the UK water and sewage companies
were first privatised via a nationwide restructuring, but they had not been very militant
about their objections because of what seemed like a plausible government sales pitch and
ambiguity about the implications. Big profits by early investors had attracted some early
complaints. What was now making customer groups increasingly hostile and potentially
very militant was a growth of ownership of water and sewage companies by banks, sovereign
wealth funds and other investors clearly in it for short horizon profitability reasons. This
was leading to high profit rates because of bond gearing. It was also leading to an increas-
ingly strident view by consumer groups that ‘their rivers and groundwater and reservoirs
being used to make people rich who were not part of or interested in the local community’.
Indeed, some of the water and sewage industry ‘bond’ funding was now offshore, mak-
ing relatively high rates of return which was exempt from UK taxation. Sooner or later
customer backlash might catch up with the water and sewage industry. Although complete
re- nationalisation was currently unlikely, it was possible, and there was a growing consensus
that the original privatisation had been misjudged and the politics which underlay current
regulation was being mishandled.
In Curt’s view, the current Ofwat privatisation proposals were just plain silly, but what
made a lot more sense was nationalisation of all these resources, plus national control of
new ways to shift water from areas in surplus to areas in deficit (water pipelines and canals
for example), a nationally determined set of undistributed water costs, and much more
effective competition among existing water companies to distribute water at the least cost.
A plausible nationalisation variant was forgetting about more effective competition amongst
existing water companies and nationalisation of the whole of the water supply and sewage
disposal industry. With his WSL ‘customer hat’ on Curt privately thought that both these
nationalisation variants were quite attractive, but with his WSL ‘executive director hat’ on
he viewed either as a very serious problem.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 365
Paul encouraged Frank, Richard and Curt to embrace the EP perspective which gave rise
to seeing the links between these concerns, and he contributed to their ongoing pursuit of
these issues, but he avoided direct personal involvement in board level controversy involving
emotive issues like privatisation versus nationalisation when he did not need to be a part of it
at this stage. He had more than enough controversy to cope with within the Projects Group,
and it needed his undivided attention.
We will return to these issues in Part 3. They are important in the context of the full set
of this book’s messages. However, for the moment their pursuit would be a distraction from
the central thrust of this chapter which is best avoided, and both Chapters 8 and 9 shed
further light on linked underlying issues, a further reason for delaying the discussion.
Technical workshops plus other parallel activities
Immediately after his first workshop, in parallel with the meetings with Ollie, Frank, Richard
and Curt just discussed, Paul began one- to- one meetings with everyone in his group. He
saw direct discussions on a personal basis as a crucial and immediate need, with careful prior
planning of each meeting to ensure that all topics needing this personal private treatment
were covered.
Paul also initiated small group working on a carefully selected and planned series of ‘dem-
onstration tasks’ within carefully chosen ‘demonstration projects’. He wanted everyone to
learn by doing, with mentoring provided by himself and supporting consultants he could
depend on, starting with relatively straightforward tasks his team was prepared for via a
series of ‘technical workshops’.
The series of technical workshops were mandatory for the PG team and optional for other
selected WSL staff. In particular, through Ollie, he invited Operations Group participation
from the start. The workshops were technical in the sense that they provided conceptual and
operational tools to support the learning by doing, but they had other objectives too, like
building teamwork and morale.
This section addresses the topics of some of these technical workshops which are imme- diately relevant to most target readers. It omits those discussing reasonably straightforward
but essential background issues covered in earlier chapters of this book, and Chapman and
Ward (2011) topics you might explore later. It begins with probability- impact grids as clar-
ity inefficient tools and ambiguity uncertainty and underlying complexity as key aspects of
estimates. The use of the same basic SPP strategy progress stage process by the OG and PG
teams in all four strategy progress stages is then considered in three subsections, followed by
three more subsections addressing plans for relationships and contracts, teamwork, and an
overview of the progress stages which follow on after strategy has been agreed on.
PIGs (probability- impact grids or graphs) as clarity inefficient tools
One of Paul’s early priorities was eliminating from ‘the WSL toolset’ any tools which were
clarity inefficient in all relevant circumstances. He wanted to make it easier for everyone to
learn how to select and use tools which were clarity efficient in the right context. He was
particularly concerned about one very simple and much loved but seriously dysfunctional
operational tool which carried ‘conceptual baggage’ with surprisingly broad implications.
Paul began one of his early technical workshops by explaining that he knew many of
the PG team used and liked what he would refer to as PIGs (probability- impact grids or
graphs), also known as probability- impact matrices (contracted to PIMs), probability- impact
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
366 Employing planning tools in practice
diagrams (PIDs), and other comparable labels. He indicated their origins lay in pre- 1960
safety management, a context in which they made sense for reasons outlined later, but they
were now used very widely in many operations and corporate management contexts in a
dysfunctional manner for reasons about to be explored. For example, Operations Group
staff might be regular users, and he knew that Enterprise Risk Management (ERM) was one
of Frank’s current interests, some ERM also using PIGs. He planned to discuss PIGs and
ERM later with Frank and Ollie. PIGs were part of current WSL practice within and beyond
the PG team, needing attention in all decision- making areas using them, but his focus that
day would be PG team usage.
Paul observed that many of the risk management experts he had met over the last two
decades defended PIGs vigorously – even though they all understood their shortcomings to
some extent, usually arguing that PIGs may have important defects but that they were easy
to explain and a good place to start. He observed that PIGs were advocated in PMI PMBoK
guides as the basis of what was referred to as ‘qualitative analysis’ which preceded directly
linked ‘quantitative analysis’, reflecting what he had observed as widespread international
common practice (e.g. see PMI, 2008). He also observed that PIGs were tolerated in some
APM PRAM chapters (e.g. see APM, 1997) and used in many other guides. Furthermore,
he acknowledged that no project risk management guides he was aware of clearly stated
‘under no circumstances use PIGs’.
Paul then put on a deliberately rather forced but broad smile and said, ‘But under no
circumstances are PIGs to be used by PG planners from now on,’ and he would be ‘arguing
for PIGs being dropped from usage by all WSL management decision making’.
He then said all members of the PG team needed to understand why, that day, and reboot their thinking using a much broader conceptual framework, and a much more powerful
associated set of tools. Given the support for PIGs by many guides, and the lack of direct
condemnation by other guides, it was not surprising that WSL staff might be confused
on this issue, but they now needed clarity without delay. In particular, he was depending
upon some of the former Risk Management Department members to provide technical EP
support to PG team led planning addressing all stages in the project lifecycle using their
modelling and facilitation skills, and they could not do that effectively if the PG team as
a whole did not understand why PIGs were a symptom of restricted and limiting thinking
within a framework which had to be abandoned. Complete and unequivocal rejection of
any framework which made use of PIGs was mandatory for anyone who wanted to continue
reporting to him.
Paul made a point of saying he believed he was not dogmatic in most circumstances, and
he generally took a very flexible approach to issues like this, but he believed a determination
to keep using PIGs despite their clear defects was a clear and endemic symptom of a failure
to fully understand the underlying basis of an EP for projects toolset, skill set, and objec-
tives. He then showed them Figure 7.7, indicating that it was a simple example of a prefer-
able alternative to any PIG variant in any context from an EP perspective. He explained that Figure 7.7, taken from Chapman and Ward (2011), had been designed
to help those who value PIGs as basic tools understand why it is clarity efficient to abandon
the use of PIGs in all contexts – provided those they have to work with can also be convinced. He observed that many others in the project management community were opposed to the
use of PIGs, for a wide variety of at least partially related reasons. Debates on this issue were
common but often confusing because different basic framing assumptions and very different
objectives were involved – stealth assumptions which really mattered were rife. They needed
an overview understanding of the issues to deal with this.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 367
Paul briefly elaborated his first workshop description of the clarity efficiency concept in
terms of its Figure 3.5 efficient frontier portrayal plus the ‘stealth assumption’ concept
introduced in Part 1. In earlier technical workshops he had covered an introduction to
the stealth assumption concept along with an overview of all the Chapter 3 estimation and
opportunity efficiency concepts. He then made several points related to Figure 7.7 which
needed clear understanding.
To begin he suggested that those familiar with PIGs could visualise any PIG structure
they preferred superimposed in Figure 7.7, but three probability classes and three impact
classes would be a convenient default assumption for present purposes. He indicated that if
he was explaining Figure 7.7 to someone who was not familiar with PIGs, he would suggest
they should imagine three rows and three columns of equal width defining nine boxes (ele-
ments), with the outer (top- right) box associated with ‘high probability’ and ‘high impact’
coloured red, the inner (bottom- left) box associated with ‘low probability’ and ‘low impact’
coloured green, with amber or yellow intermediate boxes.
Paul then suggested they picture ticking a box to represent a particular risk – a source
of event uncertainty from the perspective of the more general EP framework. From an EP
perspective, placing this tick in a box involved an implicit judgement that the plausible
range of uncertainty for the impact involved a minimum clarity estimate based on a version
of Figure 3.1 he had used earlier, where the P10 is the left- hand boundary of the box and
the P90 was the right- hand boundary, unless some alternative quantitative interpretation
Health Warning
Any source which
has a probability of 1
does not lend itself
to this portrayal, so
very important
uncertainty is
omitted by any event-
based risk
management
approach using
this framework
probability
impact
Figure 7.7 A more powerful portrayal of the information on any PIG.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
368 Employing planning tools in practice
was preferred. Similarly, the tick involves an implicit judgement that the plausible range
of uncertainty for the probability involved a minimum clarity estimate based on a ver-
sion of Figure 3.1 where the P10 was the lower boundary of the box and the P90 was the
upper boundary. If different people were going to interpret PIGs in a comparable manner,
they needed this shared interpretation to avoid misinformation and miscommunication. To
address this most PIGs now provided formal probability and impact bounds for grid boxes,
but most people using PIGs regularly did not use a minimum clarity EP interpretation of
this information, and P0 to P100 ranges were usually implicit. He next suggested that from this EP perspective there was clearly no need for a standard
grid, because if Figure 3.1 variants were used to estimate plausible ranges for probability and
impact for three sources of uncertainty like sources 1, 2 and 3 portrayed in Figure 7.7, they
could be accommodated on the standard probability and impact axes of Figure 7.7 without
forcing them to fit inappropriate boxes – the pre- specified standard grid was not just unnec-
essary, it was a seriously unhelpful restriction, because it required users to force- fit sources
like 1, 2, and 3 in Figure 7.7 into boxes they did not really fit.
With a broad and natural grin Paul observed that if a three- by- three pre- specified grid
was made more refined by its ‘expert’ users, by adding more elements to make the grid nine
by nine for example, the force- fitting was made worse unless multiple ticks were allowed.
Furthermore, in the limit a very refined grid with lots of multiple tics approximated direct
use of Figure 7.7. But this form of refinement of PIGs was a waste of time and effort when
Figure 7.7 could be used directly. Paul suggested this was a crystal- clear demonstration of a
form of complexity in an unhelpfully restrictive framework which made matters worse on all
counts relative to a clarity efficient simple approach like direct use of Figure 7.7.
Furthermore, and much more important, any minimum clarity EP quantification inter-
pretation of Figure 7.7 meant that the event uncertainty sources could be quantified directly,
using a clarity efficient minimum clarity approach or preferred alternatives, to be combined
later in a simple but effective quantitative analysis, without the confusing qualitative analysis
followed by quantitative analysis transformation process distinctions made by common prac-
tice. Furthermore, if appropriate, sources like the three portrayed could be decomposed first,
and they could be associated with responses, decision diagrams and sensitivity diagrams. This
opened a portal into a wide range of possible qualitative analysis approaches to be followed
by probabilistic quantitative approaches to analysis and synthesis which were simply not avail-
able to those using a traditional ‘qualitative analysis’ interpretation of PIGs which was treated
as a separable precursor to common practice follow- on quantitative analysis approaches. This
portal into EP qualitative and closely coupled quantitative approaches also provided a much
richer understanding of a holistic approach to all analysis concerns. It provided a ‘gateway
concept’ in the Biggs and Tang (2011) sense into a much deeper understanding of what might
be involved in any context of interest, which could be used in a clarity efficient way.
From an EP perspective, PIGs were not a sensible basis for initial qualitative analysis to be
followed by optional quantitative analysis using the traditional common practice approach
previously adopted by WSL. From an EP perspective, PIGs were not a qualitative approach at all – they were a weak approach to quantification in a very poorly structured and highly
inflexible, pre- specified qualitative analysis structure. PIGs were clarity inefficient to a wholly
unacceptable extent, even when dealing with event uncertainty in the absence of any other kinds of uncertainty.
Further still, and most important of all, the biggest flaw of PIGs was that they lead people to ignore, or attempt to treat separately, any sources of uncertainty which are not events. Indeed, many PIG users formally defined ‘issues’ as ‘sources of uncertainty with a probability of
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 369
one’, to be treated separately from risk management as they defined ‘risk management’,
arguably a dysfunctional approach of Mad Hatter proportions in the Alice in Wonderland
league as defined by the Chapter 1 discussion of terminology.
Paul emphasised that all relevant inherent variability uncertainty, systemic uncertainty,
ambiguity uncertainty, and capability- culture uncertainty, needed clear and effective inte-
grated treatment with event uncertainty, building on his earlier use of Part 1 material to
explain what these five portrayals of uncertainty involve. This was why PIGs were not just a
badly designed tool for thinking about event risks and a waste of time which failed to make
effective use of the information obtained, they were a dangerous distraction. They had to be
banned by any organisation adopting an EP approach because they ignored four of the five
kinds of relevant uncertainty, and they misconceived the relationships between qualitative
and quantitative analysis.
Paul also emphasised that the sensitivity diagram example of Figure 7.5, used in a layered
manner for any number of sources, could portray any information contained by common
practice PIGs plus subsequent common practice quantitative analysis in addition to dealing
with all relevant sources of uncertainty. If the way Figure 7.5 portrays dependencies between
sources of uncertainty which involve variability and systemic uncertainty plus ambiguity
uncertainty is used to assess common practice examples focused on independent event uncer-
tainty risks, the associated stealth assumptions became clear. Furthermore, the EP toolset
he would equip them with, involving sensitivity diagrams, decision diagrams and source-
response diagrams, could replace PIG- based analysis with simple clarity efficient qualitative
and quantitative analysis which was totally ‘off the radar’ of a PIG- based approach.
In part to ease the pain for those particularly attached to PIGs, and minimise related
withdrawal symptoms, in part to acknowledge aspects of the PIG framework which could
be useful provided the underlying basic assumptions were generalised, in part just to further
demonstrate the limitations, Paul briefly explored using variants of Figure 7.7 within an EP
perspective.
For example, PG planners could start with the Figure 7.7 probability and impact axes,
and use rectangles or rounded equivalents representing two or more sources of event uncer-
tainty on one diagram as a form of ‘heat graph’. A heat graph (or diagram) was a commonly
used portrayal of complete sets of event uncertainty risks on one page, using the same axes
as Figure 7.7, with a ‘hot’ zone coloured red at the top right, a ‘cool’ zone coloured green
or blue at the bottom left, and graduation bands in between.
If important variability, ambiguity, or capability- culture sources of uncertainty involving a
probability of one needed inclusion, they could be ordered in terms of impact and included
on the top of the same diagram, straddling the probability of one line normally treated as an
upper bound, to ensure they were not overlooked.
If systemic uncertainty linking any two or more of these sources of uncertainty needed
attention, the nature of simple one- way dependencies plus positive and negative feedback
loops could be explored on the same figure in graphical terms using influence diagrams or
other related soft systems tools for mapping relationships.
Whether or not influence diagrams or other comparable tools were used as a preliminary
to modelling systemic uncertainty, probability trees could be used to model well- behaved
dependence and more complex knock- on relationships involving inherent variability uncer-
tainty and ambiguity uncertainty portrayals in variability terms. Causal decision analysis
models to explain and manage response relationships could also be developed.
Furthermore, alternative approaches to dependence which were not addressed or even
mentioned by Chapman and Ward (2011), like using sets of ‘dependence drivers’ to capture
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
370 Employing planning tools in practice
key aspects of systemic dependence (Cooper et al, 2014), could be addressed in this kind of
framework, as they could in a more conventional EP framework.
All these possible generalisations were relevant to WSL contexts, as they are elsewhere,
even if event uncertainty is the starting place. But what is always crucially important is being
clear from the outset that sources of uncertainty need not be events. Indeed, events might
be of minor importance relative to inherent variability uncertainty, ambiguity uncertainty,
capability- culture uncertainty and systemic uncertainty. Capturing ALL relevant uncertainty at a suitable level of detail within a holistic framework is always what really matters.
At any suitable level of detail in this structured understanding of uncertainty sources,
decision trees and associated decision diagrams could be used to address risk efficient
choices. Paul emphasised this was where the clarity efficiency concept really started to
pay dividends, in terms of delivering risk efficiency and opportunity efficiency. He then
spent some time elaborating the decision- making power of the EP framework, developing
this book’s Part 1 examples and further Chapman and Ward (2011) examples. Unbiased
estimating was a necessary basic condition for good decision making, but there was a lot
more to good decision- making practice which they also needed to understand to help WSL
achieve an EP approach.
Along the way he made it very clear that the absence of PIGs did not mean ‘risk lists’
(logs or registers) would simply disappear – they would be transformed into ‘source of
uncertainty lists’, referred to using the contraction ‘source lists’, and extended to ‘source
and response lists’. These source lists would normally be much shorter than current risk
lists because decomposition of sources would only be encouraged where it paid off in clar-
ity efficiency terms. But source and response lists would be much broader in scope, much
more sophisticated in terms of their structure, and much more useful because of specific and
general response linkages between sources and responses. Some responses would be specific
to particular sources. But general responses to sets of sources would prove very important.
Some responses might involve significant secondary sources and responses – first choice
responses might not work. Diagrams to assist when the complexity involved was worth
exploring would also be important. Some sources would receive quantitative treatment, but
some would be treated as conditions.
In summary, Paul’s technical workshop on PIGs made it clear to all PG team members that the Projects Group were starting in a different place than some (not all) common-practice project risk management, certainly very different from the variant of common practice used
by the old WSL Risk Management Department. Everything they used to do would now be
done within a broader framework. This would provide them with much greater power when
needed, but with the flexibility to keep it simple in a rich range of ways appropriate to the
context. Most important, their ambitions would be very much higher, and their jobs would
be much more interesting because they would be addressing a much broader role within
WSL.
Paul indicated that current members of the Risk Management Department might feel that
they had been especially targeted on this occasion, and to some extent this was intentional.
However, all members of the PG team needed a clear understanding of the rationale for ‘no
more PIGs’.
Paul then suggested that all members of the old Risk Management Department had one
of two key roles to play in the new restructured PG team.
In some cases, their current role as facilitator and modeller might simply broaden to not
only providing a lead on model structuring for all relevant plans and all associated uncer-
tainty and complexity in a clarity efficient manner, including the elicitation of probabilities,
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 371
execution of quantitative analysis, and interpreting the synthesis of both quantitative and
qualitative analysis, but also to addressing the range of deterministic models underlying pro-
ject base plan development. They would have to broaden their skills base, but they would
be given all the help they needed. He emphasised there was an ongoing role for people with
modelling skills, both deterministic and probabilistic models, and there was a lot of scope for
helping everybody involved to use them on their own as well as assisting with formal model-
ling of novel or complex issues. Former Planning Department staff who were experienced
users of project base plan models might be interested in moving into probabilistic modelling
of all planning and playing a similar role.
In other cases, along with former members of the two other PG departments, some of the
former Risk Management Department staff might move towards a team leader role as PG
group members learned to work together in flexible teams. Team leaders would need good
interpersonal skill sets as well as relevant technical skill sets and a good understanding of EP
in holistic terms. Team leaders might have a background in project risk management, but
they could have a background in traditional project planning or estimating, and they might
have a background outside of all three.
Paul emphasised that he wanted all PG team members to clearly understand that funda- mental change was not an option if they wanted to remain in his team, but he also wanted
to be as positive and supportive as he could for those willing to make the change.
Ambiguity uncertainty and underlying complexity as a key aspect of estimates
Paul began a follow- on technical workshop by acknowledging that former Risk Manage-
ment Department members might have felt some discomfort in his last workshop, and the
current workshop might continue their discomfort. However, it would also prove uncom-
fortable for former Estimating Department members. Furthermore, former members of
the Planning Department need not feel smug because their complete absence from concept
strategy stage participation was arguably an even greater cause for concern, and he would
pick up related issues in later workshops.
Having made the overall nature of his critical position clear, to ensure everyone under-
stood they all had to embrace change, he added that none of the PG team should feel per-
sonally or professionally responsible for the problems explored last time or in the current
and later workshops, and he hoped the current workshop would start to explain why.
He reminded everyone that WSL concept strategy stage cost estimates were currently
based on a point estimate provided by the Estimating Department plus a ‘risk adjustment’
produced at a later date by the Risk Management Department.
The risk list used for this risk adjustment was generated by a common practice approach
based on PIGs, often running to 50 or more risks, with risks being ‘event uncertainty
sources’ from an EP perspective – sources of uncertainty associated with events. Each risk
had a point estimate of its probability of occurring, plus an estimated probability distribution
for its impact in terms of cost. The risk adjustment was the expected value of the complete
list, obtained via Monte Carlo simulation, usually assuming independence between events.
To understand why this was an inappropriate place to start, and how much better esti-
mates could be provided with a lot less effort, it would be useful for the PG team and the
Operations Group planners present to understand how the UK Highways Agency (HA)
started from a directly comparable position in 2007 and made its first move towards the
position WSL would adopt.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
372 Employing planning tools in practice
The first step in the HA journey began when the secretary of state for transport responded
to a very critical House of Lords review of HA estimating performance by commissioning a
report by Mike Nichols (Nichols, 2007).
The second step began when the HA accepted the Nichols Report conclusions and com-
missioned Nichols Group support for a re- estimation study. Some of the implications of
the re- estimation study were published in the Association for Project Management guide
Prioritising Project Risks – A Short Guide to Useful Techniques (Hopkinson, Close, Hillson, and Ward, 2008). They were discussed further by Chapman and Ward (2011).
Prior to the re- estimation study, HA estimators were expected to provide a point estimate
early in the concept strategy stage for new major road projects, and later a risk adjustment
was provided – a situation directly comparable to the current WSL approach. Indeed, the
pre- 2007 HA approach was directly comparable to the current WSL approach apart from
one notable difference. Mandated UK government ‘advice’ required the HA to implement
HM Treasury rules on the adjustment of all estimates for ‘optimism bias’ (HM Treasury,
2003a, 2003b) unless a formal risk management approach approved by HM Treasury was used to adjust for risk, in which case no ‘optimism bias’ uplift adjustment was required. The
HA used a risk management approach comparable to WSL’s current approach to remove
this adjustment as soon as possible or avoid it at the outset. Adjusting for optimism bias
would amount to an uplift of about 65% for most major new road projects. Optimism bias
uplift factors were dependent on the category of project, based on a table of factors pro-
duced by leading consultants derived from average cost overruns for a set of projects which
had not used ‘approved’ project risk management.
Key stealth assumptions which were implicit in the Treasury approved pre- 2007
HA approach were all approved project risk management processes were equivalent to an optimism bias uplift on average (untrue for several reasons explored shortly) and
approved risk management adjustments were preferable because they reflected the par-
ticular context being addressed instead of the average (true, in part, with important
reservations). From an EP perspective it was obvious that this pre- 2007 HM Treasury
approach was not a good idea – as clearly demonstrated by the ongoing bias in pre- 2007
HA estimates. The exact nature of reasons it was not a good idea was less obvious and
worth understanding.
Related stealth assumptions implicit in the current WSL approach included: there was no
need to even think about an optimism bias adjustment, despite the clear ongoing evidence
of unacceptably large optimism bias, and there was no need to fix an obvious problem or
explore what might turn out to be an underlying mess.
Prior to the 2007 Nichols report, HA estimators created a starting point ‘outline initial
plan’ based on their understanding of intentions associated with the business case, and
used this plan to produce a ‘base estimate’ of the construction cost, but there was no sig-
nificant planning input from the equivalent of a Projects Group, and no linked significant
design or operations input from the equivalent of an Operations Group. For example, a
new motorway project might have a type specified and a route indicated on a map but no
further design or construction planning input. This was directly comparable to the current
WSL position unless the Operations Group initiated a project, and even then, comparable
concerns remained.
To demonstrate why the estimate plus a risk adjustment approach as used by the HA
produced a biased estimate, and indicate a way forward, the Nichols led re- estimation study
started by identifying a sample of projects which had recently completed the equivalent of
the concept strategy stage and passed the associated gateway as a representative subset of
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 373
all projects from the £20 billion portfolio of HA projects currently in progress. Some WSL
projects could be sampled in the same way to explore a similar learning curve.
HA estimators revisited the outline initial plans used to represent their understanding of
intentions associated with the business case, and the associated ‘base estimate’ of the con-
struction cost was used as a starting point for the sample projects.
Nichols facilitators then used the following minimum clarity approach to all relevant uncertainty given the original estimation and risk adjustment process, explicitly working within a prototype EP framework.
The equivalent of a pre- planned version of the identify phase of a basic SPP designed by
Nichols staff employed five sources of uncertainty for the ‘construction cost’ estimate for a
new motorway. The construction cost of a new motorway was the costs paid to a contractor
to provide a detailed design and plan and deliver an operational road, comparable to what
WSL paid Paul’s firm for the sewage project he had recently delivered. The HA’s cost of land
acquisition and traffic management during construction were examples of costs excluded
from construction cost. The five sources of uncertainty in the decomposition of the ‘con-
struction cost’ estimate used by Nichols were:
1 all sources explicitly considered by the estimator originally,
2 all sources explicitly considered at least in part by the risk adjustment,
3 all other sources best considered at the project level,
4 all other sources best owned by the HA at portfolio level, and
5 all other sources best owned by the UK government.
Nichols staff clarified this structure for HA estimators who had provided the original esti-
mates in the sample as follows.
Source 1 was everything estimators had in mind at the time the estimate was made, explic-
itly or implicitly, as best they could recall, referring to notes made at the time plus any other
relevant documents if they wished.
Source 2 was everything covered by the risk adjustment process, referring to the list of
risks and the component expected values, plus associated systemic uncertainty (dependence)
they were aware of.
Source 3 was anything else which they might reasonably be held accountable for thinking
about and including in future estimates – bearing in mind the exclusions associated with
sources 4 and 5, but assuming zero tolerance of any persistent bias in future estimates. Source 4 was sources of uncertainty best owned at portfolio level by the HA, like Euro-
pean Union (EU) safety regulation changes which would affect crash barrier designs for all
motorways, and HA design rules associated with the ‘quality’ of delivered road projects,
including quality issues driven by maintenance cost expectations.
Source 5 was sources of uncertainty the HA should refuse to accept responsibility for
unless and until a contractor was prepared to accept them, like inflation. Inflation was best
owned by the UK government, with all HA cost estimates in the money of the day, unless it
was cost- effective to give inflation risk (and opportunity) to a contractor.
There was no equivalent to the basic SPP structure phase beyond what was built into
the identify phase in the sense that there was no formal testing of the robustness of the
assumptions used for qualitative analysis thus far in the process and no formal consideration
of restructuring. The focus phase equivalent explicitly selected a robust approach which
included all relevant uncertainty, and the explicit working assumption was that the use of
source 3 in the context of the other four sources would deliver sufficient robustness because
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
374 Employing planning tools in practice
it was defined to do so, and facilitators would do their very best to make sure this robustness
was delivered.
The ownership phase was also built into the focus phase in the sense that the UK gov-
ernment was allocated to source 5, the HA at portfolio level took responsibility for source
4, the estimator’s original assumed responsibility was allocated to source 1, the risk man-
agement group to the basis of source 2, and a residual of everything else of relevance was
allocated to source 3.
Having made sure HA estimators understood the nature of all five sources, Nichols facili-
tators made direct use of the Figure 7.8 format for a minimum clarity estimation process,
initially showing estimators a variant of Figure 7.8 without probability curves 1 to 3 and the
associated notes. For confidentially reasons, as well as expository convenience, Figure 7.8
uses a normalised scale to portray the sample average in approximate terms instead of indi-
vidual project cost values. For Paul’s purposes, and ours, it is simpler just to work directly
with Figure 7.8 with this in mind.
For each estimate in the sample, the Nichols facilitator first asked estimators to think
about the uncertainty they had in mind when producing their base value estimate, source 1.
They were then asked to estimate a P90 value in money of the day used for that estimate. On
average this was about 25% more than their base estimate, as indicated by Figure 7.8. Next,
they were asked to estimate a P10. On average this was slightly less than their base estimate,
as also indicated in Figure 7.8. Curve 1 was then plotted and explained in terms of the
estimates the estimator had provided and the expected value implications. Estimators soon
understood that they should have used something like a P50, which was about 15% more
than their base estimate, to provide an unbiased estimate of what they thought they had
considered. However, they were comfortable with the results of the re- estimation process,
once the observed bias was explained in terms of an underlying ‘anchoring effect’, which is
cumulative
probability
1. base value estimating uncertainty 2. uncertainty associated with previous risk registers 3. other uncertainty sources which the HA accepts
accountability for not considered at a portfolio level
cost as a % of the base estimate value
1.0
0
0.5
base estimate value
curve 1 curve 2 curve 3
probability curves 1 to 3 show the cumulative effect of
100 150 200
P90
P50
P10
Figure 7.8 Sensitivity diagram: Highways Agency (HA) example.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 375
common whenever estimators address a point estimate by starting with an optimistic esti-
mate and then adjust upwards, instead of looking at the P90 first. The sequence in which
estimation processes proceed affects bias, which was understood when it was explained.
In future, estimators would be comfortable with a minimum clarity approach or a more
sophisticated EP variant as explored in Chapter 3.
When estimators were comfortable with the location of the source 1 line on their equiva-
lent of Figure 7.8, they were asked by the Nichols facilitator to think about all the uncer-
tainty associated with source 2 in terms of 1 + 2 and identify a P90 on the same sensitivity
diagram. On average this was about 40% more than the base estimate, as indicated in Fig-
ure 7.8. Next, they were asked to estimate a P10. On average this was slightly more than
their base estimate. Curve 2 was then plotted and explained. The resulting P50 was about
25% more than their base estimate on average, a further 10% uplift on their base estimate.
When estimators were comfortable with the location of the source 2 line portraying 1 + 2
on their equivalent of Figure 7.8, they were asked by the Nichols facilitator to think about
all the uncertainty associated with source 3 and identify a 1 + 2 + 3 line P90 on the same
diagram. On average this was about 90% more than the base estimate, as indicated in Fig-
ure 7.8. Next, they were asked to estimate a P10. On average this was about 10% more than
their base estimate. Curve 3 was then plotted and explained. The resulting P50 was about
50% more than their base estimate on average, source 3 providing a further 25% uplift on
their base estimate.
The way Figure 7.8 linear cumulative distribution lines were added during the estimation
process could be associated with the minimum clarity process of Chapter 3 and Figure 3.1
assuming 100% dependence (perfect positive correlation) between sources 1 to 3, but direct
estimation in cumulative form could also be associated with indirect estimation of depend-
ence levels other than 100% – see Chapman and Ward (2011) for an explanation.
Everybody involved in the re- estimation study soon understood what Figure 7.8 por-
trayed and the implications. Figure 7.8 was a sensitivity diagram portraying sources of
uncertainty given currently assumed response choices – a key EP tool, introduced in this
chapter in Figure 7.5, used here to clarify what was missing from an HA approach because
of previously unidentified stealth assumptions, also used in a somewhat different manner in
Chapters 4 and 6.
The final right- hand curve – the line for source 3 in this case – portrayed the overall total
quantified uncertainty of interest in this example – the composite source 1 + 2 + 3. The gap
between curves 2 and 3 showed the contribution of source 3, and the gap between curves
1 and 2 showed the contribution of source 2. The contribution of source 1 was directly
portrayed. Overall, an uplift of about 50% to the base value would produce an unbiased
expected outcome estimate given the current understanding of sources 1, 2 and 3 and
ignoring sources 4 and 5, with an expectation that the 10% to 90% uplift confidence band
would be exceeded about 10% of the time. Source 3 was the biggest contributor to the uplift
and associated uncertainty of composite source 1 + 2 + 3. Source 2 was the smallest. Sources
4 and 5 were additional issues, not quantified for current purposes, and at this stage in the
lifecycle quantification of source 5 would not be appropriate. The earlier risk adjustment
approach was capturing source 2 but ignoring sources 1, 3, 4 and 5. Source 2 was about
20% of the composite source 1 + 2 + 3, perhaps 10% of the composite source 1 + 2 + 3 + 4 + 5.
The uncertainty indicated by curve 3 in Figure 7.8 and its contribution to an expected
value uplift was clearly huge compared with the variability embodied in source 2 and the ear-
lier common practice risk management approach. However, the variability aspect associated
with the 10% to 90% confidence band portrayed by curve 3 in Figure 7.8 was not risk – it
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
376 Employing planning tools in practice
was just uncertainty until commitments were made. The risk involved that really mattered was continuing with a biased estimation approach which failed to capture information and knowl- edge which was readily available providing the right questions were asked.
Paul indicated that the basic SPP evaluate phase equivalent for the HA re- estimation
study was about wider dissemination of the implications of Figure 7.8 and encouraging a
corporate- level revisiting of approaches to estimation by the HA. The HA and HM Treas-
ury responded very positively to the re- estimation study report (see HM Treasury [2014]
for example). Paul was determined WSL would learn the same lessons. WSL could take the
same approach if they undertook a comparable re- estimation exercise, an option that was
available. But he suggested they should just learn the same lessons as the HA from the HA
study as discussed in their current workshop and move on directly to a revised estimation
approach.
One obvious conclusion was they needed to base WSL estimates at a project level on a
comprehensive view of all the relevant uncertainty – to estimate the equivalent of line 3 in Figure 7.8, considering sources 4 and 5 later when appropriate.
A second obvious conclusion was source 2 was the least important of the three sources
of uncertainty needing immediate project level consideration by the HA – and clearly not
worth the extensive level of decomposition effort currently used and wasted by WSL. This
was consistent with general EP advice to use fewer ‘sources’ than common practice ‘risks’,
with a much wider scope, plus a closure with completeness approach explicitly employed
for the final component, coupled to a clear understanding of specific and general responses.
A third obvious conclusion was WSL needed to develop a form of decomposition of
sources of uncertainty that distinguished sources which would resolve themselves as the
lifecycle evolved (like uncertainty about what kind of design approach would be used) and
sources which needed direct management, or they would still be there when execution and
delivery began (like uncertainty about the most appropriate contracting structure). The
need to distinguish sources in this way was a key aspect of clarified ambiguity. Paul observed
that the Nichols Report addressed this for the HA, and WSL could take a comparable
approach. Paul also observed that WSL could use this kind of decomposition of uncer-
tainty to produce a very useful ‘control chart’ projection of how the concept strategy stage
uncertainty ought to contract as the lifecycle continued to evolve assuming that their current assessments of uncertainty were unbiased. To explain in outline what this involved and how
it could be addressed he used Figure 7.9.
Paul explained that he had used the term ‘cost uncertainty- time profile graph’ in the
title for Figure 7.9 because diagrams employing this format would not be used by WSL in
a simple classic control chart manner. This particular very simple illustrative example used
a number of assumptions which he needed to explain in outline shortly, in detail later. At
an overview level, Figure 7.9 portrayed how quantified uncertainty at the concept strategy
gateway stage might decline gradually until a fixed price contract was agreed, when a sig-
nificant step change reduction was obtained, followed by further uncertainty reductions as
the project proceeded.
The Figure 7.9 example assumed a single contractor would undertake a fixed price con-
tract for the construction of a new sewage work with an unbiased expected cost estimate of
£100 million at the concept strategy gateway stage, an estimate which remained stable over
the whole of the timeframe portrayed. At the concept strategy gateway stage uncertainty was
at its maximum, with a P90 of about £250 million. The DOT strategy progress stage was
assumed to have a big impact on uncertainty, with the P90 decreasing to about £185 million
at the DOT strategy gateway stage. The E&D strategy progress stage was associated with
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 377
cost in £m
P90
time in years from now1 2 3
0
100
200
fixed price
contract
agreed
expected value
P10
0
DOT
strategy
gateway
concept
strategy
gateway
E&D
strategy
gateway
D&A
strategy
gateway
Figure 7.9 Cost uncertainty– time profile graph for the construction cost of a sewage works based on the four WSL gateways followed by a fixed price contract.
a more modest impact on uncertainty, with the P90 further decreasing to about £170 mil-
lion by the E&D strategy gateway stage. The D&A strategy progress stage further reduced
the P90 to about £160 million. D&A strategy gateway approval meant an overall WSL
strategy for this project was sanctioned by the board with a budget approved by the board
but obtaining all necessary local government planning department permissions and contract
negotiations then took another eight months or so. Following the contract being agreed
on, there would be further modest uncertainty reductions. But uncertainty would not be
eliminated until the contract was completed at a future point off the graph.
Paul suggested that if all these assumptions proved sound, actual outcomes could be plot-
ted on the initial version of this chart as the lifecycle evolved in a simple, classic control- chart
manner. But in practice, revised charts would also be needed to reflect revisions to corporate
understanding of both expected outcomes and associated further future uncertainty. He
anticipated both plotting outcomes on earlier graphs and revising graphs would be useful at
all gateways. Furthermore, he was convinced that this approach would be crucial whenever
significant unanticipated risks were realised. The mechanics of constructing these charts was
just a matter of clarifying when in the project lifecycle each identified source of uncertainty
should be resolved, with reference to the lifecycle projection used along with the seven Ws
structure to identify sources of uncertainty.
In practice, contracting assumptions other than a fixed price contract with one contractor
might be used, and criteria other than cost might also be measured. And in practice, they
might want to look at full lifecycle costs, including maintenance costs over the life of the
assets. If they did this, important sources of uncertainty would persist until the termina-
tion stage was complete, and the importance of operating stage uncertainty would become
visible. This visibility of uncertainty over the whole anticipated project lifecycle stage struc-
ture could prove very important, even if a single contractor were used and a fixed price
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
378 Employing planning tools in practice
significantly reduced uncertainty associated with construction cost without affecting the
expected outcome, as suggested by Figure 7.9.
The concept strategy stage expected value estimates for all relevant criteria would usu-
ally increase or decrease as the project lifecycle progresses – but they should remain within
predefined confidence bounds, and the confidence bounds themselves should become nar-
rower as the project lifecycle continues, unless something exceptional happens. Four differ-
ent kinds of exceptions might lead to expected outcomes breaching confidence bounds or
confidence bounds not continuing to contract:
1 A condition or assumption underlying the expected values and confidence bounds
might become a significant realised risk – a known unknown was realised.
2 Something unanticipated happened – an unknown unknown was realised.
3 WSL was extremely unlucky (or lucky) in terms of estimates of identified and fully
quantified uncertainty which were demonstrably unbiased.
4 Bias in WSL estimates was revealed.
It might be difficult to identify what particular combination of these four exceptions they
were dealing with when confidence bounds were breached, and an ambiguous mixture was
common, but routine insistence on bad luck explanations should not fool anybody. On aver-
age all project criteria expected value and range estimates should prove unbiased in a demon-
strable manner, and all associated confidence bounds should contract fairly smoothly over
time, provided identified conditions or assumptions were not realised, unknown unknowns were not realised, or unusually good or bad luck was not involved.
Figure 7.9 confidence bound revisions might involve the spread increasing or decreasing
as the lifecycle progressed, but on average all confidence bound estimates should decrease
in a reasonably stable manner over the lifecycle, without unpleasant surprises that would call
PG team competence into question. Ongoing statistical analysis and control of bias which
ensured that this was the case would be important.
Plotting actual outcomes on a graph with a Figure 7.9 format, with a contracting
projection of remaining uncertainty, could be a very useful EP tool, the reason for elabo-
rating what is involved in this chapter, and refinements like those just outlined could be
crucial.
Paul was determined to use several variants of this kind of approach for all projects, and
he thought the uncertainty– time profile graph was an important conceptual tool, usefully
understood by all members of the PG team in broad terms at this point.
Paul finished this workshop by stimulating discussion about why WSL’s current problems
were comparable to HA problems and those of all other organisations which did not take
an enlightened view of ambiguity with clarified ambiguity in mind. He wanted the very
important practical problems resulting from biased initial estimates based on point values
and common practice risk management risk adjustments to be clear – without implying he
associated blame or incompetence with his group. The WSL Projects Group had been fol-
lowing widespread common practice, although a lot of much better practice had been used by many organisations for many years, some linked to authors drawing on Chapman and
Ward (1997, 2003, 2011) and some drawing on many other authors taking a wide variety
of approaches. Paul wanted his group to understand why a new conceptual framework based
on EP was essential before discussing how they would use a new EP framework. He wanted
his group and the Operations Group staff present to think about this discussion and then
return fresh and ready to go for the next workshop.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 379
Use of the same basic SPP by the Operations Group and PG teams in all four strategy progress stages
By the time Paul was preparing the technical workshop discussed in this subsection, Ollie
and some senior members of Ollie’s staff who had attended all the earlier workshops were
convinced that working with Paul and the PG team was a good idea. They were also con-
vinced they should start thinking about the Operations Group as the OG team with some
implications similar to those for the PG team. In addition, Paul had been invited to run
special workshops for the OG team as a whole on its own to cover the material dealt with
in his initial workshop for the PG team plus all the technical workshops so far. This led to
Paul’s persuading the OG team to treat this subsection’s workshop for the PG team as a
joint OG/PG team workshop, initiating joint OG/PG team working, an outcome Paul had
hoped for from the outset.
Paul started this joint OG/PG team workshop by emphasising the importance to WSL of
a particularly close working relationship for the OG and PG teams. He believed they needed
to work together as a joint OG/PG team in the concept-strategy progress stage. They then
needed to continue to work in a closely coordinated manner through the rest of the project
lifecycle.
To facilitate this there would be both immediate and long- term benefits if the OG team
agreed to use the same Figure 7.2 basic SPP concept tool for their management of the DOT
strategy progress stage as the PG team would use for the E&D strategy progress stage, both
groups also using their variants in simplified prototype forms during the concept strategy
progress stage, with ongoing implications in the fourth strategy progress stage addressing
angels and devils in the detail on a joint basis. There would be important differences in the
way these progress stage SPP frameworks were used by the two groups in all four stages,
and the language employed might be very different, but the use by both teams of a shared
common conceptual framework structure was crucial. In particular, once they got used to
using the same basic SPP in the DOT and E&D strategy progress stages, it would be much
easier to work as a joint OG/PG team in the concept strategy progress stage.
As an example of differences in use, the OG team would approach the basic SPP create
and identify phases by drawing on the design and operations management literature, not the
project planning literature. As a sewage works example of the different language involved,
the OG team would see key plan components and sources of uncertainty during the opera-
tions phase of the project in terms of key sewage works component availability and reliability
concerns, not key activity cost or time concerns in the project planning sense.
Paul then emphasised that neither OG planners nor PG planners were involved in the cur-
rent concept strategy stage business case preparation process for the new water mains project
in the way he thought they should be. After the concept strategy gateway was passed, the OG
team would use the DOT strategy progress stage to develop their part of the overall WSL
strategy for this project. After the DOT strategy gateway was passed, the PG team would use
the E&D strategy progress stage to develop their part of the overall WSL strategy for this
project. However, all important interdependencies between these strategies and the other
seven Ws plans need addressing in a preliminary manner during the concept strategy stage
in grounded plan terms, prior to the concept strategy gateway stage. This was not current
WSL practice. In his view, their joint involvement in the concept strategy progress stage was
usefully viewed as collaborative development of prototypes of the DOT strategy and E&D
strategy they would each need later, with a focus on the key relationships between their con-
cerns and the rest of the seven Ws plans plus associated goals– plans relationship concerns.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
380 Employing planning tools in practice
To illustrate what this involved, Paul reminded them that Frank had started the Finance
Group led concept strategy progress stage study of the water mains project by assuming
one large contract would be let to one national contractor. Paul was convinced that much
smaller work packages let to a number of medium sized local contractors as well as several of
the larger national contractors would be a much better strategy, making use of component
separable projects within a programme portfolio of projects structure. He was determined
to get a board level resolution of this issue before the concept strategy gateway approval. Part of his rationale was lower construction costs, part was more flexibility in terms of
finance costs, part was more flexibility in terms of dealing with traffic disruptions, and part
was ensuring ongoing competitive market discipline for water mains contractors. An aspect
which he believed the OG team needed to stress as part of their concept strategy stage
contribution was more flexibility in terms of design considerations related to operations
issues, which might evolve over time. There were several other important issues needing
OG team input in the concept strategy stage. For example, a 100 year planned life for water
mains instead of 50 years not only involved technical specification issues they would need
to address, but it also involved longer term water demand forecasting and capacity plan-
ning issues. This might involve different water main patterns as well as different water main
pipe sizing options – a network built on six primary feeders instead of the current four, for
example.
Paul did not want to labour what an EP approach adopted by the OG team which was
coordinated with a compatible PG team approach, or bore them with technical details. He
just wanted to make them receptive to starting to work with the PG team on a joint basis to
ensure consistent and coherent concept strategy stage input plus fully coordinated follow- on
work for both groups. But it would be useful if the OG team saw this as a joint develop-
ment project with the PG team, involving one joint OG/PG simplified prototype basic SPP
process for the concept strategy progress stage, plus two common higher clarity basic SPP
processes: one for the DOT strategy progress stage for use by the OG team, a second for the
E&D strategy progress stage for use by the PG team. He thought the OG team might like
to start this kind of prototype basic SPP development working on the new sewage project,
where the OG team were currently leading the early concept strategy progress stage.
Taking this argument a bit further, he believed a significant investment in effort now in
their joint OG/PG prototype planning for the sewage works in the concept strategy pro-
gress stage would help them learn how to use simple variants for the key issues that mat-
tered most in future concept stage studies, followed by directly dependent DOT and E&D
strategy progress stage studies. At present it was not entirely clear where their efforts would
best pay off, but he did have some ideas where they might start.
For example, he knew the OG team were already looking at urban sites A and B as options
for the proposed new sewage works, assuming a design based on the last sewage works built.
He had already suggested rural site C might be better in an earlier discussion with Ollie,
based on a different design approach. But elaborating now, he thought the OG team might
be able to identify options that were even better – designated D, E, or F. Design features
which were of immediate and obvious concern in terms of ongoing operation of the sew-
age works, which he believed the OG team would obviously address, included its ‘quality’
in reliability and maintainability terms (crudely measured by expected operating costs), its
ability to cope with low probability flooding events and the impact of such events on peo-
ple living in the vicinity, issues associated with smells, the everyday visual impact of the
proposed design, and WSL’s ability to get planning permission to build the sewage works
given the earlier considerations. However, the PG team was also directly concerned with
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 381
what might be termed ‘traditional build- ability’ issues, which included ground conditions,
access for contractors’ heavy equipment and so on, and he appreciated the OG team might
not think of all these concerns if the PG team did not raise them. Furthermore, the OG/
PG team might usefully go well beyond the usual concerns of the OG or PG teams and
offer local residents compensation for undesirable features – in the form of amenity parks
or river fishing provision on what had been private land with no public access for example.
This might have implications for objectives relevant to other WSL groups. In the concept
strategy project lifecycle stage there was room for creativity which needed to be exploited –
it could become an opportunity lost completely if not exploited in this stage. In Paul’s view,
leadership seeking this creativity was up to the OG/PG teams, seeking allies from within
WSL like the Customer Relations Director, and allies from outside WSL which included
local customer organisations.
Paul suggested the OG team might be already thinking along these lines for the proposed
new sewage works, but he thought they should also be involved in the proposed water mains
replacement programme in similarly broad design option studies as soon as possible, and all
other projects to follow.
The rest of this workshop was spent developing these ideas and beginning to use them in
terms of WSL illustrative examples.
Use of the same basic SPP in the next two strategy progress stages
Later workshops developed DOT strategy progress stage use of the basic SPP by the OG
team and comparable E&D strategy progress stage use of the basic SPP by the PG team,
plus associated collaboration, to support earlier use as prototypes in the concept strategy
progress stage as well as preparing for following stages.
Actual DOT strategy progress stage use of the process in Figure 7.2 would clearly start
with the prototype used to seek concept strategy gateway approval. It would focus on the
ambiguity uncertainty which needed reduction before DOT strategy approval. For example,
if concept strategy gateway approval for the new sewage works was based on urban sites
A and B being rejected but rural sites C, D, E, and F all being possible options, depending
on further design work, site tests, and consultation, the OG/PG team might need to pursue
the consultation jointly while the OG team pursued further design work and the PG team
pursued site ground- condition tests.
Actual E&D strategy progress stage use of the Figure 7.2 process would also start with
the prototype used to seek concept strategy gateway approval, updated as necessary by all
the progress made prior to DOT strategy gateway approval. It would focus on the ambiguity
uncertainty which needed reduction before E&D strategy approval. For example, if DOT
strategy gateway approval for the new sewage works confirmed rural site E, the PG team
would now need an E&D strategy which was refined enough to use the results for tendering
purposes once overall E&D strategy testing was complete, assuming the successful contrac-
tor would undertake detailed DOT design plus detailed E&D planning within the strategies
approved by the board.
Paul made it clear that the reason the DOT strategy progress stage preceded the E&D
strategy progress stage was the obvious need for PG planners to be clear what they were
planning to build before starting, but the reason prototypes of both had to be used jointly
in the concept strategy stage was to identify and address key interdependencies before com-
mitting to a concept strategy, and some aspects of these key interdependences might be
ongoing.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
382 Employing planning tools in practice
Another way to visualise and deal with this was a whole series of partial DOT strategy pro-
gress stages followed by partial E&D strategy progress stages, gradually evolving towards
both being jointly optimised. But in practice a period of close OG/PG joint working dur-
ing the concept strategy stage, followed by a period of more independent focus during the
next two stages, ought to be more clarity efficient. This was a working assumption which
Paul had made that could be tested, with a view to making revisions if doing so was deemed
worthwhile.
Use of the same basic SPP in the final strategy progress stage and linked earlier use
Separate later workshops addressed why there was a need to anticipate any areas where
‘devils or angels in the detail’ needed OG/PG consideration in prototype terms within the
concept strategy stages of WSL projects, linked to discussion about how these issues should
be addressed later in the actual D&A strategy progress stage.
To initiate this discussion Paul explained that the Chapman and Ward (2011) approach he
was building on did not use a separate D&A strategy progress or gateway stage, but there
were clues about why a D&A strategy progress stage might be useful when PUMP concepts
were fully integrated with the whole planning process starting in the concept strategy pro-
gress stage as recommended. He had thought about these clues and turned them into what
he now saw as a definitive case for a D&A strategy progress stage plus a following gateway
stage, plus earlier use in prototype form in the opening concept strategy progress stage.
An example used by Chapman and Ward (2011) to illustrate both the D&A stage concept
and DOT strategy progress stage planning using a prototype basic SPP involved Petro-
Canada considering a ‘pilot LNG (liquified natural gas) project’, explored in more detail by
Chapman, Cooper, and Cammaert (1984). Paul explained that this example study explored
the possibility of producing liquefied natural gas on Melville Island in the high Arctic, trans-
porting it via icebreaking tankers to US markets. Because of the speculative nature of this
venture, a single compression chain LNG production plant had been specified by Petro-
Canada. An Acres study was commissioned to look at LNG plant reliability in the context
of LNG storage and tanker arrival patterns using E&D strategy stage prototype PUMP
methodology and software developed by BP in a DOT strategy stage context. This study
demonstrated clearly that a single compression chain design plan was a false economy, a kind
of ‘devil in the detail’.
A second example used by Chapman and Ward (2011) with a more detailed discussion in
another paper (Chapman, Cooper, Debelius and Pecora, 1985) involved a study for Fluor
on behalf of the Northwest Alaska Pipeline Company. This study explored the possibility
of taking a 48 inch gas pipeline across the Yukon River on a bridge, using spare pipe- racks
provided by the bridge’s builder when it was constructed to carry a 48 inch oil pipeline
across the river several years earlier. An Acres study was commissioned to explore this option
in comparison with a separate pipe- rack or a separate suspension crossing. It used an E&D
strategy stage prototype PUMP methodology and software developed by BP in a signifi-
cantly modified form in a DOT strategy stage context. This study demonstrated that serious
drawbacks for all three of the options initially identified were very effectively overcome by
a previously unconsidered fourth option involving a separate submarine crossing, a kind of
‘angel in the detail’.
A third example involved a ‘demonstration project’ used by National Power. This study
addressed developing and implementing a simplified variant of the SCERT methodology
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 383
developed by BP in the E&D strategy stage of a National Power project. This example was
outlined earlier in Part 1 of this book, and it is not worth repeating that outline here, but
Paul explained it in detail treating the whole chain of improvements in opportunity effi-
ciency discussed earlier as ‘angels in the detail’. He emphasised that it was this whole series
of ‘angels’ which sold everyone in National Power on the value of a simplified prototype
PUMP approach which could be replicated at least in part in a comparable D&A strategy
stage by WSL. Finding these ‘angels’ was attributable to a systematic search for ‘general
responses’ involving all relevant perspectives and effectively capturing all the key interde-
pendencies between issues.
The WSL equivalents were illustrated by ideas already touched on earlier. One example
was early preliminary exploration with local residents of issues concerning them which were
relevant to choosing a location and design. Features that might make a particular sewage
works site attractive to local residents, an asset rather than a liability, included local amenity
value if the site provided a park, fishing access, or flood protection. WSL could then consider
these potential assets as well as the liability aspects in conjunction with other issues, like
ground- condition variations with construction implications.
Another example was exploring why a single contract and single contractor approach to
water main construction revised to a multiple contract and multiple contractor partnerships
approach might yield significant opportunities.
Paul suggested that all these examples were reasonably obvious possibilities for someone
relatively unfamiliar with the water and sewage industry – like Paul. He then made it clear
that these were his initial suggestions to clarify what was involved. But Paul wanted the
OG/PG team to use their deep fund of experience and that of all other relevant WSL staff
to ensure that the plans which went to the board in the D&A strategy gateway stage were
‘free of devils and full of angels’, and the earlier prototype version of the D&A stage analyses
were used effectively in the concept strategy progress stage.
They might also find it useful to draw on any experience relevant to the water and sewage
industry available from outside of WSL – perhaps using consultants, perhaps stimulating
dialogue with other comparable organisations.
Plans for relationships and contracts
Paul held a separate technical workshop on plans for relationships and contracts. He saw plans
for relationships and contracts as an important special case of planning areas where ‘devils
and angels in the detail were abundant’. He wanted to explain to the OG and PG teams why
this aspect of the seven Ws and associated goals– plans relationship linkages were crucial. He
wanted to enlist their support for well- developed prototype plans for relationships and con-
tracts which he could use to demonstrate why his concept strategy approach to contracting
for the new water mains project was opportunity efficient relative to current Finance Group
assumptions. He also wanted comparable prototype contracting plans for the new sewage
works, but this had a lower priority, and they should be relatively straightforward.
He began by explaining that formal contracts could be seen as necessary legal documents
to protect the rights of all parties, but some contract specialists were inclined to focus on
looking after the particular party who paid their invoices, with an overemphasis on legal
protection issues. He argued that preserving the spirit of a positive relationship was crucial
to any successful contract. Because the PG team would interface with WSL construction
contractors during the execution stage, and the OG team would become the interface with
these same contractors during the delivery stage, the OG/PG team needed lead roles during
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
384 Employing planning tools in practice
a period of relationship design starting in the concept strategy stage, followed by a period of
contract design which reflected the intended long- term relationships.
His starting point for relationship design was the need for a win– win philosophy. A clear
requirement was flexibility to cope with all uncertainty, and resilience to cope with signifi- cant surprises. Resilience was a core aspect of his general view of ‘robustness’. He believed
they needed to develop a ‘big team’ approach with local and national contractors which
kept all WSL contractors honest, competitive, and very keen to reliably deliver quality work
for WSL. The latter condition meant they had to be well paid for their efforts if they were
effective and efficient. Some of these relationships would involve using potential future con-
tractors as appropriately funded supporting consultants while WSL strategy was developed,
perhaps using WSL as supporting consultants in an appropriate manner once contracts were
signed. A series of fixed price contracts might be a reasonable expectation, and using fixed
price contracts in this way to keep it simple was highly desirable, but a wide range of other
‘partnering’ and ‘risk sharing’ approaches might be preferable.
Paul made it clear that he did not want to overburden them with contentious theories
and details. But he wanted their input to a prototype set of plans for the water mains project
which made sense from an OG/PG perspective, which he could then use with Frank to
attempt overall agreement within WSL before they started to explore their approach with
potential contractors. The new sewage plant’s contracting and relationship plans could fol-
low later.
The full OG/PG group approached this workshop as an initial brainstorming session,
with numerous follow- up meetings involving smaller groups to develop and test the earlier
ideas.
Later detailed planning of implementation tactics and further follow- on stages
Paul had a further separate workshop to discuss in outline what would be involved in all later
project lifecycle stages while they were still engaged with the concept strategy stage so they
could see what was coming without getting into the detail until they needed to.
He started by pointing out that detailed design and detailed planning for execution and
delivery would usually be a contractor responsibility in the ‘implementation tactics’ progress
stage. This progress stage would start immediately after contract signing, but some aspects
would continue after execution started. Detailed planning for WSL operations during the
operation lifecycle stage would be an OG responsibility, at a time to suit the OG prior to the
operation stage beginning. Furthermore, detailed planning for termination would involve
OG and PG responsibilities plus further new contractor responsibilities, starting just prior
to termination. Further still, detailed planning associated concept tactics including ongoing
updating of the business case would begin as soon as the concept strategy was approved and
continue until termination was complete.
He wanted all the OG/PG team members to understand how the whole lifecycle of each
project would unfold, and their roles in each stage. He dealt with progress stages first and
then related gateways.
Paul explained how the basic SPP of Figure 7.2 and the gateway SPP of Figure 7.6 would
need adaptation to detailed planning in the ‘implementation tactics’ stage, with further
modification to deal with the execution and delivery stages. For example, during the execu-
tion stage, the efforts of most people would have to focus on getting things done, but moni-
toring progress against plans was important, rolling detailed planning forward building in
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 385
lessons learned from monitoring was important, and dealing with any surprises was crucial,
as discussed in part 3 of Chapman and Ward (2011) and the wider project management
literature.
He did not labour the details of what would be involved, but he did emphasise that all
their current skills and experience would be made full use of in their new OG/PG team
structures, collaborative effort and learning from each other being the major changes. What
he wanted to do was shape the overall process and coordination structure they used.
Crucially, he would not be telling them what to do – but he would be telling them what
he wanted done within broad frameworks. They would be teaching him some important
aspects of how to do it and teaching each other.
They all needed strategic clarity within a common corporate basis from the outset. They
also needed tactical clarity in their own areas of responsibility, balancing the need to mini-
mise late unanticipated surprises without wasting time ‘crossing bridges before they got to
them’.
Paul’s ‘Quartet’ illustrations of enlightened planning teamwork
Before he began his series of technical workshops, Paul explained his strategy for the work-
shops and their role in his more general strategy to his father during a family Sunday walk.
Part of the feedback from his father was a suggestion that he read a current magazine review
of the film Quartet, recently released, about opera singers in a retirement home for gifted musicians, of interest to his father because the film’s director and almost all of the cast were
in his father’s post- retirement age group. Paul decided directly citing Quartet in his pres- entations was not compatible with his style, but several ideas in the review struck him as
important things he needed to emulate.
First, Dustin Hoffman, directing his first film aged 75, observed that he had never liked
directors who worked out all the scenes in their heads in advance. Despite his reputation as a
perfectionist, Hoffman insisted he wanted everybody involved ‘to surprise him’ – pleasantly
if possible, of course. Paul took this as an enlightened view of the need for creativity and
avoiding over planning which he needed to explicitly facilitate and encourage in the OG/
PG team and wider WSL teams.
Second, Hoffman significantly changed the tone of Quartet with a simple early decision to use real musicians and opera singers who were retired in supporting roles because they
brought enormous enthusiasm and passion onto the set. Paul overlooked the fact that this
might be a consequence of nobody having offered them work for several years and took it
as an enlightened view of the need to ensure passion and commitment, bearing in mind the
mature members of the current Planning and Estimating Departments might prove more
important advocates of an EP transformation than some of the relative youngsters in the
current Risk Management Department, if their enthusiasm could be stimulated.
Third, an important part of sustaining the passion and teamwork was everybody enjoy-
ing themselves and making generous allowances for the limitations and capabilities of oth-
ers. Paul took this as an enlightened view of part of what made strong teams – particularly
important when people were being asked to change their conceptual frameworks and mind-
sets as well as their toolsets and skill sets.
Fourth, Hoffman made a point of freely admitting when he was wrong and taking the
blame himself when things did not turn out right. Paul took this as part of what an enlight-
ened view of risk taking leadership involved – particularly important to remember when his
underestimation of other peoples’ reluctance to change was the underlying problem.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
386 Employing planning tools in practice
Further technical workshops
Paul never stopped what evolved into a tradition of weekly workshops, some involving other
groups, some involving a focus on OG or PG issues with attendance by others optional.
Making the learning processes and team- building exercises an enjoyable part of the overall
WSL capability- culture evolution development project, which he had initiated as part of his
broader change management project, became central to developing specific areas of compe-
tence and broader capability- culture assets as well as eliminating capability- culture liabilities.
Beyond the technical workshops – key corporate clarity concerns
A significant very specific concern for Paul was corporate clarity about provision and contin-
gency ownership issues associated with aspirational targets, balanced targets which might or
might not be expected outcomes, and commitment targets. So was his concern for a sound
corporate understanding of the ideas discussed in Chapters 3 and 4 with what needs to be
done corporate implications as well as how to do it implications. He made extensive use
of material borrowed from Chapman and Ward (2011) to emphasise what was involved at
board level as well as within the PG and OG teams.
For example, when BP adopted the SCERT prototype PUMP approach they developed
a new ‘uncertainty management group’ over a period of several years which was eventually
empowered to get involved in leading project planning and estimating, not just build on
earlier planning. BP then merged their estimating, planning, and uncertainty management
groups within a new group with uncertainty management leadership.
From the outset, BP made use of expected cost estimates approved in the govern-
ance (gateway) stage following strategy shaping for execution and delivery for corporate
budget management purposes, and plan revisions at this stage sometimes went back to
DOT strategy stage and even concept strategy stage decisions approved by earlier gateways.
These expected values were defined in terms of underlying aspirational targets plus associ-
ated provisions for expected levels of difficulty. But project budgets were defined in terms of
expected values plus a contingency sum defined by estimated potential upside cost variability
at an estimated 80% confidence level. At board level, it was clearly understood that this gave
ownership of a reasonable contingency to the project manager (which would not be spent
on average if known unknowns and unknown unknowns did not occur), as well as suitable
provisions (which would be spent on average). BP project managers found the implications
of this change in practice a major transformation. They no longer had ‘a date’ and ‘a cost’ to
meet which was repeatedly adjusted as achievement became demonstrably impossible. They
had ambitious aspirational targets, usually referred to as ‘stretch’ targets, but credible com-
pletion date commitments and cost commitment targets defined by their budget and inter-
mediate expected duration and cost estimates, with all these values understood by everyone
involved. Provided they used their provisions and contingencies effectively and efficiently,
they could deliver most projects within budget and timetable commitment targets. Paul had
insisted on a comparable starting position for his successful WSL sewage plant project when
he re- negotiated the budget before the contract was agreed. Paul was convinced that the
contract re- negotiation he had insisted upon should be avoided by WSL now that he was
Projects Director, using this form of EP approach from the concept strategy shaping stage
onwards. Furthermore, when he was approached to become Projects Director for WSL, he
had insisted on the transformation of his departmental structure to the new WSL structure,
following BP’s example immediately.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 387
One offshore North Sea project which BP was involved with in the early 1980s required
a new subsea oil pipeline that had to cross three existing pipelines. The risk of damaging an
existing pipeline was a serious concern, and for obvious reasons this risk was given consider-
able attention. The probability that such an incident might happen was estimated in order
of magnitude terms. The range of possible impacts was also estimated, in terms of additional
expenditure to repair the damage, to pay for opportunity costs experienced by the owners of
the damaged pipeline, to cover all pollution recovery and related compensation costs, and to
reflect reputational damage with widespread implications.
A key insight was the need for a board level financial provision for appropriate aspects of
the expected cost of such an incident, as was ensuring that the board was comfortable with
the associated contingency sum which might be required, and the reputational concerns,
keeping this issue outside of the project’s budget. This insight can be decomposed into
three component aspects:
1 Including any part of this possible incident’s impact costs in the project’s budget was
a clear lose– lose proposition from the board’s perspective. If this incident happened
the project manager would have to come back to the board for more money anyway
because it was a low probability high impact event which would not average out. If this
incident did not happen the board did not want the project manager to have an associ-
ated provision to spend on other problems.
2 The project manager had to take full managerial responsibility for making sure this
incident did not happen on BP’s behalf – he knew his career was riding on it.
3 The board had to own this risk financially, making a suitable provision and being pre-
pared to deal with the contingent implications if it happened, including implications
which are not just financial.
It soon became clear that the project manager could cascade a variant of this idea down-
wards within his project team, using lower level variants of his relationship with the board
to control all his project staff in terms of stretch targets, maintaining ownership of most
of his contingencies and provisions for low probability issues. He could release more time
and money without argument whenever the project was unlucky but the member of staff
involved had clearly done his or her best to meet a difficult stretch target. He could pre-
release some provision and contingencies to limit the dialogue necessary for normal varia-
tions – to avoid excessive micromanagement. Why a difficult stretch target was not achieved
would always be questioned, but the expectation on both sides would be a reasonable expla-
nation with no blame which was not clearly deserved. The project manager was cascading
the comparable treatment of his position by the board.
Paul wanted all these ideas implemented from the concept strategy stage onwards, with
a common corporate understanding of what this implied, at board level as well as within
his PG team and the OG team. In the early days of a project lifecycle, this meant significant
provisions and contingencies would be based on unresolved ambiguity uncertainty. For
example, if a new sewage works might be at possible locations C, D, E, or F and the decision
would not be made for some time, provision and contingency sums would have to reflect
this uncertainty. An enlightened approach to clarified ambiguity meant recognising the
importance of making provisions for extra cost and time which would be needed on average
from the outset, in addition addressing appropriate contingencies so that commitments had
a reasonable chance of being met, gradually refining the definition and ownership of provi-
sions and contingencies as time progressed. Before a contract was let, this uncertainty should
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
388 Employing planning tools in practice
never be confused with risk, and its ownership was not a budget issue in the post– contract letting sense. It was just ambiguity as currently assessed which needed identification and
approximate sizing with a view to accommodating it, and avoiding inappropriate decisions
because of bias until better information for making a commitment was available.
Paul pointed out that BP initiated formal statistical analysis of project outcomes about
five years after adopting an SPP prototype approach. This analysis confirmed that BP was
now bringing in projects on budget and on schedule most of the time – better than just
‘on average’ because of the role of provisions. However, this analysis suggested that it was
easy to lose sight of the importance of working assumptions which might not hold (known
unknowns) and implicit assumptions (unknown unknowns) which might provide unpleas-
ant surprises, eventually suggesting a no quibbles provision to meet some aspects of these
ongoing concerns.
Paul also pointed out that BP clearly had a corporate memory loss in terms of some of these
concepts in the planning process for their 2010 Gulf of Mexico projects, or perhaps they had
just never made the important distinction between making a project manager fully responsible
for managing all risks and the implications of a board level policy of contracting that exposed
BP to contractor failures, or perhaps something else had gone wrong. WSL had to ensure they
learned lessons from this too. WSL had to acquire, continue to develop, and maintain, all the key corporate capability- culture assets needed at all relevant corporate levels.
Enlightened approaches to governance, teams and contracts
Martha’s goal of an enlightened approach to governance, teams and contracts as explored in
Chapter 6 can be both reinforced and further clarified in Paul’s context.
One basic feature is a need for everyone involved in ‘a team’ to work with a shared
understanding of three very different values: optimistic stretch targets, unbiased expected
outcome estimates, and plausible worst- case outcome scenarios. If you clearly understand
the value and the satisfaction of working in this way and you find yourself working for a
person or an organisation which insists on a single number point estimate of cost, duration,
and other performance measures, only reveal a commitment figure with a high probability of
achievement, and be prepared to look for another job if this strategy does not look like suc-
ceeding. This is empirically tested advice which many people with confirm. The corollary is
that if you are trying to manage a team without understanding this point, you should expect
your team members to try their best to build enough fat into all estimates to facilitate an
easy life. Sometimes they will still fail to deliver even if significant fat is built into estimates,
and leave just before it all goes horribly wrong if they are able to do so.
Another key feature is a need for everyone involved to understand that ‘team commit-
ments’ should reflect the possible cost to the team of failing to meet the commitments
and the likely reward for the team of exceeding expectations. If these costs and rewards do
not align with costs and rewards for the organisations involved, this implies incentive and
broader motivation problems needing immediate attention. Capability- culture liability risks
with serious implications are involved.
Furthermore, if team members individually fail to meet team targets for good reasons
beyond their control, any ‘punishment’ is usually completely inappropriate and counterpro-
ductive, and if they exceed expectations for reasons beyond their control, ‘reward’ is usually
equally inappropriate and counterproductive. The net effect will be driving away competent
team members and promoting incompetent team members.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 389
Enlightened governance is in part about a nuanced understanding between team mem-
bers, plus those ultimately responsible for the team’s performance, which involves everyone
doing their best to pull in the same direction, accommodating delays, cost escalation and
other performance failures which may not have been anticipated if there were good reasons
for not anticipating them. Unconscious reciprocation of favours and tolerance of short-
comings across contractual boundaries as well as within organisational boundaries can be a
crucial part of associated trust, and it is trust which makes the whole thing work smoothly.
Enlightened governance in this sense involves a ‘covenant’ between all members of the
team and those the team works for. It may have some very important legal contract compo-
nents, but large parts of it will be informal and perhaps implicit. Trust is a crucial aspect, and
a loss of trust or misplaced trust can be cataclysmic.
‘Teams’ in this enlightened sense may be just small components of a much larger project
team within one organisation or a department in one organisation. However, they may
embrace a whole organisation or several organisations united by one project. Some very
small teams are narrow and dysfunctional. But some very large and extremely broad teams
work very effectively. How we choose to define the key team structures and their interfaces
matters a great deal. How we ensure each group which should work as a team does work as a
team also matters. Multiple team relationship structures may be multilayered and complex.
‘Enlightened contracts’ are frameworks for relationships which facilitate enlightened gov-
ernance. How we choose to define contracts clearly matters a great deal. The legal aspects
obviously matter. But all the informal aspects and trust issues also matter. Incentives are
part of this, especially the use of incentives to ensure the alignment of goals. Crude financial
incentives are often seriously dysfunctional, not simply ineffective. ‘Enlightened incentives’
may be highly dependent on the context and geared to behavioural issues like group pride.
A key risk that really matters is failing to fully understand these issues and their implica-
tions. A failure to act effectively based on appropriate understanding is a follow- on risk, a
part of this being a failure to understand all the options available. If realised, these risks are also usefully viewed as important lost opportunities, and seeking opportunity efficiency needs to include a focus on avoiding this kind of lost opportunity.
Paul’s overall change management strategy – target 1
The WSL board as a whole had very good reason to support Paul’s EP approach to the
concept strategy stage and rolling out its implications for all following stages. Furthermore,
Paul had done what was crucial to secure the full support of his PG team and most of the
OG team. Paul had already achieved most of what he needed to do to reach his first overall
change management strategy goal – target 1. However, as we bring this tale to an end, Paul
still had a lot of work to do to achieve his ambitions.
Target 1 loosely defined the end of the first lifecycle stage in his overall change management
project for WSL. Target 1 was ambitious, but it involved a coherent set of goals which
needed coordinated early achievement.
In summary, for all projects which ultimately involved the Projects Group, achieving
Paul’s target 1 involved WSL adopting:
1 the Table 7.2 lifecycle structure or a more complex SS3 variant;
2 the seven Ws structure of Table 7.3 and Figure 7.1 building on point 1;
3 a goals– plans relationships framework building on points 1 and 2;
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
390 Employing planning tools in practice
4 the basic SPP of Figure 7.2 and the gateway SPP of Figure 7.6;
5 initiation of enlightened planning in the concept strategy stage of all projects involving
the PG group, with companywide involvement coordinated by a PG team;
6 planning input in the concept strategy stage led by the OG/PG team using approaches
which were closely coupled prototype forms of later DOT strategy, E&D strategy and
D&A strategy development approaches;
7 planning input associated with financial resources and business case planning led by the
Finance Group using a two (or more) portfolios approach to discounting;
8 corporate understanding of the need for clarity about provision and contingency own-
ership issues and underlying aspirational targets, unbiased expected outcome values
and commitment targets (the basic ABCs of targets using expected values as balanced
targets); and
9 growing understanding of the role and importance of enlightened governance and
enlightened contracts, the importance of significant mutual support, and the crucial
role of enhanced team- working companywide.
Fundamental differences which would be noticeable when target 1 was fully embedded as
established practice included:
1 all point estimates of cost, duration, and other performance measures were replaced by
interval estimates;
2 uncertainty associated with interval estimate ranges was generally very wide but not
confused with risk;
3 an initial working understanding of ambiguity uncertainty which needed shaping and
sizing was acquired;
4 unbiased estimates were expected, with formal ongoing testing for bias an inbuilt part
of the overall SP for projects approach, and serious questions were anticipated if bias
persisted;
5 a wider range of design options with follow- on implications for other plans were
addressed at the outset of the concept strategy stage, keeping options open until it
was clear that the big opportunity strategic options which could be identified had been
identified, and as far as possible the big threat strategic options had been avoided;
6 a dynamic portrayal of how uncertainty would resolve over time in a Figure 7.9 frame-
work was understood by everyone involved, helping to clarify the difference between
risk associated with commitments and high levels of uncertainty early in a project life-
cycle which was not risk because commitments had not been made;
7 a companywide understanding of the nature of risk in terms of lost opportunities and
the role of this understanding as a capability- culture asset;
8 an overall WSL capability- culture change which all WSL staff embraced and were
proud of.
In summary, uncertainty, risk, and opportunity were clearly distinguished and related to
complexity, understanding the overlaps and interdependencies, using framing concepts and
operational tools which were clarity efficient to the extent that current capability- culture
assets allowed; aggressive further development of capability- culture assets was underway;
and ongoing ‘lessons learned’ reviews ensured relentless elimination of any capability-
culture liabilities which were exposed by reviews of what might initially look like bad luck.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 391
Paul’s overall change management strategy – target 2 and beyond
Paul’s goals for target 2 were loosely defined as the goals to be achieved by the end of the
second lifecycle stage in his WSL change management strategy. In overview terms, they
were supporting OG implementation of an EP approach to the DOT strategy stage and
leading his PG team through successful implementation of an EP approach to the E&D-
strategy stage for an initial set of projects to consolidate the target 1 achievements. His basic
guide was Chapman and Ward (2011), embedding the PUMP in the much broader SPP as
outlined in this chapter, but he also needed to add significant base plan creativity, informal
planning and formal planning process content. Paul needed his broad experience base and
his familiarity with complementary ideas to achieve his target 2 goals. Furthermore, Paul
needed the rich WSL capability- culture asset set he started with, as well as building on it and
eliminating identified capability- culture liabilities.
As additional goals and targets became the focus, concerns like companywide use of these
concepts for projects not directly involving Paul’s Project Group might become part of the
agenda – new customer- driven information systems led by Ian, Curt, and Frank for example,
which raised new issues touched on in the next chapter.
If you are interested in trying to follow Paul’s example as a Projects Director yourself, or
if you have someone like Paul reporting to you, or if you are reporting to someone like Paul,
you will need to consider these first two targets plus follow- on targets in the context of your
personal potential contributions to your own organisation’s needs, your own organisation’s
capability- culture asset and liability sets, discussion in the rest of this book, and significant
further thinking. Given the starting position this chapter associates with achieving target 1,
you should be able to work out approximately how the broadly defined EP concepts and
tools interface with the needs of projects and their partial resolution via PUMP concepts
and tools for all later targets.
The focus of Chapman and Ward (2011) is the E&D strategy stage, but earlier stages are
explored using case study examples, and how later stages flow on from the E&D strategy
stage is covered in some detail. This book, plus Chapman and Ward (2011), is a reasonable
starting point. However, this chapter is about how you might set about the kind of very
broad change programme discussed and the corporate advantages doing so might bring – it
is clearly not a definitive plan. To make its recommendations operational will require con-
siderable further effort tailored to the context of interest.
If you are not directly interested in project planning yourself, the purpose of this chapter
is clarifying the role project planning from an EP perspective could play in relation to other
aspects of organisations that are of interest to you in a preliminary way which following
chapters and later reflection can build on. For example, if you are a board level equivalent
of Charles or Michael, you might ask whether your organisation’s equivalent to Paul is cur-
rently making comparable contributions to your organisation, and if not, how you should
respond. Furthermore, you might ask related questions about your organisation’s equiva-
lent to Frank, or Ollie, or any other relevant directors.
Key credibility issues for this chapter
Interpreting WSL as a typical company which just happens to be in the UK water and sew-
age industry (specifically not a disguised variant of Southern Water), I believe all the key
features of the tale discussed in this chapter are highly plausible.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
392 Employing planning tools in practice
First, consider the discount rate issue. I developed my initial understanding of discount-
ing beyond attending undergraduate and MSc university courses discussing a mutual inter-
est in IRR controversies with Chris Hawkins and David Pearce when they were working on
an economics textbook addressing this topic (Hawkins and Pearce, 1971). David Pearce
later became a very distinguished professor of economics, and some of his later work under-
lies HM Treasury (2003a) approaches to the discount rate. My first published assessments
of discounting issues were Chapman and Cooper (1983, 1985), based on work under the
direction of Gavin Warnock of Acres for the Alaska Power Authority, trying to support the
case for proceeding with the Susitna Falls hydroelectric project, still not built. The approach
in the first of these papers was extended in Chapman and Howden (1997), based on its use
in the Nirex argument for proceeding with their plans to test the viability of disposing of
nuclear waste using a deep mine in Cumbria in their battle with HM Treasury rules during
the preparation of DoE (1994), a battle lost. It was not until after reflecting on HM Treas-
ury’s 2003 groundbreaking shift to a multiple tests approach outside the discount rate for
the public sector and publishing Chapman, Ward, and Klein (2006), that my understand-
ing of the implications for a private sector organisation discussed in this chapter started to
emerge. They were first published in Chapman and Ward (2011). Hopkinson (2016) devel-
ops some aspects of these concerns in a project management context, coupled to a directly
related concern for more project management with an uncertainty management focus at
the front end of the project lifecycle. The financial economics community has not yet been
directly confronted by these ideas on their own turf so far as I am aware. However, any pri-
vate sector organisation which uses bond funding or could use bond funding might find this
issue directly relevant, and ‘risk premiums’ in a discount rate are inappropriate with serious
implications whether or not bond funding is involved. All organisations need to review their
position on discounting. When doing so they will find the issues are highly contentious and
very complex. But they matter so much that it would be well worth attempting to properly
understand all the relevant implications and test all the key assumptions. Some people may
disagree strongly with these assertions and prefer sticking to a conventional received wisdom
position without really understanding the issues. But my contention is this aspect of the tale
is both very plausible and extremely important for a wide range of organisations in a wide
range of industry sectors. When the implications for a water and sewage industry organisa-
tion emerged, prior to publishing Chapman and Ward (2011), a draft paper was sent to
the Southern Water finance director. He responded positively and encouraged publication,
although I had been off the board for some time. It clearly has implications significantly
wider than project management or the water industry. It is certainly contentious. But there
can be no doubt that it matters greatly and needs careful attention by all organisations
affected by inappropriate choices.
Second, consider the problems associated with a common practice event- based approach
to project risk management. The event uncertainty based pre- 2007 HA approach to esti-
mating attributed to WSL was not an issue for Southern Water, a point I would like to
emphasise. While I was a non- executive director of Southern Water, the Projects Director
had previous experience which included offshore BP projects, and I was not aware of any
evidence of unusual estimation bias or other project management concerns. However, an
approach based on event uncertainty may be a significant issue for other players in the UK
water and sewage industry. In my experience it is certainly a central issue for a large propor-
tion of the UK organisations beyond the water and sewage industry, and it is a general issue
in Europe, North America, and many other parts of the world. The UK water and sewage
industry provides a useful illustrative context, but plausibility issues addressed here are not
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 393
specifically UK or water and sewage industry concerns – they are global concerns for all
organisations. My contention is this aspect of the tale is also very plausible and important in
this general context. Many colleagues share my views, and a well- established literature sup-
ports these views, but there are also wide groupings of ‘experts’ and ‘professional guidelines’
strenuously defending an opposing view. This, too, is a highly contentious issue with wide
implications.
Third, consider the problems associated with three separate departments in the WSL
Planning Group dealing with estimating, planning and risk management. Link this to a
Projects Group which does not work closely with an Operations Group during the concept
strategy stage of projects when the focus is the business case. My contention is yet again
that this aspect of the tale is both very plausible and extremely important. In some respects,
Paul’s approach to merging his three departments simply follows what BP did a few years
after adopting SCERT, which proved highly successful. The tale draws directly on this BP
example, and this helps to illustrate why the implications are in no way restricted to the UK
water industry.
Fourth, consider the problems associated with simplistic views of any aspects of any of the
four Fs – the four crucial frameworks developed in this chapter: the project (asset) lifecycle,
the seven Ws, the goals– plans relationships and the SP for projects in progress and gateway
forms. Link the implications to simplistic views of projects and all the other related concerns
addressed by the SPP concepts explored in this chapter. Yet again my contention is these
aspects of the tale are very plausible, and extremely important, with widespread implications.
The reasons for choosing the water and sewage industry in the UK as an illustrative con-
text for this chapter are various. One is it illustrates all these concerns when projects may be
one- off construction projects like a new sewage works, ongoing construction projects like
a water main replacement programme, or complex mixtures including water consumption
reduction programmes or change management portfolios. Another is it clearly demonstrates
that employing discount rates which are inappropriate involves crucial impacts that go well
beyond choosing the wrong projects. They imply shaping projects inappropriately, as in a
50 year designed life for water mains instead of 100 years, with too little concern for the
future (because it is discounted into insignificance), too much focus on a quick return as a
result. In some cases, this may drive or be driven by a culture which is inappropriate. High
quality assets with a long life, high reliability, and low operating costs in an organisation
with a long- term vision and a high quality service/products ethos can be fundamental to
a very strong capability- culture asset set. This can both support and enhance very strong
concern for maximum rates of return for shareholders, with a clear understanding of both
long- term and short- term issues. This strategic clarity and supporting ethos can be a missed
opportunity for organisations which suffer from short- term vision and quick- buck goals
driven by inappropriate discount rates leading to low quality assets which have a short life,
low reliability, and high operating costs.
Apart from the specific capability- culture liabilities addressed by Paul in the tale, WSL
enjoyed a very limited capability- culture liability set, and a formidable capability- culture
asset set. Paul had a very strong background and the knowledge plus the skill set and the
mindset needed to lead the changes involved. He also had a very supportive set of very
capable colleagues. Charles, Michael, Ollie, Frank, Richard, and Curt were all unusually col-
laborative and very skilled, with a shared corporate vision and an appetite for effective team
working, as well as achieving the goals of immediate and direct interest. If either Charles or
Michael had been interested in preventing Paul from achieving any of his key goals, Paul’s
task would probably have been non- feasible. Opposition from Frank or Ollie would pose
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
394 Employing planning tools in practice
a very serious threat. Two or more directors Paul needed support from who were simply
not up to the task would also pose a serious threat. It might be argued that such a strong
team is unlikely. As with earlier chapters, this is a reasonable view, but this aspect of the
working assumptions used to develop the tale serves to demonstrate what can be achieved if
capability- culture concerns are addressed effectively as part of the EP approach.
I have frequently encountered ‘models’ for Paul during in- house and public courses based
on Chapman and Ward (2011) and earlier versions of its approaches, in many different
countries. But the world could certainly do with many more of them, as well as many more
other board level managers like those WSL enjoyed.
Some further implications of clarity inefficient toolsets
One of the important roles of EP is using an EP perspective to explain why some common
practice tools are clarity inefficient and should be eliminated from everyone’s toolset if they
want to embrace an EP perspective. This section begins by consolidating why this is the case
for reasons addressed earlier, as a basis for then briefly considering some further implications
of crucial importance.
Clarity inefficient tools which should not be used
Figure 7.7 was used by Paul to explain why he would not tolerate further use of PIGs
within his Projects Group, and Figure 7.9 was used in the HA Nichols study discussion to
drive home one of the results of a PIG- based approach to risk adjustment estimates which
are widely employed. Some risk management experts within and beyond the project risk
management community may be seriously offended by this view. Professional bodies which
advocate PIGs (under various labels) may feel this is an unwarranted attack on their profes-
sional standards. It is a clear attack on some professional standards and many people’s pre-
ferred perspective. But in my view, it is an attack which is long overdue, very badly needed,
well founded, and fully justified. It is important everybody understands why PIGs once made sense and what has changed, especially those involved in employing people who persist with using PIGs or supporting organisations which persist with promoting PIGs.
When PIGs were first employed in a pre- 1960 safety context the only risks being addressed
were events (specific failures and accidents) happening or not. Appropriate data for conven-
tional objective data- based statistical methods were not available, subjective probabilities
were generally held to be inappropriate for any formal decision- making procedures, and a
‘decisions under uncertainty’ mode of classic decision analysis was the obvious conceptual
and operational framework. Safety managers needed a simple, rational means of prioritising
the list of threat events they were able to identify so that they could allocate resources to
risk minimisation in a defensible manner. In these circumstances PIGs were a very reason-
able basic or starting point tool. Fault trees and event trees used within an expected out-
come framework were richer tools, employing a generalisation of classical decision analysis
in ‘decisions under risk mode’. But they were very demanding in terms of data and analysis
effort, and they had some common or comparable weaknesses in practice, like largely ignor-
ing ambiguity uncertainty, systemic uncertainty and risk associated with overall variability.
Since the 1960s, authors like Raiffa (1968) have made the framing assumptions for the
view ‘proper probabilities require data’ redundant, along with classical decision analysis.
Modern decision analysis is founded on framing assumptions about subjective probabilities
shared with EP. This shared view of probabilities is a crucial issue, and it means that for
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Adapting ‘generic processes’ 395
EP purposes, all well- founded probabilities should reflect data and their objective analysis
insofar as this is both feasible and useful, but probabilities should be statements of belief
which reflect a synthesis of all relevant decision- maker knowledge of the implications of
data and analysis shortcomings. This includes addressing all relevant reasons why the future
may be different from the past, with effective treatment of valid and important differences
of opinion.
Furthermore, since the 1960s, all effective approaches to risk management have recog-
nised the importance of dependence and systemic risk, especially approaches building on a
Markowitz (1959) approach to portfolio management and the role of risk efficiency.
From this perspective PIGs are simply weak quantitative tools in a qualitative framework
which is not sufficiently general. From this perspective PIGs should be made redundant
as a starting point for all analysis in all contexts – including ERM. ERM in many different
forms has blossomed in the last couple of decades. The forms discussed in Chapters 8 and
9 are generalisations of a well- established portfolio management framework which ought
to be central to ERM but often is not. Safety and security risk management as addressed in
Chapter 9 is a special case within the Chapter 8 approach.
Even if events are the only sources of risk of interest, Figure 7.7 or some of the variants
discussed earlier using subjective probabilities are a superior tool in all respects. There is
a growing body of other people sharing this view, but the rate of enlightenment needs to
accelerate.
However, there are important further reasons for rejecting the conceptual framework
underlying PIGs – reasons which go beyond operational tool concerns illustrated directly
by Figure 7.7. For example, PIGS and related ‘risk list’ event- based approaches do not
lend themselves to dealing with sources of uncertainty which are not events, central to the
issues explored using Figure 7.8, and they do not facilitate effective assessment of alternative
responses, some of which may be specific to particular sources of uncertainty, some of which
may be a deal with sets of sources of uncertainty. The full implications of these reasons are
complex and not immediately obvious to many people who do not adopt an EP perspective,
but this does not make them unimportant.
Paul’s use of Figure 7.8 plus his use of the HA examples illustrates how estimation based
on some conventional event- based project risk management ignores information readily available which results in persistent bias. Persistent bias as displayed by the HA pre- 2007
and WSL is unacceptable. Some people see it as so widespread it is inevitable, not an avoid-
able dysfunctional risk, but a fact of life people have to live with. Just ‘get over it’ sometimes
seems to be the attitude. But I am not alone in seeing it as both dysfunctional and prevent-
able, and many share my view that it ought to be seen as incompetent in the very near
future, ultimately as professionally negligent.
Persistent bias associated with estimates is not by any means the only driver of clarity
inefficiency associated with PIGs, but it does stem from related underlying assumptions.
Furthermore, it can be exposed and avoided by adopting low clarity but clarity efficient EP
tools to capture all relevant uncertainty. As skill and understanding are gained, higher clarity
approaches will provide big dividends. The biggest benefits associated with dropping PIGs
may be driven by risk efficiency and further higher order opportunity efficiency gains.
Paul’s approach to discounting further widens the focus of some received wisdom to be explicitly avoided. Embedding issues other than an appropriate cost of capital in NPV dis-
count rates is an example of a source of persistent bias with a somewhat different underlying
assumption basis. Before we had a multiple criteria mathematical programming framework
available, single hurdle rate discount factor tests for projects seemed plausible, if not the
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
396 Employing planning tools in practice
only way to go. It is not surprising that finance texts rationalised single- hurdle rate- test
approaches with working assumptions which became invisible framing assumptions – stealth
assumptions in EP terms. HM Treasury’s 2003 advice (HM Treasury, 2003a) on discount-
ing uses a multiple test framework, the first published recognition of this issue that I am
aware of and a notable breakthrough. My belief is HM Treasury reducing the real dis-
count rate required in 2003 from 6% to 3% was about right, but the basis of its 3% choice
was not appropriate. The published argument (Chapman, Ward and Klein, 2006) that the
2003 Treasury approach correctly uses a multiple tests approach grounded on an underlying
multiple criteria framework, but it couples this to an inappropriate conceptual basis for the
choice of discount rate, which has not been accepted yet so far as I am aware. Most public
and private sector approaches have yet to even recognise the need for multiple tests to deal
with multiple objectives in a general manner.
The idea that unbiased estimates of a discount rate ought to be driven by the expected
cost of capital, and nothing else, is a very important aspect of the arguments about privatisa- tion and nationalisation of a wide range of assets. This chapter only touches on an explora-
tion of the implications, but they are clearly significant. Private sector funding rate of return
expectations are usually very much higher than public sector bond funding rates. This very
important difference will drive very different asset investment strategies. This will not just
affect which projects are selected. It will affect the way the basic nature of the projects is
shaped. It may also affect underlying culture and ethos issues in important ways.
Nobody should believe that public sector organisations funded by government bonds will have the same investment strategies if they are transformed into private sector equivalents
with 100% equity funding of capital investment, organisations like WSL occupying a mid-
dle ground. But also very important, nobody should believe that nationalisation of private sector organisations is a simple transformative change with impacts centred on discount rate
considerations, whether or not a reversal of previous decisions is involved. The implications
of this kind of change are extremely complex, and context specific assumptions need careful
attention, with some examples explored in Chapters 8, and 10.
One of the most fundamental concerns this book addresses is stimulating an understand-
ing of the need to establish a widely accepted set of general framing assumptions which can
be used to identify and overcome stealth assumptions which cripple capabilities in many
planning and decision- making contexts. Project planning is a particularly good context to
illustrate what this might involve. But the evaluation of conceptual and operational tools
addressed in this chapter just scratch the surface.
Some organisational structure and professional body implications
EP for projects as explored in this chapter not only builds on the central ideas in How to Manage Project Opportunity and Risk (Chapman and Ward, 2011), but it also builds on a vast project planning literature, with many important influences not even referenced.
There are particularly important implications associated with managing the creation and
enhancement of plans as well as shaping plans in all stages of a project from the concept
strategy stage onwards. Using the SPP approach implies that risk management in the broad
opportunity plus risk and underlying uncertainty management sense should never be pur- sued in a separate silo, separated from associated planning and estimation, which should
themselves never be in separate silos. The focus is a project planning context, but the SPP
approach links project planning to operations and corporate planning directly, crossing
silo boundaries in the process. Furthermore, this perspective is also clearly relevant in any
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
398 Employing planning tools in practice
projects approach illustrated in this chapter. The SP for bidding of Chapter 6 illustrates a
mid- range position. If the ‘select and focus the process’ phase of the SPP used by any organ-
isation does not accommodate what is known about useful processes for all the contexts to be addressed by that SPP, then someone directly involved needs to use a UP equivalent to assess how to modify the SPP to accommodate new contexts beyond the current scope
whenever this may be appropriate. This is true of all SPs.
Common ‘techniques’ of all kinds can be seen as SPs in this sense and assessed for critical
stealth assumptions using an EP perspective. A key EP capability is testing common practice
techniques for stealth assumptions which make them inappropriate for some contexts, suit-
able for other contexts, to facilitate enlightened choices.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
The questions raised by the brief discussion of Ajit’s corporate strategy formulation con-
cerns towards the end of Chapter 5 need responses now. Building on Ajit’s observations,
one set of issues which needs to be addressed involves developing a clear top- down view of
the full scope of corporate goals, starting with key ownership issues. A second set of issues
involves formulating a top- down view of appropriate corporate strategy directions given
corporate goals. A third set of issues involves deciding which bottom- up propositions for
corporate changes with strategic implications need further attention.
The focus of this chapter is structuring top- down formulation of corporate strategy in a
way which facilitates integrating bottom- up strategy formation and collectively testing the
resulting overall corporate strategic plans. Bottom- up corporate strategy formation trig-
gered by the strategic needs of specific operations has already been addressed in several ways
in all preceding Part 2 chapters, and the role of enlightened project planning in the integra-
tion of some corporate- wide issues was considered in Chapter 7. However, top- down cor-
porate strategy formulation and testing have so far been discussed to a very limited extent.
‘Strategy’ is used in the usual plain English sense of ‘choices involving decisions with
important, broad or long- term implications’. ‘Corporate strategy formulation’ assumes a
broad plain English interpretation of ‘formulation’ in the ‘setting forth in a systematic man-
ner’ sense. This ‘setting forth in a systematic manner’ is interpreted in terms of creating,
enhancing, shaping, testing, and interpreting prior to implementation, in a manner consist-
ent with the UP of Figure 2.1. A narrow interpretation of ‘formulation’, like ‘reducing to or
expressing in a formula’, is explicitly avoided. The key underlying framing assumptions are
‘corporate strategy formulation’ should involve a formal corporate top- down overview from
the strategic end of a strategy- to- tactics scale, but integrating bottom- up strategy formula-
tion is an essential part of the holistic approach required. This is consistent with earlier usage
of these words and common practice.
Most organisations distinguish between short- term and long- term corporate planning,
and many use a variety of additional categories. The enlightened approach to corporate
planning advocated in this book adopts a formal separation of the approaches taken to short-
term, medium- term and long- term corporate planning horizons as a basic default position.
It also assumes that all three of these ‘categories’ of corporate planning are focused on
‘feasible’ options, where ‘feasible’ is assumed to mean ‘known to be technical feasible’,
explicitly excluding ‘currently technically non- feasible’ options. For reasons clarified later, it
may be important to include some technically feasible options which are not economically
feasible at present, and it may also be crucial to include some options which are currently
non- feasible in terms of the organisation’s capability- culture assets and liabilities. The com-
plex set of underlying issues clarified by this interpretation of ‘feasible’ in technical terms,
8 Corporate strategy formulation – an electricity utility example
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
400 Employing planning tools in practice
with economic feasibility and capability- culturally feasible nuances, are important in prac-
tice. They are handled as simply as possible by restricting the long- term, medium- term,
and short- term planning concepts to feasible options in this nuanced sense plus the use of
two further separate ‘categories’ of corporate planning, referred to as ‘futures planning’ and
‘goals planning’.
‘Futures planning’ provides a focus for addressing what might be feasible but is not cur-
rently feasible for technical, economic or capability- culture reasons. Futures planning may
look beyond the long- term planning horizon to a very distant ‘futures horizon’, but it
may also work within the range of the long- term, medium- term and short- term planning
horizons. ‘Futures planning’ includes a number of different common practice approaches,
many with labels which indicate their nature – like ‘horizon scanning’, ‘futures research’,
and ‘technological forecasting’. But a core component which is less well- known is Ackoff ’s
‘designing desirable futures’ perspective. ‘Futures planning’ in an EP sense uses a ‘designing
desirable futures’ perspective to get beyond a ‘predict and prepare’ mindset when address-
ing the gap between currently feasible options and options which are currently judged or
implicitly deemed non- feasible but might be made feasible, identifying what is involved,
where attempting to close the gap may or may not be a valuable option, and how to pursue
the chosen options. The ‘gap’ may be a technical, economic, or capability- culture issue.
‘Goals planning’ is about a fully integrated formal view of corporate goals in an operational
goals– plans relationships structure. Corporate goals may be initially expressed in a wide vari-
ety of ways, using words like ‘vision’, ‘core values’, ‘mission’, ‘aspirations’, ‘aims’, and ‘objec-
tives’. But however corporate goals are expressed, they need to be formally structured for
operational purposes within and between four nominal planning horizons: the short- term,
medium- term, long- term and futures planning horizons. Goals planning is about addressing
where an organisation wants to be at any point in the future in terms of basic corporate goals
relevant to all planning horizons, giving the vision which shapes goals operational content.
Some goals may be associated with measurable criteria, but some crucial goals may not be
measurable in any meaningful sense. The achievement of some goals may not be feasible
currently, but trying to make their achievement feasible via futures planning may be crucial.
‘Long- term planning’ addresses where the organisation should aim to be in terms of key
capability- culture assets and liabilities plus all other relevant assets and liabilities at a suitable
long- term planning horizon. This form of planning is based on the organisation’s corporate
goals and what is currently known to be feasible, acknowledging all relevant uncertainty,
including opportunities associated with the possible transformation of currently non- feasible
options into feasible options via futures planning.
‘Medium- term planning’ is about given where the organisation is now and where it is
aiming to be at the long- term horizon, what is the best way to plan to get there over a
medium- term horizon, making commitments when appropriate. Medium- term planning
has to address commitment and lead- time issues in terms of both expectations and associ-
ated uncertainty. It also has to address changes in objectives which might reasonably be
expected in the process of getting to the long- term horizon, including those driven by
possible successes with attempts to transform currently non- feasible options into feasible
options via futures planning.
‘Short- term planning’ is about making the best use of current resources in the usual
short- term- planning sense, responding to new feasible options as they arise if appropriate.
Assuming separable treatment of these five different categories of approach in four dif-
ferent planning horizons does not imply assuming independence. Addressing all relevant
decision- making interdependencies is a central concern. However, formally acknowledging
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 401
separable categories of approach facilitates using different operational frameworks within
each of the five categories of planning, a practical concern of central importance.
It can be useful to associate different categories of approach and their associated hori-
zons with different ‘levels’ of planning, putting goals at the top level, followed by futures,
long-term, medium-term, and short-term. Using this view there are obvious ‘downward
dependencies’: goals planning decisions shape futures planning decisions, futures planning
decisions shape dependent long- term decisions, and so on. These downward dependen-
cies are clearly important. But there can also be very important ‘upward dependencies’,
which are often overlooked. Examples illustrating ‘upward’ as well as ‘downward’ depend-
encies are developed in this chapter.
Strategy formulation and testing also involves ‘management levels’ and associated ‘man-
agement structures’ to deal with the integration of top- down and bottom- up strategy for-
mulation and testing processes. Boards of directors and supporting strategic planning staff
have to shape strategy from the top down, but effective strategy management also requires
‘emergent’ strategy processes driven from the bottom up or ‘sideways’ within corporate
management structures which have to serve a number of functions. This chapter empha-
sises the structure provided by an effective top- down approach, but it also recognises and
addresses the importance of integrating bottom- up and sideways approaches in whatever
management structures an organisation adopts.
Within the EP corporate strategy formulation and testing framework developed in this
chapter, all aspects of corporate capability- culture and other relevant asset/liability issues
need explicit attention. The capability- culture aspects are given limited attention, and they
are not addressed until later, but they are much more important than suggested by the
ordering used and the extent of the discussion.
At a corporate strategy formulation and testing level, it is crucial to remember that ulti-
mately everything is potentially connected to everything else – and key interdependencies are a major source of both opportunities and risks. This makes keeping strategy formula-
tion and testing simple and focused for everyone involved particularly difficult but particu-
larly rewarding. And it means that not doing so effectively is particularly damaging. The
notion of enlightened simplicity in process terms will receive very limited direct considera-
tion, mostly deferred until towards the end of the chapter, but this seriously understates
its crucial role.
Within the overall formal corporate planning framework developed in this chapter, it is
useful to routinely remind ourselves that the form and interpretation of formal corporate
planning by all participants is always heavily influenced by the context details. Furthermore,
‘context’ includes all other planning activity, both formal and informal. In an implementa-
tion context, background informal planning enhances and shapes plans as well as filling in
necessary detail. Prior to implementation, context assumptions raise important questions
about communication and motivation. For example, will all the parties involved interpret
plans in the same way, and are they all motivated to pursue aligned objectives?
This chapter’s tale involves a Canadian electricity utility owned by its provincial gov-
ernment, labelled ‘Canpower’. In some respects an electricity utility has strategy formula-
tion and testing concerns which are particularly difficult. This gives the tale useful richness.
In other respects strategy formulation and testing for an electricity utility is particularly
straightforward to structure, which helps to make the tale effective.
The basis of the tale is a consulting contract undertaken over the period from 1991
to 1993 for the Independent Power Producers Society of Ontario (IPPSO). IPPSO was
an official ‘intervener’ in an Ontario government formal enquiry process set up because
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
402 Employing planning tools in practice
Ontario Hydro (now OPC, Ontario Power Corporation) wanted approval for the construc-
tion of ten nuclear power stations over a 25 year planning period. IPPSO was opposed to
these plans, and because it was an official intervenor, it obtained funding from Ontario
Hydro for my contract. I acted as a consultant and expert witness for IPPSO, producing
Predicting and Dealing with an Uncertain Future (Chapman, 1992), a report published by the government commission. My report was a highly critical assessment of Ontario Hydro’s
overall approach to corporate strategic planning for the next 25 years, including underlying
forecasting. The Ontario Hydro corporate plans assessed were built around the proposed
ten new nuclear power stations, justified using just two planning horizons – short-term,
plus what Ontario Hydro called ‘long-term’, referred to in this chapter as ‘medium/long-
term’. Ontario Hydro used a largely deterministic approach to medium/long- term plan-
ning, employing demand and cost forecasts which were seriously optimistically biased but
associated with a wholly unwarranted degree of confidence.
My report argued that the deterministic treatment of medium/long- term planning within
a two horizon approach involved a serious set of flaws; the forecasting approaches which
Ontario Hydro’s medium /long- term strategic planning approach was based on, including
cost estimation approaches in a broad view of ‘forecasting’, were seriously unreliable and
would probably fail in an empirically demonstrated manner before long, a second serious set
of flaws which underlay the first; Ontario Hydro’s whole approach to strategic planning and
associated forecasting should be scrapped, and its recommended strategy should be rejected,
as a consequence of the combined effect of these two serious sets of flaws; a very much bet-
ter approach to strategic planning based on separate medium- term and long- term planning
approaches should be developed, built on a revised interpretation of current forecasting and
a much better approach to all cost estimation aspects of forecasting.
To avoid criticising current practice without explaining the nature of better alternatives
my report outlined what an appropriate approach ought to look like in ‘what needs to be
done terms’, the starting point for drafting this chapter.
Ontario Hydro’s forecasting failed as I had predicted, and this failure was picked up by the
Canadian press, shortly before I was due to appear in person as an expert witness. Ontario
Hydro then withdrew its 25 year corporate plan.
Three key sets of flaws are attributed to Canpower in the tale of this chapter. The basis
of the two sets of flaws identified near the outset of this chapter and the nature of the
proposed treatment of them draw on my Ontario Hydro report with enhanced illustrative
examples enabled by EP developments over the last 25 years. Some of Canpower’s char-
acteristics underlying the third key set of Canpower flaws are based on my understanding
of hearsay evidence about Ontario Hydro capability- culture liabilities in the early 1990s.
In brief, this third set of flaws involves capability- culture liabilities driven by assumptions
about how a public sector utility might reasonably address feasible options which were
arguably flawed in a way which put the organisation’s public-sector status at risk. Dealing
with the issues associated with this third set of flaws was not part of my IPPSO contract,
and the approach to this third set of flaws adopted in this tale would not have occurred to
me 25 years ago.
The story told by the tale of this chapter is grounded in a very real situation. However,
the plot of this chapter’s tale and all the characters in the tale are fictions. I have no wish to
embarrass Ontario Hydro (OPC) beyond the public critique in my 1992 report.
That said, I would urge all the citizens of Ontario to ensure that flaw sets one, two, and three have all been dealt with effectively if the current momentum behind OPC privatisation does not make it too late to do so, in which case a more complex but directly related set of
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 403
problems will have to be addressed or endured. Furthermore, my plea to anyone with elec-
tric power supplied by a government utility is to see resolving the third Canpower flaw set,
or comparable issues, as the key to avoiding the very serious damage to a national treasure
caused by the 1980s privatisation of the UK Central Electricity Generating Board (CEGB)
and the politics of subsequent regulation of energy utilities in the UK. Further still, my plea
to all those who prefer having their electricity provided by private sector utilities is to make
sure that there are no features of the approach to this chapter’s tale and the linked discus-
sions in later chapters which might improve private sector performance from a consumer
perspective.
You may regard my view that privatisation of electricity generation in the UK was a seri-
ous mistake as contentious and misguided and perhaps as irrelevant to a book on planning
in general terms. However, if you currently support private sector ownership as a political
preference in all contexts, I hope you might wish to reconsider your position on the basis
of the approach developed in this chapter plus the next three chapters. Furthermore, if you
currently support a more conventional approach to public ownership of utilities than that
proposed for Canpower by the end of this chapter, I hope you might also reconsider your
position.
The position on public versus private sector ownership choices argued by this book is
there is no such thing as a ‘good’ or ‘bad’ choice in all contexts – the overall merits of the
options involved depend on the context and how the selected choice is likely to be imple-
mented. Chapter 7 touched on this issue but avoided dealing with it directly. Chapter 9
develops further related issues, and Chapter 10 provides a synthesis of the earlier discus-
sions, briefly added to in Chapter 11.
The capability- culture issues underlying the third key flaw set in Canpower’s approach
were not addressed until the third draft of this book. The insight needed to identify and
address this third flaw set is attributed to the joint use of the capability- culture concept
and the revised front end of the UP when using the UP concept in top- down planning
mode. The basis of this additional insight had a related impact on my third draft treatment
of TLC’s approach to EOQ (Economic Order Quantity) models in Chapter 5. However,
in this chapter the UP concept functions in the background. There are brief clarifying
comments about top- down UP use at the end of the chapter, but directly illustrating the
role of the UP in the development of this chapter’s position as the chapter progressed
did not seem feasible. A very simple process to make use of the top- down structuring is
described, which should be interpreted as a sketched outline of an SP which needs further
development and tailoring by any organisation adopting a variant of the approach advo-
cated in this chapter. This further development will require the use of some variant of the
UP concept.
The critique of Canpower’s approach in terms of flaw sets one and two starts with a sim-
ple explanation of what the pre- 1992 approach adopted by Canpower was doing instead of
the more technical treatment in my 1992 report on Ontario Hydro. This is closer to what
I would have liked to have said in a private conversation with the board of Ontario Hydro,
had I been asked for such a confidential discussion instead of the formal public report.
Most of the key concepts developed in this chapter are relevant in all corporate strategy
formation contexts, but some important details may need significant adaptation in other
contexts because context is so central to the shaping of strategy formulation processes. As
with the context of preceding tales, you may have a limited interest in electricity utilities, but
to fully understand how to adapt the key ideas which this chapter explores, you will need to
engage with the electricity utility context and examples.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
404 Employing planning tools in practice
An EP mantra particularly relevant to this chapter is a generalised and holistic interpreta-
tion of the common saying
If it ain’t broke don’t fix it.
This is a very widely used saying, but the generalised and holistic interpretation intended
is not the usual emphasis. It is attributed to Paul Henderson, because he introduced me
to both the basic mantra and its holistic interpretation in the 1950s. I was learning to race
small sailing boats on Toronto Harbour in my early teens, and Paul was the senior instruc-
tor at the junior club that I was very lucky to attend every weekday over two months for
four summers. Paul represented Canada in three Olympics in the 1960s and 1970s, sailing
as helmsman in the Flying Dutchman class, just one of his many contributions to Cana-
dian sailing. Paul’s take on ‘if it ain’t broke don’t fix it’ in terms of racing dinghies was
‘understand what lies behind your boat speed and that of your competitors in terms of a
holistic systems view, and only make changes which make you more likely to win overall.
Furthermore, don’t fiddle with the tuning of your boat (e.g. adjusting mast rake) unless
there are good reasons to make you think that you know what you are doing and may cur-
rently have it wrong’.
I remember the mantra and Paul’s broad interpretation because when I was about 14 he
used the mantra with characteristic sarcasm to admonish me when I was guilty of fiddling
(inappropriate fine- tuning). Having grabbed my full attention he then explained, with trade-
mark kindness and clarity, a systems theory view of winning sailing races which a 14 year old
could understand and implement. Several years later when I was an undergraduate engi-
neering student, a professor at the University of Toronto explained the importance of safety
factors as provisions for uncertainty to deal with complexity we do not fully understand
which should not be fiddled with because engineers had a habit of reducing safety factors
until bridges and buildings started to fall down more often than expected, a notion which
when generalised intersects with Paul’s mantra. Many years later, Stephen Ward suggested
I read an article with the memorable title ‘Beware Fine- Tuning’, the same basic idea from
a further perspective, exploring issues like making organisations more fragile by removing
‘organisational slack’.
This seems a particularly apt mantra for EP in this chapter, reinterpreted in terms of the
achievement of corporate aims and objectives in holistic terms, recognising that everything
is potentially interconnected, much of the complexity is difficult or impossible to really
understand, and the difference between unhelpful fiddling and continuous improvement is
very subtle but absolutely crucial. It can be interpreted as a top- down take on ‘keep it sim-
ple systematically’ aspects of enlightened simplicity, linked to but not quite the same as the
bottom- up take on enlightened simplicity developed in earlier chapters.
Constructive tension between the mantras ‘keep it simple systematically’ and ‘always “go
the second mile” when appropriate’ is useful, as argued earlier. Comparable constructive
tension between the mantras ‘if it ain’t broke don’t fix it’ and ‘if it’s worth doing it’s worth
doing right’ is important in a top- down strategic planning context, as is understanding the
subtle but important differences between inappropriate fine- tuning or ‘fiddling’ and robust development and implementation of a continuous improvement policy.
All these concerns and their interdependencies are arguably central to developing strategic
clarity in terms of your personal perspective and strategic clarity for organisations of interest.
As in previous Part 2 chapters, the credibility of the tale raises issues discussed towards the
end of this chapter, the issues are different from those addressed in earlier chapters, and EP
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 405
language is used throughout. Like Chapter 7, the emphasis is not quantitative models, but
a few illustrative numerical examples are employed to illustrate key issues.
Carl’s situation and all the key characters in this tale
The key characters in this chapter’s tale are
Clive, the recently appointed chairman of the board of Canpower;
Ivor, the Canpower director made responsible for innovative thinking by Clive;
Carl, a consultant hired by Canpower for a study initiated by Ivor;
Larry, the director responsible for Canpower’s long- term planning and forecasting;
Dick, the director responsible for Canpower’s ‘day- to- day’ operations and associated
short- term planning and forecasting; and
Pat, the director responsible for Canpower’s project planning.
In the early 1990s, Carl was a corporate strategy consultant in the UK. He was contracted
to advise ‘Canpower’, a Canadian electric utility owned by its provincial government, with
a remit summarised by Ivor as ‘How should Canpower embed all key management issues
associated with uncertainty, risk, opportunity and underlying complexity into the formal cor-
porate strategy formulation processes employed, revising current approaches as necessary?’
Carl believed he had the background needed for this assignment, but he knew from the
outset that Ivor, the director who had insisted on his appointment and created the study
brief which he had to address, was not fully supported by Canpower’s board. Three execu-
tive directors were openly hostile: Larry, Dick, and Pat – led by Larry.
Larry was responsible for what he viewed as long- term strategic planning. Larry was also
responsible for what he viewed as Canpower’s long- term corporate forecasting, including
electricity demand and all relevant costs. His title was Corporate Planning Director. He
used a framework for long- term strategic planning involving a deterministic approach with
a mathematical programming optimisation basis, which was unusual. He used conventional
approaches to long- term forecasting which formally acknowledged uncertainty in probabil-
istic terms, but his understanding of the implications was seriously flawed. Carl will refer
to Larry’s long-term as ‘medium/long-term’. Carl’s terminology is this chapter’s default interpretation of ‘long-term’ and ‘medium-term’ from here on.
Dick was responsible for day- to- day operations management of electricity production
and distribution. He was also responsible for associated short- term forecasting, plus all gen-
eration and distribution plant maintenance, repairs, and minor refurbishment projects. He
viewed his remit as short- term corporate planning and short- term plan execution. His title
was Operations Director. His role was comparable to Ollie’s role in Chapter 7.
Pat was responsible for the planning and delivery of all new electricity power station and
distribution grid projects plus major refurbishment projects – a Projects Director with a role
comparable to the Projects Director Paul replaced in Chapter 7.
This trio had oversight of three largely independent empires, whose senior management
liked it that way, and wanted to keep it that way. They saw no reason to change what, in their
view, they had done successfully for some time. Larry was incensed that anyone should even
suggest that significant change to his approach to corporate planning might be necessary.
Ivor, the Canpower director who initiated Carl’s study, was ‘the board’s innovator and
catalyst for new thinking’, as he described his role. Ivor had just been appointed as an execu-
tive director on the insistence of Clive.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
406 Employing planning tools in practice
Clive was a recently appointed new chairman of the board. Clive had insisted on Ivor’s
appointment, to test the status quo by challenging it where there might be room for
improvement. He was convinced there was room for improvement, but he wanted clarity
about where, and what needed to be done. Clive sometimes referred to Ivor as ‘the board’s
devil’s advocate’.
Ivor had yet to prove that he was not just ‘a devil of a nuisance’, as Larry referred to him
in private conversations with Dick and Pat.
Luke Johnson provides a contemporary summary of what Ivor’s role might imply in the
short article ‘Every Board Needs a Member of the Awkward Squad’ (Johnson, 2017), using
the ‘devil’s advocate’ term. However, it is useful to see Clive, Ivor, and Carl as a team who are
focused on designing a desirable future for Canpower, with Ivor’s responsibility for innova-
tion involving a much wider and deeper interpretation than ‘devil’s advocate’ might suggest.
The hunch which motivated Ivor to pick on corporate strategy formulation in relation to
uncertainty, risk, opportunity, and underlying complexity management was aspects of the
Canpower culture driven by the way Larry, Pat and Dick managed their groups, as portrayed
by a number of ‘cultural clues’ which Ivor had observed. Several relevant strong ‘revealed
preferences’ were observed by Ivor:
1 ‘Big was beautiful’ in terms of plant sizes and developments. Larry and all his managers
seemed to favour big developments. Dick and all his managers liked big plants. Pat and
all his managers seemed to favour big projects.
2 ‘Deterministic detail was desirable’ in terms of all plans. They all liked detailed deter-
ministic plans for many years ahead which assumed things would go as planned.
3 ‘Control was crucial’ in terms of keeping operations in- house.
4 ‘Collaboration with others and projecting influence to other organisations which mat-
tered, but Canpower could not control, was not a priority’.
Revealed preference 4 was a particularly perverse preference in Ivor’s view, in the sense that
it amplified the difficulties generated by the first three preferences.
In terms of some obvious consequences and knock- on implications, when things didn’t
go as planned it was never anybody’s fault within Canpower – it was always external events
or organisations beyond Canpower’s control – and sometimes it wasn’t clear who was
responsible for sorting it out.
In terms of a sample of some specific recent developments, Ivor observed the following:
1 Several small output (and low hydraulic pressure) hydroelectric plants owned by the
utility for many years had been sold because they were ‘unprofitable’, but the new
owners were making money with them, in addition to achieving a high- profile ‘green’
reputation, and Canpower badly needed a greener image for a number of reasons.
2 The current medium /long- term plan was built around a proposal to build a series of six
new nuclear power stations over a 20 year period, but the business case ignored uncer-
tainties about full lifecycle costs which Ivor thought needed much more attention.
3 There was a degree of Canpower resistance to private power producers within the prov-
ince which seemed both counterproductive and politically dangerous.
4 There were political movements afoot to break up the public sector monopoly position
currently enjoyed by Canpower, with an international history of similar pressure on
other publicly owned utilities, but nothing had been done within Canpower to even
think about how to react effectively to such a development.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 407
Ivor explained to Carl that he had not indicated to anyone other than Clive what his hunch
was based on. He saw no point in challenging what he called ‘the terrible trio’ on such issues
directly at this stage. These issues were simply symptoms of an underlying culture and inter-
dependent capability problems which Ivor wanted to diagnose and cure.
Ivor wanted Carl to test Canpower’s current frameworks and models for strategy formu-
lation with respect to corporate uncertainty, risk, opportunity, and underlying complexity
management. He wanted Carl to address all relevant concerns and suggest enhancements or
alternatives which would provide immediate improvements in corporate performance and
help Canpower to address the associated culture problems. Ivor wanted Carl to proceed
with as much tact as possible but with clarity and frankness. Carl would report to Ivor in the
first instance, but subject to Ivor’s approval, Carl would then brief Clive. Subject to Clive’s
approval he would then report directly to the board.
Carl began by talking to Larry, Dick, Pat, and their staff. He reviewed a sizeable col-
lection of documents they provided to illustrate the way their groups operated. He then
prepared a presentation for the board to outline his preliminary findings and a recom-
mended way forward. He tested his presentation for the board on Ivor, section by section
as it was developed. He then reported to Clive and incorporated Clive’s feedback over a
cycle of meetings. What follows outlines the form Carl’s presentation to the board took,
over a full day at a suitably isolated and comfortable location. It does not include the
preliminaries or the comfort building or the explanations of issues discussed earlier in this
book – just a distillation of the day’s proceeding which concentrate on what you might
find interesting. It complements earlier tools and insights, building on them to provide
further new tools and insights. The discussion assumes you are interested in how boards
need to deal with these issues, whether or not you currently have or aspire to this kind
of role.
Interjections link the discussion to earlier tales and broaden the discussion to other kinds
of organisations when this seems useful, primarily towards the end of the chapter.
Corporate planning horizons and associated planning categories
Carl began his presentation by indicating that all Canpower staff should find it useful to
think in terms of four ‘nominal planning horizons’ for formal corporate planning, each
linked directly to an associated ‘category’ of corporate planning which is focused on dif-
ferent sorts of decisions using different kinds of planning tools within each of the four cat-
egories of planning. They would also find it useful to understand a fifth category of formal
corporate planning which addresses the ‘goals’ of all of the other four planning categories
in a fully integrated operational manner. He then outlined what each category involved as
discussed in this chapter’s opening section, using Figure 8.1 to provide a specific opera-
tional form.
Carl suggested that in some contexts some organisations might find fewer than five cat-
egories appropriate as special cases, and some organisations might want more categories.
Different nominal horizons might be appropriate in some cases. However, Figure 8.1 pro-
vided a suitable framework for Canpower in terms of their current discussion.
By ‘short- term planning’ in a ‘short- term planning horizon’ he meant what he believed
all Canpower staff currently meant by both these terms, and everyone involved could agree
without controversy or discussion to put an outer bound on the short- term at about a year
for most formal short- term planning purposes. The exact boundary was flexible, and differ-
ent approaches could use variations.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
408 Employing planning tools in practice
By both ‘medium- term planning’ and ‘long- term planning’ he meant new approaches to
corporate planning for Canpower which he would start to explain shortly, the first associated
with a 10 year planning horizon, the second associated with a 20 year planning horizon.
By ‘medium/long- term planning’ he meant medium and long- term planning addressed
jointly as a single category over a 20 year medium /long- term planning horizon without any
separability, Canpower’s current approach, referred to as just ‘long- term planning’ by Larry.
Carl made it clear that he appreciated Larry would not like this terminology change, but
Carl and Larry meant very different concepts when they used the term ‘long- term plan-
ning’, and it was essential for Canpower to use one agreed- on definition for one agreed
concept which was clearly understood by all relevant people. In Carl’s view Larry’s defini-
tion of ‘long- term planning’ had to be scrapped to make way for a better interpretation of
this very important term and its underlying defining component concepts, linking a new
and very different Canpower definition of ‘long- term planning’ to a separate ‘medium- term’
planning concept.
Elaborating on this observation, Carl said he appreciated this meeting was going to be
a very uncomfortable experience for Larry, and some other directors would also feel some
discomfort. However, he had reviewed his messages carefully with Clive as well as with Ivor,
and they all believed there was no point in him not being completely open about all the key
changes he believed Canpower had to make, starting with some of the more straightforward
fundamental concerns.
Both goals planning and futures planning were implicit in Larry’s corporate planning
approach, in terms Carl would explain. Initially he would use Larry’s implicit assumptions
as he understood them – stealth assumptions in EP language.
He would now develop a brief overview of Canpower’s current practice in the form of
an outline of a first pass approach to top- down planning in the Figure 8.1 framework using
current Canpower corporate planning assumptions. He would start with goals planning,
followed by futures planning, medium /long- term planning, and short- term planning. He
would then consider underlying short- term and medium /long- term forecasting, offering a
short-term planning
futures planning
medium-term planning
nominal planning horizons
now 1 year 10 years 20 years 50 years
long-term planning
context – including all other planning which significantly influences the way participants interpret formal corporate planning plus other relevant capability-culture concerns
goals planning
Figure 8.1 Formal corporate planning portrayed in terms of five different planning categories using four nominal planning horizons plus a comprehensive view of all the relevant issues.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 409
critique of the latter and a summary explanation of what it was doing and why this meant
that neither the current medium /long- term planning approach nor its underlying forecast-
ing could be trusted. He would finish this part of his presentation by briefly outlining the
implications of separating medium- term and long- term planning. This will take us to the
end of this section.
He would then start to address how to replace the key features of the current approach in
some detail – the following sections which are central to this chapter.
Goals planning
To address Canpower planning in an effective and efficient manner, they had to start with
initial goals planning, and then systematically link this initial goals planning to futures,
long- term, medium- term, and short- term planning in operational terms. To address the
underlying complexity associated with Canpower’s goals in a clarity efficient manner, they
had to begin with a clear and comprehensive understanding of the basic nature of Can-
power’s ‘goals’. These ‘goals’ might be described using words like vision, core values, aspira-
tions, mission, key aims and objectives. Underpinning their approach, they needed robust
assumptions, a sound appreciation of critical corporate strengths to be preserved, and clarity
about important weaknesses to be overcome. This category of planning was crucial, but
Carl indicated it would get limited attention initially that day, only modest attention later,
for convenience in terms of ordering their discussion. In practice, goals planning provided
the foundations for all four of the other categories, so it should always be given extensive
attention before moving on. However, starting to probe what was involved would raise sig-
nificant complexities better left until later that day, so they would begin with a very simple
minimum clarity view of what was currently being assumed.
He would interpret this minimum clarity starting point as his current inference about the
‘primary criteria’ which underlay Canpower’s current approach to corporate planning plus
several ‘secondary criteria’, recognising that Canpower had many other criteria which were
also very important to the overall set of criteria defining their goals.
He thought a non- controversial and simple initial understanding of Canpower’s primary
criteria over the planning period of interest was a minimum expected value for the standard
tariff (price) per kWh (kilowatt- hour) for electricity, based on a minimum average expected
cost per kilowatt- hour in each of the 20 years between now and the long- term horizon,
with a reasonably smooth expected tariff profile over time. Canpower’s current approach to
corporate planning clearly implied that this was the central concern because expected cost
over the medium /long- term horizon of 20 years was the only criterion used in their current linear programming medium /long- term corporate planning model.
He thought a non- controversial and simple initial understanding of Canpower’s most
important secondary criteria over the planning period of interest was an acceptable level
of associated risk of higher costs and tariffs than the expected outcomes, plus acceptable
levels of risk associated with availability and reliability. Canpower’s current corporate plan-
ning approach did not measure the risk associated with expected cost and tariffs, or address
it effectively even in qualitative terms, so their currently recommended strategy implicitly
assumed that an acceptable level of cost risk was involved, and he was assuming that Larry
and the Canpower board recognised the importance of this assumption.
This minimum clarity perspective meant that they could start their discussion today by
looking at medium /long- term planning in terms of expected cents/kWh costs associated
with all through- life costs, including depreciation of capital expenditure, the cost of capital,
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
410 Employing planning tools in practice
fuel costs, other costs like maintenance and repair, checking on associated uncertainty impli-
cations including risk. After they had developed an initial first pass understanding of futures,
medium /long- term and short- term planning given this minimum clarity perspective of ini-
tial goals planning, plus underlying forecasting concerns, they could start to consider some
of the key implications of the changes he would propose. Addressing further goals could
follow later.
Futures planning
Carl indicated that it was important to limit what he meant by short- term, medium- term,
and long- term planning to feasible options in a ‘technically feasible’ sense. However, it was also important to look beyond proven technology which was currently economically viable
to decide where investment in technology development and close scrutiny of developments
by others might pay. One way to look at this was testing current context assumptions about
feasible technology.
A useful generalisation of the testing of current technology assumptions in both technical
and economic terms was the testing of all other current corporate context assumptions for
new approaches better suiting corporate goals. He had in mind some key capability- culture
liabilities, the third key flaw in Canpower’s approach he wanted to address later in the meet-
ing, not elaborated at this stage.
He did not want to spend too much time on futures planning at this point in his pres-
entation, but he wanted to introduce the idea of a comprehensive set of formal planning
approaches addressing a mix of planning horizons ranging up to about 50 years, in all cases
looking beyond existing technology and beyond existing approaches in all other respects.
In particular, he wanted to explicitly avoid framing assumptions which locked Canpower
into an approach which was passive, in the sense that it began by predicting a future which
must be accepted, like it or not, what some people call a ‘predict and prepare’ paradigm.
He wanted Canpower to attempt ‘the design of desirable futures’ (Ackoff, 1974) – to adopt
a proactive approach to shaping the future as well as accommodating aspects beyond Can-
power’s control.
Carl emphasised the importance of comprehensive uncertainty and complexity manage-
ment processes that can be used to help organisations plan and then achieve desired futures,
as well as to help foresee and then avoid undesirable futures. Uncertainty and underlying
complexity involved very serious problems and risks, but they also involved very impor-
tant opportunities. Opportunities must be identified and then seized to exercise control
over Canpower’s destiny to the extent this was feasible, and ensuring this happened was an
important role for the board.
In terms of his overview of current practice, they could just adopt what he inferred was
Canpower’s current corporate attitude to futures planning – Canpower should keep an eye
on possible developments but focus on medium/long- term and short- term planning in
relation to feasible options they were comfortable with in current corporate culture terms.
Canpower’s current medium/long- term planning approach
Carl expressed the view that the central formal long- term corporate planning issue for any
electricity utility anywhere in the world should be ‘What mix of hydro, nuclear and thermal
power units of different kinds (gas, oil, coal) by type (large high- head hydro, small low- head
hydro, and so on) plus “other” sources (renewables for example) should be aimed for at a
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 411
nominal long- term planning horizon like Canpower’s 20 years?’ The separate companion medium- term planning issue which needed addressing was ‘How should we plan to get
there in terms of “action plans” we can make commitments to as necessary in the near- term
with associated longer- term provisional plans plus contingency plans to deal with reality
departing from predictions?’
His interpretation of Canpower’s current medium/long- term planning approach was
they did not actually ask either question, which meant they had no explicit sense of long- term direction for their variant of medium/long- term planning, and no explicit sense of commitment timing. Instead, they derived a mix of unit types at the 20 year horizon indi-
rectly, by assuming they could use a deterministic detailed planning process over the 20 year
medium/long- term planning horizon employing:
1 point estimates of the expected supply available from current existing power stations in
each year over the medium/long- term planning horizon until they were retired because
they were no longer economically competitive or other reasons, plus
2 point estimates of how long it would take to get new power stations operational from
the point when permission to build was given, plus
3 point estimates of the expected costs per kilowatt- hour for all existing and possible new
sources in each year over the medium/long- term planning horizon, plus
4 point estimates of the expected demand for electricity in each year over the medium/
long- term planning horizon, plus
5 a linear programming computer software system to minimise the expected overall cost
over the medium/long- term planning horizon, with
6 no integrated contingency planning for the medium/long-term.
This linear programming model approach optimised medium/long- term planning assum-
ing that the measure of expected cost employed was their primary criteria and uncertainty
associated with all the point estimates plus the indirectly derived outcome at the long- term
and futures horizons was not an issue. It produced a 20 year schedule of current power sta-
tion retirements and new power station first availability dates.
The current Canpower approach to medium/long- term (and short- term) planning dis-
tinguished between ‘base- load’ requirements associated with electricity demand 24 hours
per day 365 days per year and ‘peak- load’ requirements. Peak- load requirements involved
demand over and above the ‘base load’, very important distinctions in practice because
no significant energy storage was currently available. Turning power stations on and off
involved different degrees of difficulty and cost – nuclear being the most difficult and small
hydro and gas- powered stations being amongst the easiest. However, current medium/
long- term planning did so assuming predictable base- load and peak- load variations within
predictable overall load growth patterns over the 20 year horizon.
This current Canpower approach to the medium/long- term assumed forecasting elec-
tricity demand, current power station availability, and all the relevant costs involved so lit-
tle uncertainty a simple deterministic corporate planning model could be used for their
joint medium- term and long- term planning approach. In Carl’s view this was an unusual
and wholly unacceptable approach which needed replacing immediately. He was aware of
no other electricity utility explicitly adopting such an inappropriately simplistic ‘optimised’
approach.
That said, it was crucial to understand that in practice any utility which used a largely deterministic rule- based or heuristic approach to medium/long- term planning which focused
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
412 Employing planning tools in practice
on point value estimates without explicitly addressing associated uncertainty in effective for- mal terms would have directly comparable performance problems. This was because they were actually just doing the same thing without trying to optimise in a formal manner. Many
other utilities used largely deterministic approaches which were comparable in terms of per-
formance problems, so Canpower’s problems were more visible and demonstrable because
they used linear programming, but they were by no means unique. It was important the
board understood that the fundamental underlying flaw in the Canpower approach was not
the use of linear programming per se – it was the deterministic approach based on point
estimates which failed to address relevant uncertainty in an effective and efficient formal
planning framework.
Relevant uncertainty had to be addressed in terms of two very different concerns. One
was the need for a portfolio theory approach to define their direction of travel for long-
term planning purposes. The second was their need for a separate medium- term plan-
ning approach to address commitment timing decisions including associated contingency
planning.
Linear programming was just a way to optimise given seriously flawed underlying assump-
tions about the relevance of uncertainty. The use of linear programming might have been
a ‘good thing’ because it indicated a desire to achieve a ‘best solution’, if a deterministic
approach based on point estimates had been appropriate. However, in the Canpower con-
text the use of linear programming was an unequivocal symptom of a ‘very bad thing’
because it implied (and highlighted) a failure to grasp the implications of significant uncer-
tainty which rendered any deterministic approach which did not address uncertainty effec- tively inappropriate. An appropriate set of effective and efficient formal planning approaches
for dealing with all relevant uncertainty and underlying complexity was essential, simplifying
in the right way.
Canpower’s current short- term planning approach
Canpower’s current short- term planning was conventional in approach and well executed,
needing no further attention from Carl other than modest clarification of what it involved
when explaining forecasting issues for those directors who were not familiar with the details
of what was involved.
Dick was obviously pleased to hear this, as was everybody else apart from Larry, who
could see the potential for one of his usual allies deserting him.
Short- term forecasting of costs, power station availability and electricity demand
Short- term planning for Canpower depended on four very different kinds of short- term
forecasting, approached in four very different ways.
One kind of forecasting addressed power station generation availability, interpreted by
Dick in traditional availability and reliability terms. Models used for forecasting short- term
availability and reliability of individual power station units were well developed and estab-
lished, drawing on conventional state- of-the- art approaches. The short- term action plans
Dick used were deterministic and assumed no power station failures, but contingency plans
were in place, and the spinning reserve and standby reserve which Dick’s plans included
should provide the cushions Dick needed to implement these contingency plans when
needed in most circumstances.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 413
The second kind of forecasting involved estimating the costs associated with electricity
generation to make short- term planning decisions about which power stations to operate.
The dominant issue was fuel costs. In the short- term fuel cost estimates were based on
known contracted prices or reasonably predictable forecasts just days or weeks ahead. Start-
up and shutdown costs were also relevant to managing peak demand satisfaction, but they
were reasonably straightforward to estimate using well- established accounting practices.
The third kind of forecasting addressed estimating the electricity tariffs needed for fore-
casting demand for electricity plus all other key explanatory variables used to drive their
electricity demand forecasts. All demand forecasting had to accommodate the effect of the
cents/kWh tariffs on demand, incorporating the underlying effect of all relevant Canpower
costs. However, in a short- term planning context tariffs could be treated as known or rea-
sonably predictable. Other key explanatory variables (like temperature) could also be fore-
cast a short time ahead.
The fourth kind of forecasting considered the demand for electric power on a time- of-
day, day- of- week, and week- of-year basis. This captured daily, weekly, and monthly (sea-
sonal) cycles. Models used for short- term demand forecasting typically included exponential
smoothing (adaptive expectation) models which track a variable, predicting future values
based upon past observed values and their related forecasts; time- series models, which iden-
tify systematic variations relating to trends and cycles; and econometric regression models,
which also use external or exogenous explanatory variables identifying causal relationships,
like the effect on the demand for electric power of temperature, the current level of economic
activity, the current price of electricity, the current price of substitute energy, and so on. All
these approaches were well- developed and established within Canpower, drawing on a widely
used and well- established literature. The most effective short- term demand forecasting mod-
els were typically hybrid regression models incorporating all useful features of the preceding
models. Effective use required intelligent manual adjustments when factors not in the model
were recognised, such as industrial action against major industrial consumers. Such models
provided expected value point estimates and associated fit (forecasting) error probability dis-
tributions which could be used to define confidence bands. For short- term use, such mod-
els were well understood and reasonably straightforward to use effectively. Canpower made
effective and efficient use of state- of-the- art approaches to these models.
Short- term planning built on these four different kinds of forecasting to match demand
and supply at minimal expected operating cost – planning decisions involving minute- by-
minute implementation, in terms of which electric power generation units to run within
each station and how much spinning reserve was appropriate to cope with unexpected
surges in demand and unplanned outages, plus planning decisions with slightly longer- term
concerns about next week or next month in terms of planned maintenance and putting units
on or off standby status. The basis was a ‘merit ordering’ of all power stations available by
operating cost per kilowatt- hour, selecting the cheapest first and then working up towards
the most expensive to meet expected demand with spinning reserve and standby reserve to
cope with unexpected surges in demand or plant failures.
In Carl’s view Canpower’s approach to short- term forecasting and associated planning
was exemplary with just one key proviso – short- term power station outages which proved
to be longer- term failures relevant to medium/long- term planning were not addressed with
the degree of formality he would argue for later. A more general view of ‘availability’ was
needed to accommodate this concern because it was not currently addressed as part of the
medium/long- term planning agenda either. It involved what might seem to be a short- term
incident initially which turned into a medium- term or long- term concern.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
414 Employing planning tools in practice
To illustrate what he was getting at, a worst- case scenario was a nuclear power station fail-
ure which investigation showed was caused by a fundamental design problem which would
require shutting down all their nuclear power stations with comparable designs permanently.
Larry knew this possibility was not addressed by his approach because he had in effect
deemed the probability of it happening ‘too small to worry about’ for formal planning pur-
poses. He could see Carl saw it as his problem, not Dick’s, and he correctly suspected Carl
was going to return to this issue later, but neither said anything more at this stage.
Medium/long- term forecasting of costs, availability and electricity demand
Medium/long- term planning for Canpower depended on the same four kinds of forecasting
as short- term planning, but each of these four kinds of forecasting had to be approached
in very different ways relative to their short- term equivalents, with a clear understanding of
some crucial differences.
One kind of forecasting addressed power station availability on a year- by-year basis given
current supply capability, provisional medium/long- term plans to release current supply
sources by taking them out of service, and a provisional view of the implementation of the
medium/long- term plans to make new power stations available. This meant the reliability of
new power stations and their technology was of concern as well as existing power stations,
and the reliability of the estimated time between deciding to build a new station and having
it operational was also an issue. Carl indicated these issues were all very important, but they
were not at the top of his agenda for the moment.
The second kind of forecasting considered cost estimates to make medium/long- term
planning decisions. An obvious central concern was fuel costs – easy to predict a week ahead
but sometimes getting difficult to predict a year ahead, becoming virtually impossible 5, 10,
or 20 years ahead. A less obvious but also central concern was the full capital cost of all new
power stations, expressed in an amortised form so that the capital costs for proposed new
power stations could be combined with operating costs to derive the estimated full lifecycle
overall costs estimates per kilowatt- hours delivered to the grid for new power stations to
make capital investment choices. These costs depended upon usage assumptions. Assuming
a new power station would run for 90% of the time it was available for medium/long- term
planning purposes, but then only using it for 45% of the time for any combination of rea-
sons, doubled the amortised capital cost per kilowatt- hour, a crucial source of uncertainty
and risk currently ignored.
The third kind of forecasting addressed the electricity tariffs and other key explanatory
variables used to drive their electricity demand forecasts. All demand forecasting had to
accommodate the effect of the cents/kWh tariffs on demand, and the effect of all relevant
Canpower costs on the cents/kWh tariffs. Larry’s approach involved two key unrealistic
assumptions with effects which became visible at this point. First, the expected capital cost
of new nuclear power stations was optimistically biased to the extent that it needed uplifting
by a significant amount in Carl’s view, for reasons he would explain shortly. Second, all fossil
fuel costs were assumed to rise at a steady rate of climb from current levels, with gas main-
taining its current cost advantage relative to oil and coal. These two assumptions confirmed
Larry’s current revealed preference for building new nuclear power stations for base-load
as soon as current base- load power stations reached the end of their planned life or became
relatively uneconomic, plus Larry’s current revealed preference to build large new gas- fired
power stations as soon as current peak- load stations reached the end of their planned life or
became relatively uneconomic. Without looking any further, the board could assume these
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 415
revealed preferences meant that Larry’s linear programming framework would drive Can-
power to preserve current big hydropower stations, but otherwise move Canpower towards
an all nuclear base load plus all gas- fired peak- load position. They could get exactly the same result using a simple variant of Dick’s ‘merit- ordering’ approach, achieving this position by
the 50 year futures horizon if all aspects of the plan held.
Carl assumed Larry was probably aware of this, but most other members of the board
probably were not. He did not want to linger on or labour this issue at present, but he
wanted to start to trigger a crystal- clear understanding by all board members by the
end of this meeting that the sophisticated linear programming basis for Canpower’s cor-
porate planning was in practice just a simplistic device for delivering Larry’s revealed
preferences – his ‘desirable future’ in Ackoff ’s terms. Larry’s ‘desirable future’ was an
existing large hydro plus nuclear base- load future, plus gas- powered peak- load stations.
The use of linear programming simply produced a variant of Dick’s merit- order approach,
without Dick’s associated recognition of uncertainty and the need for effective contin-
gency planning.
The fourth kind of forecasting addressed the demand for electric power on a year- by-year
basis 1, 2, 3, . . . , 20 years ahead. Forecasting demand for electric power for the short term as
undertaken by Dick was exemplary in Carl’s view, as observed earlier, and adaptations of the
same forecasting model approaches could serve as a starting point for medium/long- term
forecasting. Hybrid regression models incorporating smoothing effects, causal factors like
population growth and economic variables like price tend to fit annual past data reasonably
well, although not as well as short- term forecasting models, and this kind of long- term fore-
casting models were generally well understood by electricity utilities and other experienced
users. Relative to their short- term forecasting use, such models required reformulation for
the peaks associated with different time period structures, and re- estimating, with significant
changes in some of the individual variables involved. For example, seasonal temperature
variations would cease to be relevant, but population growth might become a key explana-
tory variable, and economic variables like price would become crucial. Such changes might
seem to pose no immediately obvious difficulties in terms of the econometrics involved in
estimating the models used. Given stable relationships and appropriate data, these top- down
forecasting models can work quite well over the data period used to estimate the models in
the sense that they fit the past data set. However, they cannot cope with a future involving
significant new issues or structural changes. Furthermore, the error terms associated with
their ‘fit’ to past data ignores errors associated with predicting explanatory ‘exogenous’
variables used to drive the forecasts of interest beyond the data period into the future – fuel
prices, capital costs in amortised form, population growth, underlying industrial demand,
technical changes, the load levels new power stations achieve, and so on. A failure to deal
with the implications was seriously negligent.
A very important complement to these top- down demand forecasting models, widely
used by Canpower and other electric utilities, was bottom- up or ‘end- use’ models, which
looked at how much electricity refrigerators or air conditioning or other consumer and
industrial uses involve, together with trends and cycles in associated use levels and efficiency
levels. These bottom- up models provided significant insight into the nature of some changes
in the structure of the market and the instruments available for ‘active demand manage-
ment’ (encouraging consumers to use less electricity). However, they did little to help fore-
cast when structural changes might take place.
A residual significant difficulty posed by both top- down and bottom- up demand forecast-
ing models used in combination for medium/long- term planning purposes was predicting
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
416 Employing planning tools in practice
the values of all the explanatory economic variables used to drive the forecasting process and
underlying structural changes not addressed by the models.
Carl emphasised that it was crucial everybody understood now, with complete clar-
ity, several reasons why medium/long- term forecasting was not the same as short- term
forecasting.
First, matching supply and demand at the heart of medium- term planning had to focus on
peak demand within months or years over the medium/long- term planning horizon rather
than hourly or daily demand in the immediate future. Larry’s forecasting team worked with
Dick’s forecasting team, and both teams understood these differences and adapted their
approaches accordingly. This issue was not a problem, unlike all the others.
Second, forecasting up to one year ahead using the confidence bounds associated with
goodness of fit over the data period assuming known values of all the exogenous explana-
tory variable values was optimistic but with a modest bias adjustment could be reasonable.
However, forecasts 2, 3, 4 or 5 years ahead were problematic, and forecasting 10 or 20 years
ahead was just wishful thinking because of structural shifts not addressed by the models
even if correct values of all the explanatory variables were known. Canpower’s projected
confidence bounds 5, 10 or 20 years hence were just fantasy because of underlying structural
changes not modelled top- down econometrically or via bottom- up models even if explana- tory variables could be correctly predicted – which they could not. For example, a social change towards greener consumption patterns might have profound effects on electricity demand,
but predicting the nature of all relevant shifts and whether or not such shifts would happen
was not feasible.
Third, while the cents/kWh tariffs could be assumed to be known in the short-term,
tariffs needed forecasting in the medium/long- term, along with the direct effect of changes
in tariffs on demand. This meant amortised capital costs and operating costs including fuel
needed medium/long- term forecasting, along with the costs of buying power instead of
generating it. These forecasts had to be translated into standard tariff levels which would
drive demand, a further source of electricity demand unpredictability associated with a key
explanatory variable. Canpower’s current medium/long- term planning models assumed all
new base- load power stations would be nuclear because extremely optimistically biased esti-
mates of the cost of nuclear power stations had been used, driving up the assumed demand
for electricity in a very optimistically biased manner which became increasingly unrealistic
as time went on. The assumed dominance of gas- fired peak power at higher but still modest
prices reinforced this growing bias as forecasts approached the 20-year planning horizon.
Fourth, further explanatory variables of importance also need forecasting, like population
growth. Population growth had been rapid in the near past, and Canpower’s medium/long-
term forecasts assumed this would continue, probably too optimistically, further biasing the
forecast for electricity demand on the high side.
The net result was a corporate planning model seriously overestimating rapid growth in
demand which Canpower was planning to meet at the very low tariffs helping to generate this
demand. The reality was Canpower’s actual costs would be very much higher, and the tariffs
needed to cover these higher costs would lower demand relative to forecasts, causing second
order cost increases and third order demand decreases which continue to have further higher
order effects. Canpower would build a lot of new nuclear power stations to meet base load,
plus new, big gas- fired stations to meet peak loads, and then find that the actual costs of both
were much higher than anticipated and demand was much lower than anticipated.
In Carl’s view they would be setting out on this disastrous journey with no properly
grounded sense of direction, no understanding of the serious risks associated with a planned
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 417
future dominated by nuclear power stations for base- load and big gas- fired stations for peak
load, and no effective contingency planning in place when their ‘plan A’ started to unravel.
The linear programming approach Larry was using was capable of going well beyond the
simple model currently used by Canpower in terms of additional detail, following estab-
lished practice by oil majors running refineries and other sophisticated deterministic plan-
ning users. For example, they could use the linear programming model to decide when to
take on potentially useful non- utility generation (electric power purchased from outside the
province or private sources within the province), when to negotiate the sale of power to
other provinces or utilities in adjacent US states, and when to start and expect results from
new conservation programmes (selling low energy consumption light bulbs for example).
But the focus would still be all relevant costs in a deterministic framework, with no feedback
to corrective action when forecasts did not hold, and nothing to stop the current momen-
tum towards a seriously disastrous situation until it became blindingly obvious to some
more insightful observers that they were heading for serious trouble.
Carl’s overview of what separating medium- term and long- term planning involved
Carl repeated the view expressed earlier that the central issue for formal long- term corporate
planning for any electricity utility anywhere in the world should be the addressing the ques-
tion, ‘What mix of hydro, nuclear, thermal of different kinds (gas, oil, coal) and “other”
units by type should be aimed for at a nominal long- term planning horizon like Canpower’s
20 years?’ This question needed addressing explicitly, with considerable care and attention,
deferring until later the companion closely coupled medium- term planning issues. Only
after Canpower was clear about their intended long- term planning direction should they
consider ‘how should we plan to get there’. This would involve ‘action plans’ requiring
commitments, with associated provisional plans to deal with any aspects of reality departing from base plans and parameter predictions like demand for electricity.
To keep the long- term planning approach as simple as possible, the mix of unit types tar-
geted at the horizon could be addressed in proportionate terms first. This initial proportion-
ate approach could be used for an initial assessment of the average cost per kilowatt- hour at
the 20 year horizon. This could be used to predict tariffs. These predicted tariffs could, in
turn, be used to predict the demand for electricity. At this point they could start to move
beyond a mix to aim for in proportionate terms, addressing issues like the need to retire
some specific current power stations but retain others. Specific choices would require revised
tariff estimates. However, they should be able to converge to a planned set of specific new
power stations to meet demand the forecast 20 years ahead with minimal iteration.
Looking at the long- term planning issue in proportionate terms first, Carl suggested it
was important to acknowledge that many organisations in a wide variety of industries had
done a great deal of pioneering work on the use of what he called ‘robustness- driven sce-
nario’ approaches to long- term planning to cope with significant uncertainty. These scenario
approaches involved ‘qualitative’ or ‘semi- quantitative’ scenario approaches to planning
which used forecasts but focused on planning a robust approach which can cope with any plausible future. Such approaches were particularly relevant for organisations with a strat-
egy largely driven by uncertainty about one parameter, like the cost of oil. Marsh (1991)
provided a current synopsis of Shell International practice, Shell being an early pioneer of
this kind of scenario approach. This kind of scenario approach abandons the notion of point
estimates with an associated error or confidence band. The focus is two scenarios which
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
418 Employing planning tools in practice
represent in a consistent way the credible extremes of an infinite number of possible futures,
plus a middle of the range possibility to formulate base plans for some form of ‘central fore-
cast’ future.
This kind of robustness- driven scenario approach and its concern for robust planning
were a significant positive step which Canpower had yet to exploit, unlike some other elec-
tricity utilities. However, such approaches did not on their own resolve the difficulties facing
electricity utilities in a ‘risk efficient’ manner. He believed Canpower needed a ‘portfolio-
driven’ framework with a focus on both robustness and risk efficiency via explicit pragmatic use of robustness- driven scenario approaches plus portfolio theory ideas, effectively linked to an understanding of all relevant systemic uncertainty as well as ambiguity. He would
explain what this involved shortly, but the basis of his generalisation of the Marsh/Shell
scenario approach was recognising that the key driver was not the uncertain price of one key
factor like oil, but the uncertain cost per kilowatt- hour of electricity produced by a port-
folio of possible choices. This portfolio included hydro, nuclear, gas, oil, and coal, buying
power from non- utility sources (any provider outside Canpower); conservation (persuading
customers to use less); and any other options thought relevant. Canpower should develop
a formal planning approach to long- term planning which would produce a target mix of sources to aim for in portfolio terms with requisite robustness. They might usefully see the
needed approach as directly comparable to a financial investor picking a risk efficient port-
folio of stocks and shares which would maximise return for an acceptable level of clearly
understood risk.
It was also very important to acknowledge that many electricity utilities had done a great
deal of pioneering work on the use of informal portfolio planning approaches to seeking
a reasonable approximation to both risk efficiency and robustness. For example, before its
recent privatisation the UK’s CEGB had a formal policy of seeking efficient diversification
of power sources, including a very clear formal policy to avoid only one approach to nuclear
power station technology, pursued, as far as he knew, by informal planning means.
Looking at the separate medium- term planning issue, once Canpower had a clear idea
of their approximate target mix of power stations 20 years hence, they could look at the
difference between that target and where they would be if they built no new plant, use the
gap to draw up priority lists of new power stations, and then start building commitments to
begin construction and further commitments to get ready to begin construction as well as
other provisional plans up to 20 years ahead, using a very different but compatible formal
planning approach.
With nuclear power station construction high on Canpower’s current agenda, and a mini-
mum construction period for nuclear power stations of about ten years, this defined the
ten year potential commitment horizon in terms of the maximum lead time they needed to
consider for medium- term planning commitments. Other organisations would, of course,
define their horizons on a basis appropriate to their context, and might use different hori-
zons for different kinds of planning. Within this ten year horizon, medium- term planning
decisions might be based on estimated dates to the nearest month. More tentative provi-
sional plans might be to the nearest year for years 11 through 20, to profile plans to move
towards the target mix of generation types by year 20 for the demand expected by the time
they reached the long- term horizon.
Contingency plans could be developed to look at medium- term departures from expecta-
tions, with costs linked to prior option choices when relevant, as they often were. Every year
Canpower would have to review their long- term plans and underlying forecasts, updating
their target mix at the 20 year horizon and medium- term plans.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 419
An important feature of Canpower’s current corporate culture was the mantra ‘we must
keep the lights on’, and meeting this objective was clearly very important. But simply relying
on significant surplus capacity, which seemed to be the current corporate strategy to achieve
this goal, was both an expensive approach and a potentially unreliable approach.
When closing his introductory overview discussion of corporate planning horizons and
associated formal approaches to planning categories as a whole, Carl emphasised that Can-
power’s current approach to short- term planning was very good, and not in need of any
changes. However, they had to confront their inability to forecast with any certainty beyond
the short-term, and the quite profound difficulties associated with long- term forecasting of
cost per kilowatt- hour for different sources of electric power. He suggested long- term plan-
ning had to focus on uncertainty relevant to targeting the mix of supply sources they needed
in 20 years given what they knew was feasible now, and medium- term planning needed to
focus on uncertainty relevant to timing movements towards the 20 year horizon targets,
emphasising commitments within the first 10 years of the medium- term planning horizon.
Futures planning needed to start on a sound basis provided by goals planning, addressing
aims and objectives over the whole of the 50 year planning period of interest.
When he had made sure everyone understood this overview in broad terms, Carl moved
on to more detailed consideration of what was involved, starting with long- term planning.
Long- term planning
Long- term planning in the Figure 8.1 framework started by asking the question, ‘Given
what was currently known, where did Canpower want to aim to be in 20 years, in terms of
an appropriate mix of hydro, nuclear, thermal, and “other” units by type, assuming the use
of feasible technology options and other robust working assumptions?’ Canpower should be
aiming for a target mix of sources of electricity at a planning horizon of 20 years, recognising
that within 5 years the target would probably have moved, and annual reviews with adjust-
ments whenever needed would be necessary. For example, ‘working to the nearest 10%,
should the current base- load contribution of nuclear power of 50% be increased to 70%, as
currently implied by Larry’s plans, or should it be decreased to 40% (which he was inclined
to believe might make more sense although he did not say so at this point), and what did
this imply for other target proportions?’
When Canpower had the approximate target proportions assessed, they could forecast
their cents/kWh tariffs at the 20 year target horizon, and use a 20 year electric power
demand forecast to convert target proportions into target gigawatt (GW) capacities.
Each year both long- term and medium- term plans would have to be reviewed in the
light of emerging information about economic, political, and social issues and in the light
of futures planning. Significant changes in long- term plans might occur, and some knock-
on changes in medium- term plans might become essential. However, medium- term plans
involving action plan commitments ought to be fairly stable, and this ought to be an explicit
objective of the overall approach. For example, it should now be very clear to all Canpower
directors that it was crucially important to never delay new nuclear power station projects already under construction because demand had fallen, but how much it was worth spend-
ing to avoid doing so, and how best to address both proactive and reactive contingency
plans to implement this insight, had not yet been assessed. This was a deliberately unsubtle
reference by Carl to a current nuclear power station project which had recently revealed
that factor of ten cost increases were now anticipated, largely attributable to such delays by
Pat and Larry. Carl knew that other important factors were also in play, but Pat and Larry
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
420 Employing planning tools in practice
blamed the whole of the increases on changes in safety regulations which occurred during a
deliberate pause in construction following a downturn in the demand for electricity. In Pat
and Larry’s view, neither the demand downturn nor the change in safety regulations could
have been anticipated. The consequences included a need to redesign core aspects of the
power station halfway through construction, with obvious serious knock- on implications.
Ivor and Carl saw the safety regulation changes as a development which should have been
anticipated, along with the downturn in the forecast demand, and the resulting debacle as
the result of a particularly glaring example of incompetent medium- term planning in con-
junction with incompetent long- term planning. Carl wanted to indicate to all the directors
what they were addressing without dwelling on it at this stage. Pat and Larry flushed when
most of the other directors looked at them for a reaction, but they remained silent.
Keeping their attention to the long- term for the moment, Carl suggested that a number
of ingredients were needed before they could evaluate where they currently thought Can-
power should aim to be in 20 years. These ingredients were a full set of feasible options,
unbiased full lifecycle cost estimates for each option addressing all relevant uncertainty, and
unbiased estimates of the key interdependencies between all relevant uncertainties in an
appropriate form. He would deal with these ingredients one step at a time.
A full set of feasible options
A full set of possible new power stations was an obvious place to start. The currently pro-
posed series of six new nuclear power stations to begin construction over the next 20 years
were obvious examples. In addition to hydro, coal, oil, and gas alternatives, proposed ‘new
supply’ which provided equivalent power without construction cost investment by Can-
power also needed attention – like buying power from other utilities. Furthermore, renew-
able sources like wind power and solar power should be included as possible options in
their best currently feasible technology forms, even if their inclusion was believed to be
inappropriate at this stage of their development because of current costs. Canpower needed
to be able to explain why they were not going down any potentially useful technically fea-
sible routes if that was the choice, and they needed to understand the exact nature of the
gap between exclusion and inclusion in the portfolio of priority new supply which would
emerge. Put slightly differently, the feasible/non- feasible boundary in economic terms was
uncertain. This uncertain boundary needed explicit formal recognition in terms of including
all potentially feasible options and measuring cost gaps associated with exclusions for cost-
based reasons. The size of these gaps was important input to futures planning, and a good
example of one way interdependencies between the five categories of planning portrayed by
Figure 8.1 could be approached.
To compare different types of power stations and their equivalents, and to build trial
portfolios of options, Canpower would find it useful to work in terms of 1 GW ‘new units’
of additional nuclear power and 1 GW ‘new units’ of all other ‘new supply’ types, initially
addressing base- load and peak- load portfolios separately.
For those not familiar with electricity power station planning, planning assumptions used
at that time meant that 1 GW ‘planning units’ would provide power for about a million homes, with a significant proportion of total demand servicing industry and commerce
(non- domestic demand). Carl chose 1 GW ‘new units’ in a ‘planning unit’ sense to facili-
tate a simple discrete value illustrative HAT (histogram and tree) approach which avoided
excessively detailed planning without excluding low- output possibilities dealt with on a col-
lective basis. Large nuclear power stations normally use multiple ‘physical generation units’
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 421
in addition to involving multiple new units in our planning- unit sense. For illustrative sim-
plicity this tale assumes that Canpower’s current plans for six new nuclear power stations
involved 1 GW physical and planning units, with three of these units per power station, 6 × 3 = 18 new units of nuclear power in total. Conventional thermal power stations might
be significantly less than 1 GW, and very low output sources dealt with on a collective basis
might involve several dozen physical generation units to provide a 1 GW new unit of power
in these planning terms.
Carl suggested they defer consideration of new hydro, gas, oil, coal, and a range of other
further technically feasible options that day, and focus on simple illustrative numerical exam-
ples for just three base- load portfolio possibilities to explore simple portfolio models:
1 a 1 GW new unit of nuclear power,
2 a 1 GW new unit of ‘conser vation’ (using ‘active demand management’ to reduce
demand), and
3 a 1 GW new unit of ‘non- utility power’ (buying electric power from other organisa-
tions, including the currently very limited number of privately owned electricity utilities
within the province plus electricity utilities in other adjacent Canadian provinces or US
states).
Nuclear needed consideration that day because it was Canpower’s currently preferred
option for their base- load portfolio and he wanted to illustrate key asymmetric distribution
concerns.
Canpower was under significant political pressure to consider conservation. Simple forms
of active demand management which they might adopt included the ‘supported sale’ of
low energy consumption light bulbs, implying ‘subsidised sale’ in the sense that full retail
marketing levels of profit would certainly not be required, and a further subsidy might be
considered. But a rich set of further options could be considered. Carl knew that Larry was
considering some options, but a much more proactive stance was an immediate requirement
in his view, one reason Carl had included conservation in his illustrative set of three options. Canpower also needed to give very close consideration to non- utility power options in
Carl’s view. He believed an increase in the attention paid to privately owned utilities within
Canpower’s province was essential, as was a much more collaborative approach to utilities in
other adjacent provinces and US states, one reason Carl included ‘non- utility power’ in his illustrative set of three options.
A further key reason Carl included both non- utility power and conservation examples was
the plausibility of simple symmetric probability distribution examples plus associated simple
dependence issues facilitating a usefully straightforward initial discussion of portfolio model-
ling to address risk efficiency concerns, as you will see shortly.
Unbiased cost estimates for each option as a starting point
Carl began by indicating they needed to use a full lifecycle approach for each option of inter-
est, spreading the capital cost over the asset lifecycle on a ‘levelised’ cents/kWh basis. This
would involve annualised average cost using a variant of the discounted net present value
approach outlined in Chapter 7 to amortise the capital cost component.
He would focus on assessing ‘direct’ costs initially. For the moment, he would explicitly
exclude ‘indirect’ costs like reserve backup units in case of unit failures, electricity transmis-
sion lines, overhead management costs, and further potentially complex and contentious
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
422 Employing planning tools in practice
items, deferring consideration of these concerns until later. He would use 1992 Canadian
cents/kWh ‘money of the day’, prior to adjusting for general inflation and further ‘escala-
tion’ (differential inflation) of items like non- renewable fuel costs. They could later consider
adding provisions for all relevant indirect costs and all relevant inflation and escalation,
including economic, political, and social effects on markets and prices. Initially they could
assume specific sites were not being considered. Ignoring site- specific issues introduced
some uncertainty requiring a provision, and they would have to make the initial estimate
unbiased given what was known about the sources of ambiguity in these provisions.
In the context of nuclear units, the resulting initial cost estimates would include amor-
tised capital costs, fuel, all further operating costs, decommissioning costs, and any federal
government levies or incentives specific to nuclear power. In the context of non- utility gen-
eration, this would involve market costs reflecting electricity supply and demand balance as
well as underlying fuel supply and demand balance plus amortised capital costs, other oper-
ating costs, and government subsidies or levies for the supplier. In the context of conserva-
tion, the cost structure was less conventional, but consistency would be important.
One key aspect of these estimates was an unbiased view of the expected cost. A directly
related concern was an unbiased view of the potential for variability associated with this
expected cost. A further directly related concern was an unbiased view of the interdepend-
ence of variability when individual units were combined into a portfolio of units. Pursuit of
risk efficiency, a core uncertainty and complexity management concept from an EP perspec-
tive, required clarity about all three of these component concerns and unbiased estimates
of all three components. Pursuit of clarity efficiency in conjunction with risk efficiency plus
related aspects of opportunity efficiency required a reasonably simple practical approach to
addressing the complexity underlying uncertainty in this context, and this demanded a care-
fully judged approach to keeping it simple systematically.
Carl said he assumed most members of the board were convinced that they had a very
good grasp of expected values and their relationships with associated probability distribution
spread and shape, including the role of interdependence when portfolios were involved, but
he wanted to explore some very simple illustrative numerical examples with them to ensure
they all had a shared understanding of the key concerns in the Canpower context. He would
use a HAT approach which was designed to help them all to understand both simple and
complex forms of statistical dependence in a way which would facilitate agreeing to a col-
lective board level Canpower understanding of key issues which could be communicated
effectively to others.
Constructively simple models of asymmetric nuclear power cost distributions
They would focus their initial attention that day on what he would refer to as the ‘example
A’ estimate of Table 8.1 and the related Figure 8.2, which he showed them.
Carl began by suggesting that Larry might see some aspects of this example A estimate as
pessimistic, but he believed all aspects were actually very optimistic and roughly consistent with the numbers currently used by Larry. However, his focus for that day was construc-
tively simple numerical examples to illustrate key concepts, not reliable estimates. They
should defer the issue of reliable estimates until a later date.
They would use the Table 8.1 and Figure 8.2 example to consider the levelised full lifecycle
cost of delivered power at the 20 year long- term horizon in 1992 Canadian cents/kWh for
1 GW of nuclear power. They should see this example as an illustration employing just three
scenarios for simplicity of interpretation using a HAT approach which he would explain.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Table 8.1 Example A cost distribution for one GW of nuclear power, tabular form.
Scenario value in 1992 cents/kWh (range)
Scenario probability (probability density per unit over the range)
Product Cumulative probability
4 (3– 5) 10 (5– 15) 25 (15– 35)
0.5 (0.25) 0.3 (0.03) 0.2 (0.01)
2 3 5
0.5 0.8 1.0
Expected cost (‘product’ column sum) 10 Most likely (most probable or modal) cost 4 cents/kWh, and median (P50 or fifty percentile)
cost 5 cents/kWh
1.0
outcome cost in cents/kWh
0
0 3 5 10 15 20 25 30 35 40
0.2
0.4
0.6
median value (50 percentile or P50 value)cumulative
probability
0.8
0.25
outcome cost in cents/kWh
probability
0 3 5 10 15 20 25 30 35 40
0.03
expected value
(point of balance)
most likely value
(most probable value)
cumulative format diagram (piecewise linear trapezoid)
density format diagram (rectangular histogram)
Figure 8.2 Density and cumulative probability format diagrams for the Table 8.1 probability distribution.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
424 Employing planning tools in practice
To relate the Table 8.1 approach to earlier HAT based approaches you might like to see
it as an example of enlightened simplicity which used just three scenarios to capture the
important implications of asymmetry with the minimum level of complexity needed for the
initial messages which Carl wanted to communicate in this particular presentation context,
a form of ‘enlightened’ simplicity in a context when reverting to Stephen Ward’s initial
‘constructive’ simplicity terminology (Ward, 1989) is arguably useful to illustrate the basis
or starting point.
The ‘4 cents scenario’ was the ‘most likely’ scenario, associated with a probability of 0.5
in Table 8.1, bigger than the probability associated with any other scenario. The first col-
umn of Table 8.1 associated the central scenario value of 4 cents/kWh with the range 3 to
5 cents/kWh, putting the range in brackets. The second column of Table 8.1 indicated that
a probability of 0.5/(5 – 3) = 0.25 was associated with both 3 to 4 and 4 to 5, putting the
0.25 in brackets. The density form of Figure 8.2 showed all values in the 3 to 5 range were
assumed to be equally likely at the 0.25 probability level per x- scale unit (cents/kWh).
This most likely scenario was also the ‘minimum plausible’ scenario, a plausible lower
bound, not a central value. Although Carl did not overemphasise the full implications of
this potentially counter- intuitive dual nature, some emphasis was evident and deliberate
because a most likely estimate of 4 cents/kWh was often cited in his conversations with
Canpower staff. The point he was making was some people were prone to confusing most
likely estimates with other ‘central value’ estimates, including ‘best estimates’ or ‘expected
values’. This kind of confusion was never a good idea, and it was potentially very seriously when significantly asymmetric distributions were involved, as in this case.
The ‘25 cents scenario’ was an illustrative ‘maximum plausible’ scenario, associated with
a probability of 0.2 in Table 8.1. Table 8.1 associated this scenario with the range 15 to 35
cents/kWh, and the density form of Figure 8.2 showed all values in this range were assumed
to be equally likely with a 0.2/(35 – 15) = 0.01 probability per x scale unit (cents/kWh), as
indicated in the second column of Table 8.1 and in Figure 8.2.
The ‘10 cents scenario’ should be interpreted as an illustrative ‘mid- range’ scenario, asso-
ciated with a probability of 0.3. Table 8.1 associated this scenario with the range 5 to 15
cents/kWh, and the density form of Figure 8.2 showed all values in this range were assumed
to be equally likely with a 0.3/(15 – 5) = 0.03 probability per x scale unit, as indicated in
Table 8.1 and in Figure 8.2.
Both Table 8.1 and Figure 8.2 also showed that 10 cents/kWh was the expected out-
come, 150% more than the most likely estimate of 4 cents/kWh.
Carl had portrayed what he saw as a deliberately crudely structured but defensible initial
view of the relevant uncertainty at this stage. He had chosen the three scenario structure
and his example’s numeric values carefully to illustrate, in simple but dramatic terms, the
importance of the long right- hand tail associated with the density format. The 150% uplift
from the most likely value of 4 to the to the expected value of 10 was a clear preliminary
demonstration of the implications of this level of asymmetry, which some people might pre-
fer to read as a 250% ‘increase’ (100% × 10/4 = 250% instead of 100% × (10 – 4)/4 = 150%)
‘more than’.
The cumulative form in Figure 8.2 followed directly from the rectangular histogram den-
sity form. The cumulative form showed the P50 or median value of 5 cents/kWh, which in
this particular case could also be observed directly in Table 8.1.
More scenario values could have been used to provide greater precision with associated
smoother curves in both the Figure 8.2 formats. This could have been based on probability
values in Table 8.1 estimated to the nearest 0.01 instead of the nearest 0.1, using 30 or more
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 425
scenarios instead of three if a high level of precision was worthwhile. But a limited level of
precision and a simple structure were useful places to start to maximise clarity by keeping
it as simple as possible, adding precision when this became useful later. Carl emphasised
the role of Table 8.1 and Figure 8.2 was illustrative, different probabilities could have been
argued for, and he did not wish to defend the values used. However, both the 4 cents/
kWh most likely value and the approximate factor of ten range between 3 and 35 cents/
kWh was consistent with the shape and spread of Canpower’s current planning values as he
interpreted them.
The low most likely (modal or most probable) value of 4 cents/kWh, a 25% higher P50
(median) value of 5 cents/kWh, and a 150% higher expected value of 10 cents/kWh, were a
consequence of the overall spread combined with the long right- hand tail in Figure 8.2. There
was a 0.25 probability of a value below 4, but very limited scope for costs very much below
the most likely value of 4. There was a 0.75 probability of a value above 4, and considerable
scope for much higher costs, albeit with a relatively low likelihood of such costs being realised
as the right- hand tail is approached. There was a very wide range of potential reasons for these
higher costs, some of which they could explore later. The simple model nature of this portrayal
relative to a more complex reality could be explored later using Chapter 3 ideas.
The expected value was a best estimate of what should happen on average. Because of this
and because of its additive properties, expected values should be the basis of all estimation
processes at a portfolio component level as well as at a total portfolio estimate level. Most
likely values were of interest at a component level, especially because they tend to be the
implicit basis of deterministic estimation processes but also because they sometimes approxi-
mated to reasonable aspirational targets (such as an initial ‘no major problems’ budget given
to the manager of a project component). Median values were of less interest. With skew to
the right as indicated, median values always lie between most likely and expected values, and
the probability of an outcome less than the expected value was always greater than 0.5, as
illustrated in Figure 8.2.
A key aspect of quantitative evaluation of uncertainty was the identification of an unbi- ased estimate of the overall expected value. If uncertainty was only partially considered, the spread of the probability distribution would be less, especially to the right in terms of Fig-
ure 8.2. As indicated by this numerical example, reasonably full consideration of uncertainty
involving a significant skew leads to dramatic differences among most likely, median, and
expected values. Underlying conditions and dependence can be part of and amplify these
effects. Carl knew this illustrated one reason why Larry and Pat’s cost estimates for nuclear
power were generally out by a factor of two or more, a factor of ten having been recently
experienced with respect to construction cost estimates – Larry was not addressing all rel-
evant sources of uncertainty.
Carl was not looking at low or medium levels of bias here. Relative to the levels of bias
which Paul had to confront in the Chapter 7 context, or BP had to address in the Chap-
ter 3 context, in a cost of nuclear power context Carl was addressing very high levels of bias because of the combined effect of asymmetry (skew) and omitted sources of uncertainty.
Carl then drove home the basis of this preliminary demonstration. The 150% uplift from
nuclear power’s most likely cost of 4 cents/kWh to an expected cost of 10 cents/kWh was
large, this difference was generated by the long right- hand tail of Figure 8.2, the ‘product’
column of Table 8.1 showed the way extreme values influence expected values, and the sim-
ple three scenario approach facilitated sensitivity analysis he would now pursue.
He had deliberately used an illustrative example A which he believed was very optimistic
to avoid being accused of being alarmist. He would now consider further examples B and C.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
426 Employing planning tools in practice
As a simple illustrative example B, if the 25 cents/kWh scenario really ought to be a ‘35
cents/kWh scenario’ associated with a range from 15 to 55 cents/kWh, with no changes to
the other two scenarios, then a 200% uplift from the most likely 4 cent/kWh to an expected
value of 12 cents/kWh was involved (0.2 × 35 = 7 instead of 0.2 × 25 = 5).
As an illustrative example C, if the 25 cents/kWh scenario of example A really ought to
be a ‘65 cents/kWh scenario’ associated with a range from 15 to 115 cents/kWh, with no
changes to the other two scenarios, then a 350% uplift from the most likely 4 cent/kWh to
an expected value of 18 cents/kWh was involved (0.2 × 65 = 13 instead of 0.2 × 25 = 5).
Examples B and C involved the same most likely and median values as example A, but the
expected values were clearly very sensitive to the length of the distribution tail.
The implications associated with the long right- hand probability distribution tail were
also of crucial importance for reasons which go beyond biased expectations of expected
outcomes, taking us into enlightened prudence concerns associated with the risk of higher
cost outcomes.
You may feel you already understand why long tails associated with highly asymmetric
distributions are crucial in terms of both expected outcomes and risk, perhaps drawing on
your understanding of some of the ideas discussed in Chapter 3. However, what Carl did
not explain here – which you may find useful to think about briefly now – is that ‘experts’
involved in estimating these tails may be very poor at it because they may suffer from both
conscious and unconscious bias at serious levels, some aspects involving highly contentious
issues. Carl would later suggest a number of reasons, not discussed earlier in this book,
which you may not be familiar with which clarify why an even more pessimistic view of the
impact of this tail than example C might be appropriate. As fairly simple extreme examples,
there is a case for including ‘indirect costs’ which incorporate the possible costs of a nuclear
accident like that at Chernobyl, plus other concerns like failing to find a cost- effective way
of dealing with nuclear waste. There is no provision in Table 8.1 for the possible indirect
cost implications of concerns of this kind. Many people who are pro- nuclear would be very
concerned if there were, but many people who are anti- nuclear would argue that is just one
very good reason why ‘experts’ employed by the nuclear industry should not be trusted.
Chapter 9 will touch on these issues, and later discussion in this chapter will suggest exam-
ples B or C might be more appropriate than example A for much less contentious reasons.
Constructively simple models of non- utility and conservation cost distributions
Carl suggested that Table 8.2 provided a useful example A estimate for 1 GW of either non-
utility power or conservation, comparable to the example A estimate of Table 8.1.
Table 8.2 would let them begin to explore building portfolios without suggesting any
preferential ordering of nuclear, non- utility, and conservation in terms of expected cost.
A figure based on Table 8.2 comparable to Figure 8.2 could have been provided, but its
form was fairly obvious, and the cumulative format would be used in a comparative diagram
to follow shortly.
Again, a finer definition with more scenarios could have been used, but the immediate
concern was constructive simplicity to illustrate key issues, this time with a focus on depend-
ence relationships. The numbers were for illustrative purposes and could be argued with,
but the shape and spread of the distribution illustrated in Table 8.2 were reasonable. There
was scope for sizeable departures from the expected value in both directions, but a reason-
ably symmetric distribution shape without extremely long tails either side, for reasons he
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 427
Table 8.2 Example A cost distribution for one GW of non- utility power or conservation, tabular form.
Scenario value in 1992 cents/kWh (range)
Scenario probability (probability density per unit over range)
Product Cumulative probability
6 (4– 8) 10 (8– 12) 14 (12– 16)
0.2 (0.05) 0.6 (0.15) 0.2 (0.05)
1.2 6.0 2.8
0.2 0.8 1.0
Expected cost (column sum) 10.0 Most likely (most probable) cost 10 cents/kWh, also equal to the P50 (median) values
would develop shortly. For illustrative purposes when considering dependence concerns in
a discussion focused on diversification, assuming complete symmetry was a helpful simpli-
fication, recognising that in practice modest asymmetry would be likely. As a result of this
assumed symmetry the most likely, median, and expected values were all the same.
Carl then emphasised that estimation of distributions like those of Tables 8.1 and 8.2 in
practice would require considerable effort. The uncertainty involved should include a very
wide range of technical and economic issues. In practice it would be important to use a care-
fully structured approach to building these distributions from component factors and provi-
sions. It would also be important to use this uncertainty composition process to develop an
understanding of the relationships between different sources of uncertainty associated with
the overall probability distribution for each potential type of electric power, or its equivalent
in the case of conservation. He had in mind the use of nested sensitivity curves like those dis-
cussed in Chapters 4, and 7 as well as the full range of approaches discussed in Chapters 3,
6 and 7, but he did not develop examples at this point.
Building a portfolio of two units – with a focus on the importance of dependence
Carl began exploring portfolio concepts by arguing that all members of the board needed a
shared conceptual framework for building and testing their joint understanding of the impli-
cations of different kinds of causal dependence between distributions plus different forms
and levels of statistical dependence.
‘Causal’ dependence implied systemic uncertainty relationships which were understood
and modelled in cause– effect terms, while ‘statistical’ dependence implied underlying causal
relationships which were not fully understood or not worth modelling. An absence of both
causal dependence and statistical dependence implied ‘independence’.
He would start with a focus on statistical dependence between probability distributions
like those portrayed by Tables 8.1 and 8.2 for different types of electric power or its conser-
vation equivalent. For example, the extent to which the unit costs of non- utility power and
conservation might depart from current expectations in opposite directions would require
careful study by Canpower, with all directors understanding the implications. The HAT
approach he would use was a very simple conceptual framework in its basic forms which
could be generalised to visualise and model complex causal and statistical dependences asso-
ciated with very important practical issues – like economies of scale, common- mode failures,
and cascade effects. He would use it initially at a relatively simple illustrative level to portray
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
428 Employing planning tools in practice
a limited number of complexity issues, but practical adaptations they did not need to con-
sider until later could cope with any complexities worth analysis.
What follows makes basic causal and statistical dependence issues as simple as possible in
a framework which is as general as possible. It is relevant to EP as discussed in this book as a
whole, not just portfolio modelling. It builds on earlier use of the HAT framework, and it is
also directly relevant to implementing some of the approaches discussed in earlier chapters.
To begin to illustrate what was involved, and provide a basis for numerical models used
later, the board were asked to assume that the immediate concern was restricted to the pro-
vision of 2 GW of power from the three options identified earlier, as the starting point for
developing a portfolio of proposed new units to meet anticipated load growth and current
unit retirements at a 20 year long- term planning horizon. For illustrative convenience, this
provision would involve two choices from the three options listed earlier – both of one kind
or any mix.
They would look at the independence and dependence issues in illustrative terms by first
considering independence, using a 1 GW unit of conservation plus a second 1 GW unit of
conservation to provide 2 GW of conservation in total. They would work with the discrete
scenario values and a probability tree framework, assuming that these two units of conserva-
tion involved very different types of conservation approaches, with reasons for potential cost
variations which might be assumed to be independent.
Figure 8.3 portrayed the nature of the independence assumption and the calculation of
the joint cost distribution in a standard probability tree format, which Carl explained and
then related to Table 8.3.
The first number on each branch of the Figure 8.3 probability tree portrayal was the
assumed scenario value, the second the probability of that branch occurring. The Figure 8.3
joint cost column entries were averages and the probabilities were products.
Table 8.3 used a tabular format to convey the same information, but it also combined the
probabilities for common outcome values, and it included the implied cumulative distribu-
tion which was useful for the graphical portrayals to follow.
‘Independence’ required the probability distributions for the second unit to be exactly
the same whatever the outcome of the first, all first unit probability tree branches leading to
identical sets of second unit branches in Figure 8.3. Sometimes it would be convenient to
refer to independence as portrayed in Figure 8.3 and Table 8.3 as ‘0% dependence’, equiva-
lent to a ‘coefficient of correlation’ of zero. A working assumption of independence in this
case was reasonable if the two units involved very different uncertainties, perhaps because
very different approaches and technologies were involved, as assumed in this case.
The next step was considering the +100% dependence case (perfect positive dependence
equivalent to a coefficient of correlation of +1) by considering a 1 GW unit of non- utility
generation plus a second 1 GW unit of non- utility generation, with prices set in the same
marketplace to provide 2 GW of non- utility generation in total, working with the same
discrete scenario values and the same probability tree framework. Figure 8.4 portrayed the
implications of assuming +100% dependence and the calculation of the joint cost distribu-
tion using the same probability tree format as Figure 8.3. The joint cost calculations were
dropped for simplicity, but probability calculations were maintained.
In this case the +100% dependence assumption implied that the second distribution out-
come was known given any outcome for the first, and directly positively correlated, inter-
preting ‘directly positively correlated’ to mean that all percentile values correspond. If both
units involved prices set in the same marketplace, it was reasonable to use this working
assumption. It corresponded to a coefficient of correlation of +1.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
conservation first unit
conservation second unit
joint cost
6
14
10
0.2
0.6
0.2
0.2
0.6
0.2
6
10
14
0.2
0.6
0.2
6
10
14
0.2
0.6
0.2
6
10
14
probability
(6 + 6)/2 = 6
(6 + 10)/2 = 8
(6 + 14)/2 = 10
(10 + 6)/2 = 8
(10 + 10)/2 = 10
(10 + 14)/2 = 12
(14 + 6)/2 = 10
(14 + 10)/2 = 12
(14 + 14)/2 = 14
0.2 × 0.2 = 0.04
0.2 × 0.6 = 0.12
0.2 × 0.2 = 0.04
0.6 × 0.2 = 0.12
0.6 × 0.6 = 0.36
0.6 × 0.2 = 0.12
0.2 × 0.2 = 0.04
0.2 × 0.6 = 0.12
0.2 × 0.2 = 0.04
Figure 8.3 Example probability tree specification for two units of conservation, assuming independence.
Table 8.3 Cost distribution for 2 GW of conservation assuming independence as in Figure 8.3.
Average of scenario values
Probability calculation components Probability (density)
Cumulative probability
6 8 10 12 14
0.04 0.12 0.04
0.12 0.36 0.12
0.04 0.12 0.04
0.04 0.24 0.44 0.24 0.04
0.04 0.28 0.72 0.96 1.00
non-utility first unit
non-utility second unit
joint cost
6
14
10
0.2
0.6
0.2
1
0
0
6
10
14
0
1
0
6
10
14
0
0
1
6
10
14
probability
6
8
10
8
10
12
10
12
14
0.2 × 1 = 0.2
0.6 × 1 = 0.6
0.2 × 1 = 0.2
Figure 8.4 Example probability tree specification for two units of non- utility power assuming +100% dependence (all percentile values coincide).
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
430 Employing planning tools in practice
The Table 8.3 format could be used too. He would not bother because the result was
obviously the same as the distribution for a 1 GW unit as specified in Table 8.2 scaled up by
a factor of two.
To demonstrate that two units of nuclear power, 1 GW each, would have a scaled- up ver-
sion involving exactly the same distribution as one if +100% dependence was assumed, they
could use exactly the same approach as their two units of non- utility power example. That
is, if one nuclear power unit costs more for fuel cost, labour cost, technical reasons, or any
other reasons, an identical unit on the same site owned by the same utility was very likely
to have exactly the same problems, so assuming +100% dependence would be a reasonable
working assumption if a portfolio of two planning units of nuclear power was modelled,
directly comparable to Figure 8.4.
The next step was addressing the −100% dependence case (perfect negative dependence
equivalent to a coefficient of correlation of −1). They could do this by considering a 1 GW
unit of non- utility power plus a second 1 GW unit of conservation to provide 2 GW in total,
assuming that higher market prices for electricity would make it cheaper to persuade people
to reduce demand via low energy consumption light bulbs and similar devices. Figure 8.5
portrayed the nature of the −100% dependence case and the calculation of the joint cost
distribution in the same probability tree format as Figures 8.3 and 8.4.
Again a Table 8.3 format could also be used. In this case the −100% assumption implied
a P0 outcome for one unit meant a P100 outcome for the other, a P1 for one unit meant a
P99 for the other, and so on, with the result being a 10- cents scenario with certainty – con-
servation was assumed to work as a perfect hedge for non- utility power.
In practice −90% or −80% might be involved rather than −100%, and Canpower might
want to model this departure from the obviously overly optimistic −100% assumption.
Indeed, the 0% assumed earlier might be better estimated as +20%, and the +100% assumed
earlier might be better estimated as +90%, to correct for optimistic bias in the first case and
pessimistic bias in the second case.
non-utility first unit
conservation second unit
joint cost
6
14
10
0.2
0.6
0.2
0
0
1
6
10
14
0
1
0
6
10
14
1
0
0
6
10
14
probability
6
8
10
8
10
12
10
12
14
0.2 × 1 = 0.2
0.6 × 1 = 0.6
0.2 × 1 = 0.2
Figure 8.5 Example probability tree specification of one unit of non- utility power plus one unit of conservation assuming −100% dependence (P0 and P100 values coincide, as do P1 and P99, P2 and P98, and so on for all percentile values).
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 431
More generally, negative or positive dependence need not be perfect, and interpolations
need not be simple. A key reason for using a HAT portrayal of uncertainty framework
was the generality it provided for alternative forms of dependence and alternative levels of
sophistication in the modelling used to facilitate clarity efficiency. To demonstrate the basis
of this generality using a simple initial example Carl employed Figure 8.6.
This example illustrated a situation when there was a variable degree of dependence. If
the market price for non- utility electricity increased above the expected value, this might
stimulate stronger negative dependence than would be the case if the price decreased below
the expected value, illustrated by the example probabilities. That is, if electricity consum-
ers were more sensitive to price rises than they were to price falls, the level of dependence
experienced would be different for variability above and below expected outcomes. Again
the Table 8.3 format could also be used.
Carl explained that they could model much more complex relationships in practice, but
he was deliberately keeping his examples as simple as he could for the moment.
He then indicated that Canpower had to address the portfolio management implications
of the issues raised by dependence using a generalisation of Markowitz’s mean– variance
approach within the HAT framework he had just illustrated. The basic ideas which EP port-
folio models can capture which the board needed clarity about were:
1 ‘don’t put all your eggs in one basket’,
2 ‘diversify using a risk efficiency concept’,
3 ‘use an opportunity efficient approach which includes the right risk– reward trade- off’, and
4 ‘use a clarity efficient approach which includes the right trade- offs between clarity of
various kinds and the cost of further clarity in terms of analysis effort’.
Choosing the mix of sources of electric power by an electric utility was a classic portfolio
analysis problem, whether or not the board appreciated this was the case. Many electricity
non-utility first unit
conservation second unit
joint cost
6
14
10
0.2
0.6
0.2
0
0.5
0.5
6
10
14
0.1
0.8
0.1
6
10
14
0.8
0.2
0
6
10
14
probability
6
8
10
8
10
12
10
12
14
0.2 x 0.5 = 0.10
0.2 x 0.5 = 0.10
0.6 x 0.1 = 0.06
0.6 x 0.8 = 0.48
0.6 x 0.1 = 0.06
0.2 x 0.8 = 0.16
0.2 x 0.2 = 0.04
Figure 8.6 Example probability tree specification of one unit of non- utility power plus one unit of conservation assuming variable dependence.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
432 Employing planning tools in practice
utilities recognised this explicitly and formally, even if they did not model the issues formally.
For example, as noted earlier, prior to the privatisation of electric power in the UK, the
CEGB policy of building a limited number of nuclear power stations, and doing so based on
several different technologies, was explained to the public as a deliberate attempt to spread
the risk to ensure that if nuclear power was required in the future, at least one successful
UK- developed technology would be available.
One very simple diversification policy to achieve a robust balance of generation types was
just using equal proportions of all available options. Some early investment advisors (well
before the 1929 stock market crash) recommended ‘keep a third of your assets in property,
a third in fixed interest vehicles (bonds) and a third in stocks (shares)’. These early invest-
ment advisors had much more wisdom than those who put everything into stocks. However,
a very simple diversification policy like this would probably decrease expected reward in an
inefficient manner well worth improving on. The concepts of risk efficiency and risk/reward
balance applied to Canpower as elsewhere, and the issues involved were both important and
complex, so a degree of sophistication in the diversification strategy was essential.
Elaborating on item 1 in his list earlier, one key corporate goal should be to diversify, to
avoid ‘putting all the Canpower eggs in one basket’. This was just basic common sense. He
might have added that this aspect of common sense seemed to have escaped Larry’s atten-
tion, and the board had apparently not noticed. He refrained, but some board members
clearly got the message anyway.
A second key corporate goal should be to use the risk efficiency concept to ensure that
they were delivering the maximum level of expected reward for whatever level of risk and
associated diversification the board believed was appropriate.
A third key corporate goal should be to diversify by moving along the risk efficient bound-
ary to achieve the least increase in expected cost for the greatest reduction in risk, stopping
this diversification process when risk was reduced to an acceptable level, understanding the
implications of the level of risk accepted and the risk– reward trade- off.
The fourth key corporate goal should be to achieve the first three goals in a clarity effi-
cient manner, with no more effort than was necessary to understand how diversification
was working and what options should be taken, in a manner which could be explained to
all interested parties.
The overall opportunity efficiency Canpower should be seeking involved clarity efficiency
plus risk efficiency plus an optimal risk– reward trade- off.
To start to illustrate what was involved making use of his earlier examples, Carl showed
the board Figure 8.7.
Figure 8.7 portrayed a 0% dependence curve plotted from Table 8.3 based on Figure 8.3.
In addition, it showed a −100% dependence curve plotted from Figure 8.5 using the Table 8.3
format to compute a cumulative distribution. It also included a +100% dependence curve
plotted from Figure 8.4 using the Table 8.3 format to compute a cumulative distribution.
The +100% dependence curve illustrated no decrease in cost variability when two units
were combined and no diversification effect was a reasonable assumption. The clear implica-
tion was no decrease in variability if further units were added when perfect positive correla-
tion was involved. It was crucially important to understand that 100% positive dependence
meant a zero diversification effect no matter how many units were added together in a
portfolio. For example, if Canpower moved to an all non- utility power portfolio, buying all
its power in a perfect marketplace involving producers of electricity within their province,
neighbouring provinces, and US states, it had a zero- diversification policy. Simply relying on
the market to provide the electricity required by its province involved a zero- diversification
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 433
policy in terms of this model. Also, if Canpower had a policy that decreed all new base- load
power would be Canpower- owned nuclear, it was adopting a zero- diversification policy for
all new base- load power.
The 0% dependence curve illustrated a significant decrease in variability when two units
were combined and zero correlation was a reasonable assumption. This clearly implied fur-
ther significant decreases in variability if further units were added and independence was
involved. The 0% dependence curve in conjunction with the +100% dependence curve illus-
trated what interpolating at something like +60% or +90% dependence might yield.
The −100% dependence curve illustrated the huge impact of negative correlation, in the
limit just two units with perfect negative correlation eliminating uncertainty completely.
Furthermore, what −80% dependence might imply was indicated by the −100% curve taken
together with the 0% dependence curve. Even very modest levels of negative dependence
could be extremely useful, a very important feature of conservation which was clearly a
considerable asset.
As indicated when discussing Figure 8.6, analysis of partial dependence was not restricted
to interpolation, and results based on various forms of variable dependence modelling could
be compared using an approach like Figure 8.7. He had not plotted the Figure 8.6 example
in Figure 8.7, but if they wanted a clear understanding of what Figure 8.6 implied and they
wanted to communicate that understanding to others, then showing the cumulative curve
defined in Figure 8.6 between bounds portrayed by the 0% and −100% curves of Figure 8.7
might be very helpful.
Carl did not want to develop these ideas further for the board at this point, but you might
find it useful to understand now that the HAT approach and the Figure 8.7 portrayal of
results also provided a basis for generalisation to portray the implications of more detailed
and sophisticated modelling, including those discussed in Chapters 3, 4, 6, 7, and 9. BP
software developed in the late 1970s provided support for all these forms of dependence
modelling, and they were employed in the Canadian and US studies discussed briefly in
Chapter 7. The toolset for a rich set of alternative approaches to understanding dependence
0
+100%
cumulative
probability
outcome cost in cents/kWh
0 5 10 15 20
–100%
1.0
0.2
0.4
0.6
0.8
0%
Figure 8.7 Cumulative probability diagrams for the +100%, 0% and −100% dependence examples.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
434 Employing planning tools in practice
via formal modelling has been available for some time, and it has been used successfully in a
range of contexts, some illustrated directly in earlier chapters.
To link this discussion to common experience, you might reflect that independence
is the basis of spreading risk in insurance operations. For example, if one life insurance
policyholder dies earlier than expected, it is likely that another will die later than expected,
this cancelling out effect producing an acceptable average claims experience provided
a reasonably large number of policies is involved. Positive dependence means that the
cancelling- out effect of independence is reduced or even lost altogether. If an insurance
company specialises in earthquake risk in San Francisco, positive correlation puts it and
its policy holders at risk because the cancelling- out effects insurance companies have to
achieve to be effective are lost. Reinsurance practices and insurance regulation have to
address these kinds of issues along with other concerns. Negative dependence means that
the cancelling- out effect of independence is enhanced. Bookmakers attempt to ensure
negative dependence on their portfolio of bets by changing the odds so that they get a bal-
ancing of bets on either side of an outcome: a skilful bookmaker always keeps a balanced
book. Hedge fund operations try a comparable approach in some respects, although dif-
ferent in others.
Understanding how all these concepts operate in terms of one very general framework
with a rich set of implementation approaches which can be tailored to the context is not easy,
but this understanding is central to the aspects of corporate strategy formulation addressed
by Carl via portfolio- driven long- term planning. Larry’s approach to medium/long- term
planning simply ignored all these considerations, implicitly assuming they did not matter.
Comparable deterministic merit- order approaches also ignore these considerations. There is
a reasonable case for suggesting that ignoring issues as important as this in Larry’s context
involved very serious incompetence, on his part, and with respect to the board as a whole
because they did not understand what Larry was doing, but Carl was careful not to say so
in a direct manner. Nor did he bother to develop the implications for any organisations
adopting a simple robustness- driven scenario approaches which focus on a single source of
uncertainty like the price of oil when some aspects of diversification might be important and
warrant explicit formal attention, an issue you may find relevant in other contexts of interest.
Building a portfolio of different unit types to define the long- term target proportions
Carl began building on this Figure 8.7 discussion by making it very clear that developing
suitably robust operational portfolio models in any context was difficult, for a number of
technical reasons. Even in the context of managing a portfolio of financial securities, the
simplest classic portfolio problem, simplifying the complications involved in a clarity effi-
cient manner was a demanding task. However, Canpower could address a portfolio- driven
long- term planning approach using practical tools in a manner he would outline.
The illustrative numerical example distributions used so far all assumed a common
expected cost of 10 cents per kilowatt- hour. This was in part for expository convenience, in
part, to avoid seeming prejudiced in relation to any particular option.
Canpower’s current medium/long- term plans assumed nuclear power units had the low-
est expected cost per kWh. Carl did not believe this was true, but assuming for the moment
that it was, it should be clear that six new nuclear power stations would involve very signifi-
cant risk of costs well above the expected cost. The board needed to understand and deal
with this risk in an effective and efficient manner.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 435
If Canpower wanted to make a case for this kind of significant nuclear power increase at
the 20 year horizon, they might start by refining Table 8.1 and Figure 8.2 and understand-
ing the importance of components of the uncertainty which drove different degrees of
dependence. They might then use a generalisation of the HAT based approach of Figure 8.6
and interpolation structures like Figure 8.7 to consider likely actual locations for nuclear
power stations, the grouping of multiple units within power stations, economies of scale,
and related dependence concerns as more units of nuclear of the same or different designs
were added to a portfolio. They could add to these basic approaches a range of additional
tools to build a clear understanding of the expected cost trajectory, and the associated cost
variability risk trajectory, as the proposed nuclear portfolio grew.
In terms of new sources of electric power, there was nothing wrong with the Canpower
board starting where their instincts suggested they should concentrate, but they needed an
unbiased first- cut estimate of all key sources of uncertainty, including an unbiased estimate of associated variability about expected outcomes, plus how these sources of uncertainty
were related. They needed a robust understanding of the effects of these sources of uncer-
tainty on both expected costs and spread, including the dependence issues involved when
units were combined, acquired in a clarity efficient manner.
All other technically proven sources of electric power which were relevant then needed
comparable treatment – conservation and non- utility power as just discussed, plus the obvi-
ous hydro and thermal options, plus currently technically feasible renewables options they
ought to consider, even if they did not expect to include them on economic grounds, so they
could be clear about the current cost gap preventing their inclusion. These gaps would be
relevant to all futures planning, and any discussions about government sponsored subsidies.
They could use this starting point to take a view on the lowest expected cost ‘initial core’
of new power stations for their target overall portfolio at the planning horizon. It might be
all nuclear, it might be all gas- fired power, or it might be something else.
Instead of an initial focus on new power stations, they might start with all the existing
power stations which they expect to still have at the 20 year horizon, adding currently
preferred minimum expected cost new power stations to work towards the total capability
needed at the horizon. This would let them start by dealing with dependence between new
and existing power station cost variability.
They could then use their initial starting point portfolio to start replacing some of their
initial core with alternative options which offered negative dependence or a minimum
degree of positive dependence together with a minimum increase in expected cost. Their
initial focus on lowest expected cost could be used as a basis for starting to explore risk effi-
cient ways of reducing risk.
Whichever route they took, there was nothing wrong with starting with simple models
like those illustrated by Figures 8.3 to 8.7, but to develop a more complex understanding of
dependence relationships between identified components of uncertainty, they might draw
on a wide range of more sophisticated models. All the contingency planning models devel-
oped by BP for offshore projects as discussed in Part 1 and Chapter 7 used exactly the same
HAT approach, so all the associated modelling capabilities could be brought to bear on this
kind of analysis, plus further alternatives, including those discussed in Chapter 9 and a range
of approaches making use of quite general Monte Carlo simulation techniques.
The first order objective would be seeking risk efficiency. In this context, one obvious
view of risk efficiency involves minimising cost risk defined in terms of cost which exceeds
expectations for any given level of expected cost. A more convenient view was minimis-
ing reward risk in terms of a reward which fails to meet expectations for any given level
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
436 Employing planning tools in practice
of expected reward. Carl preferred this less obvious alternative because it links the current
discussion to the Part 1 discussion, and he provided a background explanation of the reason
he preferred a reward- based approach to risk efficiency, drawing on the ideas discussed in
Part 1 as a preliminary to explaining that if the board viewed Canpower’s basic corporate
strategy goal as providing provincial electricity needs at a minimum cents per kWh tariff,
then ‘reward’ and risk efficiency in reward terms might be assumed to involve expected
cost savings relative to just buying power from other utilities in the province or in adjacent
provinces or US states.
He then showed the board Figure 8.8, a variant of Figure 3.8.
If an all nuclear base- load portfolio provided the lowest expected cost option, and if cost was deemed to be the only relevant objective so an all nuclear base- load portfolio
equated to highest expected reward option, then point ‘a’ in Figure 8.8 could be associ- ated with an all nuclear base- load portfolio. However, even if this was the case, and it made sense to limit the scope of ‘reward’ to cents/kWh cost, an all nuclear base- load portfolio
would be extremely risky, because of the very long tails associated with distributions like
that illustrated by Table 8.1 and Figure 8.2, or more extreme possibilities, plus the positive
dependence relationships between the units in a portfolio of nuclear power units. Canpower
needed a corporate understanding of how this level and kind of cost risk could be reduced
efficiently, so it could move around the risk efficient boundary towards a point like b1 or b2,
accepting a risk efficient decrease in expected reward (increase in expected cost) and choos-
ing an appropriate level of risk given the cost of reducing risk. As they progressed around
the risk efficient boundary, they would be using the insight developed earlier, plus new
insight about the dependence between types of units already in their provisional core and
new candidates for inclusion.
risk in terms of reward which fails to meet expectations
Canpower’s
expected
reward In terms of expected cost saving per kWh relative to just buying power from other utilities plus other positive objectives
a–c efficient boundary target, with
a maximum expected reward approach
c minimum tolerable expected reward approach
bi intermediate expected reward approaches
c–d inappropriate reward levels
ei inefficient approaches
key:
a
b3
c
d
competent management area
(usefully interpreted as an opportunity management area)
e1
non-feasible area
b2
b1
e2 e3
incompetent management area
Figure 8.8 Canpower’s expected reward and associated risk in an efficient frontier portrayal.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 437
If nuclear power did not provide the highest expected reward (lowest expected cost) option, which Carl believed was highly probable, then an all nuclear new plant core might be a risk inefficient point like e1, e2, or e3. In this case their first task might be viewed as
getting to a point like b2 to get on the risk efficient boundary, or they might reconsider the
starting- point core.
It was crucial that Canpower had a clear idea where point ‘a’ was in expected reward and associated risk terms, and what point ‘a’ involved in terms of the proportions of different unit types in its proposed portfolio. Furthermore, Canpower could not afford to miss the opportunity to reduce risk to some extent, whatever type of unit had the lowest expected cost. Further still, Canpower could not afford to fail to understand its options in terms of an unbiased view of expected outcomes plus all associated variability, including the crucial role of relevant interdependencies.
Canpower needed risk efficiency, generalised to opportunity efficiency, and it needed to
explore what this meant in practice.
Carl did not labour the ‘competent management area’ associated with ‘opportunity man-
agement’ or the ‘incompetent management area’ where they currently resided to the extent
he was tempted to, but he used BP, IBM and other examples comparable to this book’s
Part 1 discussion to make sure they understood what was needed, in the language of elec-
tricity generation in the 1990s.
He did emphasise that if Canpower continued to reduce risk by moving around the risk
efficient boundary from point ‘a’ to b1 and then b2 in the way just described, the expected
reward would decrease (expected cost would rise), and the value of further diversification
would decrease, so Canpower must select a point of balance when the diminishing returns
from further diversification made further risk reduction inappropriate. Those who were
unwilling or unable to undertake any more than a minimum level of risk might wish to
continue to reduce risk until risk was very low, but this could involve a significant decrease
in expected reward (increase in expected cost). Electricity utilities like Canpower should
not take an extreme position on either minimising risk or maximising expected reward
(minimising expected cost), and Canpower needed to understand clearly what this meant
in practice.
Carl also emphasised that the portfolio- driven long- term planning modelling he was
advocating would not itself yield a precise prescriptive ‘optimal’ solution. It would provide
the insight required to make a reasoned set of robust judgements which achieved risk effi-
ciency in approximate terms with an appropriate trade- off between expected reward and
associated risk, plus an appropriate trade- off between effort expended on analysis and clarity
achieved via this analysis – opportunity efficiency. It would help Canpower to get approxi-
mately correct answers to the right questions. He was tempted to suggest that a precisely
correct answer to the wrong question was an accurate characterisation of Larry’s approach,
but resisted the temptation. Larry understood the unstated message, as did several other
directors.
At this point Carl did point out that a high nuclear base- load future exposed Canpower
to risks he had yet to address, like the identification of a design failure that might necessitate
shutting down all nuclear power stations with the same or similar design features.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
438 Employing planning tools in practice
An introjection to clarify your interpretation of Figure 8.8
Carl was convinced that Canpower was currently at a point like e3 in Figure 8.8, with plans
to move farther to the right in terms of their six new nuclear power stations core strategy.
He thought starting to lead the board towards this view from the outset was important, but
he did not think it would be helpful to say so in a direct manner at this stage in the meet-
ing. The reasons Carl believed this will become clearer as this chapter’s tale progresses, but
the basis was Carl’s view that Canpower’s current levels of bias in favour of nuclear power
relative to other sources, plus its total failure to address all relevant risk or the concept of
risk efficiency, meant that moving from the current 50% of base- load nuclear power towards
40% rather than 70% was a much better bet given his current understanding of the issues. In
Carl’s view, the case for nuclear was a degree of diversification away from an overly fossil fuel
dependent approach in the economic climate of the early 1990s, not minimum expected
cost per unit of electricity. This was a position many energy experts would have agreed with
unless they were anti- nuclear for reasons which went beyond cost as portrayed by Carl’s
example A. Issues associated with examples B and C are addressed later in this chapter, and
Chapter 9 develops further related concerns.
The importance of size and variety to risk reduction
While 3 GW was a practical size for a nuclear power station, many hydro, thermal, and other
power stations would be smaller. In particular, most renewable sources, conservation meas-
ures, and many sources of non- utility generation available within province would involve
‘operating units’ which were much smaller than 1 GW ‘planning units’. This meant that 1 GW planning units of sources of new electric power like conservation
or within province non- utility power would usually themselves be a portfolio of components
with dependence/independence characteristics. This would tend to reduce cost variability
and centralise the expected value of the distribution, as assumed for the Table 8.2 illustration.
More generally, a large number of relatively small operating units with a variety of tech-
nologies were inherently very much less exposed to cost variation uncertainty and associated
risk than a limited number of relatively large operating units involving similar technology.
Put the other way around, any new very large power stations with a common technology
would significantly increase the overall cents/kWh risk relative to diversified smaller units,
be they nuclear, gas- powered thermal units, or anything else.
Alternative portfolio modelling approaches to risk efficiency and separability
Carl explained that the formal modelling approach developed by Markowitz (1959) for
any number of possible allocations (securities in a financial portfolio context, electricity
generation units in the present context) used quadratic programming (a form of mathemati-
cal programming) to consider all possible allocations simultaneously. Quadratic program-
ming algorithms were, in fact, first developed to provide solutions to this model. Such
models involved technical complications which his presentation did not need to address.
However, the simple approach to considering a portfolio of two allocations he had used
could be generalised to any number by building up a nested structure, a pairwise approach
to nesting being particularly useful. He had in mind a framework described at that time
(Chapman and Ward, 1992), but more relevant and recent treatments are now available
(e.g. Chapman and Ward, 2002, chapter 10).
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 439
This framework could be used to provide a basis for overlaying a ‘softer’ (more intuitive)
approach to portfolio risk management which addressed issues not measured numerically in
a direct manner. The consideration of these softer concerns might manifest itself in terms
of constraints on the proportions of a portfolio which could be associated with any single
allocation or any set of allocations. For example, this framework could be used to assess
regulatory or planning constraints of the form ‘no less than X1% of the supply capability of an electricity utility should be nuclear, no more than Y1% an electricity utility should be nuclear, no less than X2% of the supply capability of an electricity utility should be non- utility power from sources within the province, and so on, for all possible sources of electric power’.
These constraints could reflect the board’s intuitive view of the portfolio risks as well as
external views of this risk. These constraints could also reflect political or regulatory views
based on quite different considerations which were superimposed on the cost- driven risk
efficiency basis for a portfolio approach. For example, if the Canadian federal government
was offering ‘support’ for Canpower’s nuclear programme in terms of subsidies for Cana-
dian Candu nuclear technology used by Canadian utilities, the political price might be an
agreement that no less than X1% of Canpower’s supply capability must be Candu nuclear
power stations. Other constraints might be driven by security of supply concerns, including
concerns like a common fault in nuclear power stations of the same kind leading to a major
loss of supply, as viewed by a regulator.
Considerable development of the interface between intuition- driven approaches based
on constraints on proportions of the Xi minimum and Yi maximum form and more direct
modelling frameworks might be useful, and this development was a practical proposition,
but it might not be needed. Whether or not it was used, translating the results of various
direct forms of portfolio- driven modelling into a policy result in the form of proportion
minimum and maximums should be very attractive to a board which wanted to understand
what really mattered in a long- term planning context. Carl suggested this would be the form
of output from the long- term planning process which most board members would probably
find particularly useful.
Using decision diagrams and sensitivity diagrams
Carl might have spent some time explaining that the nested composition structure he had
in mind lent itself to using decision diagrams for making choices and sensitivity diagrams to
understand the composition of the uncertainty portrayed by decision diagrams. He did not
have time, and we will leave even overview discussion these issues until later in this chapter.
Decisions involving timing considerations
Carl made the point that this portfolio modelling approach could, in principle, be extended
to a multiple period model which explicitly looked at the timing of all new plant decisions at
the same time, in a manner comparable to Canpower’s current linear programming models
for long- term planning. Indeed, doing so on a year- by-year basis using a Markowitz mean–
variance portfolio model with zero concern for risk was a special case which equated to
Larry’s approach. However, he would not recommend this approach.
Most timing issues should be deferred to the proposed separate medium- term planning
approach for clarity efficiency reasons. However, any potential source of power at the hori-
zon which involved important timing considerations could be treated as a set of options
which allowed consideration of key timing relationships. For example, there were important
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
440 Employing planning tools in practice
cost savings and associated simplifications in approvals processes if new thermal or nuclear
power stations were built on the sites of retired power stations. If an existing plant was due
for retirement, starting to replace it on the planned retirement date was an obvious choice,
with related options involving early retirement, postponed retirement, or retirement fol-
lowed by site preparation for new plant at a later date.
Concluding summary for long- term planning
Carl finished his discussion of long- term planning by emphasising four basic deliverables
must be provided by the Canpower approach to long- term planning:
1 Estimates of maximum and minimum proportions of each type of currently feasible
electric power generation capability which Canpower should be aiming for at its 20 year
long- term planning horizon – an interval or range- based set of target proportions.
2 An estimate of electricity demand at the 20 year horizon used to convert the propor-
tions provided by point 1 into generation capacities for each type of power station and
associated demand driven uncertainty.
3 Clarity about the risk of possible variations from the expected cents/kWh cost of elec-
tricity associated with the target portfolios defined by points 1 and 2.
4 Clarity about the robustness of all key assumptions underlying points 1 through 3.
From an EP perspective, all four of these deliverables were essential and would be provided
by a practical operational form of his recommended approach.
Medium- term planning and integration upwards
Carl began his treatment of medium- term planning by reminding everyone that medium-
term planning must deliver Canpower commitments to agreed plans over the next ten years.
But it had to relate these commitments to other tentative plans for the rest of the 20 year
horizon so that the commitments plus the tentative plans fully bridged the gap between the
current situation and the long- term planning targets. He would begin with the basis of this
complete set of bridging plans – what he would refer to as the ‘provisional overall plans’.
Generating provisional overall plans in approximate time- based profiles
Overall corporate strategic plans had to clarify current assumptions about the total portfolio
of existing sources of electric power which would still be operational at the 20 year planning
horizon, when any current sources would retire in the interim, plus when all new units would
become operational. When developing plans defined in these overall terms they would have
to be decomposed into different types of units, and all associated uncertainty would need
clarification. For example, all new units would need provisional ‘start construction’ and
‘start operation’ time- based profiles to meet anticipated load growth plus retirements. As
a specific illustration, the board would want to know that if 18 GW of new nuclear power
units for base load were planned to be in service at the 20 year horizon, this meant some-
thing like ‘all new base- load units would be 3 GW Candu power stations with construction
starts for the six new power stations planned at one year intervals beginning in about four
years, starting operation about ten years after beginning construction’. If this information
had not been created to clarify the long- term planning target for the board, it would need to
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 441
be created as the first step in the medium- term planning process. He thought the develop-
ment of this kind of component of the ‘provisional overall plan’ was usefully seen as part of
the medium- term planning process for that day’s discussion purposes, but the board could
see it as part of the long- term planning process to help interpret the long- term plan if they
preferred.
He indicated this kind of approximate time- based profile for each type of new unit, plus
associated plans for retiring units linked to load growth assumptions, should match expected
demand and supply at the 20 year horizon and for interim years and clearly portray obvious
priorities if a currently imbalanced portfolio needed rebalancing as soon as possible.
Actions plans and associated medium- term provisional plans
The rest of medium- term planning was about transforming part of the first ten years of this
provisional overall plan into ‘action plans’ which reflected agreed commitments set within
the context of the provisional overall plans. The commitments aspect of the medium- term
plan as a whole would have to include immediate commitments to nuclear power stations
required within the decade if Canpower wanted to be clear about when commitments to
build nuclear power stations should translate into realised capacity.
‘Action plans’ and ‘action horizons’ would be associated with each commitment to action.
An action horizon was the period requiring commitments to implement associated action
plans because of the need for lead times and the ‘irreversible’ or ‘reversible at a considerable
cost’ nature of associated decisions. Long lead times required long action horizons.
The medium- term planning horizon would not be equated to an action horizon in a gen-
eral sense. For example, it was now clearly recognised that giving the ‘go- ahead’ to build a
nuclear power station was a decision which ought to be seen as irreversible in the sense that
it was not sensible to plan a significant slowdown, and a construction period of the order
of ten years was involved, the reason that a ten year action horizon for nuclear power sta-
tions was needed plus a ten year medium- term planning horizon overall. However, a project
involving building a gas- turbine power plant with all permissions in place and the site pre-
pared might be deferred for six months without comparable risks, and once the go- ahead to
begin construction was given, action horizons might be reasonably long for key components
like the turbines and generators but fairly short for the civil construction building materials
and virtually zero for the landscaping supplies.
This implied a need for action plans which involved commitments appropriate to each
part of the plans, embedded in provisional medium- term plans using a nominal planning
horizon of ten years to cope with the longest lead times needed. Canpower needed to plan
commitments as far ahead as necessary to optimise the action plans, but no further, recog-
nising that it could later alter aspects of the other provisional medium- term plans if unfold-
ing events suggested this was advisable. Indeed, it might have to alter action horizon plans,
but this should be avoided to the extent that the costs warranted doing so.
Detailed planning for implementation of an action plan was appropriate, but other parts of
the medium- term plans which did not involve commitments might not receive any detailed
planning effort until it was needed.
It was in this medium- term planning context that Canpower had to address when to start
the construction of new units, when to implement a new conservation programme, when to
retire or mothball existing units, and so on.
This was part of the interface between corporate planning in top- down terms and project
planning, as discussed in Chapter 7, which Carl would have to return to later.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
442 Employing planning tools in practice
Modelling load growth uncertainty
What Carl wanted to focus on next was how modelling load growth uncertainty could help
medium- term planners to begin to understand and deal with some of the important issues
they needed to address. Figure 8.9 could be used to start exploring some of the features of
interest.
period 1 period 2 growth, GW
0
2
0.2
0
1
2
0.4
0
1
2
1 0.4
0
1
2
0
2
0.3
0
1
2
0.3
0
1
2
1 0.4
0
1
2
0
2
0.4
0
1
2
0.2
0
1
2
1 0.4
0.1
0.4
0.5
0.4
0.4
0.2
0.3
0.4
0.3
0.2
0.4
0.4
0.4
0.4
0.2
0.3
0.4
0.3
0.2
0.4
0.4
0.5
0.4
0.1
0.3
0.4
0.3
0
1
2
0 0.3
1 0.4
2 0.3
period 3 probability
0.3 × 0.2 × 0.1 = 0.006
0.3 × 0.2 × 0.4 = 0.024
0.3 × 0.2 × 0.5 = 0.030
0.3 × 0.4 × 0.3 = 0.036
0.3 × 0.4 × 0.4 = 0.048
0.3 × 0.4 × 0.3 = 0.036
0.3 × 0.4 × 0.4 = 0.048
0.3 × 0.4 × 0.4 = 0.048
0.3 × 0.4 × 0.2 = 0.024
0.4 × 0.3 × 0.2 = 0.024
0.4 × 0.3 × 0.4 = 0.048
0.4 × 0.3 × 0.4 = 0.048
0.4 × 0.4 × 0.3 = 0.048
0.4 × 0.4 × 0.4 = 0.064
0.4 × 0.4 × 0.3 = 0.048
0.4 × 0.3 × 0.4 = 0.048
0.4 × 0.3 × 0.4 = 0.048
0.4 × 0.3 × 0.2 = 0.024
0.3 × 0.4 × 0.2 = 0.024
0.3 × 0.4 × 0.4 = 0.048
0.3 × 0.4 × 0.4 = 0.048
0.3 × 0.4 × 0.3 = 0.036
0.3 × 0.4 × 0.4 = 0.048
0.3 × 0.4 × 0.3 = 0.036
0.3 × 0.2 × 0.5 = 0.030
0.3 × 0.2 × 0.4 = 0.024
0
1
2
1
2
3
2
3
4
1
2
3
2
3
4
3
4
5
2
3
4
3
4
5
4
5
6 0.3 × 0.2 × 0.1 = 0.006
Figure 8.9 Example load growth probability tree for three three-year periods, nine years in total.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 443
Figure 8.9 employed the same basic HAT framework introduced earlier. It portrayed load
growth in three sequential periods, each period three years long, for nine years total. Each
period employed three probability tree branches associated with load growths of 0, 1, and
2 GW. Some correlation was assumed over the nine years. For example, on the top branch
a load growth of zero in the first period was assumed to have a 0.3 probability, but if it
was zero in the first period the chance that it would also be zero in the second period was
assumed to be 0.2, lower than it was in the first period. This interdependence extended to
third period relationships in a similar manner. For example, on the top branch the chance
that it would be zero in the third period if it was zero in the first two dropped to 0.1.
Figure 8.9 reflected random shocks which might or might not be sustained over time,
as Canpower should expect in practice. That is, over time there was a general tendency
to return to underlying trends, although such returns might not occur, and trends might
alter. More generally, Figure 8.9 illustrated a probability tree approach for uncertain load
growth which allowed the portrayal of statistical dependence structures which might take
any plausible form. This included the tendency to return to a long- term growth pattern
after random shocks, sometimes referred to as ‘regression to the mean’, but a possibility
that this would not happen and that sustained changes in the trend would occur. A very
general modelling approach to explore understanding what was involved could be used
by Canpower. The net effect over nine years of the three levels of branches in terms of
outcomes was computed by summing the growth over the branches. For example, zero
growth in all three years yielded zero growth over the nine years. The probability of each
outcome was the product of the associated branch probabilities. For example, the prob-
ability of zero growth was 0.3 × 0.2 × 0.1 = 0.006 over nine years, obtained following the
top branch to its end point.
This three level tree could be collapsed to an equivalent single level tree, as shown in a
tabular form in Table 8.4, a generalisation of Table 8.3.
Each of the probabilities associated with an end point of the Figure 8.9 probability tree
was allocated to its GW value in the ‘probability calculation components’ column. Summing
across rows yielded the density form of the distribution. What was involved was the repeti-
tive addition of conditionally specified dependent probability distributions, a slight generali-
sation of the treatment of adding conditionally dependent distributions in Figure 8.6. The
symmetry of the dependence pattern resulted in an expected growth in each of the three
three year periods of 1 GW, equating an expected growth of 3 GW over the nine years, as
indicated in the final row of Table 8.4. In practice asymmetric dependence patterns might
be extremely important, a complication avoided for this example.
Table 8.4 Load growth probability distribution for nine years.
Load growth GW
Probability calculation components Probability (density)
Cumulative probability
0 1 2 3 4 5 6
.006
.024
.030 .036 .048 .036
.048
.048
.024
0.24 .048 .048
.048
.064
.048 .048 .048 .024
.024
.048
.048 0.36 .048 .036
.030
.024
.006
0.006 0.084 0.246 0.328 0.246 0.084 0.006
0.006 0.090 0.335 0.664 0.910 0.994 1.000
Expected growth 3 GW, 1 GW every three years
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
444 Employing planning tools in practice
One message that Carl wanted to begin to convey to the board, using the example por-
trayed by Figure 8.9 and Table 8.4, was load growth (demand) uncertainty was an impor-
tant component of the overall uncertainty associated with matching demand and supply for
medium- term planning. For some purposes, load growth could be treated simply via a single
level tree like Table 8.4, but load growth for a new power station was intended to meet
also needed understanding using context specific decomposed models underlying the very
general model of Figure 8.9. They needed to link econometric demand model approaches
with presumed future possible price changes and other market and political changes beyond
the end of the construction period. For example, if a demand growth of 3 GW over nine
years was expected, but 0 GW was the outcome, an ongoing zero or low growth outcome
might have a very high probability. To have confidence in a Table 8.4 summary, with likely
ongoing trends part of that understanding, it would be important to understand some fairly
complex underlying relationships.
A second message was the possible relevance of asymmetric penalty cost functions associ-
ated with the direct use of Table 8.4 together with inferences about longer- term trends.
This was a concern which could be addressed using approaches he would illustrate using a
numerical example in a moment.
A third message, to be further developed later, was large power stations with long con-
struction lead times and high capital costs involved a significant contribution to the system’s
cost per kilowatt- hour of electricity which was associated with Canpower’s inability to fore-
cast precisely when its power would be needed, and this cost was not normally attributed to
it. A set of small power stations with short lead times and proportionately lower capital costs
involved much less system cost per kilowatt- hour of this kind for the same total amount of
power because the uncertainty associated with load growth during construction was very
much less. In decision analysis (decision theory) terms, this cost was known as ‘the value of
information’. In the present context, it was more appropriately termed ‘the cost of uncer-
tainty about load growth’, a cost which was a function of the size of the power stations being
considered and the lead times involved. For present purposes, Carl was not considering
uncertainty about lead times between a commitment to build and power being available – a
further complication considered later.
Modelling the cost of uncertainty about demand with an asymmetric penalty function
As the basis for a simple illustrative example, Carl suggested the board assume that nine
years was the minimum construction period for a 3 GW nuclear power station, given design
had been completed, permissions were in place, and so on; construction would take the
nine year minimum duration; they wanted to model the implications of committing now
to the construction of this nuclear power station with the probability tree of Figure 8.9 as
summarised by Table 8.4 representing their current view of load growth over the next nine
years, with trends established over these nine years probably carrying on; and contingency
plans to cope with a departure from the expected load growth of 3 GW were reflected in
the costs per kilowatt- hour associated with the possible load growth outcomes over the GW
range shown in Table 8.5.
The second column of Table 8.5 associated a load growth of 3 GW over the nine year
period with 10 cents/kWh, consistent with the expected cost discussed earlier.
Working outwards from this 3 GW expected load growth outcome in an upwards direc-
tion, gigawatt load growths of 2, 1 or 0 were associated with cents/kWh costs of 12, 15
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 445
Table 8.5 Cost outcomes linked to load growth outcomes.
Load growth, GW Cost, cents/kWh Probability (density) from Table 8.4
Product
0 1 2 3 4 5 6
20 15 12 10 11 12 13
0.006 0.084 0.246 0.328 0.246 0.084 0.006
0.120 1.260 2.952 3.280 2.706 1.008 0.078
Expected cost, cents/kWh (column total) 11.404
and 20 respectively. The 12 cents/kWh reflected an assumption that if a load growth of
2 GW occurred, the most effective contingency plans were expensive, but not hugely so.
However, if a load growth of only 1 or 0 occurred, the whole lifecycle cost per kilowatt- hour
increased in a non- linear manner. The zero load growth scenario assumed a new nuclear
power station or other existing stations intended for base- load generation (24 hours per
day) being used for shoulder periods and peaks (perhaps 10 hours per day) for many years,
perhaps shutting down the lower merit order plant completely, only recovering from this
position when existing plant retired if the zero growth continued. There was a less extreme
variant on this theme if 1 GW of growth occurred. All three of these scenarios assumed the
full cost was attributed to the new nuclear power station in question because it gave rise to
a cost which alternatives might avoid.
Similarly, working outwards from the 3 GW expected load growth outcome in a down-
wards direction, gigawatt load growths of 4, 5 or 6 were associated with cents/kWh costs
of 11, 12 and 13, respectively. The 13 cents/kWh reflected an assumption that if the 6 GW
growth case occurred, the most effective contingency plans were expensive but much less
expensive than the zero load growth case, and the intermediate 11 and 12 cents/kWh
assumptions reflected a simple linearity assumption. In the 6 GW growth case, he assumed
Canpower would have to use a source like outdated base- load plant due to retire or new
peak- load gas turbines to meet the unexpected load until the nuclear power station under
consideration was on stream, the additional cost during this period being attributed to the
nuclear power station cost per kilowatt- hour. However, a second nuclear power station or
some other new base- load equivalent would be initiated part way through the nine year
period if load growth looked as if 3 GW might be exceeded, reducing the impact. There
were less extreme variants on this theme if 4 or 5 GW growth scenarios occurred.
As with earlier examples, more precision could be used, different numbers could be
argued for, and Carl indicated that he would not wish to defend the numbers used for
illustrative purposes. But the importance of the asymmetry of the penalty costs associated
with being early or late with a nuclear power station reflected in Table 8.5 was worth under-
standing, as was the increase in expected cost even if symmetry was assumed for the penalty
costs. Indeed, whatever the nature of the penalty cost distribution, in practice Canpower
medium- term planners would need to undertake extensive contingency planning exercises
to both estimate penalty costs without undue bias and prepare robust plans to deal with the
implementation of their plans, a Canpower capability not currently in evidence.
The 14% increase in expected cost per kilowatt hour portrayed in Table 8.5, from 10
to 11.4 cents/kWh, was clearly relevant. It implied a 14% increase in expected cost per
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
446 Employing planning tools in practice
kilowatt hour associated with imperfect knowledge of demand growth in conjunction with
a nine year lead time for a nuclear power station, given a policy of building nuclear power
stations in line with expected load growth.
The asymmetry of the penalty costs was also important because it suggested that a policy
for the medium- term planning process which involved never attempting to meet the whole
of anticipated base- load growth with nuclear power stations was the optimal approach. The
expected cost was lower, meeting such anticipated load growth at least, in part, with non-
nuclear, and then bringing in nuclear stations to take over established base load. That is,
deliberately lag the introduction of the new nuclear plant because if being late is less expen-
sive compared to being early, it pays to be late, although there are still costs involved. The
extent to which it pays to be late is a function of the relative penalties associated with being
early or late.
Furthermore, whether or not asymmetry of the penalty costs was involved, this example
made it clear that if there was a way of meeting commitments which did not involve large
units and lags, it would be worth paying a premium to use it.
Table 8.6 illustrated what the modelling of a policy using alternative lags might imply.
The 0 years (0 GW) ‘target capacity lag’ column corresponded to planning to meet the
expected load growth of 3 GW over the assumed nine years of construction with zero
planned lag in terms of years or gigawatts. In this case the Table 8.5 results were repeated,
including the expected cost of 11.404 cents/kWh.
The 3 year (1 GW) target capacity lag column corresponded to starting three years later
with a planned lag of 1 GW. In this case an ‘on target’ 2 GW load growth implied a cost of
10 cents/kWh, consistent with Table 8.4. The rising costs if load growth was 0, 1, 3, 4, 5 or
6 GW (2 or 1 under target or 1, 2, 3 or 4 over target) were consistent with Table 8.4, with a
slight extrapolation. The expected value of 11.294 involved a decrease in expected cost rela-
tive to the 11.404 of Table 8.4 and the 0 years (0 GW) case of Table 8.6. Planning as if 2 GW
of growth will take place over nine years was an improvement on assuming a mean growth
rate of 3 GW, as might be expected given the assumed asymmetry of the penalty costs.
The 6 year (2 GW) target capacity lag column corresponded to starting six years later
with a planned lag of 2 GW. In this case the rising cost if load growth was 6 GW (5 over
target) was also consistent with Table 8.4, with a longer extrapolation. In this case the same
approach had been used. However, comparable treatment of planning as if only 1 GW of
growth would take place was not an improvement, and it was worse than planning for 3 GW.
Table 8.6 Cost outcomes linked to alternative levels of ‘target capacity lag’.
Load growth, GW
Probability (density) from Table 8.4
Cost, cents/kWh, given a target capacity lag in years (GWs) as indicated
6 years (2 GW) 3 years (1 GW) 0 years (0 GW)
0 1 2 3 4 5 6
0.006 0.084 0.246 0.328 0.246 0.084 0.006
12 10 11 12 13 14 15
15 12 10 11 12 13 14
20 15 12 10 11 12 13
Expected cost, cents/kWh 12.018 11.294 11.404
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 447
These results suggested an optimal strategy involved a deliberate lagging of load growth
in terms of new nuclear capacity for about three years, assuming a nine year construction
lead time and an expected increase in demand of 3 GW over these nine years, 1 GW every
three years. This lag would involve an expected cost saving per kilowatt hour of about 1%
(the difference between 11.404 and 11.294) relative to scheduling construction of a 3 GW
capacity to meet to expected demand at the end of year nine. A 1% cost saving was clearly
not of great importance, but even with this optimised lag policy, the nine year construction
lead time involved, in conjunction with demand uncertainty, added about 13% (the differ-
ence between 11.294 and 10) to the expected cost of 10 cents/kWh. A 13% bias favourable
to nuclear power might be significant. In particular, if a short lead time option with a lower
expected cost than nuclear was available, it should be seen as preferable.
The load growth distribution used for this example was assumed to be symmetric, to
focus attention on the asymmetric penalty costs. If the load growth distribution was not
symmetric, this would complicate the situation, and two sources of asymmetry reinforcing
each other might become much more significant. The same principles would apply, but the
1% improvement via the lag optimisation might become a more significant 3% or 4%. Carl
made the point that asymmetric functions, such as the one for cost illustrated here, were
quite common. Sometimes they could be dealt with effectively on an intuitive basis, but
sometimes it was worth modelling them to test intuition. It was not intuitively obvious that
planning to bring nuclear power stations in later than needed was sensible. Indeed, it con-
flicted with Canpower’s traditional bias in favour of trying to bring them in early ‘to ensure
the lights never went out’.
Integration upwards
Even the preferred option indicated by Table 8.6 involved a significant increase in the
expected cost and some increase in risk. Carl suggested that the lag optimisation effect
as portrayed so far was not of huge importance. However, what was very important was
embedding this optimised expected cost and risk increase in Table 8.1. It was crucial to
consider these optimised increases in expected cost and cost risk in the long- term plan-
ning process discussed earlier. Current Canpower practice would not involve attributing the
additional costs associated with temporary use of gas- fired units to a nuclear power station
as assumed for Tables 8.5 and 8.6. Nor would current practice attribute the full costs at
the margin to the system of completing a nuclear power station prematurely as assumed
for Tables 8.5 and 8.6. But if such costs must be borne by the system and they stem from
the particular nuclear power station decision, they needed to be considered in the context
of long- term planning if alternative sources of electricity did not involve comparable costs.
Small hydro units or small gas- fired units that could be constructed very quickly did not
involve such large additional expected costs and risk of this kind. In the latter case, relatively
low capital cost and savings in fuel costs if the gas- fired units were not operated contribute
to a reduction in penalty costs (over and above the shorter construction period). What this
implied was the need for an ‘integration upwards’ process which adjusted the long- term
strategy formulation process to reflect medium- term strategy formulation impacts. This, in
turn, implied a need for a second pass through the long- term strategy formulation process
to incorporate what was learned in the medium- term planning process.
Linking this to the approach outlined earlier based on Table 8.1 for nuclear power, with
comparable distributions for all other power sources considered, suggested that this ear-
lier modelling ought to be extended to consider these additional system costs and risks
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
448 Employing planning tools in practice
associated with new supply implementation lags. This might or might not involve embed-
ding them directly in the cost per kilowatt hour distributions of individual sources. The
effect would have to be the same: a shift in the preferred mix, away from sources involving
large power stations with high capital costs and long lead times, towards sources involving
smaller power stations with proportionately lower capital costs and shorter lead times.
In practice, models of this kind to assist an electricity utility with medium- term system
planning would have to examine reasonably complex decision rules associated with contin-
gency responses, to develop appropriate rules to manage the medium- term planning deci-
sions involved, and to provide a basis for capturing the dependence upward illustrated by
the numbers of Tables 8.4 to 8.6. And in practice, as further medium- term planning com-
plications were addressed, the use of commercial Monte Carlo simulation software might
become a clarity efficient replacement, to generalise the computational aspects of the simple
illustrative models used here. However, for some exploratory and presentational purposes,
at a conceptual level it would be very worthwhile avoiding simulation, exploring trees and
tables using HAT models in the forms employed by these examples. They clarified the struc-
ture of the models required and the key working assumptions which the board needed to
understand.
Modelling other medium- term planning uncertainty concerns
Medium- term planning also needed to consider other sources of uncertainty which had
important implications for the timing of commitments to action plans plus linked provi-
sional plans, and some could involve further dependencies upwards. For example, supply
uncertainty driven by construction period uncertainty was particularly important for nuclear
power stations, because significant uncertainty over long construction periods clearly exac-
erbated the demand uncertainty issues of the last section. Canpower had significant difficulty
with completing all big projects on time. This situation would need dramatic improvement,
which was feasible. However, even assuming this was done, the risk of significant departures
from expected dates would remain an important source of uncertainty in terms of matching
supply and demand, perhaps increasing the 13% just observed to 15% to 20% or more.
Supply uncertainty driven by reliability concerns also needed attention, perhaps further
increasing the 13% to 20% to 30% or more. For example, Carl suggested the board con-
template the implications of an expected load growth of 3 GW becoming a realised growth
of 6 GW plus the planned 9 year construction period becoming 12 years plus an extended unplanned outage period for a base load power station currently delivering 3 GW.
He would not further develop medium- term uncertainty models today, but he hoped
that it was now very clear that all relevant uncertainty needed explicit attention when mak- ing commitments to the full range of medium- term decisions involved in matching supply
and demand. The examples he had used illustrated the use of a HAT framework in a very
different manner to its earlier use for long- term planning, to address a very different set of
questions. Canpower currently lacked the capability to address medium- term uncertainty
issues effectively, in addition to its long- term planning failures.
Carl finished his discussion of medium- term planning by emphasising the usefulness of
explicit uncertainty modelling in a framework focused on timing, the bias associated with
currently omitted ‘upward integration’ adjustments to expected costs and associated risk,
and the very different nature of the issues addressed and the models required for medium-
term planning relative to a portfolio- driven approach to long- term planning. He made
a particular point of emphasising that the long- term planning process had to focus on
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 449
‘portfolio issues’ in the sense of trade- offs between cents/kWh expected costs and risk expo-
sure associated with different proportions of hydro, nuclear, thermal and so on, leaving the
medium- term planning process to focus on timing issues and minimising the expected cost
of implementing the long- term strategy. However, links and interdependencies between
them needed addressing via multiple pass approaches, and a HAT- based framework with
Monte Carlo simulation options was a flexible toolset for addressing all the relevant issues.
Short- term planning and integration upwards
Carl began his treatment of short- term planning by reminding everyone that short- term
planning was about ‘keeping the lights on at the lowest feasible cost given the generation capability currently available’. This involved short- term demand forecasting plus short- term supply availability issues which had to include reliability concerns.
Canpower’s approach to demand forecasting for short- term planning purposes was ‘at
the top of the class’ in terms of industry standards, and posed no concerns. Canpower’s
approach to availability planning was generally good by industry standards so far as he
was aware, but there were two closely coupled serious problems which had an impact on
medium- term and long- term planning needing attention. The availability of generation
capability was assessed for the whole system in a way which needed revision, and the way the
cost for reserve generation capability was attributed needed revision. His next step was to
illustrate how modelling could illuminate the nature of these issues, and help Canpower to
resolve them using simple models.
One important message was large power stations units, and large numbers of units of a
common type exposed to common- mode failure, both involved a greater need for system
spare capacity than small units of different types. A closely coupled message was this increase
in spare capacity involved an increase in system cost which ought to be attributed to the
large units, or large numbers of units of a common type, in terms of an increase in their
expected lifecycle cost when choosing the desired mix of units. Another related message was
large power station units, and large numbers of common units, also increased the risk of
major losses of availability which have to be borne by the consumers. Unlike smaller capacity
losses, it would not be economic to provide spare capacity to eliminate the risk of failures
which affected a large proportion of a system. This risk might be fully borne by customers
rather than Canpower, in the form of unreliable service, but it might be partially borne by
Canpower using compensation payments. But in either case it ought to be reflected in the
long- term portfolio modelling discussed earlier, because it was an important additional type
of risk, unless it could be effectively mitigated by mutual backup arrangements with other
utilities within the province or in other adjacent provinces or US states. For example, a con-
straint of the form ‘no more than X1 percentage of the supply capacity of Canpower should
be of nuclear power station type A1’ might be appropriate, entirely driven by this kind of
reliability/availability risk considerations, perhaps within an overall restriction of nuclear
power as a whole being driven primarily by the cents/kWh cost- risk issues discussed earlier,
perhaps with other background secondary constraints.
Modelling unplanned outage within one multi- unit nuclear power station
For illustrative purposes, Carl asked the board to consider one of the 3 GW nuclear power
stations used earlier, for simplicity, assuming it consisted of three 1 GW units, all to the
same specification, the probability of a 1 GW unit becoming unavailable in any month for
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
450 Employing planning tools in practice
unplanned reasons which were independent across units (such as damage caused by operator
error restricted to a single unit) was 0.01, and the probability of a unit becoming unavail-
able in any month for unplanned reasons which involve common- mode failures across units
within the power station was 0.005. If one unit was unavailable they would all be unavailable
in a common- mode failure context, as would be the case if common systems failed. In prac-
tice, both these probabilities would be functions of time since operation of the units began,
so for illustrative purposes, he was assuming that these probabilities were both averages
over time. Intermediate levels of dependence might also require modelling in practice. Carl
acknowledged that both these probabilities might not be correct, but revisions would not
alter the substance of what these numbers would illustrate, and he had deliberately chosen
very simple round numbers.
Figure 8.10 portrayed a three level probability tree associated with these assumptions, and
Figure 8.11 illustrated the equivalent single level collapsed tree.
Key: n = no outage, i = independent outage, c = common-mode outage
0.99
0.01
n
i
0.99
0.01
0.99
n
i
n
0.01
0.01
0.005
i
i
0.99n
unit 1 unit 2 unit 3
n
n
i
0.99
0.01
0.99
c
i 0.01
i 0.010
n 0.985
0 0.9653985
outages probability
1 0.0097515
1 0.0097515
2 0.0000985
1 0.0098010
2 0.0000990
2 0.0000990
3 0.0000010
3 0.0050000
Figure 8.10 Example outage probability tree for a 3 GW nuclear power station-with three 1 GW units.
outages
0
1
2
3
probability (density)
cumulative probability
rounded probability
rounded cumulative
expected number of outages 0.0449000 rounded to 0.045
contribution of common mode 0.0150000 rounded to 0.015 (33%)
0.9653985
0.0293040
0.0002965
0.0050010
0.9653985
0.9947025
0.9949990
1.0000000
0.965
0.995
0.995
1.000
0.965
0.030
0.000
0.005
Figure 8.11 Collapsed single level probability tree associated with Figure 8.10.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 451
These two probability trees and the associated rounded values indicated a high probabil-
ity of no outage (0.965), a significant probability of one outage (0.030), and virtually no
chance of two or three outages because of independent outage sources (rounded to 0.000),
but a probability of three outages due to common- mode failures of 0.005. The probability
of three outages due to common- mode failure of 0.005 was small, but large enough to be
of serious concern given the consequences. The bimodal nature of this distribution might
be counter- intuitive, and it was worth trying to understand why it arose, as a general feature
of all common- mode failures, using mode in the sense of a ‘mechanism’ or ‘set of causes’. Earlier in this chapter mode was used in the sense of a ‘most likely’ value associated with a single ‘peak’ value, a somewhat different meaning which can be confusing.
In addition to the risk associated with losing all three units at the same time, Figure 8.11
showed that these common- mode failures contributed about a third of the expected number
of outages in any month, a perhaps surprisingly high proportion.
Carl then assumed that any of these outages might be associated with the duration of out-
age probability tree associated with Table 8.7, to illustrate the overall impact.
Table 8.7 used a HAT approach which made explicit use of continuous variable ranges
plus a non- linear scale for outage duration. At this point Carl indicated that he could explain
the rationale and implications explored in Part 1 for this kind of modelling at a later date
if they were interested, linked to but slightly different from that used for Table 8.1 and
Figure 8.2.
The expected values from Figure 8.11 and Table 8.7 could be used to compute the
expected unit months of outage per month of operation on start- up, 0.045 × 4.4 = 0.198.
This approximated the expected unit months of outage per month of operation once the
system’s operation reached a steady state (which could be computed using a semi- Markov
process). It was slightly higher than the result a steady state calculation would provide,
because when units were not available the probability of additional units becoming unavaila-
ble falls. Availability per unit using this result was 93.4% (obtained by 100 × (3 – 0.198)/3),
about 95% if a steady state calculation was used. This implied an increase in expected cost
per kilowatt- hour of just over 5% to meet the expected level of outage. This increase was
normally associated with the units in question through the standard treatment of availability.
The risk of losing large units required a further increase in expected cost associated with
additional reserve capacity provision to that required if smaller units were used, which
was not normally associated with the units in question. Furthermore, it required another
increase in cost associated with the risk of this increased reserve proving inadequate and
requiring contingency demand management or other responses with significant costs borne
by electricity consumers, also not normally associated with the unit in question. The risk of
being unable to meet demand was inevitably increased if larger units were used because it
would be uneconomic to scale up reserve capacity in direct proportion.
Table 8.7 Example outage duration probability distribution.
Outage duration (months) Associated range Probability Product
0.5 2 6 24
(0 to 1) (1 to 3) (3 to 12) (12 or more)
0.4 0.3 0.2 0.1
0.2 0.6 1.2 2.4
Expected value 4.4
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
452 Employing planning tools in practice
Modelling unplanned outage for a set of nuclear power stations of a common type
Carl then turned his attention to the situation if this 3 GW nuclear power station was one of
four similar stations, each of 3 units, making a total of 12 units. He assumed that the same
probabilities of independent and common- mode outage within a single station applied.
He also assumed that one in ten of the within- station common- mode outages involved a
common- mode outage across all four stations. Such common- mode outages would include
the discovery of design defects with safety implications requiring immediate shutdown of
all stations using the same design, for example. One in ten is a monthly probability of only
0.0005, which is very low. However, the impact would be substantial.
The implications of these across- station common- mode outages included a direct contri-
bution to the expected cost of outages already costed in relation to station availability and
normally included in the cost attributed to a nuclear power station, plus all the additional
costs considered in the last subsection. The implications also included a very much larger
system reserve to keep the consequences of such an outage acceptable, and the cost of this
reserve was directly attributable to the use of a number of nuclear power stations of the same
type, although it was not usually associated with them in this way. More important was the
large increase in expected cost associated with such common- mode failures which would
be borne by customers which should be considered when choosing the mix of sources of
power, although this was not usually done. More important still, if the consequences of
such a very large loss of availability were unacceptable for any reasonable increase in reserve
capacity, from a regulator’s perspective it might be critical to restrain a utility from undertak-
ing any developments which implied a plausible possibility of common- mode failure on this
scale. That is, externally imposed constraints of the form ‘no more than X1 percentage of
the available supply of a utility should be from nuclear power stations of type A1’ may arise
because of availability risk rather than the cost risk considered earlier. Such constraints might
be imposed after a major incident, if the regulator was not aware of the possible implications
in advance.
The additional reserve costs, and the high costs of responses borne by consumers, clearly
needed to be considered by Canpower, and both the expected cost implications and the risk
implications needed to be considered and attributed to the nuclear power stations in the
context of long- term system planning. In practice, the ‘consumers’ who suffer most may be
electricity- dependent industries. Their boards ought to be aware of the issue and be politi-
cally vocal and proactive if appropriate, as should their employees and customers.
Modelling unplanned outage for small hydro units
Carl then considered the situation if a 3 GW nuclear power station was replaced by 300
relatively small units of hydropower. For simplicity he assumed the same probabilities of
independent and common- mode outage applied, and the same outage duration distribu-
tion. In this case common- mode outage would include a drought affecting all 300 units,
for example.
The impact of the large number of units in relation to independent sources of outage was
a marked increase in the predictability of outage, but the common- mode outage issue was
not changed. Using the 95% availability figure, they could expect about 15 of the 300 units
to be out at any one time, with very little variation relative to the nuclear power station risk
of having one or more units out at any one time for reasons not involving common- mode
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 453
outage. This meant that Canpower did not have to provide as much reserve for individual
unit failures, with lower expected cost per kilowatt hour implications. However, Canpower
would still have to worry about the risk of common- mode outage and its expected cost
implications.
To consider the impact of the larger number of units in relation to the common- mode
outages Carl assumed the same 0.005 probability of losing all 3 GW, with the risk that
reserve capacity provision would prove inadequate, requiring contingency demand manage-
ment responses with costs borne by electricity consumers. If Canpower reduced reserve
provision because of the larger number of units and the more stable outage pattern, then the
cost implications of common- mode failure in terms of costs borne by electricity customers
become even more serious than in the nuclear power station case. The costs associated with
restoring hydropower are likely to be rather less than those associated with common- mode
failures within a nuclear power station, and very long outage periods are less probable, but
the short- term impact on the consumer is similar. This strengthened the case for inter- utility
collaboration across provincial and national boundaries.
As with nuclear power stations, to some extent common- mode outages needed to be
considered in relation to all existing units, as well as all proposed new units. That is, a serious
nuclear incident could lead to the recognition of design defects which resulted in all nuclear
power stations being shut down, and a long- term drought might seriously affect most of an
electricity utility’s hydro facilities. In practice the probability of common- mode outages on
this scale, and on a smaller scale, might be very different for different sources of power. If,
for example, nuclear power was very exposed to this kind of common- mode outage relative
to hydro (or vice versa), then failing to address these issues biases the case for one option
and against the other.
Modelling unplanned outage for a portfolio of electricity power sources
Carl next considered the situation if the four 3 GW nuclear power stations considered
earlier were replaced by one 3 GW nuclear power station, the 3 GW of small hydro as
just discussed, 3 GW of gas- fired power, and 3 GW of conservation. What were the
implications?
If the gas units were midway between the nuclear and small hydro units in terms of size,
and the same probabilities applied, the situation with respect to the 3 GW of gas power
will be intermediate in terms of the characteristics described earlier in relation to 3 GW of
nuclear and 3 GW of small hydro. In this case, common- mode failure might involve loss
of gas supply, for example. On the other hand, the 3 GW of conservation raises quite dif-
ferent issues. It might be argued that if conservation was unpredictable it should contrib-
ute to reserve requirements, but most demand management concerned with conservation
was aimed at smoothing demand variations as well as reducing average consumption. This
implied an expected cost reduction as well as a risk reduction in terms of the contribution
of conservation to the portfolio considered by the long- term system planning process com-
parable to a hedging effect.
A key difference in the situation for the 12 GW as a whole was the absence of common-
mode failures outside the nuclear, gas, and hydro subsets. It was for this reason that regula-
tors might wish to constrain the proportion of electricity produced from any single source
which was subject to the possibility of common- mode failures.
Again, this realisation might follow a major outage. However, it would clearly be prefer-
able for everybody if it was anticipated before a major outage.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
454 Employing planning tools in practice
Explicit risk sharing with customers and other utilities
Carl now made the point that outage costs borne by customers should be considered by
Canpower when determining strategy. However, sometimes it might be in the customers’
interests to share outage costs as part of an efficient approach to risk sharing. When this was
the case it might be important to make it clear to both customers and regulators that to do
otherwise would involve increasing the cost per kilowatt hour charge to an extent which
most customers would not support – they would prefer the outage risk.
Developing this a bit further, if customers wanted Canpower to insure them against the
consequences of major outages by paying compensation, Canpower would have to charge
a premium for that insurance which would involve expected costs plus administrative costs.
Canpower might argue that this would not be good value for money for customers.
Alternatively, if customers wanted Canpower to avoid the risk of outage by increasing
spare capacity, Canpower might have to charge even more. Canpower might argue this
would be even worse value for money for customers.
One way to mitigate this effect would be explicit risk sharing arrangements with other
major utilities in other adjacent provinces or US states. Such arrangements might help to
cope with shortages for combinations of reasons – outages in conjunction with delayed
project completions and unexpected load increases – a form of general response to build in
robustness which includes ‘any other unspecified problem’. However, common- mode fail-
ures and common reasons for other problems would still need consideration.
Explicit costing of regulatory constraints
Carl also made the point that costs associated with regulator/government constraints on
the make- up of Canpower’s portfolio imposed for reliability/availability reasons would be
passed on to the customers. Such costs could be identified and measured in a shadow cost
framework, to test the appropriateness of the constraints as well as providing a simple sys-
tematic basis for embedding them in the long- term planning process.
An overview of integrated short /medium/long- term planning
Modelling short- term availability issues as addressed in the last section was clearly a useful
basis for considering short- term corporate planning issues. However, in the context of an
overview of what integrating short- term, medium- term and long- term planning involved
for Canpower, the implications for long- term planning were also extremely important – the
primary reason for raising these issues that day. The need for reserve capacity was a function
of the size of the units involved and the extent to which common- mode outage could occur,
plus the availability and effectiveness of general responses based on collaboration across
provincial and national boundaries. As a consequence, all reserve capacity costs should be
attributed to types of unit, as should the risk of costs borne by electricity consumers as a
consequence of involuntary demand management (or other responses), in terms of both
expected costs and associated risks. Outage risk might also require limits on the proportion
of power provided by any single source in the overall system. As an extreme example, if one
kind of nuclear power station had the lowest expected cost in cents/kWh, and it was the
most reliable of all available sources, diversification might still be absolutely essential just to
avoid the risk of a design fault being found after the system was 100% this source. A shut-
down of the whole system because of a design fault affecting 100% of the system would
clearly be catastrophic.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 455
A portfolio view of the system in reliability terms was as important as a portfolio view of
the system in cents/kWh cost terms or a national security perspective based on fuel avail-
ability if potential hostilities threatened fuel supplies or technical dependence if the country
of origin of a supplier of part of the generation portfolio became belligerent. These issues
should be considered explicitly by Canpower within its long- term planning framework. If
they were not, it would fall to a perceptive regulator or government to take appropriate
action, if either was aware of the issues and motivated to act effectively.
Carl’s portfolio- driven approach to long- term planning accommodating more than one
attribute plus more than one criterion per attribute facilitated this. Although the analysis
needed was clearly made much more complicated as additional attributes were addressed,
whatever really mattered could and should be considered.
It is always crucial to be able to both recognise and deal with everything that really mat-
ters, a central feature of Carl’s recommended EP approach. Carl might have observed that
people who do not know how to deal with difficult issues tend to avoid looking for them,
and sometimes they do not recognise them even when they are blindingly obvious or they
bump into them. Carl did not do so to avoid antagonising the board. But I will, because you
should find it useful to remember this observation in a wide range of contexts. To see this
tendency in ourselves as well as others, and deal with it as effectively as possible, is arguably
a core aspect ‘enlightenment’.
Futures planning of all kinds and its integration at all levels
Goals planning plus futures planning jointly inform long- term, medium- term and short-
term planning. Carl suggested that once a mature approach to goals planning was in place,
futures planning was usefully seen as a layer of planning primarily driven directly by goals
planning, but making use of cost- gap information feedback involving shadow costs from
long- term, medium- term and short- term planning.
Futures planning could itself be associated with a top- level to bottom- level internal struc-
ture which was comparable in some respects to the long- term, medium- term and short- term
horizons of planning currently feasible options but very different in others. It could also be
associated with project lifecycle stage structures with features comparable to the bidding stages
of the Astro tale in Chapter 6. For example, ‘horizon scanning’ could be seen as a largely pas-
sive process of assessing technology developments by others with a view to possible Canpower
usage or research and development for usage; ‘futures research’ could be seen as a higher
level of resource commitment for selected areas needing clarification research prior to possible
‘futures development’, which, in turn involved significant development expenditure to make
new technology feasible or economically viable. Gateway stages between ‘scanning’, ‘research’
and ‘development’ futures planning process stages could ensure unbiased assessment of the
options to progress propositions to a following stage or not.
However, there were a number of reasons why a much more complex interdependent per-
spective might be relevant in any context. In a Canpower context, Carl was convinced that
an approach addressing Canpower’s capability- culture assets and liabilities, as well as related
external concerns, was crucial, with an overall emphasise on ‘designing a desirable future’
and exploring the feasibility of delivering it with an in- house variant of entrepreneurship
some people call ‘intrapreneurship’ driving Canpower’s innovation agenda. One feature of
this kind of futures planning he believed especially important was a concern for develop-
ing awareness of threats to Canpower’s public ownership status. Carl believed these threats
might be transformed into Canpower opportunities.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
456 Employing planning tools in practice
What Carl believed was crucially important for the Canpower board to begin to under-
stand immediately was the need for a proactive top- down approach to ‘designing desirable
futures’ based on the premise that all organisations needed a clear understanding of the
opportunities and risks associated with their ‘business model’ which summarised their most
important goals, in particular, their core values, in terms of their key operating rules and
policies. This ‘business model’ needed robust assumptions and clarity about key corpo-
rate strengths and weaknesses. Understanding Canpower’s key strengths and weaknesses
required corporate self- knowledge. It also required a well- founded appreciation of their
potential political allies and similar knowledge of crucial political opponents, fully under-
standing that potential allies could prove indifferent bystanders or enemies if not treated
appropriately.
What Carl had in mind was starting to address Canpower’s key fault set number three – a
serious capability- culture liability which he believed needed fixing immediately. He did not
put it quite so bluntly, but he made no effort to hide his concerns. To briefly illustrate what
he had in mind he suggested they started by looking at some of the key criticisms of Can-
power currently receiving public attention and media focus.
Influential international ‘green’ movements like Greenpeace were strongly opposed to
nuclear power. In Canpower’s province, these groups were also strongly opposed to public
sector ownership of Canpower, although they did not state this opposition explicitly. Their
opposition to a public sector ownership position arose, in part, because the low cost of capi-
tal associated with provincial government bond funding of Canpower assets made electricity
cheaper than it would be if Canpower used private sector funding, which made consumers
relatively wasteful. This effect was amplified by the implicit national government subsidy of
nuclear power associated with further hidden indirect costs linked to nuclear waste disposal
and potential nuclear accidents. Green movement supporters would not openly criticise
Canpower for its low costs because they did not want to draw attention to the benefits of
public sector ownership in terms of low electricity costs, but they were indirectly support-
ing calls for the privatisation of Canpower as a route to removing all subsidies for electricity
costs and openly attacking Canpower for lack of leadership on green concerns. Some of
Canpower’s owners/customers (the citizens of its province) were beginning to sympathise
with the Greenpeace position.
Vocal private sector power producers and potential members of this group wanted less
nuclear power and higher electricity prices for straightforward commercial reasons. They
would not openly attack Canpower for its low costs, but they were indirectly attacking
the public sector status of Canpower and openly attacking Canpower for its restrictive and
obstructive approach to private power (within province non- utility power) because Can-
power were visibly unhelpful when small rural communities might benefit from small scale
local private power projects. When renewable sources were involved, this exacerbated Can-
power’s current lack of ‘green credibility’.
Many economics, finance, and political media commentators who were critical of Can-
power were convinced that public ownership of utilities was not a good idea for purely
ideological reasons, and they tended to see the low capital cost of public ownership as a
threat to private enterprise, with scope for potential application in many other industry sec-
tors, not as an opportunity for private enterprise which needed electric power to produce
their services and products, arguably much more relevant. These critics of Canpower usually
did not even mention the low cost of capital benefits, seeing this as an argument put by its
ideological opposition which needed ignoring or discrediting, a variant of the strategic-
misrepresentation euphemism discussed earlier. Some of these critical media commentators
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 457
actually believed economic theories which suggested public and private sector cost of capital
should be seen as the same, a reasonably clear indication of the effectiveness of the misrep-
resentations involved. Many of Canpower’s owners/customers were either misinformed or
confused by the media or the textbook treatment of the cost of capital discussions they were
exposed to in their earlier education.
Some of Canpower’s owners/customers who worked as non- union tradespeople saw the
people doing similar jobs employed by Canpower as overpaid and overprotected from the
marketplace due to overly powerful union pressures as a direct result of its public sector
status. Some of Canpower’s owners/customers who worked as private sector managers or
entrepreneurs saw both the unionised employees of Canpower and their managers as over-
paid and overprotected from the marketplace as a result of its public sector status.
Most of Canpower’s owners/customers were ambivalent about the strengths and weak-
nesses of public versus private sector status for Canpower, but a combination of adverse
pressure group rhetoric, misinformation, political opportunism, and growing public con-
cern about green issues was building a groundswell of hostility to Canpower in general, and
its public sector status in particular.
Canpower’s obstructive approach to private power producers, its weakly grounded plan
for six new nuclear power stations, and growing evidence of its lack of competence when
costing and planning other new developments were crucial issues the board needed to
address. They were key components of Canpower’s dangerous reinforcement of a complex
set of factors making privatisation a serious and growing threat. Carl believed privatising
Canpower would be a needless social disaster, a catastrophe easily avoided provided Can-
power could effectively meet all genuine criticisms and disarm all unfounded prejudices.
He had no strong personal prejudices or feelings about private sector versus public sector
ownership in general terms, but he was convinced that in most respects Canpower was a
national treasure – an excellent example of what a good public sector utility ought to be,
and the recent privatisation of the CEGB in the UK provided a directly comparable example
of a fundamentally flawed approach which Canpower’s province should avoid. He believed
Canpower needed to design a future which would eliminate all the genuine criticisms of
Canpower which contributed to the case for its privatisation by its enemies, make allies of
some of those currently hostile or indifferent, and plan the realisation of this future.
Goals planning as a starting point for all planning
Carl reminded the board that at the outset of this meeting he had suggested goals planning
should be the starting place for all planning. The rationale was quite simple – whenever goals
planning was not the explicit starting point, by default it was the implicit starting point, with
an implied failure to test the implicit planning assumptions.
Goals planning should be followed by futures planning in designing a desirable future
mode, followed by other aspects of futures planning. Again, implicitly doing so was the impli-
cation of not doing so explicitly, with an implied failure to test all associated assumptions.
Futures planning should be followed by long- term, medium- term, and short- term plan-
ning. Goals planning also had to drive long- term, medium- term and short- term planning
directly. Complete downwards passes should then be followed by ongoing iteration in a
planned and orderly manner, integrating bottom- up dependencies, along with other adjust-
ments and elaborations which testing the robustness of the analysis as a whole suggested
were needed. He would return to the idea of orderly cycles later. His immediate concern
was starting to revisit the minimum clarity position on goals planning which he had used
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
458 Employing planning tools in practice
earlier. He had delayed beginning to address more sophisticated goals planning because the
changes implied were radical and likely to generate significant controversy. He thought the
changes discussed earlier needed clarity first because they were relatively straightforward and
unequivocal.
He believed the board now needed to begin addressing goals planning for Canpower
from the perspective of its owners, explicitly asking what Canpower’s key goals and core
values ought to be from this perspective, and what crucial operating rules and policies might
flow from this perspective. Before long it would be crucial to ensure that there was board
level agreement to a formal summary of the results of an effective development of this exer-
cise which they could agree with the government and their regulator and then publish, with
a view to responding positively to any constructive public criticism by Canpower’s ‘owners’.
Canpower’s ‘owners’ were ‘the citizens of the province’. For the most part, their ‘own-
ers’ were also their ‘customers’, although some ‘owners’ produced electricity for their own
use and sale to others and some ‘customers’ who operated businesses within the province
were not citizens of the province. He suggested they use the term ‘owners/customers’ to
start talking about key goals and core values that day, treating owners and customers as the
same for most practical purposes, but testing the assumption that there might be differences
between owners and customers that matter when relevant.
It seemed reasonable to Carl to assume that their owners/customers wanted Canpower’s
key goals and core values to be ‘serving the needs of owners/customers in terms of a low
cost, reliable and secure electricity supply provided in a responsible manner and charged for
in a fair and transparent manner – today and for the foreseeable future’. They might debate
and refine this statement – but it was not worth doing so that day.
He suggested they started from this position and ask the question ‘What are the crucial
operating policies and rules this suggested?’ These crucial policies and rules would in effect
define the basis of ‘Canpower’s desirable future’ at a corporate strategy level.
They might begin with the crucial policy understanding that Canpower had to provide a low cost, reliable and secure electricity supply, but Canpower did not have to generate all that supply.
They might then link this to the crucial rule that Canpower had to maintain responsibility
for planning and implementing an ongoing coordination of the portfolio of electric power
sources of all kinds for their province, operating in close cooperation with their provincial
regulator under the supervision of the provincial government and working closely with the
federal government and relevant national regulators (on nuclear safety for example).
He thought that an equally crucial set of rules should be based on the assumptions that
Canpower needed to maintain responsibility for owning and operating all electricity trans-
mission capability and connections to the grid, as well as all customer- facing operations,
including dealing with billing, relationship management during outages, and all aspects of
safety and security. He pointed out that all these rules were broken by the privatisation of
the CEGB, so all the assumptions involved were clearly debatable, but he believed they were
crucial working assumptions underlying the value of public sector status for Canpower to
its owners/customers.
However, private sector companies might take over some aspects of the operation of
power transmission assets and some aspects of customer- facing operations – Canpower’s
employees and their trade unions did not own Canpower as a consequence of their jobs,
although employees were also citizens of the province, and as employees they were inter-
ested parties who needed very careful consideration for a number of reasons. Although
outsourcing in any Canpower context would need considerable care and caution, and there
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 459
was a good case for avoiding even raising the issue as a possibility in the immediate future,
outsourcing was an important option which needed thorough exploration in the longer
term. When it was discussed they should start from a clear acknowledgement of the ethos
of service to the public which all Canpower employees were justly proud of which should
be preserved and enhanced – treating it as an important capability- culture asset. But they
might embrace a possible ‘friendly float- off ’ variant of the Astro ‘friendly takeover’ which
Martha discussed in Chapter 6, and float- off strategic partners that might be owned by their
employees, along the lines Ajit suggested TLC ought to consider, as well as accepting that
direct market competition in some areas made sense for its owners.
For example, Canpower did not need to own and operate all power stations. In simple
‘traffic light’ terms there was a blue– green– red light spectrum with intermediate shades in
terms of private sector generation of power they should vigorously encourage (assign a blue
light), welcomed (assign a green light), or firmly resist (assign a red light).
The extreme deep- red light end of the spectrum might be characterised by the own-
ership of existing large hydropower stations which usually involved exploiting important
geographic assets of the province which the citizens of the province should continue to own
for the benefit of their children and grandchildren as well as current generations. As with
transmission assets, private sector companies might take over some aspects of the operation
of these assets, but he believed outsourcing these jobs would be on the light- red end of a
red– green spectrum of job outsourcing priorities.
Equally deep-red was the case for Canpower maintaining full responsibility for both own-
ing and operating all existing nuclear power stations. It could make an equally strong case
for building and operating all new nuclear power stations, given a working assumption that
Canpower would want to maintain some nuclear power and Canada’s Candu reactors and
associated nuclear technology would remain internationally competitive. However, if the
Canadian federal government did not ensure that the Candu technology remained competi-
tive, the international market for nuclear technology would need consideration. Further-
more, if the federal government made nuclear power too expensive for reasons explored
towards the end of Carl’s presentation, other futures involving less or possibly no nuclear
power would need exploring.
The blue light to deep- green end of the spectrum was small scale local cogeneration facili-
ties. For example, the board might begin immediately by mandating efforts to foster, not hinder, the kind of cogeneration facilities which were currently the basis of private power
in their province, like combined heat and light facilities in timber processing towns which
might produce more electric power than normal local needs by burning wood waste. It
might be crucial to formally and explicitly adopt Martha’s Chapter 6 ‘blue light’ for this
kind of option to emphasise much stronger support than a ‘green light’ might imply.
A jointly optimal blue light approach might go well beyond Canpower just buying surplus electricity to coordinated concentration of the surplus to reduce peak- load requirements
for new Canpower stations, with contracts explicitly facilitating this focus. Canpower could
effectively advertise explicit ‘support’ by visibly erring on the generous side with their con-
tractual arrangements. For example, they could provide free consultancy based on pooling
the experience of other private power producers, treating private power producers as valu-
able members of an extended Canpower team – not competitors with a high nuisance value.
Canpower could extend this supportive approach to local low- head hydropower and
combined cycle gas- burning facilities which might become economically attractive given
Canpower technical support, supportive joint finance, and risk sharing on prices for fuel as
well as electricity purchase prices.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
460 Employing planning tools in practice
Furthermore, they could fund research into and prototype developments of wind and
solar and small scale low- head hydropower. This could be followed by assisting those inter-
ested in developing them in a locally accepted manner when cost- effective contributions to
meeting provincial base- load and peak- load power demands were feasible. The objectives
could include minimising provincial and national expenditure on subsidies for renewables
but helping to ensure their province was at the leading edge of cost- effective use of these
emerging technologies as their cost- effectiveness was established.
Further still, they might encourage, not discourage, appropriate private power produc-
ers to build and operate some significant new facilities they believed the system needed.
Canpower ownership and funding could be preserved, using contracts which reflect enlight-
ened risk sharing. For example, the globally used traditional BOOT contract approach (the
contractor Builds and Owns, Operates, and then Transfers at a fixed horizon like 20 years)
should be seen as a form of contract which was wholly unsuited to any public sector organi- sation, because it means relatively expensive private sector ownership of the capital needed,
also true of the UK Private Finance Initiative (PFI) approach and related Private– Public
Partner (PPP) variants. However, if the BOOT approach was adjusted to reflect the lower
cost of public sector capital which could be used by Canpower, an adjusted version might be
highly suitable. He suggested what was needed might be reflected in the modestly modified
acronym BOLT (the contractor Builds and Operates a new power station using a Loan from
Canpower secured on the asset, and then Transfers the operating contract to Canpower or
another competitor contractor at a review gateway if Canpower is not fully satisfied with
the contractor’s performance, with an agreed- on routine and exceptional gateway structure
built into the initial contract). Candidates might be non- Canadian or out- of- province utili-
ties with new technologies, perhaps working in concert with gas- turbine manufacturers or
gas producers or wind- turbine manufacturers. Appropriate risk sharing contracts would be
crucial, even though private sector capital was not used for the assets themselves because
issues like gas price rises in conjunction with electricity price decreases would remain a
crucial risk issue for private sector participants in public sector managed electric power gen-
eration. There was no point in making private sector partners carry risks they would have
to charge for if those risks were better carried as part of the Canpower portfolio of risks.
Managing all relevant risks for their owners in opportunity efficient terms at a portfolio
level should be a core Canpower role which should not be weakened or threatened by any
changes Canpower supported.
And further still, Canpower might use federal government support and the support of
all the agencies behind the Candu technology to develop Canpower’s nuclear power pro-
gramme but be open to approaches from competing non- Canadian sources if this provided
better value for Canpower’s owners/customers in the long-term.
Wherever new approaches might benefit its owners/customers, Canpower should be
proactively seeking opportunities, recognising that its owners and customers had interests
beyond electricity which might involve useful synergies. To illustrate the scope for such
synergies, Canpower might help the city authorities responsible for rubbish removal from a
city to find a suitable technology and site for incineration facilities which would also produce
community heating plus electric power for the grid, perhaps to be loan funded in part by
Canpower but owned and operated by the city, as a service to Canpower owners in this city
in terms of waste management as well as a benefit to Canpower customers buying electricity
in this city and elsewhere, but perhaps owned by Canpower if the cities involved preferred
this route. A degree of ‘support’ for this kind of activity which had elements of ‘subsidy’
could be socially beneficial for Canpower’s owners in terms which were directly comparable
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 461
to low cost low energy consumption light bulbs, more efficient heating and cooling systems,
and better building insulation. All these changes would involve opening Canpower up to
significant competition plus a degree of diversification, but only where competition and
diversification made sense for Canpower’s owners/customers. Terms which were attractive
to private sector ‘big team’ collaborators would have to be designed – but leadership of the
design of the system and its incentives should remain firmly in the hands of a Canpower
board which protected the best interests of Canpower’s owners/customers.
One way to look at what Carl was suggesting was avoiding the risk of a complete privatisa-
tion breakup like that visited on the CEGB and its owners/customers in the UK by trans-
forming a possible threat into a realised opportunity. Canpower’s board members could do
this effectively if they recognised that their core concern was a reliable supply of electricity at
low cost for the private and corporate consumers of electricity in their province for the fore-
seeable future, produced in a socially responsible manner, overcoming as far as possible all
the relevant weaknesses of public sector monopolies in general, and Canpower in particular,
and capitalising as far as possible on Canpower’s relevant strengths.
The move to supporting some power stations operated by other organisations, in place
of blanket resistance, and outsourcing work on Canpower assets and functions which was
previously not even contemplated, should be seen as part of an associated culture change.
It was crucial to unobtrusively eliminate any evidence that Canpower was managed for the
benefit of its employees. He did not emphasise the obvious fact that ‘employees’ included
Canpower’s board and other managers, but this was clearly implied. They could start by
trying to develop a ‘big team’ approach to the sharp end as soon as possible. For example,
the next time storms brought down power lines and caused local outages, they might give
high public visibility to local contractors working alongside Canpower employees, drawing
on experience gained assisting with routine construction, replacement, and maintenance
work, and they might start planning for this now, involving local contractors much more on
suitable routine maintenance work with immediate effect and scaling back internal teams
as this capability was developed. This kind of approach could be central to the broader aim
of breaking down ‘them and us’ attitudes and union- driven wage differentials. However,
Canpower could also demonstrate the same kind of culture change in terms of some highly
visible board level decisions as soon as possible. He did not elaborate using any of ‘cultural
clues’ Ivor had discussed earlier, but his message was clear.
In summary, Canpower could disarm all their pro- privatisation critics by doing a much
better job of opening themselves to appropriate competition and new environmentally
friendly technology than any fully privatised utility or set of utilities could, saving the gov-
ernment the risk of trying to create and then regulate private sector generation of electric-
ity to do better. They would have to work with their regulators and both provincial and
national governments, but a collaborative approach to all these players should be best for
all relevant parties if Canpower could formulate an appropriate approach. Canpower could
continue to do what it could do well, getting better at it where needed, but it could col-
laborate effectively with those who could help it do even better in some areas.
Goals planning with a capability- culture focus
Carl indicated that he had used criticisms of Canpower as a basis for discussing goals
planning in terms of a redesign of Canpower’s ‘business model’ and all other planning
which flowed from this. All ongoing strategy formulation and testing for any organisa-
tion had to flow from a broadly stated ‘business model’ which defined the organisation’s
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
462 Employing planning tools in practice
key goals and core values in terms of implied crucial operating rules and policies from
the perspective of its owners bearing in mind the interests of other parties. ‘Bearing in mind the interests of other parties’ implied a broad view of the implications of actions
by and pressures from all relevant players needed careful attention, not just the direct
and obvious interests of their owner/customers. A key issue this broad view raised was
‘privatisation risk’.
Privatisation risk was a fundamental threat to Canpower’s business model, but it was
only a threat because Canpower had allowed the way its business model was interpreted
to become distorted by managerial preferences and employee practices which lost sight
of its core purpose to a significant extent. Canpower did not need a radically different set
of values, but Canpower did need to readjust corporate focus and emphasis. To this end,
the board members needed to be very clear about all the key assumptions underpinning
their business model, testing them for robustness, and changing them to more appropriate,
robust, and resilient forms when appropriate.
In terms of a holistic take on the mantra ‘if it ain’t broke don’t fix it’, three things were
badly broken and needed fixing immediately. The public sector status of Canpower was decid- edly not broken and did not need fixing. However, Canpower’s capability-culture was broken in the sense that capability- culture liabilities were a direct reason for current attacks on Can- power’s public sector status, the third of the three crucial flaws which he was dealing with directly that day.
Somewhat different capability- culture liabilities also underlay Canpower’s other two key
flaws, and all three of these capability- culture flaws required immediate attention, dealing directly and explicitly with interdependencies that mattered.
Current practice compared to Carl’s proposed approach so far
After discussing goals planning with a capability- culture focus, Carl indicated this was an
appropriate time to compare his outlined proposed approach with current practice in Can-
power and other utilities. The following subsections provide outlines of the key points which
he developed in more detailed summaries for the board.
Underestimation of variability in long and medium- term forecasts and cost estimates
Canpower had well- developed top- down and bottom- up demand forecasting models, as
Carl had noted earlier. Canpower used these models to make ‘long- term demand forecasts’
in Larry’s current Canpower terminology, ‘medium/long- term demand forecasts’ in Carl’s
terminology. Carl believed these forecasts needed reframing in separate medium- term and
long- term categories of planning and associated planning horizons.
Canpower currently associated these demand forecasts with confidence bands which
were too narrow. They were misleadingly narrow because of the framing assumptions
underlying the statistical models plus the seriously misleading assumption that exogenous
explanatory variables in these models could be predicted without errors. In Carl’s view,
these confidence bands needed massive adjustment to make them much wider, with rap-
idly growing ranges the farther they extrapolated from the data period. Cost estimates
associated with forecasts of energy prices and capital cost estimates had similarly narrow
and unrealistic confidence bands, requiring massive increases, and important interdepend-
encies were involved.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 463
Rejection of robustness- driven scenario approaches
Most other electricity utilities had demand forecasting models which were comparable to
but not as well developed as Canpower’s. Most other electricity utilities supplemented these
models by robustness- driven scenario building approaches, an approach explicitly rejected
by Canpower.
The reason given by Canpower staff for rejecting robustness- driven scenario approaches
was their lack of precision. One option was to believe Canpower staff and draw the conclusion
that they were confusing precision and accuracy. Another option was to reject the Canpower
staff explanation as a deliberate cover- up and assume that what actually motivated Canpower
staff to reject of this kind of scenario approach was a misguided attempt to cling to a largely
deterministic view of a very uncertain situation that they were not equipped to deal with.
There might be other explanations, but whatever the rationale, Canpower’s whole
approach to long- term planning was premised on the validity of narrow confidence bands
for their associated demand forecast, and similarly narrow confidence bands for cost esti-
mates and duration estimates, and neither of these premises was realistic.
Carl chose his language very carefully but made it crystal clear that Canpower staff dem-
onstrated an unacceptably low level of competence in terms of managing uncertainty and
underlying complexity in what they currently called the long-term, usefully reframed as the
medium-term plus the long-term.
Estimation bias for expected values and associated error bands
Canpower’s excessively narrow error bands for demand forecasts over the medium- term
and long- term, similarly narrow error bands for the cost per kilowatt hour estimates for a
preferred set of options, and similarly narrow error bands for project duration estimates,
should be interpreted as direct evidence of optimism bias associated with potential variabil-
ity. This kind of bias implied that expected value estimates were also seriously biased on the
optimistic side. Interdependencies would scale up this bias, long durations and high costs
for projects driving up electricity prices which would depress demand. Demand estimate
expected values were much too high, while cost and duration expected value estimates were
much too low. That is, excessively narrow confidence bands implied both expected outcome
and associated variability estimates were consistently biased on the optimistic side, and inter-
dependencies significantly magnified this effect.
Given these biased expected outcome estimates and biased error bands, Canpower used
elaborate linear programming models to minimise the expected cost of meeting demand
over the medium-term and the long-term. These models simultaneously ordered the
choices in Carl’s long- term planning sense and timed their start of construction in Carl’s
medium- term planning sense. They provided an optimal solution which involved massive
new plant commitments, exclusively large nuclear power stations for base load. However,
the assumptions used involved a massive expected cost estimation bias in favour of nuclear.
Carl’s presentation had suggested this bias mattered. He was not in a position to measure it
immediately, but he would be deeply surprised if future work on estimates, which he would
recommend, did not confirm it was significant.
Simply looking at realistic probability distributions for direct costs, as he had illustrated
by Table 8.1 example A plus the example B and C variants, might uplift currently assumed
nuclear costs per kilowatt- hour by 100% or more. Other sources of bias might make an
appropriate uplift of the expected cost of nuclear power 200% or more.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
464 Employing planning tools in practice
Canpower’s linear programming models would give an entirely different solution if
expected cost estimate bias was corrected. Carl believed nuclear power stations might play
no part at all in such a revised solution, although he personally believed there was an impor-
tant role for nuclear power in Canpower’s portfolio.
Cost risk ignored
Bias associated with expected cost estimates and associated variability at the level of indi-
vidual units was just the ‘tip of the iceberg’ of the problems associated with the Canpower
approach as a whole. Even if the biased expected costs issue was put on one side, and it was
assumed that large nuclear power stations really did provide the lowest expected cost per
kilowatt- hour over the long term as perceived now, if correlated cost variability and other
interdependences were addressed effectively it would be obvious that to move Canpower to
a predominantly nuclear portfolio of electric power sources was very risky.
The conceptual basis provided by the linear programming framework completely ignored
risk in the risk efficiency sense portrayed in Figure 8.8, limiting itself to always looking
for the point on the risk– reward boundary defined by point ‘a’ because risk– reward (risk-
expected cost) trade- offs are ignored. Implicitly the Canpower approach assumed that
nuclear power would remain the lowest actual cost solution over the long term and that there was no need to worry about ANY risk associated with a Canpower portfolio of base- load power sources which was 100% nuclear, perhaps 100% one particular kind and design of large nuclear power station.
There was no need to take such an extreme position. Most utilities used a robustness-
driven scenario analysis basis to test a range of demand and cost growth scenarios, and
test proposed plans for robustness against the range of ‘optimal’ solutions which different
assumptions produce. This could be a reasonably effective approach to managing risk. How-
ever, it was not always effective and it was rarely efficient, because it does not address risk
efficiency or opportunity efficiency in a direct clarity efficient manner.
Some utilities used informal portfolio analysis based approaches, which could be a consid-
erable improvement. But in his view formally addressing all risk in a framework based on risk
efficiency was crucial. It was also crucial to seek associated clarity efficiency and opportunity
efficiency.
No effective integration of dependence upward concerns
Carl suggested that full consideration of uncertainty within the short- term, medium- term
and long- term planning frameworks had to address important dependence upward issues
currently ignored. The contribution to significant biases was part of this, but so was the
exposure of Canpower to serious potential supply shortfalls. A simple iterative approach
would allow them to adjust their overall plans for these upward dependence concerns.
No formal futures planning or goals planning
Formal futures planning with underlying formal goals planning omitted in Canpower’s cur-
rent approach in the privatisation risk sense just developed was arguably as serious as any
of the other concerns addressed above. Carl left this set of issues until last and gave it less
space and emphasis than it deserved simply because it was much more complex, difficult to
discuss, and controversial, and he did not want to get bogged down in controversial issues
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 465
until the two flaws with much clearer implications and potential fixes were discussed. In
practice, Canpower should address these formal futures and goals planning issues first, as its
top priority.
A summary of Canpower’s current position and proposed changes discussed so far
In summary, Carl suggested that his approach would allow Canpower to start by clarifying
its goals. It could then plan how to shape a desirable future and deal with hoped for pos-
sibilities as well as currently feasible options. This would provide a basis for an enhanced and
very different approach to long- term planning. Having established a set of current targets
at its long- term planning horizon, it could develop detailed action plans within a medium-
term planning framework using quite different processes to address timing concerns in a
clarity efficient manner.
Taken as a whole, this approach would allow Canpower to identify where it ought to
be going first, and how best to get there second. This could be viewed as a very complex
portfolio of corporate change management projects in a practical framework for addressing
what was involved. A risk efficient set of feasible solutions for a range of potential demand
growth patterns would be central, assessing in approximate terms a preferred mix of sources
of power to aim for at the long- term horizon. Also central would be a closely coupled
medium- term planning approach to minimising the cost of meeting medium- term demand
fluctuations while moving in the direction of the desired balance of sources at the long- term
planning horizon.
If Canpower adopted Carl’s approach, it could move ‘from the bottom of the class to
the top of the class’ in terms of industry standards in one step. What would flow from a full
implementation of Carl’s approach was not just a different process for making decisions.
Nor was it just a different solution. It was a different kind of process and a different kind of
solution. Most of the differences should be self- evident already, but some further profound
differences had yet to be discussed.
Organisational separability for integrating corporate planning
Carl indicated that the board had been extremely patient through a very long initial pres-
entation with minimal breaks, and he was conscious that minimal discussion beyond simple
clarification concerns had been involved. But before he concluded, he believed it would be
useful to provide a very brief introduction to how Canpower’s corporate planning might be
structured in operational terms to achieve full integration of all Canpower planning, intro-
ducing the simplest structure which he believed would work effectively in practice.
He began by suggesting that he had significant concerns about important features of
the approach to project planning and costing taken by Pat, but he believed that he and Pat
should try to resolve these concerns without radical changes to Pat’s role or to his Project
Planning Group approaches and structure before more significant changes were contem-
plated. He had in mind the kinds of changes discussed in Chapter 7. He believed it was
highly likely that Pat would need replacing by somebody like Paul in a 1990s prototype
form, but he did not want to be distracted by these issues that day.
He suggested that relatively minor changes associated with Dick’s role and his Operations
Group approaches and structure were implicit in his earlier comments, and other directors
would also feel very minor less obvious effects of the overall change required.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
466 Employing planning tools in practice
He then indicated what he wanted to focus on briefly was the massive changes to the
current role, approaches, and structure of the group reporting to Larry as Canpower’s Cor-
porate Planning Director. Without having said so directly to this point, he was sure Larry
would need immediate replacing. He now wanted to lead up to saying so unequivocally,
having first addressed the implications in terms of what Canpower should expect of their
new Corporate Planning Director, assuming for the moment minimal changes elsewhere,
and reassuring everyone but Larry to avoid them feeling threatened.
Larry currently assumed his role as Corporate Planning Director equated to what Larry
called long- term planning and associated long- term forecasting. Viewed from Carl’s per-
spective, this amounted to a focus on medium- term plus long- term planning in terms of a
framework which was not fit for the purpose, based on forecasting which was interpreted
in an inappropriate manner, without addressing goals planning or futures planning in the
explicit way required. In addition to executing these roles differently, Larry’s current role
needed expansion and more than one person to deal with its oversight management in a
closely integrated team, and that team needed much better integration with the areas of
responsibility of all other board members.
In Carl’s view it would be useful to start discussing all the necessary changes to the role of
Larry’s current Corporate Planning (CP) group by visualising its decomposition into four
subgroups – CP1, CP2, CP3 and CP4 for that day’s purposes – these subgroups could be
given more appropriate labels later. Each of these subgroups would require leadership by
one of four new ‘Corporate Planning Managers’ reporting to the person holding the new
Corporate Planning Director post, with separable roles for each of these four new managers
and the groups reporting them outlined in the following. This separability would not imply
independence – it would serve the same kind of role as separability between the five different
categories of planning portrayed by Figure 8.1.
It would then be useful to discuss the roles of two new additional Canpower executive
directors with their own small supporting groups to address some of the wider implications.
CP1: Goal alignment with owners using an enhanced regional customer focus
Addressing the goals which Canpower owners were concerned about at a regional level
using a customer focus in a coordinated manner seemed a sensible place to begin. For
example, say they looked at one particular region involving a dozen small towns plus the
surrounding rural region in a far northern part of the province. Each town might be based
on producing timber or mining, with significant industrial electricity demand driven by
these industries as well as some manufacturing, retail, other business and domestic demand.
Some of the timber producers might use wood waste to generate some of their electricity,
but this might be a minor issue. A key local concern might be the long transmission lines
to this region which were vulnerable to winter storm failures, with particularly concern-
ing implications for the regional hospital, schools, and other key services. If a Canpower
planning advisor looked at Canpower customer goals for this particular region in terms
of a generalised and enhanced service from Canpower, what might emerge if the opening
question to a panel of local people was, ‘What would local Canpower customers like from a
locally focused electric power company prepared to go beyond just supplying electricity in a
supportive partnership relationship?’
A local multiple- unit combined- cycle gas- fired electricity generation station whose
waste heat could be used for the regional hospital and schools plus other nearby users of a
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 467
community heating service might be a central proposition, reducing the impact of possible
winter transmission line losses being a core aspect of the rationale. The community heating
idea might be a way of making local gas- fired electricity generation cheaper as well as serv-
ing heating needs more effectively. If it made economic sense, the community heating idea
might be extended to waste wood and other joint purposes like waste disposal. This electric-
ity generation station might be owned and operated by a company owned by the community
with Canpower loan funding, a third party which also provided similar facilities elsewhere, or
Canpower. If Canpower did not own the station, it might still export power to the grid as and
when this was optimal for the Canpower system, at cents/kWh rates tied to the cost of gas
but reflecting dual use, so it was part of Canpower’s gas- fired portfolio for long- term plan-
ning purposes, and its owners would not be squeezed by rising gas prices coupled to falling
electricity prices or peak /base- load distinctions. Its owners might in effect be subcontracting
with Canpower to operate a small scale local gas- fired power station with linked other service
facilities on behalf of Canpower. But the CP1 approach to the contract could be designed
to look after the best interests of the regional Canpower customers, so long as this did not
disadvantage Canpower customers in other regions without acceptable trade- offs.
If a key local customer benefitting from these arrangements was a company not owned
by local citizens, but this company provided local jobs, associated complications could be
accommodated explicitly as simply and effectively as possible.
Other rural regions might be encouraged and assisted to develop local low- head hydro-
power facilities, with complementary seasonal variation in demand and supply balance impli-
cations. Major urban area regions might be very different again, with a possible focus on
converting rubbish into electricity or electric powered public transportation for example.
The CP1 manager responsible for this subgroup would see his or her group’s goal as
serving the best interests of Canpower customers in terms of an enhanced service in each
of the regions Canpower had to serve, recognising that each region might have a unique
combination of issues in addition to important areas of common interest. The CP1 focus
would be the best interests of those the company were serving in each region, not the details
of how the other groups within the CP group as a whole might meet those needs. However,
explicitly starting with goals planning followed by futures planning followed by long- term
planning, this group might generate specific project propositions in conjunction with the
proposed collaborators.
This is a form of bottom- up strategy formation process driven by regional ‘owners as cus-
tomers’ concerns. Because CP1 was focused on each region’s individual concerns in terms of
an enhanced service for customers, satisfaction with Canpower’s service to customers should
be maximised by creating this new subgroup – provided this subgroup’s manager held an
appropriate balance of power within the CP group as a whole, and provided the CP group
as a whole was competent, suitably empowered and managed in an enlightened manner.
CG2: Goal alignment with owners focused on new collaborative non- utility generation
New collaborative non- utility generation of power like that associated with specific regions
identified by CP1 might be championed and prospective players encouraged and facilitated
on a province wide basis by CP2, a second subgroup who might see the focus of their role
as serving the needs of local community owned or third party owned electricity genera-
tion companies. CP2 might also work with third party companies who do not have specific
regions in mind and might be based outside the province. These third party organisations
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
468 Employing planning tools in practice
might have specific technologies in mind, like low- head hydropower or solar or wind power,
which Canpower did not currently have the expertise for. In part, the role of CP2 might be
seen as looking after the best interests of citizens of the province who wanted to produce
electricity. But a broader interpretation would be ensuring a level playing field for all poten-
tial providers of non- utility generation in terms of their competitive relationship with electric
power generation provided directly by Canpower. And the underlying role was ensuring the
best deal for Canpower owners who are electricity purchasing customers, explicitly eliminat-
ing any bias towards keeping electricity generation in- house when this was not in the best
interests of owners/consumers. Explicitly starting with goals planning followed by futures
planning followed by long- term planning, this group might explicitly address a portfolio
approach to generic types of non- utility power, to ensure that a balanced portfolio approach
to non- utility sources and power generated by Canpower. It might also generate specific pro-
jects to develop new collaborations, in conjunction with the proposed collaborators.
Because CP2 was focused on the interests of potential providers of non- utility electricity
in all regions, satisfaction with Canpower’s treatment of non- utility providers should be
lifted to an appropriate level by creating CP2 – provided the CP2 manager held sufficient
power within the CP group as a whole, and provided the CP group as a whole was compe-
tent, suitably empowered and managed in an enlightened manner.
CP3: Long- term and medium- term planning of all feasible options
CP3 would be the subgroup within the overall CP group responsible for long- term planning
in portfolio terms of all feasible options as addressed earlier in this chapter, incorporating
generic consideration of non- utility and conservation options as well as Canpower owned
and operated electricity generation stations, plus follow- on medium- term planning incorpo-
rating new proposals generated by CP1 and CP2 and agreed by the Canpower board. CP3
would have to reflect the interactions with short- term and medium- term planning discussed
earlier, the reliability/availability concerns also associated with short- term planning, and any
other relevant links. Goals and futures planning would be directly relevant in terms of a stra-
tegic view of appropriate trade- offs between cost per kilowatt hour tariffs driven by a policy
of minimising expected cost and associated cost risk, risks associated with security of supply
for economic, political or military reasons, and concerns like sustainability or environmental
implications with or without government taxation or subsidy implications.
Because CP3 was focused on the best interests of owners/customers of all regions, bal-
ancing the overall and regional concerns of the CP1 and CP2 subgroups, satisfaction with
Canpower’s treatment of the overall planning of all currently technically feasible options
should be maximised by creating CP3 – provided the manager of CP3 held sufficient power
within the CP group as a whole, and the CP group as a whole was competent, suitably
empowered and managed in an enlightened manner.
CP3 should not assume that its plans or those of CP1 and CP2 would be approved by
Canpower’s board, but jointly sought approvals following negotiation coordinated by CP3
would be the focus of their effort.
CP4: Potentially feasible opportunity realisation and related threat neutralisation
CP4 would be responsible for goals planning and futures planning plus follow- on project
proposal planning and delivery in terms of addressing the gap between what is currently
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 469
feasible and economic and what might become feasible and economic given the imple-
mentation of plans for direct investment or subsidies. For example, if solar power was not
currently economic, plans to help make it an economically viable future proposition might
involve collaboration with a company interested in developing a manufacturing capability
for the ultimate products, and it might also involve provincial or national government direct
investments and subsidies.
CP4 plans would require board level testing, and funded projects would feed- back into
the other three group’s plans in a lagged manner, with provisions and contingency plans
associated with possible outcomes when appropriate.
Goal and gateway integration and testing processes with a new director responsible
Routine corporate planning updates from CP3, plus specific project proposals from CP1,
CP2 or CP4, plus any bottom- up project proposals from Dick, plus more general current
Operations Group plans, all needed effective and efficient integration involving Pat and
his Projects Group as appropriate. Furthermore, this coordinated portfolio of propositions
needed integrated testing in preparation for oversight by the board, using a suitably devel-
oped variant of the kind of gateway structure discussed in Chapter 7.
In Carl’s view, this kind of integration and testing needed a small team with suitable
expertise led by a new board level director with a title indicating a ‘Gateway Director’ role.
This integration and testing in preparation for oversight by the board should probably be
extended to include proposals from other directors with comparable implications, like a
new information system proposal from Canpower’s Information Systems Director. Whether
or not its remit included all significant Canpower changes, such a group needed to use an
approach to overall goals planning and overall futures planning, plus all other forms of
corporate planning reflecting the preceding discussion, to provide an operational gateway
structure designed for use by the board to approve any significant changes. This was a key
role which should be separated from the Corporate Planning Director’s role. The person
fulfilling this role would be reporting to the board with a proposed balanced portfolio
of recommendations for what would amount to all change management decision making
associated with Canpower electricity generation assets plus other comparable assets. They
needed a degree of distance between themselves and those generating all the proposals to
ensure freedom from bias. Advocates of particular proposals are usually inherently biased
in favour of those proposals. The balance of power between this Gateway Director and all
other directors, including the Corporate Planning Director, would be very important.
A board level manager for this gateway group, separating this function from the groups
preparing proposals, would be very useful in Carl’s view. He did not say so directly, but help-
ing to avoid the acceptance of a lack of competence in any proposals from any directors was a central role for this new Gateway Director in Carl’s view, with Canpower’s recent approv-
als of proposals from Larry clearly illustrating why this was needed.
A corporate processes group with a new director responsible
The four Corporate Planning Director’s subgroups plus the new Gateway Director’s group
discussed earlier were all focused on planning the operation, renewal and extension of elec-
tricity generation and transmission assets, other comparable assets, and associated plan test-
ing processes.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
470 Employing planning tools in practice
To achieve their goals these groups all needed appropriate people with the right skills and
knowledge, and these people needed to be provided with all the appropriate processes if
they do not bring them with them as part of their toolset.
Canpower and most other organisations did not have a Corporate Processes Director.
Carl suggested that Canpower needed one, with an initial mandate to redesign the processes
needed to implement the revised approach to corporate planning this chapter has outlined,
working with Canpower’s Human Resources Director to recruit the people needed to help
develop and implement all the new processes. External consultants should be used when-
ever this was the most effective and efficient way forward, but in- house expertise needed
well- planned personnel selection, development and retention, and overall management in
a coordinated manner by a suitably capable and empowered board level manager seemed a
sensible working assumption from Carl’s perspective.
Concluding comments about the Corporate Planning Group
In Carl’s view, none of Larry’s current staff group should be replaced until it became clear
that specific replacements would be necessary, but he anticipated a major transformation,
and he did not see how any of his proposed changes could proceed with Larry still in post.
Carl finished his presentation on this note, put as tactfully but clearly as he could, apart
from a brief closing summary.
Limited inferences instead of further developments for this tale
A plausible Canpower board reaction to Carl’s presentation is highly debatable, and for a
number of reasons attempting to conclude this tale with an outcome relevant to the 2020s
and beyond in a wide range of contexts did not seem a useful strategy. However, briefly
consider some of the additional advice that I would be inclined to offer Clive, the Canpower
board chairman, if he had your understanding of EP as discussed so far in this book. Assume
that you and Carl might be a part of this conversation, and both you and Clive might be
interested in how the ideas discussed so far in this chapter relate to organisations other than
Canpower.
A useful starting point might be interpreting the UP of Figure 2.1 as the overall guide to
the analysis structure adopted by Carl, the basic SPP of Figure 7.2 and the rest of the SP for
projects as an additional supporting more focused operational and conceptual framework
for managing change. From this perspective, with the ‘context – including all other plans’
in mind, the impact of the ‘capture the context’ opening phase of the UP is clearly visible
in Figure 8.1.
For any organisation, a top- down view of the role of the Corporate Planning Director in
the Figure 8.1 framework might reasonably start with clarification of the owners’ goals and
then planning a desirable future from the perspective of the organisation’s owners, drawing
on the seven Ws framework of Figure 7.1, and using a suitable variant of the overall four Fs
concept. This implies taking into account the relevant views and concerns of customers for
the basic service currently provided whether or not they are also ‘owners’. It also implies
taking into account competitors, suppliers, employees, regulators, and other relevant inter-
ested parties and working with a separability structure which suits the nature of the different
interests involved.
With his identification of a separable CP1 group with a manager given appropriate power,
Carl was suggesting that goals planning ought to start with Canpower owners – citizens of
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 471
the province – with a focus on their closely coupled role as electricity customers, decom-
posed region by region to ensure that a local perspective is given due weight, including an
important potential creative influence. Carl was suggesting that the past Canpower failure to
effectively respond to the best interests of their owners as customers with needs which were
specific to where they lived was a crucial board level failure. It was arguably directly attribut- able to Larry, but it had been inappropriately tolerated by the board as a whole. Canpower
needed a much more effective way to address its owners’ best interests as domestic supply
customers, employees of local firms using electric power, and users of local services requiring
electric power, recognising that there may be important differences in different geographical
regions as well as important commonalities.
There was no need for Canpower to get directly involved in running small gas- fired local
electricity stations providing linked heating and other related local services if this was not
economic for all Canpower owners/customers, but facilitating a better overall solution at
no extra cost to other Canpower owners/customers should be part of Canpower’s agenda if
regions wanted this additional dimension to the role played by Canpower. It might not be
convenient for Larry or some other Canpower employees to bother, and Larry might prefer
to interpret private power producers as unwelcome competition rather than as part of a big
team. But that was one of the key reasons for replacing Larry immediately, and for making
the CP1 manager part of his replacement package.
To link this CP1 rationale to a private sector example using the Astro tale of Chapter 6,
the sensitivity of Sian’s Systems group to outsourcing some of their roles was directly com-
parable to outsourcing some electric power generation. Furthermore, the roles of Trevor
and Martha within Astro were comparable in ‘customer- facing dialogue’ terms in some
respects. However, Astro did not need to link this to Astro owners (shareholders) in the
same way, and the comparison cannot be pushed too far.
What Carl was doing with his identification of a partially separable CP2 group with a
manager given appropriate power was suggesting competitors might be a nuisance for Larry
and some other Canpower employees, just as external providers in competition with Sys-
tems within Astro Wales were a concern for Sian and her Systems staff. However, if they
provided better value for Canpower owners/customers than Canpower, they needed to be
encouraged and protected by a suitably empowered Canpower manager, a role Rhys took
on in the Astro context. A level playing field was essential, but not enough in itself. Astro’s
friendly takeover of a supplier faced by a hostile takeover illustrated the kind of enlightened
self- interest the manager of CP2 might advocate. This might raise important concerns for
Canpower employees, who might see outsourcing their jobs as a direct threat requiring
union action. However, this concern needed management in the context of the overrid-
ing need to serve Canpower owners/customers. The unions should be respected, but they
should not be allowed to wield inappropriate power. When I observed the outsourcing in
IBM UK which inspired the Astro approach which Martha took, IBM colleagues were very
concerned, as should be expected for the employees of any large and successful organisation
which needed to restructure to cope with a changing environment, be it in the private sec-
tor or the public sector. It is not an issue specific to public sector organisations, unless they
have excessively powerful trade unions which need better management by the organisation’s
board. Supervision of board appointments by a government which understands the nuances
of these kinds of issues is a closely coupled concern.
What Carl was doing with his identification of a partially separable CP3 group with a
manager given appropriate power was suggesting that the new options generated by CP1
and CP2 which were approved jointly or separately within overall Canpower goals planning,
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
472 Employing planning tools in practice
futures planning, long- term planning plus medium- term planning, needed effective inte-
grated treatment by a group capable of the kind of approaches he had outlined earlier.
This kind of planning was not discussed in an Astro context. However, a very different but
comparable variant clearly had a role in a broader view of Astro operations and corporate
planning at Astro UK and Astro Inc board level.
What he was doing with his identification of a partially separable CP4 group with a man-
ager given appropriate power was suggesting that new technology developments leading to
expenditure related to currently non- feasible options required separate encouragement and
facilitation, avoiding distracting or confusing those concerned with planning in the CP1– 3
sense. However, when appropriate, the long- term, medium- term and short- term planning
processes for feasible options needed contingency plans in place to capitalise on successful
new technology developments. Part of the focus of the CP4 group might be new business
model developments in collaboration with other parties, like working with urban centres
on waste incineration to produce electric power and using the power for local transport to
reduce emissions, as well as costs and congestion.
What he was doing with his identification of two additional board level directors – a
Gateway Director and a Corporate Processes Director – was arguing that implementing
the overall changes that Canpower needed to make required two further roles which did
not simply replace tasks Larry arguably should have been doing. These roles were outside
the current Corporate Planning Director’s role, and they should remain outside the future
Corporate Planning Director’s role.
What he was doing with his identification of the partially separable CP1– 4 groupings plus
the two additional directors and their supporting groups was suggesting that the Canpower
overall corporate planning process needed separability (not independence) for these differ-
ent roles, facilitating the separate ‘creation of plans’ to address all the relevant concerns and
their integration, because different kinds of issues are involved. The same rationale underlay
the separation of goals planning, futures planning, long- term planning, medium- term plan-
ning, and short- term planning.
These assumptions of separability did not mean these considerations were truly separable
and could be treated independently. It did mean they needed separate pre- integration treat-
ment which adopted different framework structures and goals to achieve an effective and
efficient understanding of different issues before they were considered jointly. For example,
long- term planning which addressed an optimal mix of currently feasible sources of electric
power at the 20 year horizon needed an analytical framework which was very different from
the framework required by medium- term planning to construct new electricity generation
stations and then bringing them on- stream in conjunction with maintaining and retiring
existing plants. Using the same framework for considering both at the same time might be
feasible in principle, but it was not clarity efficient in practice, requiring more effort for less
clarity at best. Producing seriously misleading results if inappropriate framing or working
assumptions were adopted was a serious threat, as illustrated by Larry’s approach, which
was inappropriate for both long- term and medium- term planning. Short- term planning was
different again, as were futures planning and goals planning. The CP1– 4 structure plus two
new directors with supporting groups just extends the use of this ‘operational separabil-
ity’ concept without assuming independence, explicitly addressing coordination via iterative approaches to full integration.
Building on this initial discussion of the role of separability, the need for separate SPs plus
the separate use of UP concept features leading to processes not developed in this book
might be useful. For example, the application of Soft Systems and other soft OR approaches
in a strategic planning context has enjoyed considerable success, and a wide range of other
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 473
very different approaches to top- down organisational planning contexts might be integrated
into a common broad framework.
Building on this further, a more general conversation about the role of the Corporate Pro-
cess Director and the Gateway Director plus their groups might be useful. Without these two
additional directors plus their groups, directors like Larry can attempt to build walls around
separate empires which can prove seriously dysfunctional. One of the many very useful roles
these two new groups might serve is effectively testing the clarity efficiency and the associated
risk– reward trade- offs plus other goal trade- offs for all corporate processes, not just those
associated with Corporate Planning. Helping to build the bridges needed to reach across the
gaps created by separability working assumptions and avoid separate silo effects is a crucial role
which both these new groups might help to fulfil in any organisation’s context, whatever the
nature of the business being pursued, regardless of its public or private sector status.
The need to effectively and efficiently test all proposed corporate changes via gateway
process structures might be another useful area for focused general discussion. The need
for an explicit focus on bias as part of this could be very important. This is about required
features of specific decisions, not processes in general terms, although specific gateway pro-
cesses would need development and ongoing testing with feedback. This is where the goal
programming framework associated with multiple test approaches to discounted cash flow
analysis discussed in Chapter 7 intersect with the goals planning framework discussed in this
chapter, further intersecting with the discussion to follow in Chapter 9.
Goals planning and futures planning each need leadership and coordination, with specific
people given coordinator roles but a wide range of Canpower staff involved. What the Gate-
way Director and supporting group needed to do is distil what might be thought of as ‘the
current corporate business model’ as approved by the board, operating gateway processes
to test all proposed changes to Canpower’s current operations and explain the implications
to the board. Apart from approving some of the recommendations and rejecting others, the
board needs to consider adjusting the corporate business model when appropriate.
At a more specific level, if I were advising Clive in the 1990s or now, I would start by
stressing that Clive was very clever suspecting the need for change as soon as he arrived as
the new Canpower chairman of the board, praise the way he insisted on Ivor exploring what
needed doing as an opening move, and support everything Carl had said. But I would then
suggest that Clive and the rest of the board were going to have to be both seriously clever
and exceptionally brave to make the significant changes outlined by Carl without serious
mishaps, despite the very clear need to fix what was obviously broken.
I would emphatically acknowledge the importance of the ‘if it ain’t broke don’t fix it’
line of holistic thinking, but suggest that I fully agreed with Carl that Larry has to go with
immediate effect, and Pat needed to be put on notice that keeping his job depended on
rapid changes as advised by Carl. There was a very good case for suggesting that Pat should
go immediately too, an important issue. Everybody else did need clear assurance their jobs
should be safe assuming they could rise to the overall changes necessary.
Carl could be offered the new Corporate Processes Director role, at least on an acting
basis as a short- term plan. Getting his agreement to take on this role before he made the
presentation just outlined, or making sure someone else who could take on this job could
be recruited, would seem prudent if not essential.
Clive might want to take on the role of Gateway Director on an initial interim basis, with
immediate effort given to finding a suitable person to initially help him, develop and shape
the role, and then take over the role.
Carl and the Canpower board were going to need the direct assistance of an additional
director in charge of the corporate change process – preferably someone like Paul from
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
474 Employing planning tools in practice
Chapter 7 with additional relevant change management experience. Their long- term role
could be seen as comparable to a long- term role Paul might aspired to within WSL, with
a Corporate Projects Director label. Pat or the individual taking Pat’s place might then
become one of several programmes directors reporting to the Corporate Projects Director
with responsibilities in specific areas, like electricity generation and distribution assets, infor-
mation system assets, and so on.
The Corporate Processes Director could use a UP concept plus SP for projects concepts
plus a wide range of further SP concepts, plus any other compatible approaches or the
equivalents, as a basis for developing the processes needed within each of the corporate plan-
ning groups. Each of these groups might need very different approaches, but one overall
supporting process development group with a rich mix of different talents and skills would
clearly be valuable.
The person replacing Larry as Corporate Planning Director obviously needed extensive
knowledge and experience relevant to this post. I make no pretence of being competent to
judge what this should include. However, they and a new Gateway Director and Corporate
Processes Director plus Clive clearly need to be aligned in their views.
Carl’s advice might not be taken in the sense that Clive might not prepared to agree to
with Carl’s presentation in the first place and then argue that the board needed to act on it
and ensure that he won the argument. It is certainly easy to see why someone in Clive’s posi-
tion might feel the propositions put by Carl were too radical, or the board as a whole might
not be convinced. For a start, Carl’s propositions imply that Canpower’s staff numbers
might shrink significantly if outsourcing some electric power generation and other current
in- house roles took off without compensating growth in other activities, implying the risk
that all directors and senior managers might face salaries shrinking in real and relative terms
if not in actual terms.
However, my working assumption when drafting this chapter was that a government in
charge of a publicly owned company like Canpower should be capable of appointing a new chairman of its board who would want to act in the best interests of the citizens he or she serves, they should have the required competence and passion, and suitable incentives for all key staff could be engineered by a competent senior management team. Given these assump- tions, all the changes suggested were a reasonable basis for discussion, without assuming
that you or anyone else is going to be 100% convinced about all aspects now or later, includ-
ing the likelihood that current and near future political realities make these assumptions as
robust as most of us would wish them to be.
A brief exploration of the implications of context differences
The implications for Canpower of grasping all the feasible opportunities available and neu-
tralising all the relevant threats could be quite profound. For example, Canpower might
transition from a traditional public sector electricity generator and distributor which was
heavily criticised for a range of reasons and at risk of privatisation into a reinvigorated
national treasure involved in the operation of a portfolio of businesses which included waste
disposal via incineration, community heating and electrically powered public urban trans-
portation in addition to providing low cost electricity for provincial businesses and consum-
ers. Doing so might involve a range of collaborative arrangements and a variety of new
ventures. It might also involve outsourcing some current in- house operations. The overall
approach might ensure that provincial ownership was a matter of ongoing pride and satisfac-
tion for all its citizens, with no serious legitimate complaints about its operations which did
not get resolved.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Corporate strategy formulation 475
One obvious issue is the extent to which this kind of significant potential transformation
might be feasible for other organisations in very different contexts if they used appropriate
variants of some of the approaches considered in this chapter. But a more general question
is, ‘To what extent are the conceptual and operational tools explored in the very special
Canpower context portable?’
If we start with the framework provided by Figure 8.1 and its five categories of planning
(goals, futures, long-term, medium-term, and short-term) in four planning horizons, some
variant of this approach would seem essential for all organisations operating in the private or
public sector, and all of the key features look essential in most contexts. For example, failing
to be very clear about how the goals of an organisation’s owners relate to those of other
relevant parties and the goals– plans structure for all four other categories of top- down cor-
porate planning including a ‘designing a desirable future’ aspect of futures planning could
cripple any organisation if associated risks were realised and important opportunities missed.
Any aspect of this framework which was broken would seem to be in need of fixing. All
associated capability- culture liabilities needing attention would seem to be an issue for any
organisation, and most organisations might find scope for improvements in performance
using this starting point for a reflective analysis review.
If we relate this to the need for separability to facilitate different people working in dif-
ferent groups in order to use some different conceptual and operational tools with a clear
understanding that independence is not involved, and an iterative approach is needed to
accommodate upwards, downwards, and sideways dependencies, again this seems univer-
sally applicable. Silos need avoiding, as everything is interdependent to some extent, but
suitable working assumptions about separability need to be made and the operational impli-
cations tested to ensure effective and efficient functioning.
Also very general is the need for everyone involved understanding the role of diversifica-
tion to reduce some forms of risk in a risk efficient manner and select risk– reward trade- offs
which are suitable. This may not relate to major long- life capital investments like the port-
folio of hydro, nuclear, thermal, and other sources of electric power decisions faced by Can-
power. But diversification may be an issue in futures planning contexts involving research
and development, or in terms of product ranges, product types, geographical markets, and
input component sources, for example. Different contexts may require very different use of
the formal operational tools provided by a flexible HAT framework, or just conceptual use
of these frameworks for informal planning. However, ignoring issues of this kind may be
risky, and not being capable of considering them effectively may be incompetent. Further-
more, this diversification concern usually needs to be balanced with the benefits of a clear
focus on doing what an organisation does best, ‘sticking to the knitting’ in some people’s
jargon when the fashion for diversification via conglomerates ran out of steam – as many
fashions do.
The role of ‘balanced targets’ which are not expected outcomes because of asymmetric
penalty functions is a core component of the very basic ABCs of targets concept, conveni-
ently explored with an illustrative numerical example in this chapter, but potentially relevant
in all organisations in all contexts.
The scope for very high levels of optimistic bias driven by asymmetric probability distri-
butions with very long tails if all relevant uncertainty is not considered appropriately is also
relevant in most organisations in many different aspects of their operations. The need to
illustrate its importance using numerical illustrations will come up again in the safety context
of the next chapter, but it also underlies many other operations management contexts and
many project management contexts.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
This chapter addresses complex concerns deliberately avoided in Chapters 5 to 8 which need
to be confronted before completing Part 2. The central concern considered is the concep-
tual and operational toolset and mindset needed to develop well- founded trust between
different parties with different perspectives on complex concerns needing formal planning.
The trust issues addressed in this chapter are particularly demanding in terms of the need
to avoid the wrong kind of simplicity but leverage the benefits delivered by the right kind
of simplicity.
A central feature of the tale used in this chapter is addressing potential catastrophes in
an enlightened manner at a high level of clarity. ‘Catastrophe’ is used in the plain English
sense of a very high negative impact incident which may have a very low probability and may have implications which go well beyond financial concerns. From an EP perspective, all potential catastrophes need to be approached using plausible worst- case scenarios defined
with a particularly carefully considered systematic approach to achieving simplicity. Being
especially careful to achieve the right kind of simplicity in a potential catastrophic context is
enlightened because the wrong kind of simplicity matters much more than usual. However,
the approach to developing well- founded trust used in this chapter’s tale requires concep-
tual and operational tools which can be useful even if the kind of very high negative impact
catastrophic outcomes initially addressed are not an issue – they have important implications
and areas of application that are much wider than starting to read this chapter may initially
suggest.
A defining feature of this chapter’s approach to potential catastrophes is using a scenario-
based approach to explicitly confronting both the extreme events associated with catastro-
phes and a full range of related but less serious intermediate events. The difficult trade- off
concerns all these scenarios may give rise to is addressed using a practical holistic top- down
framing assumption perspective which avoids common practice restrictive assumptions. The
rationale for the approach as a whole includes avoiding inappropriate optimistic bias and
eliminating serious blind spots, coping with very complex interdependencies, facilitating a
‘big team’ approach led by board members with strategic clarity who are supported effec-
tively by all other relevant players, and the need to earn and maintain well- founded trust
from different parties with very different perspectives working within a formal planning
framework everyone involved can understand with appropriate clarity.
Defining features of the context are always crucial. Safety concerns in a railway system
require planning approaches which confront a usefully rich set of complexities that need
attention by a range of different people with different concerns. It is also a usefully simple
context in some respects, which helps to keep the discussion as simple as possible before
attempting to generalise to other contexts where different kinds of complexity are important.
9 Building well- founded trust about complex concerns – a railway safety example
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 477
Additional significant analysis complexity associated with concerns like extensive oil pol-
lution or nuclear contamination is addressed briefly in the last section of this chapter. Quite
different security issues which involve comparable concerns in some respects are also con-
sidered briefly in the last section, using cybersecurity as an example. Exceedingly complex
political or military goals may be relevant if complete generalisation is considered.
A central part of the experience base for the tale of this chapter was a strategic level review
of Railtrack’s approach to safety for the UK railway system undertaken in the late 1990s.
Views about relevant approaches were significantly updated and revised using a much more
general conceptual and operational framework initially developed over the period 2010– 13
for the UK MoD. The MoD work involved consulting support for the development of an
operational procedure for the MoD to justify high levels of expenditure on a portfolio of
approaches to preventing and mitigating the implications of low probability but potentially
very high- impact non- conventional weapon attacks on British troops, given broader NATO
airing subsequently. A number of other relevant background studies shaped some of the ideas,
mentioned when directly relevant. What proved crucial was being forced to look for practical
and robust new simplifications in order to cope with new complexities associated with the
MoD context, followed by a period of reflection supported by feedback from colleagues.
A key feature of this chapter’s approach is a formal acknowledgement that the nature of
the relevant decisions and objectives means that the issues being addressed cannot be con-
sidered as separable ‘add- ons’ to operations or project plans at a tactical level. The nature of
relevant objectives and the scale of potential impacts mean that both safety and security must be seen as a fully integrated ‘designed- in’ feature of corporate planning for operations and
project planning concerns at a corporate strategy level. One implication is that this chapter’s
approach has to build on the framework developed in the last chapter, following instead
of preceding it for expository reasons. We are dealing with a form of ‘enterprise opportu- nity, risk, uncertainty and underlying complexity management’, but the approach advocated
should not be confused with ERM as this term is usually interpreted.
The central role for fatalities as a core metric in the railway safety context of this chapter
means that the number of people killed in railway incidents is one key basic metric which
is unambiguously of central importance. However, injuries with a wide range of degrees of
severity are also extremely important, and trade- offs between all key metrics, including cost
and revenue measures, have implications which are inherently ambiguous.
One significant aspect of railway system accidents is the very wide range of numbers of
people who might be killed or injured in a single incident. Small numbers of people are fre-
quently killed or injured, perhaps several hundred per annum; 30 or 40 fatalities in a single
incident is not a rare event, perhaps every other decade on average; and very large numbers
of people might be killed or injured in very low probability incidents of a kind not yet
experienced, perhaps not yet even contemplated, with a return period expectation like once
every 1,000 years. The railway system examples which are central to this chapter address
‘incidents’ which include both accidents and malicious attacks.
Deciding whether to make strategic changes to a railway system with safety and security
implications in the framework developed in this chapter requires board level leadership. It
involves operational decision making which needs attention in the gateway processes dis-
cussed in Chapters 7 and 8. This means key strategic decisions must be part of the top- down
strategic management processes addressing all issues relevant to corporate strategy, building
on the frameworks developed in all earlier chapters.
The nature of the competing objectives involved means that attribute value trade- offs
central to this chapter involve very different attributes (like fatalities, injuries, and financial
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
478 Employing planning tools in practice
costs), and very different perspectives on values associated with these trade- offs (like those
of railway passengers and those of private sector railway company shareholders). This, in
turn, means that trust is a core issue. An organisation’s owners, employees, customers, and
regulators need a robustly grounded basis for well- founded trust in the decisions made.
Trust which can be assumed to be well founded invariably depends on corporate ethics
as well as corporate competence. It may also depend on many other factors, like media
responses to incidents and the way information is managed by regulators.
This book argues that all the relevant parties have a particularly strong need for a shared common basis for dialogue because their perspectives are so different and trust is so important. This makes designing a formal planning framework and using it for associated management
decision making inherently very difficult, but that does not excuse not addressing the really
difficult central issues with an enlightened level of competence. It simply makes finding an
appropriate framework for effective dialogue about answers to the right questions more
important, arguably essential. Furthermore, any relevant party that does not have well- founded trust needs to take decisive and effective remedial steps, reinforcing the need for
widespread understanding of the central issues. In the context of a railway company, its
contractors and subcontractors, this includes all relevant company board members, plus other employees associated with safety and security concerns, passengers, railway company
shareholders, regulators, and governments.
When seeking simplicity, avoiding becoming confused by the wrong kind of simplicity is
always important – clarity inefficiency involving seriously inappropriate assumptions can be
very costly. Clarity efficiency concerns explored in Chapter 5 were at their simplest. This
chapter’s clarity efficiency concerns are unavoidably complex – the most complex addressed
in this book. This is a central part of the rationale for deferring addressing them until the
final chapter of Part 2.
This chapter has been written to provide those who are not safety and security experts
with an accessible basis for deciding what really matters in terms of the overall approaches
to complex trust concerns appropriate for their organisations. If safety and security are not
part of your current responsibilities, much of this chapter’s more technical discussion can be
read without too much attention to the detail, so long as you develop a reasonable feel for
‘what needs to be done’. However, selected specific technical issues which are contentious,
and the reasons why associated differences in opinion matter, are discussed in this chapter
because strategic clarity matters to all those who are not experts but have to live with the
consequences of expert judgements. That includes all target readers for this book in terms of their roles as citizens and consumers, as well as their current or potential roles in terms of
contributing to management decision making.
The primary readership target for this chapter is board level managers in private and pub-
lic sector organisations, plus the regulators and politicians who influence their approaches,
plus those aspiring to these roles. It is not aimed directly at safety or security experts or
ERM experts. Nevertheless, the material of this chapter should be of direct interest to the
professional safety and security community and the ERM community, and I certainly hope
it will be.
Anyone even considering aspiring to board levels of responsibility, plus those involved
in direct supporting and regulating roles, should find it useful to understand all the issues explored in this chapter at an overview what needs to be done level. This includes the
very complex nature of the interconnected set of controversial issues which no doubt will
continue to surround safety and other contexts associated with potentially catastrophic inci-
dents for the foreseeable future.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 479
Some of the key issues are relatively straightforward to deal with – like how should an organisation approach the framing of safety and security as a top- down strategic consid-
eration fully integrated with the top- down and bottom- up treatment of all other strategic
concerns. This is addressed in reasonably straightforward terms near the outset. However,
other issues of central importance are relatively complex. Of particular importance, these
more complex concerns include what should we mean by ‘risk efficiency’, ‘enlightened cau- tion’, ‘enlightened gambles’ and ‘enlightened prudence’ which embraces a safety and secu-
rity context when ethics and well- founded trust are crucial and ‘revealed preferences’ are
central concerns, and how do we make these concepts operational in a robust but toler-
ably simple practical manner. Although these issues are of central importance, they are not
addressed in a direct manner until towards the end of the chapter because providing req-
uisite clarity involves building on earlier illustrative numerical examples and discussion in a
layered manner.
Significant departures from conventional approaches to safety management include:
1 a holistic top- down approach to extreme and low- level and related intermediate- level
incidents, with an initial focus on the extreme end of the range;
2 addressing ‘risk’ in terms of the full implications of departures from expectations in
addition to expected outcomes, using a risk efficiency framework as for all other man-
agement decision making in an EP framework;
3 understanding the full implications of an opportunity efficiency view of concerns like
clarifying the difference between good luck and good management, bad luck and bad
management, and requisite enlightened prudence when very large numbers of people
might be killed or injured with very low probabilities of this occurring;
4 employing an ‘avoided fatality’ concept with shadow price and revealed preference
implications which are of crucial importance and need widespread understanding,
including understanding why a common value independent of the number of people
involved may appropriate in some contexts but not in other contexts;
5 a major clarity efficiency improvement which can deliver much more clarity at a much lower level of cost and effort, driven by the insight that considering potential fatalities in conjunction with correlated injuries plus physical damage to the system plus all other
correlated impacts whenever scenarios involving fatalities are addressed delivers a game-
changing increase in clarity efficiency; and
6 using the approach to point 5 to clarify what the process of earning and maintaining
trust between all relevant parties in very complex contexts must involve.
Context and main characters
Only two named characters play a central role in this chapter’s tale:
Oscar, the operations director of a railway company;
Sophie, a corporate safety manager hired by Oscar.
Sophie was an experienced management consultant who worked throughout Europe from
her base in Copenhagen when she accepted an assignment to help Oscar rescue Northern
European Railways (NER) from a crisis triggered by a major railway accident in 2013.
NER is a fictitious private sector organisation which owns and manages all aspects of a
national railway network. NER is not a real company, and the country NER is located in is
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
480 Employing planning tools in practice
not Denmark. Both these points are explicitly emphasised because all the fictions involved in this tale are grounded on real cases which it would be inappropriate to discuss without
unambiguous changes in context. This chapter will use US dollar monetary units in a Euro-
pean country throughout, in part to emphasise the maintenance of a deliberately fictional-
ised status for NER in a simple and convenient manner.
Sophie had 20 years of experience over a broad range of topics and contexts. Strategic
change leadership on safety and security management was Sophie’s current passion. The
connecting thread in her career development was the routine use of her variant of the UP
of Figure 2.1 to develop specific models and interpret the results creatively in the tradition
illustrated by the tale of Chapter 5. Sometimes the creative interpretations led to an SP in the
tradition illustrated by the tale of Chapter 6. Sophie used project management approaches
including generic processes as discussed in Chapter 7, and she had attended the 2012 IPMA
(International Project Management Association) Copenhagen advanced training workshop
Managing project risk, uncertainty and value in new ways based on Chapman and Ward (2011). However, Sophie had never seen herself as a project management professional. She
regarded project management as just one of the many toolsets needed for all her consulting
roles. She had attended other courses provided by a wide range of other organisations on a
comprehensive range of operations and corporate management topics, including corporate
strategic planning, with a current focus on safety and security.
Oscar was an experienced and senior member of the Operations Group of NER when an
accident in 2013 killed 40 people and injured several times that number. Two years earlier
30 people had been killed and more than 100 injured in another incident. The second
accident triggered a corporate crisis, in part because most of the population of NER’s coun-
try believed that two serious accidents in two years were unlikely to be just bad luck. The
initiating fault for the first incident was a signalling malfunction, leading to a front- to- rear
collision between one train travelling at the lower end of a ‘medium speed range’ and a sta-
tionary train on the same track. The initiating fault for the second incident was a track main-
tenance failure, leading to a higher end of medium speed derailment with the train falling
down an embankment. Any maintenance failure was a matter of negligence for some people.
For about a month after the second incident, the Operations Director of NER struggled to
contain the fallout from these two incidents, with support from his board. However, knock-
on operations problems and media pressure eventually led to both the Operations Director
and the Head of Safety ‘resigning’. Oscar became the new Operations Director on the same
day. Oscar immediately asked Sophie to become his new Head of Safety. She had started
work the following week on a short- term contract for a trial period.
Oscar had attended the same 2012 IPMA Copenhagen advanced training workshop as
Sophie. Like Sophie, he regarded project management as just one of the many toolsets he
needed. He had been seriously impressed by Sophie’s approach to a case study (Transcon,
based on the tale in Chapter 6 in this book), but the reason for appointing Sophie Head
of Safety was a conversation in the bar after the case study session. Sophie had asked Oscar
about his role within NER. As part of his response Oscar had explained his frustration with
the NER board because they had turned down his proposal for a new generation signalling
system before the first incident, caused by a signalling malfunction, and there had been
ongoing procrastination about a new signalling system despite the incident. Oscar had then
asked about Sophie’s current work as a simple matter of reciprocating interest in a convivial
conversation. He had been deeply impressed when she described recently completed work
for the defence research section of Oscar’s country’s armed forces. Sophie had explained
why her defence work was directly relevant to railway safety decisions like Oscar’s signalling
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 481
system example, and he saw the connection immediately. He hired Sophie to help him trans-
form the safety aspects of his role as Operations Director of NER.
Oscar had initially assumed that Sophie would start with the track maintenance issues
underlying the second accident and the ongoing signalling system issues. However, she
quickly persuaded him that they had to start with much more fundamental concerns.
A few weeks after Sophie joined NER, Oscar held an extended workshop for the NER board.
He used this workshop to introduce Sophie, ask her to explain how she would set about trans-
forming the NER approach to safety, generate initial feedback, and make sure that he had board
support for the radical programme of changes which he and Sophie now envisaged.
Generalised foundations for NER safety and security
Oscar began his workshop by indicating that the basic foundations of NER approaches to
safety and security had to change radically. This would have to include board level corporate
decision- making process changes and linked corporate culture changes. The board had a key
leadership role in this transformation. The workshop that day had to communicate why, and
an overview of what was involved.
Oscar said he had hired Sophie as Head of Safety for NER because he was convinced that
she was the best person available. He observed that Sophie had accepted this post for a trial
period on a short- term contract. He might have added that the trial period was a condition
of Sophie’s. From Sophie’s perspective, the NER board were on trial, not Sophie. But he
thought it best to let the board come to this realisation themselves. Oscar did stress that
both he and Sophie were convinced that radical changes, starting at the board level, were
essential, and if he and Sophie were going to resolve NER’s current crisis board, feedback
from today’s workshop which was positive ‘on balance’ was also essential.
Oscar then outlined Sophie’s credentials to the board before turning the meeting over to
her. He emphasised the broad nature and importance of her current work for the defence
research section of their country’s armed forces, but also indicated her limited experience
with railways.
‘As low as reasonably practicable’ (ALARP) as a starting point
Sophie began by reminding the board that the generally accepted starting point for the man-
agement of safety in their country was the risk of death or injury should be ‘as low as reason-
ably practicable’ (ALARP), and broadly comparable approaches were widely employed in
the rest of the world. She asserted there was nothing contentious about this until you asked
‘What exactly does “reasonably practicable” mean?’ Many safety professionals saw this issue in a legal framework, and some might begin with
the practicability comments in what she believed was the original court judgement using
this term by Lord Justice Asquith in the UK case of Edwards v National Coal Board 1949:
‘Reasonably practicable’ is a narrower term than ‘physically possible’ and seems to me to imply that a computation must be made by the owner, in which the quantum of risk is placed on one scale and the sacrifice involved in the measures necessary for averting the risk (whether in money, time or trouble) is placed in the other; and that if it be shown that there is a gross disproportion between them – the risk being insignificant in relation to the sacrifice – the defendants discharge the onus on them. Moreover, this computation falls to be made by the owner at a point of time anterior to the accident.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
482 Employing planning tools in practice
A large international body of safety experts had developed a rich body of relevant knowledge
over many years. This knowledge base had to be applied in a coherent and comprehensive
way to address this question effectively. Some of this body of knowledge addressed broader
issues than NER had to face, including environmental risk and associated health risk.
Reports on the widely reported 2010 BP Macondo oil well blowout accident in the Gulf
of Mexico by the US- based Centre for Catastrophic Risk Management (CCRM, 2011)
outlined key lessons learned which echo many earlier reports in many different countries,
as explained in more recent reviews. Viewing these reports and the wider global literature
collectively, it was crystal clear that many of the crucial underlying issues had been under-
stood, at least, in part, for many years, but they had not yet been fully grasped in a holistic
manner by all those directly responsible for practice. This raised some very complex issues,
including the necessary corporate strategy aspects of addressing potential catastrophic risks
in conjunction with what were increasingly referred to as ‘corporate culture’ issues.
A central issue which NER had to address could be linked directly to a quote from the
final CCRM (2011) report – ‘It has been observed that “BP forgot to be afraid ” ’. In her view, NER had never learned to be afraid, and addressing this had to be a starting point. They needed to use their initial meeting to start to explore what needed to be done to deal
with a plausible maximum level catastrophic risk incident, plus all related incidents involving relatively low and intermediate levels of related risk, using a systematic framework for asking
the right questions in a practical way which was creative and optimistic in a realistic manner.
Her recent work addressing the safety of soldiers in terms of non- conventional weapon
attacks, including large scale attacks on their bases, suggested that understanding the impli-
cations of a coherent and enlightened approach to all the stealth assumptions associated with
key aspects of common practice provided new insights which NER needed, starting at the
board level. She would first outline, in just a matter of minutes, some of the key NER con-
cerns involved at a very high overview level (the remaining few subsections in this section). To demonstrate the level of impact implied by her proposed changes in approach, she
would then (in the following section) briefly outline one example of the very different
NER strategy which might result from this approach. She would next (in another section)
address a relatively simple context, using her proposed approach in a simple form, to provide
an example of catastrophically bad management of safety issues in conjunction with con-
ventional good practice management of directly related concerns, both interpreted in the
best practice EP framework she was proposing. This would illustrate the basis of some key
aspects of her approach in fairly simple terms. These two brief discussions (the next two sec-
tions) would be followed by filling in some of the key aspects of the detail of her proposed
approach (an extensive series of sections), using a more sophisticated version of her basic
initial framework and illustrative numerical examples so they could gradually build a layered
understanding of how she believed they ought to proceed at a what needs to be done level,
the main body of her presentation (and this chapter). She would finish by making a number
of observations on what she believed should be the next steps.
A holistic base- case model set using a carefully structured top- down perspective
The starting point for an NER approach to ALARP had to be a ‘holistic base- case model set’ which captured their railway system’s current safety and security status with strategic
clarity. In Sophie’s view there was no alternative. This was a central aspect of replacing their
current ‘add- on’ approach to safety by a fully integrated ‘designed- in’ approach to safety
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 483
in terms of both accidents in the usual safety management sense and security in terms of
malicious incidents comparable to accidents. She made this point forcefully and emphasised
that this foundation level framing assumption change implied a mindset shift and a culture
shift which she believed they would soon understand clearly and respond to positively, with
follow- on toolset changes which were complex and would take longer to digest.
The purpose of this base- case model set would be to inform the assessment of all planned or imposed changes to the NER system with potential safety or security implications. Some
potential changes to the NER system might be initiated by the safety team, to improve
safety and security. However, the majority of the changes needing assessment would be
suggested by others, to improve operational efficiency and profitability, like bigger or faster
trains. Furthermore, some changes would be externally imposed, like new evidence that a
key subcontractor could no longer be trusted or new information about increased terrorist
activity threat levels.
The effect of Sophie’s approach would be full integration of safety and security concerns
with all other relevant issues at a corporate strategy level whenever this was relevant. Addi-
tionally, the proposed approach would provide a framework for integrated top- down and
bottom- up safety and security planning for NER. She would emphasise the top- down per-
spective today, but full integration with bottom- up planning was crucial.
This holistic base- case model set would use a set of ‘incident’ scenarios which had to be
addressed. She would use the word incident instead of accident for two reasons. The primary reason was NER needed to consider malicious acts and accidents simultaneously in the same framework because the effects of malicious acts may be directly comparable to the effects of accidents and the most effective preventative or mitigating responses may involve common
issues. The importance of understanding any differences in causation could also be crucial when relevant, but this did not diminish the need for joint consideration of some responses to comparable incidents caused accidentally or maliciously. A secondary reason was that for
some purposes it was convenient to consider ‘near misses’ as incidents using a very general
interpretation. Near misses were an example of a form of highly relevant information which
clearly needed consideration and sophisticated subjective probability perspectives.
For assessment to be complete, the set of incident scenarios addressed had to be exhaus-
tive, with no blind spots. Furthermore, when addressing these scenarios, it was crucial to
address all sources of uncertainty, including systemic uncertainty, making effective provi-
sions for all relevant ‘partially and wholly unknown unknowns’.
The NER Safety Department she was now responsible for need not change its name, but
including a remit for security in her department’s role in the sense that she was proposing
was crucial. It was not clarity efficient to treat safety and security separately in terms of mali-
cious acts leading to incidents in terms of her department’s role. The existing NER Security
Department responsible for security in the policing sense would need to contribute to her
Safety Department analysis, and she anticipated much more coordinated effort between the
two departments, but no really new issues of significance now. She would refer to ‘her safety
and security team’ today, assuming that there was no need to even consider changing any
departmental labels for the time being.
Before NER began any detailed safety and security analysis, it was essential to ensure that
their potential incident set was structured so that they could address the most serious NER
incident scenarios first, with the greatest level of effort. This was, in part, a clarity efficiency
concern, but it was also linked to complex corporate bias and blind spot concerns.
This top- down strategic focus would involve significant corporate process changes and
linked culture shifts for the board as well as for her safety and security team and other
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
484 Employing planning tools in practice
groups within NER. These corporate process changes and linked culture changes were badly
needed. The 30 fatalities incident two years ago, followed by the 40 fatalities incident a
month ago, had caught them all by surprise. At board level they needed to completely aban-
don their usual focus on small numbers of fatalities at any given time in an accident, which
was what happened most years. They also needed to abandon an occasional crisis focus on
one or two major incidents in the recent past, the current situation. They needed to start
thinking clearly and systematically all the time every year about ‘what- if ’ scenarios involving levels of fatalities which they had never experienced, and hopefully never would experience,
possibly caused by wilful acts of terrorism, plus all intermediate- level incidences. For exam-
ple, they should start thinking about scenarios like 350 fatalities in one incident, perhaps
with several times that number of people seriously injured, as a ‘plausible catastrophic inci-
dent’ scenario which she would elaborate shortly. They could then address a representative
set of relevant lower level incidents.
This revised perspective suggested a top- down decomposition of NER safety and security
concerns into four ‘types of incident’, associated with four ‘grades of incident’, defined in a hierarchical nested end- event type- grade structure.
The four ‘types of incident’ were
1 ‘collisions’ – when two or more trains are involved,
2 ‘derailments’ – when a train leaves the track but a collision is not involved,
3 ‘level- crossings’ – a level- crossing incident without derailment or collision, and
4 ‘other incidents’ – any other relevant kind of incident.
The four ‘grades of incident’ were
1 ‘fatal incidents’ – when one or more people are killed,
2 ‘injury incidents’ – there are no fatalities but one or more people are injured,
3 ‘property incidents’ – no fatalities or injuries but property is damaged, and
4 ‘other reportable incidents’ – any other relevant observations including near misses.
For this initial meeting she would focus on what she would call the ‘collision with fatalities’
scenarios, 1– 1 incident scenarios in her type– grade structure because the 1– 1 scenarios were
potentially the most serious and the board needed to start with what might matter most. Sophie said she was very aware that ‘derailment caused by maintenance failure with fatali-
ties’ was a subset of her 2– 1 type– grade combination of special interest to the NER board.
However, at their initial meeting she wanted the board to treat ‘collisions with fatalities’ as
the first priority for a board level top- down view, with direct links to derailments and other
types and grades of incidents which she would start to develop in a few moments.
She appreciated that her overall top- down approach using types and grades of incidents
might look baffling for the moment. All the models and data and other knowledge associ-
ated with types of incident 2, 3 and 4 and grades of incident 2, 3 and 4 would have to be
addressed as underlying components of NER consideration of 1– 1 incidents, providing a
foundation for understanding 1– 1 incidents. However, in their first meeting she wanted
them to ‘see the wood’ (the forest as a whole) ‘without being distracted by the trees’ (at
an individual tree level). She would start to develop the underlying thinking now, and she
believed they would begin to see the overall rationale as the day progressed.
She thought it might be helpful to observe now that the models used to understand all
‘level-crossings’ incidents had to underlie some of the models used to understand ‘derail-
ments’ to explain why a level- crossing incident might turn into a derailment which might
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 485
turn into a collision if another train became involved. In practice, analysis undertaken by the
safety and security team would have to start at the bottom and build a pyramid of under-
standing from the bottom up. The reason they were starting their discussion that day at the
top was that 1– 1 scenarios were the most serious potential end- event scenarios, exposing to
danger in the limit all the passengers on two or more trains, plus any other people who were
not on the trains but could be killed by the consequences of a collision.
The top- down perspective the board needed had to see the pyramid built by the safety
and security team from the top. The bottom- up building of that pyramid by the safety and
security team needed to understand the architecture of the whole structure needed by the
board. In this framework, a 2– 1 scenario was a derailment involving fatalities which did not
progress to a collision, and a 3– 1 scenario was a level- crossing incident involving fatalities
which did not progress to a derailment and then a collision.
Building on this comment, the collision with 30 fatalities two years ago might seem
much less important than the derailment with 40 fatalities a month ago. In some respects,
this was the case. In particular, the maintenance failure triggering the derailment had NER
incompetence and possible negligence implications which were crucial, requiring immediate
attention.
However, it was important to avoid too much focus on the specifics of recent major
events. The low end of medium speed nature of the front- to- rear collision involving 30
fatalities two years ago might be viewed as a ‘very lucky’ type of incident relative to a head-
on collision with both trains moving at very high speed in opposite directions, when many
hundreds of fatalities might be involved. The derailment a month ago involving 40 fatalities
might be viewed as a ‘moderately lucky’ type of incident relative to a train leaving the rails
at even higher speed and colliding head- on with another train. They might observe that
level- crossing incidents when a train strikes a vehicle are common and often cause the train
to leave the track.
Finally, and crucially, they might formally and explicitly acknowledge that a very obvi- ous and relatively easy ‘soft target’ for terrorists was a level crossing when two very high
speed trains moving in opposite directions were due. Terrorists might smash a heavy vehicle
through the level- crossing traffic gates just before the first train arrives, in a way calculated
to send the first train into the path of the second train, perhaps using an articulated tanker
full of fuel fitted with impact explosives, with a wide range of possible variants.
In summary, NER needed to ‘learn to be afraid’ and develop the capability and culture to
use all relevant information and knowledge to develop a robust approach to managing all relevant uncertainty, not just focus on trying to making sure specific types of failures in the
recent past were not repeated.
Building this line of thinking further, she had been unable to uncover any evidence that
anyone within NER had ever considered the sort of terrorist threat she had just used as an
example, or a head- on collision involving very high speed trains caused by defective track
maintenance directly attributable to the kind of track maintenance outsourcing currently in
place as recently approved by the board. If incidents of this nature occurred in the context
of no fully documented effective prior analysis, the board would probably face negligence
claims which could permanently destroy the credibility of each and every director as well
as bankrupting NER. She assumed that the board would like her to eliminate this possibil-
ity, using a well- balanced approach which looked after the interests of all stakeholders in a robust and reasonable manner, including shareholders, passengers, staff, and everyone else with related interests.
Sophie indicated serving the NER board in this way was central to her role as she saw it,
and it required ‘opportunity efficiency’, including underlying ‘risk efficiency’ and ‘clarity
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
486 Employing planning tools in practice
efficiency’ in respect of associated analysis, which she would explore that day. One aspect of
this was an understanding of ‘sources of uncertainty’ of a composite nature viewed from the
top down, making due provisions for all relevant ‘partially and wholly unknown unknowns’,
to avoid bias associated with what is particularly difficult and ambiguous.
Many of these concepts have been discussed in rather different terms earlier in this book.
When this is the case, they will not be discussed again here, apart from revisiting some par-
ticularly useful graphical tools later in this chapter. But let’s assume that Sophie would find
the most appropriate way to convey all the relevant ideas to the NER board without unnec-
essarily repeating material discussed in earlier chapters in the tale of this chapter, without
worrying you about the details of how she achieved this.
Two basic probability distribution components of the proposed stochastic model
The holistic base- case model set which NER needed involved stochastic (probability based)
models with two key component probability distributions. Sophie said she would assume for
the moment that both were defined in terms of fatalities associated with collisions, but later
generalisation would address all other types and grades of incidents.
One was a Poisson probability distribution defining Pi, the unconditional probability of i
incidents per annum, with i = 0, . . . , n, and n defining the maximum number of incidents per annum. NER would use this Poisson probability distribution directly to calculate P0, the
unconditional probability of zero incidents (i = 0) in a given year, and NER would use P0
to compute (1 – P0), the probability of one or more incidents in a given year. This Poisson process working assumption would be their starting point for modelling the implications of
the incidents addressed.
The second component probability distribution would define CPS, the conditional prob-
ability of S fatalities, given the occurrence of one or more incidents, for S = 1, . . . , m, with m defining the maximum number of fatalities per annum. This second component would use a HAT- based probability distribution approach which would not require any specific
probability distribution assumptions.
For S = 1, . . . , m, the unconditional probabilities of S fatalities per annum would be
calculated using the relationship PS = (1 – P0) × CPS and the (1 – P0) value obtained via P0.
P0 defines the unconditional probability of zero fatalities per annum (PS for S = 0) as well
as zero incidents per annum (Pi for i = 0). The only use made of the Pi for values of i other
than zero that day would be background understanding for the PS estimation process (for
those interested indirect use of the Poisson process for Pi with i = 1, . . . , n). For all direct
modelling purposes P1 or P5 or P350 implied an S value subscript, not i.
These two component probability distributions would be associated with two key metrics,
L and S.
The metric L determined the value of P0. Underlying the P0 value was the single param-
eter which defined a Poisson probability distribution and process, a metric with an expected
value they would have to estimate. She would define this metric as ‘L’, the ‘level of inci- dent occurrence’, more formally ‘the expected number of incidents involving fatalities per
annum’, interpreted as an unbiased estimate of the average number of incidents involving
fatalities to be expected per annum. Remember ‘L for level’, measured as an expected value for the number of incidents per annum.
The metric S was a ‘size of incident measure’, formally defined as ‘the number of fatali- ties per annum given one or more incidents involving fatalities occurs in any given year’ for
values of S = 1, . . . , m. Remember ‘S for size’, with S = 0 if no fatalities occur.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 487
For illustrative purposes that day she would use a ‘five size scenarios’ portrayal of the S
value range given fatalities occur, with S = 1, 5, 20, 70, and 350. This involved simplifying
working assumptions to avoid working directly with the general size metric S = 1, 2, 3, . . . ,
m, where m was the maximum number of fatalities which was feasible.
She was treating the S = 350 scenario as a maximum plausible number of fatalities in any
given year. This scenario covered, in conditional outcome terms, a range from 200 fatalities
to the maximum number of fatalities physically possible in one or more incidents within a
year, primarily associated with one catastrophic incident, with working assumptions and
implications which she would have to explain in more detail later.
If a 1– 1 scenario occurred, the conditional probability that this incident was an S = 350
scenario, CPS for S = 350, would be a particularly important conditional probability in terms
of defining the high S value conditional probability distribution tail. In overview terms the
key probabilities include the P0 value and the PS value for S = 1 as well as S = 350. The PS
for intermediate S = 5, 20, and 70 values were relevant but constrained by the S = 1 and
350 values.
For all other types of incidents involving fatalities (2– 1, 3– 1 and 4– 1 scenarios), she would
assume the use of the same illustrative example size scenario values, although zero probabili-
ties might be assigned to the higher S value scenarios. A credible narrative involving all the
incident types and grades would have to underlie a credible set of PS for collision incidents
involving fatalities.
For ongoing operational purposes, she would assume that the probability an incident
might occur was constant over time. This implied a stable Poisson process – a random
process described by a Poisson probability distribution function with L constant over time.
This in turn implied that the occurrence of one incident would not change the probability
of another incident, a working assumption that in practice could be modified as soon as
circumstances were deemed to have changed.
A stable Poisson process assumption was common for this kind of modelling, because it
provided a very powerful and robust model based on realistic assumptions, provided the con- stant probability of an incident assumption, with a mean as currently estimated, was routinely tested. This needed to be coupled to ongoing testing for a stable set of CPS values.
The probability of i incidents per annum using a Poisson distribution was defined by the
function
P = e L / i! ,
for i = 1 . . . n, where e was the exponen i
L i− −
ttial number.
L defined both the mean and the variance of this single parameter probability distribution,
a useful characteristic of the Poisson distribution. Tables provided in textbooks could be
used to look up Pi values for particular values of L. In most textbooks, her L (for ‘level’) was
replaced by a Greek letter lambda (or lambda × t with t = the number of time intervals). The Roman L given an average ‘level’ (frequency of incidents) per annum definition was more
convenient for NER purposes.
A two part transformation of the fatalities metric S into a cost metric CS
A transformation which converted fatalities as portrayed by S into an appropriate assumed
monetary equivalent could be associated with all management decision making with safety
and security implications by any organisation which wanted to minimise fatalities and mini-
mise associated cost, whether or not all relevant parties chose to recognise this was the case.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
488 Employing planning tools in practice
Sophie emphasised the existence of such a transformation was not an option – it was an
unavoidable fact of life.
What was an option was those involved recognising this reality explicitly and using asso-
ciated insights effectively and efficiently. In her view, it was essential that the NER board
explicitly and formally acknowledged a leadership role in the definition, use, and interpreta-
tion by others of this transformation. Furthermore, the board had to require explicit and
formal acknowledgement of two inherent and inescapable properties of this transformation
by all parties involved in NER safety and security decision making or concerned about the
decisions made.
First, this transformation was a shadow cost concept, because NER wanted to minimise
fatalities and minimise the cost of achieving an appropriate minimum number of fatalities, which implied an inescapable trade- off between fatalities and cost. NER had to explicitly rec-
ognise and use this trade- off consistently for all decisions which influenced safety and security
to avoid wasting lives, as well as to avoid wasting money. This could be viewed as being ‘fatal-
ity efficient’, with a minimum level of fatalities for any given level of expenditure on avoiding
fatalities. A shadow price (instead of shadow cost) alternative interpretation would useful if minimising fatalities and maximising profit was the preferred conceptual framework, but she
recommended avoiding this perspective, for a number of reasons, some fairly obvious.
Second, this shadow cost transformation was also inescapably a ‘revealed preference’ con-
cept. A trade- off rate between avoided fatalities and cost revealed by NER decisions revealed
board approved trade- offs. These trade- offs ought to reflect actual preferences for all parties with a legitimate interest in the implications of the value being used, for ethical reasons with
legal implications in ALARP terms. The ultimate decision makers responsible for the use of
this transformation were dealing with the lives and money of other people, with inescapable
ethical and legal implications in terms of articulating the different interests of different par-
ties and taking all relevant legitimate concerns into account. An explicit requirement for all
relevant parties to have an appropriate level of understanding of what this implied was an
implication.
The transformation employed had to involve some form of ‘value of an avoided fatality’
concept. The ‘value of an avoided fatality’ was a widely used concept with a range of inter-
pretations which she would discuss in terms of current practice within NER and elsewhere.
This discussion would include her interpretation of key areas of contention associated with
different approaches to decision- making practice which were relevant to NER.
At a basic working assumption level within these framing assumptions, she advocated
using a two part transformation involving the option of further decomposition of the second
part. She would begin with the basic two part transformation now and explain the further
decomposition option for the second part shortly.
The notation she would use for a ‘cost associated with fatalities’ measure CS and a ‘two
part transformation factor’ TS was:
C S T , with
T V E ,
V defined as an appropriate benchm
S S
S 1 S
1
= ×
= ×
‘ aark value for one avoided fatality ,
E defined as the evS
’
‘ eerything else factor when S fatalities were involved .’
Her starting position was that NER needed a carefully selected benchmark value for V1 asso-
ciated with a trusted organisation independent of NER which would be generally accepted
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 489
as a reasonable basis. She had in mind V1 = $2,000,000, the value of an avoided fatality
currently used by the National Highways Agency (NHA) in NER’s country to make all
road planning decisions when safety issues were involved. This provisional choice assumed
verification by the NER safety and security team that this was a sound benchmark choice
and approval by the board. Assessment of other country’s comparable values and alternative
benchmark possibilities would be part of the verification process.
She assumed that when only one fatality was involved, ES = E1 was a factor for transform-
ing the V1 benchmark value chosen by NHA into a suitable value of one avoided fatality in
an NER context, reflecting all relevant views about the implications of one fatality in this dif- ferent context. One obvious key difference if NER used the NHA $2 million for V1 was the
NER association with railway travel rather road travel. Another was that NER was a private
sector organisation while the NHA was a public sector organisation. E1 had to reflect these
two differences plus all other relevant differences.
She assumed that when S was greater than one, ES was a factor which reflected E1 plus an
appropriate increase to embody all relevant views about the implications of more than one person being killed in the same NER incident. In her view, NER had to acknowledge that
most people believed that 1,000 people killed at the same time in one incident was much
more serious than 1,000 people killed one at a time in 1,000 separate incidents, with inter-
mediate valuations of 100 or 10 fatalities in one or multiple incidents. This was reflected in
the observable differences in approaches to safety by highways agencies and air travel indus-
tries and public acceptance of those approaches. It was widely accepted that risk of death or
injury when travelling by road, rail, and air was not the same, nor should it be assumed to
be seen as the same by the travelling public, for a very complex set of reasons. One reason
was how many people tend to be killed or injured in any given incident. Other concerns
included the level of trust in the organisation and the nature of the death involved.
The starting point for defining TS = V1 × ES equivalents in all contexts ought to be clarity about what she would refer to as ‘revealed preferences’ associated with:
1 preferences expressed by each of the parties with a legitimate interest other than the NER
board, as interpreted by the NER safety and security team for use by the board;
2 a synthesis of all the components of the set of revealed preferences associated with point
1, as interpreted and reflected in revealed preferences defined by NER board level deci- sion making.
The board was responsible for prudent decisions involving the lives and welfare of other
people as well as other peoples’ money. This implied that gathering and interpreting the
implications of revealed preferences associated with all the parties directly involved other
than the NER board ought to be a crucial aspect of the board’s decision- making process,
effectively supported by the NER safety and security team, and tested by other involved
parties. To fail to take proper account of the preferences of all relevant parties was arguably negligent, even if members of the board did not appreciate this was the case, and not all other relevant parties appreciated this was the case.
The choices the board members approved would reveal the board’s preferences, and each individual board member was accountable for those revealed preferences, both legally and in a more general ethical sense. In the EP framework that she would explain, this was the case
whether or not they chose to understand that this was the case, and an important feature of
the EP framework was making this clear. In her view every member of the board needed to
understand why this was the situation and all the key implications by the end of their initial
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
490 Employing planning tools in practice
meeting. Subsequently, everyone else with an interest in NER decisions would also have to
understand this was the situation, and the implications, at a level of clarity appropriate to
their interests.
‘All the parties directly involved other than the NER board’ clearly included rail passengers
plus on- board NER staff and the families of both these groups. It also included those paying
for the railway travel, which included most rail passengers plus all citizens of NER’s coun-
try because of significant government railway travel subsidies. Furthermore, it included NER
shareholders. However, many approaches to estimating TS values began with cost– benefit
analysis, and some organisations never really got beyond this starting point. In practice, the
values of TS = V1 × ES employed by board level decision making should also pay close atten-
tion to interpreting the revealed preferences of other indirectly interested parties not yet men- tioned. For example, the direct cost implications of compensation awards by the courts, plus
the amounts involved in ‘out of court’ settlements, should be important practical concerns.
The board also needed to consider the direct and indirect knock- on implications of hostile
media attention and the consequences of commuters switching to the use of cars instead of
trains, with immediate revenue effects, plus knock- on political and regulator implications.
NER had to recognise that small numbers of fatalities in frequent NER incidents might
be seen as comparable to road traffic accidents in some respects, but very large numbers of
fatalities in one incident would be seen by many people as comparable to airplane accidents,
whether or not NER safety experts and other safety experts currently shared this view. To be
more specific, plausible NER incidents could span a range of fatalities which was so wide that
NER could not expect to get away with their current assumption of a single constant ‘value
of an avoided fatality’ relevant to any feasible S value. There was a good case for highways
agencies using a single value appropriate to the limited number of fatalities usually needing
consideration in road traffic accidents, viewed as a relevant average. There was a somewhat
different but reasonable case for aircraft travel organisations using a single value associated
with average numbers of fatalities in major airplane crashes, despite the larger range involved
than that associated with road traffic accidents. But NER had to explicitly confront an unu-
sually wide range, with most incidents clustered at the bottom end, some very low probabil-
ity but very high impact incidents at the top end, plus infrequent but moderate probability
intermediate level incidents like the 30 and 40 fatality incidents in the last two years. In her
view, moving to an S dependent ES was an essential generalisation of NER current practice,
with very important implications the board needed to understand.
To understand the implications of her TS = V1 × ES framework when used by NHA or
any other highways agency, a simplified T = V1 × E form could be assumed, assuming E was
not dependent on S. When any highways agency first used an avoided fatality transforma-
tion, V1 might have been based on a cost– benefit analysis approach to a single fatality, and
E might have been a constant uplift which was implicit and set to unity. But later on, the E
value might have included an uplift to reflect feedback from road users, road funders, and
other interested parties, plus inflation since V1 was first disclosed, and possibly plus a further
uplift to reflect more than one person being killed in the average incident involving fatalities.
The highways agency might have never actually used T, V1, ES, and E parameter equivalents
explicitly, simply employing a subscript- free T = V equivalent for the average number of
fatalities when incidents occur which was updated and redefined in the updating process on
a regular basis, with a subscript t which was implicit to indicate the period, perhaps made
explicit when relevant. Her current working assumption was that this was the case.
This notation for this kind of approach would be useful for their discussion shortly, and
she wanted to emphasise now that she thought most highways agencies took a reasonable
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 491
approach on the whole, viewed within her TS = V1 × ES basic framework, recognising that
they probably actually used another framework.
She thought aircraft industry use of related approaches, insofar as she understood them,
were also very reasonable, and it was worth saying so now, although the plausible range of
fatalities which might be involved was clearly very different.
Furthermore, there were a number of reasons why she thought aircraft industry safety and
security approaches might offer NER some very useful ideas not usually found in approaches
to road traffic safety or railway safety, but that was a matter for future discussions.
A key board level strategic issue needing clarity
Before getting into building the requisite ‘how to do it’ detail needed for strategic clarity by
the board, Sophie thought an overview of the nature of the strategic issues which she was
convinced needed attention might be useful. She suggested a very brief discussion of one
example of a potential new strategic view of ‘where NER should be going’, to motivate the board’s interest in understanding ‘what needed to be done’ in terms of ‘how it might be
done’ and why the degree of complexity being advocated was necessary.
She began by indicating that the current NER strategy in terms of high speed intercity
services 20 to 30 years into the future, as she understood it, was trains of about the same
passenger capacity as current trains operating on modestly modified track layouts with com-
parable timetables with top speeds about 50% faster than the current trains. She believed
that the approach she was advocating might lead them to the view that even current high speed trains involved a level of risk NER should not be taking, and enlightened prudence
required much smaller trains in the future to reduce the scale of a potential catastrophic
high speed collision incident. If no other changes were made, the additional costs might
seem unacceptably high. However, if more frequent trains as a result of smaller trains were
seen as a way of improving the service as well as reducing journey times, this was done in
conjunction with a coordinated approach to improving service levels for other connecting
train services, and an approach to all ticket pricing which significantly reduced crowding at
current peak times, they might increase passenger satisfaction and NER profitability while
increasing safety and security at the same time. This might or might not involve significant
increases in passenger numbers and changes in their plans in terms of track layouts and other
relevant interdependent issues.
One key reason for suggesting this possible change in strategic direction was NER’s need to embrace ‘enlightened prudence’. Enlightened prudence for NER might involve being
prepared to reduce the size of all high speed trains in order to reduce the chance of cata-
strophic incidents involving deviations from expected costs which might put NER out
of business. The concern might be reducing ‘risk’ as risk is viewed in both plain English
and EP terms, by altering the risk– reward trade- off while achieving and maintaining risk
efficiency. Currently NER used a definition of ‘risk’ for formal safety risk management
planning purposes which did not facilitate NER addressing enlightened prudence because
it did not address risk in the generalised Markowitz based risk efficiency plus opportunity
efficiency EP sense. The NER current definition of ‘risk’ in a safety context for corporate,
operations and project planning purposes limited what they formally addressed to expected
outcomes, which was unacceptably myopic from an EP perspective. In all relevant contexts
when safety was considered, NER needed to address risk efficiency in a general sense.
NER also needed to address risk– reward trade- offs and clarity efficiency in an opportunity
efficiency framework.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
492 Employing planning tools in practice
A second key reason was NER’s need to understand that a failure to address enlightened
prudence in a risk efficiency plus opportunity efficiency framework was aggravated by their
failure to address an S dependent ES concept or understand the implications of these inter-
dependent framing assumption failures.
A third key reason was NER’s need to understand why the distortions associated with her
first two points earlier were aggravated and amplified by biases associated with the way they
approached long tailed probability distributions.
A fourth key reason was a further compounding of these concerns associated with NER’s
need to understand the importance of ensuring that all negative aspects of catastrophic inci- dents correlated with high numbers of fatalities needed joint consideration at the same time to avoid underestimating the importance of such incidents and to capture clarity efficiency
benefits.
The fifth reason she would mention was NER’s failure to consider passengers’ preferences
for safer travel on all NER train services in conjunction with passenger’s preferences for more frequent trains with crowding at peak periods reduced, even if they had to pay more.
In some cases, these preferences might be linked to a willingness to travel more at different
times if travel was more comfortable, as well as safer, on more frequent trains, allowing con-
venient transfers from high speed to local stopping services, as just one example of complex
interdependencies which matter.
Sophie emphasised this particular possible change in strategy example was based on her EP perspective leading to immediate recognition of fundamental flaws in the current NER
approach to conceptual and operational tools for strategic planning involving safety, and the
nature of some reasonably likely implications. However, her currently limited understand-
ing of railways in general, and NER in particular, meant that she could not yet be definitive
about the exact nature of any needed NER changes in this kind of passenger service strat-
egy. What she could be sure about was the need for all relevant NER staff to reframe NER
understanding of their railway business using the EP approach she would provide, and then
use this enhanced perspective in a creative way based on their extensive railway industry
expertise.
It would take NER some time and effort to achieve this full integration of expertise. She
would focus the initial board level discussion on gradually building a layered understanding
of the key framing and working assumptions underlying the operational form the new NER
approach would take. The board would then be able to see in more detail where they were
going, how they could get there, and why a new approach was necessary.
The 1970s Ford Pinto controversy in the US
At this point Sophie indicated the board would find it useful to explore a context which was
much simpler than the NER situation. They would consider using the simplified T = V1 × E
version of the basic TS = V1 × ES framework in the context of road travel safety. This would
clarify some basic issues without being distracted by the complications associated with an S
dependent ES.
To begin to explore the implications of her TS framework at a T = V1 × E level of complex-
ity, the Ford Pinto ‘controversy’ as set out in The Gift (Hyde, 2006) provided Sophie with a useful example which could be linked to the NHA basis of her V1 = $2 million proposal.
Most North Americans who were adults in the 1970s knew about the Ford Pinto con-
troversy, which severely damaged the reputation of Ford as well as cost– benefit analysis as
used in this context. NER board members might not have heard about it, but it was an
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 493
internationally renowned example of how not to approach safety which still attracted com-
ment, and it provided a useful example for them to consider briefly.
‘Controversy’ was Sophie’s choice of word to keep the discussion neutral and avoid emo-
tive language. However, some people called the Ford Pinto ‘controversy’ a ‘catastrophe’,
‘disaster’, ‘debacle’, ‘travesty’ or ‘outrage’, and she had some sympathy with all these views.
It was a catastrophe for those who lost loved ones in Pinto accidents involving fires resulting
from low speed rear- end collisions which could and should have been avoided. It was also a
catastrophe from the perspective of Ford Motor Company shareholders and the Ford board
when the US press got involved. From a current perspective, she believed the Ford board
was lucky to have escaped very serious charges.
Hyde, in The Gift, started his discussion of the Pinto controversy by observing that it was ‘a classic example of both cost– benefit analysis and the confusion between “value” (as cal-
culated by economists) and “worth” (as understood in broader frameworks which recognise
the importance of “gifts” and other non- market- based concepts)’. Hyde’s basic concerns
about important limitations of cost– benefit analysis and Ford’s use of it were well- founded
and still current issues. NER did not want to replicate ‘the Ford Pinto experience’.
Hyde noted that when Ford introduced the Pinto in 1971 (a relatively small ‘compact’
car of a kind not previously produced by Ford in the US), they were aware that the designs
involved a fuel tank prone to rupture in rear- end collisions, even at low speed, with conse-
quent avoidable fire risk. Ford had several tested options to reduce the risk, the most expen-
sive (presumably the most effective) costing $11 per vehicle.
According to Hyde, an internal Ford memo estimated that if the Pinto was sold without
the $11 safety feature, 2,100 cars would burn every year, 180 people would be hurt but
survive, 180 people would burn to death. To assess the option of installing this safety fea-
ture Ford used a US National Highway Traffic Safety Administration (NHTSA) value of
an avoided fatality of $200,000. In Sophie’s terms, $200,000 was a T = V1 × E value, an
average value for all relevant S, because NHTSA was aware that 2 or more people might be
killed in the same accident, and it was using a common value for all of them. For simplicity
she would assume that E was equal to unity, all V1 components were in the money of the
day, and T = V1 = $200,000.
Hyde itemised the NHTSA components of T to show that it included $10,000 for ‘vic-
tim’s pain and suffering’, along with $900 for ‘funeral’, and $173,000 for ‘future productiv-
ity losses’. Ford used $67,000 for the average T equivalent for people injured (presumably
related to a comparable NHTSA figure), $700 for each lost vehicle. Ford used a market
estimate of 12.5 million vehicles per year. The resulting ‘benefit’ (money saved by a safer
car) was $49.5 million. The corresponding ‘cost’ (money spent on safety devices) was
$137.5 million. Based on these figures Ford decided not to install any safety devices. It
would seem that the obvious option of reconsidering the whole design concept was not
addressed.
What brought these internal Ford memos into the public domain was in part the reliabil-
ity of the Ford forecasts of fatalities. Hyde noted that according to a Mother Jones magazine article by Mark Dowie, by 1977 at least 500 people had burned to death in Pinto crashes.
It was important for the NER board to understand that from her current perspective, she
would be very comfortable defending the NHTSA approach if it had formally recognised, in a way that Ford in the 1970s and Hyde in 2006 plus everyone else involved would have
subsequently been aware of, that this kind of T value was a plausible minimum to be used as a starting point for a discussion about the role of shadow costs, revealed preferences and all ethical issues appropriate to the NHTSA context; but it should never be used in other contexts, like the
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
494 Employing planning tools in practice
Ford Pinto decisions, without a very carefully considered set of changes, including an E factor much greater than unity.
It was even more important for the NER board to appreciate that Ford’s decisions could not be defended from any perspective, at any point in time, by anyone who had a moral compass and modicum of common sense. The reasons were very obvious in her TS = V1 × ES frame- work, and they should have been obvious enough to Ford even without her EP framework.
To ensure that the board understood the EP perspective which NER had to adopt, she
would use her TS = V1 × ES framework to review all the key issues which NHTSA got abso-
lutely right, plus those where improved practice would have been helpful. She would then
comment on the Ford position and finish by linking this discussion to the NER position.
The key point to begin this discussion was that the NHTSA needed some way to shape the
allocation of public funds to public road safety improvements, and they needed some basis for
a benchmark starting point, a V1 in her terminology. If they simply gave up on TS concepts,
they would have no way of influencing the level of fatalities for whatever safety budgets those
spending the money on public roads in the US could make available. In the absence of other
organisations that took over the public’s concern about road safety in the USA, the NHTSA
needed a V1 value or it was doomed to wasting lives as well as money. The NHTSA was an
important pioneer in this area and deserved full credit for formally and explicitly embracing a
pioneering version of her TS based approach and publishing its methodology and value esti-
mates. A cost– benefit analysis basis for a V1 concept was arguably the only obvious option at
that time, and the NHTSA did implicitly allow it to acquire a shadow cost status via adjust-
ments incorporating feedback on public exposure of its initial approach and value estimates.
Her second point was assuming that an S dependent approach was not necessary was a
very reasonable NHTSA assumption, because their focus was small numbers of people being
killed in each of a very large number of road traffic accidents. Although high speed bus colli-
sions or major pile- up accidents would make S dependent generalisation worth considering
if motorway crash barriers were being addressed, this was not an obvious initial concern.
So, using an initial simplified framework with T = V1 × E and V1 = $200,000 with E = 1 was
very reasonable. This T or V1 value needed to be interpreted as a per fatality transformation
based on the average number of fatalities in road traffic accidents in the US around 1970,
using a cost– benefit basis which was assumed to be suitable for the average person killed in
typical road traffic accidents.
Her third point was that injuries and physical damage had been treated separately by the
NHTSA, with injuries treated via a transformation linked to V1 like the Ford $67,000. In
the 1970s NHTSA context was reasonable. It was still global common practice.
Her fourth point was the NHTSA explicitly adopted a full disclosure approach and
implicitly moved towards a revealed preference interpretation of their shadow cost concept.
They published the details of their approach to V1 as well as the $200,000 value. This full
disclosure approach allowed road users and road funders (largely but not exactly the same
set of people) who thought $200,000 was too low, or too high, to provide an argument
for change which allowed a very crude revealed preference adjustment to the initial value
selected. Implicitly, the NHTSA associated a revealed preference status with its T value and
embedded an ethical content in E by responding to feedback.
Her fifth point was that NHTSA contributed to the Pinto controversy and failed to escape
criticism itself by four omissions, with closely coupled implications:
1 failing to formally treat T as a shadow cost concept, 2 failing to formally acknowledge a revealed preference status for a T concept,
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 495
3 failing to formally acknowledge that the relevant revealed preferences were always inherently ambiguous and difficult to capture in a fair and unbiased manner, and
4 failing to formally acknowledge a need for an E parameter adjustment to V1 which had an ethical content which those determining T had considered and addressed in con-
junction with the shadow cost and revealed preference issues.
Elaborating on the fourth omission, which addresses embedded responses to the other
three, if the need for ethical content in a T value, which was both a shadow cost concept
and an effectively managed revealed preference concept, had been formally addressed, then
an E = 1 assumption would have clearly implied that T = V1 = $200,000 calculated in this
way was usefully interpreted as a ‘plausible starting point’ for NHTSA use. It would have
been clear that this value was based on a particular economic analysis approach in a par- ticular context, not an evaluation which organisations like Ford should ever use directly and not a value NHTSA could expect others to trust until feedback had been considered. This
particular economic approach could have no real ethical validity until the values involved
were reviewed by those affected by associated decisions. It did not reflect what any of the
people involved thought about the approach used, or the numbers produced, and its nature
was totally unsuitable for direct use by Ford.
If used directly by NHTSA, with no acknowledgement of the need for an operational
shadow cost trade- off or of a balanced and fair way of identifying relevant revealed prefer-
ence concerns and then embodying related ethical concerns, it was not likely to receive a
majority vote or engender trust. It was not a revealed preference concept unless and until all relevant parties agreed that its use in this way was appropriate. However, if it had been
made clear that lives and money would be wasted unless a shadow cost was agreed on and
applied consistently by the NHTSA and a formal approach for receiving and balancing feed-
back from all involved parties in a fair way had been put in place to move towards a robust
revealed preference E value, arguably the NHTSA would have escaped all serious criticism,
and Ford might not have initiated the Pinto controversy in the first place.
In a 1970s context, all these NHTSA shortcomings from an EP perspective were regret-
table but understandable. The NER board might note that adjusted for inflation, and other
relevant factors, the T value used by the NHTSA of $200,000 was of the same order of
magnitude as the $2 million currently used by NHA, and both values seemed plausible sums
in the intended context as far as she could currently judge, given suitable caveats about their
interpretation.
Ford should have approached the role of trust and ethics with a clear recognition that
their position was entirely different from that of NHTSA, despite NHTSA not making the need to do so explicit in formal terms. People could buy other cars if they felt Pintos were
not safe. Ford had no intention of full disclosure. Ford’s use of any T = V trade- off suit- able for NHTSA use without significant E factor uplifts with a clear ethical basis and other
careful revisions to the approach was unacceptable for very obvious reasons from an EP
perspective. If Ford employees did not have the moral compass to see this difference in
ethical terms, they should have had the insight to recognise the trust aspects in legal terms,
and the inescapable fact that all the other interested parties involved would see their behav-
iour as inept, immoral, or criminal. Even a hardened cynic ought to have seen the chances
of Ford not getting caught and being accused of being stupid at best, arguably criminal,
was negligible. If Ford had added $11 to the cost of a car (implying an E value of about
3 or more), they would perhaps have sold marginally fewer Pintos and possibly made less
money, but there would have been no public sector trade- offs between expenditure on roads
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
496 Employing planning tools in practice
and expenditure on other public services. The ethical issues were very different. Somebody within Ford should have understood this, and there was a good case for arguing ‘the buck
stops with the board’.
As an interjected aside, we might observe that from an EP perspective it is obvious that
Ford should have spent the $11 per Pinto as a minimum. But a redesign would have been
much better, and getting the design right at the outset would be better still. Martha’s
Chapter 6 approach could provide the foundations of a basis, treating safety as a feature of
cars which purchasers want, along with reliability, fuel economy, acceleration, and so on,
the non- price value added or quality features which Trevor and Sian were concerned about
in Chapter 6 which need building into products offered to potential purchasers. There was
a clear need for Martha’s ‘red face test’ built into Ford’s design and engineering processes,
along with suitable ‘halo tests’, ‘smiles tests’ and ‘frowns tests’. On top of this basis, some
features of a simplified and significantly revised variant of Sophie’s NER approach as devel-
oped in this chapter might be used.
Returning to the tale, linking the NHTSA and Ford cases to NER, Sophie suggested that
the way forward was a variant of the NHTSA approach. The shortcomings she had identified
needed to be fully rectified. Furthermore, an S dependent ES concept incorporating a four
part decomposition generalisation was needed. They had to embrace what was common as
well as what was different when comparing the NHTSA, NHA, and NER. For example,
despite the private sector nature of NER, they needed an approach moving towards full
disclosure with their regulator, as well as other parties, which she would comment on later.
As she had suggested earlier, given adjustments for inflation and other relevant factors, the
$200,000 figure used by NHTSA in the 1970s was roughly comparable to the $2 million
currently used by NHA, the highways agency responsible for road traffic safety in NER’s
country. She was not an expert in the road safety area, but she understood that the road traf-
fic authorities in most countries had started with T values significantly below the NHTSA
figure, experienced T value inflation which might be attributed to ‘revealed preferences’ of
road users and other citizens, coupled to growing concerns about safety plus growth in real
incomes as well as inflation. This had led to a revised T which could now be treated as a
reasonable approximation to the kind of T = V1 × E she would recommend.
She wanted the NER safety and security team to look very closely at the provenance and
robustness of the $2 million NHA value, but it was clearly relevant, and she would assume
an NHA based V1 = $2 million working assumption for NER to use that day.
Using numerical examples to clarify the board’s understanding
Sophie started the next section of her presentation by explaining that some of the key con-
cepts the board needed to understand were not going to be considered until near the end
of her presentation – ‘enlightened prudence’ for example. This was because she believed
the easiest way for board members to understand such concepts was via illustrative numeri-
cal examples demonstrating the use of decision diagrams plus underlying sensitivity dia-
grams to consider trade- offs when increasing the expected cost to reduce risk in terms of
departures from expectations was worthwhile. But before she could demonstrate this, she needed to explain how all NER decisions involving safety and security concerns needed
to address making risk efficient choices using decision diagrams and sensitivity diagrams.
Before she could do this, she needed to build example numerical models which provided plausible portrayals of expected outcomes and associated variability in terms of all the costs
relevant to choices between the current system and alternative proposed new systems. Before
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 497
she could do that, she needed to build underlying plausible illustrative numerical models which transposed fatalities into ‘cost related to fatalities’. And before that, underlying models which represented a plausible illustrative portrayal of the probability of any given number
of fatalities per annum needed consideration. So, they were going to start at the bottom
of an analysis structure using illustrative numerical examples, and gradually build towards
a full understanding of ‘what needed to be done’ and ‘why an EP approach made sense’,
deferring discussion of key considerations like enlightened prudence until they were reach-
ing the top. Sophie had developed her layered approach to keep the discussion in this initial
meeting as simple as possible in a systematic manner. But the issues were very complex, and
she had not been able to make her presentation as short and simple as they might all wish
had been feasible.
As an interjection, this is going to be the first of several significant sections in a long chap-
ter, and in practice Sophie would have had to provide breaks, which will not be mentioned,
and explain EP concepts which we will gloss over when the issues have already been covered.
Your patience will be stretched as well as the NER board’s endurance. But we are addressing
the most complex interdependent concepts this book considers, the route taken is the most
painless I could evolve, and I believe that you will begin to understand the rationale and find
the journey worth the effort before the chapter’s end.
Estimating key parameter values with an initial focus on L, P0 and (1 – P0 ), then CPS
Sophie indicated that the first step was addressing an estimate of the incident ‘level’ L – the
metric for collisions with fatalities in terms of the expected frequency level of incidents per
annum. L defined the P0 and (1 – P0) associated with a Poisson process of incident occur-
rence, the probability of no incidents and the probability of one or more incidents per
annum.
Associated CPS defined the conditional probabilities of S fatalities given an incident
occurs. They could use P0, (1 – P0) and CPS estimates to define PS for S = 0, 1, 5, 20, 70, or
350, the unconditional probability of no fatalities or one of five ‘size’ scenarios she would
use to illustrate a general EP approach. PS = (1 – P0) × CPS is the relationship they would
use to convert conditional probability estimates into unconditional probability estimates.
Any estimated values of L and the CPS used to compute the associated PS were defined
given current physical systems (equipment), procedural systems (including information
systems, maintenance regimes and associated contract structures), culture (including staff
morale) and external factors (including terrorist threat levels). L, CPS, and associated PS
would need estimation in terms of unbiased expected values. Variability in the number of
fatalities each year, given these expected values, needed direct modelling which she would
consider shortly. Variability associated with errors in the expected values for L and the
CPS required explicit consideration, but formal modelling of this higher order uncertainty
beyond stating ranges and listing component issues of sources of associated uncertainty
would be deferred.
To clarify their initial discussion, she had reviewed NER estimates related to L provided
by Oscar. She had concluded that 0.05 was a plausible estimate of L for illustrative purposes.
She would associate this estimate of an expected value of L with a plausible uncertainty
range of approximately one order of magnitude, from 0.01 to 0.10.
She made it clear that this uncertainty range might be a good deal wider than some board
members anticipated, but she believed it was realistic until a lot more analysis had been
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
498 Employing planning tools in practice
undertaken. Furthermore, L = 0.05 might be higher or lower than anticipated by board
members, depending on the extent to which they believed the recent collision incident was
bad luck or as yet unresolved bad management. She might want to make significant adjust-
ments to this illustrative expected value for L, up or down, after further analysis. However,
for illustrative purposes L = 0.05 would serve their current needs.
In practice, it would be crucial to decompose uncertainty associated with estimating the
expected value of L and associated CPS into at least five components when they did pursue
further formal analysis:
1 uncertainty associated with data- based estimates using NER past experience;
2 uncertainty associated with data- based adjustments using relevant experience from
other railway systems, plus a clear understanding of all related competency levels;
3 uncertainty associated with adjustments based on judgements of NER system change
effects;
4 uncertainty associated with adjustments based on judgements of malicious incident
threats; and
5 all other uncertainty which might be relevant.
Her current views on the expected value of L = 0.05 were dominated by item 1 and discus-
sions with Oscar because she had no basis for items 2 to 5. But her current views on the
uncertainty range were dominated by her uncertainty about items 2 to 5.
The fourth component, associated with malicious acts, should not be too significant in
expected value terms for L. Even if they took a pessimistic view, a reasonable estimate might
be of the order of 5% of the assumed L = 0.05, perhaps less. However, this fourth com-
ponent needed estimation with great care, and it raised causal dependence issues with the
CPS expected values which might prove important. If malicious attacks took place, the con-
ditional probabilities of large S values might be much bigger than they were for ordinary accidents because that would probably be the intent.
Some directors might intuitively associate her illustrative example L = 0.05 with a bino-
mial process, which could be interpreted as a special case of the Poisson process when
only zero or one incident was possible. She thought those inclined to this binomial model
interpretation might like to know that in terms of her Pi notation, an example textbook
table of Poisson distribution values (Yamane, 1973, page 1106) would show that for
L = 0.05:
P 0.9512, which might be rounded to 0.95,
P 0.0476, which 0
1
=
= could be rounded to 0.05,
P 0.0012, which could be round2 = eed to 0.00,
P 0.0000.3 =
This clearly implied no chance of three or more incidents in any given year working to four
places of decimal, a binomial approximation being reasonable for this fairly small L value if
two places of decimal were used.
However, NER could use the Poisson model as their basic model without requiring bino-
mial distribution approximations, and the 0.0012 probability of two incidents in a single
year could be captured by the CPS. This generality would be important if much bigger L
values were relevant in other contexts and they did not want to use shorter time periods.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 499
To keep her illustrative numerical examples reasonably simple, she would use the assump-
tion P0 = 0.9500, working to four significant figures, implying (1 – P0) = 0.0500 but a
slightly smaller L than 0.0500, still 0.05 to two places of decimal. This would be convenient
because P0 and (1 – P0) were the probabilities NER need to work with, L underlying the
definition of both, and it was useful to distinguish L = 0.05 and (1 – P0) = 0.0500.
The value of L = 0.05 and (1 – P0) = 0.0500 was uncomfortably large, and plausible values
in the uncertainty range were larger still – collisions involving fatalities in the sense being
addressed were reasonably low probability events, but they were not rare events.
A useful feature of the Poisson model was 1/L defined the average time between inci-
dents. Instead of thinking about L as an estimate of ‘the average number of incidents per
year’, it was often useful to think in terms of 1/L as ‘the average interval between incidents
in years’, commonly referred to as ‘the return period’. Assuming L = 0.05 implied a 1/L
return period of 1/0.05 = 20 years. Many people found it easier to think about a context
involving an L = 0.05, (1 – P0) = 0.0500 and P0 = 0.9500 as for their current example in
terms of a return period of 20 years using the 1/L interpretation.
It was important to understand that the ‘20 year return period’ perspective implied one
incident involving a collision with fatalities every 20 years on average, but for all practical purposes any realistic notion of the average would rarely happen. Even with the usual ‘ran-
dom process means Poisson Process’ assumption that the probability of an incident remains
constant over time, bunching was likely to be observed – in popular parlance ‘buses often
come in threes’. Two incidents in two years did not mean no more for a while. Tomorrow
could be the day their third bus arrived, and the day after might be the fourth.
Different board members might prefer different ways of thinking about what was involved,
but several related issues needed clarity for everyone:
1 The uncertainty associated with an expected value estimate of L or P0 was currently
significant, and it would be both disingenuous and dysfunctional to suggest otherwise.
2 The subjectivity associated with estimates of L or P0 and associated uncertainty required
visibility, to clarify all working assumptions, and to avoid the dysfunctional stealth
assumptions inherent in any pretence of ‘objectivity’ in a simplistic textbook ‘objective
probability’ sense. If ‘objectivity’ was interpreted as ‘reasonableness’, in the sense that
equally well- informed and enlightened people could agree proposed decisions without
necessarily agreeing on the grounds for those decisions, then the only way to achieve
‘objective decisions’ was via ‘subjective probabilities’, with high clarity about all rel- evant assumptions. This applied to L and (1 – P0), their current concerns, but it also
applied to the CPS, and it applied to both the V1 and the ES values underlying TS
transformations.
3 The NER ‘safety and security team’, in the narrow sense of ‘the set of experts she
managed directly’, needed to discuss the models and analysis underlying the range and
expected value and agree they were ‘reasonable’ before this expected value and range
became the basis of a board approved ‘NER estimate’. The board would have to under-
stand in broad terms the basis of these recommendations.
4 It was important that everyone involved understood that more effort could reduce the
uncertainty range but not eliminate it, and the implications of the cost of using scarce
resources to reduce uncertainty. Uncertainty about all relevant probabilities and all rel-
evant transformations needed consistent clarity efficient treatment.
5 The board members needed to confirm that they understood what the safety and secu-
rity team was doing well enough to confirm the NER estimate was ‘reasonable’ before
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
500 Employing planning tools in practice
the associated conclusions became ‘the NER policy’. The board were directly responsi-
ble for NER policy, including, in broad terms, the approach adopted by the NER safety
and security team, the level of effort involved, and the competence of those undertak-
ing the analysis.
6 All NER staff and contractors plus subcontractors whose work had an impact on NER
safety and security had to be part of the NER ‘safety and security team’ in the broad
sense of ‘working together to deliver NER safety and security’. Crucially, they all needed
an appropriate level of understanding of the way everyone had to work together to
achieve shared goals.
7 It was also important to understand that if L or P0 were viewed in a Bayesian frame-
work, integrating ‘prior’ views with ‘posterior’ views based on recent data, many people
would see P0 = 0.9500 for all collision with fatalities scenarios as far too optimistic in
the light of the last two years. Whether or not a formal Bayesian perspective was evoked,
a very pessimistic view of NER safety by the public and relevant government agencies
should be understandable.
8 It was important not to panic. There were very serious issues which needed addressing,
but to some extent ‘bad luck’ might underlie the two NER incidents in two years. To
most people, ‘too much bad luck’ was the same thing as ‘careless’, or even ‘negligent’,
a reality they had to deal with. But they had to keep their nerve and focus on achievable
changes which made matters better.
Her discussion that day would focus on NER collisions with fatalities, but all type– grade
scenarios involving fatalities plus those involving injuries but no fatalities and those just
involving damage needed an NER unbiased expected value estimates of L in this top- down
framework as soon as possible. Initial first- cut estimates would be a high priority for her
group, to be refined on an ongoing basis.
Estimating the CPS for a suitable set of incident size scenarios
The next step in their discussion was estimating the expected values of the conditional prob-
abilities of a suitable set of incident size scenarios and then using these conditional prob-
abilities to compute the unconditional probabilities.
A parametrically defined probability distribution approach was one option, and given the
power and simplicity of the Poisson probability distribution just discussed, using a compa-
rable parametrically defined probability distribution approach to estimating a full set of inci-
dent size scenarios was a very common and very tempting option. For example, a negative
exponential distribution was a popular choice in some circumstances, like estimating the size
of a seismic event (earthquake) given a seismic event occurs or estimating the depth of an ice
scour in the ocean floor given a scour occurs, or estimating the duration of a telephone call,
given one begins. The first two examples I have relevant direct consultancy experience with.
The third involves an early application area for ‘queueing theory’ a century ago which most
people familiar with traditional basic OR will have been exposed to.
There were two good reasons behind choosing to use a parametrically defined probability
distribution like the negative exponential in the NER context, and Sophie believed that
both were worth understanding by the board and everyone else involved in interpreting the
analysis required.
First, a parametric approach allowed the estimation of the defining parameters using lim-
ited data. For example, the single parameter defining a negative exponential distribution
could be estimated in terms of the average size of the incidents observed, which meant that
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 501
even if only limited observations of small to medium sized seismic events over a 100 year
history were available, assuming a negative exponential distribution provided a feasible
approach to estimating the 10,000 year return period seismic event incident which might
define the design approach to a nuclear power station. The ‘price’ for this ‘power’ was an
extreme extrapolation. But the need to extrapolate was not going to go away, and ‘tests
for robustness’ associated with the extreme extrapolations could be pursued by looking at
history prior to seismic event data in addition to the usual tests for ‘goodness of fit’ using
interpolation data. A key issue was any available evidence that there was a maximum plausi- ble event size or the complete lack of such evidence.
Second, if some data were missing and systematic bias was involved, analysts could revise
the estimation process to correct for systematic bias associated with any patterns in the miss-
ing data. For example, Chapman (1988) demonstrated how this kind of missing data adjust-
ment process used in conjunction with a Poisson arrival process and a negative exponential
depth distribution was employed to estimate the depth of ice scouring on the floor of the
Beaufort Sea when the gradual silting- in of scours over time beginning immediately after a
scour occurred meant that observations of very small scours were always missing, and all the
scours observed using sonar had been partially filled in to various degrees. This approach
allowed Gulf to assess oil pipeline bury depths when ice scours down to depths of the order
of 10 metres with return periods in excess of 100 years were the concern and limited sonar
scour depth data over about 5 years was all that was available.
Despite the advantages in terms of ‘power’ provided by using a member of the exponential
family of distributions or a related distribution like the log- Normal, in the NER context the
‘cost of that power’ led her to believe they should not take this route. NER needed generality
and a related ability to test a complex set of assumptions which meant that it was much easier for all relevant parties to follow what was involved using about five ‘size scenarios’ which were not defined using a specific distribution function. Alternative specific parametrically defined
probability distribution functions could be tested within the framework this approach pro-
vided, and one or more specific parametrically defined examples could be used as a starting
point, but they did not need to start with a commitment to a specific parametrically defined
probability distribution function. She believed the additional generality allowed by this flexible
scenario based approach was of crucial importance to NER. In broad terms the board should
understand why in terms of the implications of testing and discussing assumptions about:
1 extreme distribution tails (maximum plausible values in particular),
2 security assessment information,
3 any missing data when other relevant information was available,
4 the two-part transformation they needed to consider in order to relate fatalities to
expenditure on reducing fatalities, and
5 the four- part decomposition of ES they needed to consider other correlated issues like
injuries.
She did not think it would help the board to delve into the technical details, but she did
think that an outline of some of the key working assumptions involved in her first cut at five
suitable scenarios for her presentation that day might help them to understand the complex
issues which NER needed to address, using some variant of these five size scenarios after
they had been further developed.
Her starting position was the need for two ‘benchmark scenarios’ which facilitated talking
about the transformations over the whole range and the extreme distribution tails, especially
the high S value tail.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
502 Employing planning tools in practice
Her choice of ‘lower benchmark’ was simple. What she would call the ‘S = 1 scenario’ was
unequivocally one fatality. This was simply the minimum feasible number of fatalities given
fatalities occur. For transformation benchmarking purposes, it was comparable to a road
traffic accident involving only one fatality in some respects, although it was different in other
respects, for reasons she would touch on shortly.
Her choice of ‘upper benchmark’ was more complex. She had briefly experimented with
several approximations to logarithmic fatality scales, to cope effectively and efficiently with
the highly asymmetric probability distributions with very small probabilities of very large
numbers of fatalities they had to address. She had decided that successive intervals or class
marks increasing by a factor of about five for each successive increase led to an upper bench-
mark scenario which was appropriate given the physical upper limit on fatalities defined by
the number of people on the trains colliding or otherwise directly involved who might be
killed plus data available on relevant past incidents which went beyond NER experience.
A quick web search had revealed a French railway derailment accident 100 years ago
involving 800 to 1,000 fatalities (there was uncertainty about the number). A physical upper
limit for all fatalities resulting from a very high speed head- on collision involving two of
NER’s current trains might be in the range 1,000 to 2,000. But these were very extreme
possibilities, with very low probabilities, and outcomes much closer to 200 fatalities were
much more probable. She had concluded that a reasonable first- cut approach for illustra-
tive purposes that day would use an ‘S = 350 scenario’ as the ‘catastrophe scenario’ which
represented the ‘maximum plausible number of fatalities incident’, formally interpreted as
‘a conditional expectation of 350 fatalities associated with a scenario when more than 200
fatalities occurred in a single year, for most purposes associated with a single incident’.
Strictly speaking, the P0 = 0.0500 and (1 – P0) = 0.9500 assumption implied a very small
probability of a second incident in the same year (0.0012 as noted earlier, about 1 in a
1,000), but there were several ways of accommodating the implications in practice, and for
present purposes the board need not worry about these technicalities when thinking about
an S = 350 scenario in terms of a single incident.
Both the 350 conditional expectation and the 200 lower bound on the associated range
were illustrative values that she would adjust after the safety and security team better under-
stood the full range of potential catastrophic incidents and their probabilities defining this
‘upper benchmark’. For benchmarking purposes this ‘catastrophe’ scenario was comparable
to an airplane crash in some respects, different in others. NER had never experienced any
incidents in this category, and there were very few worldwide. It would be reasonable to
hope that for NER an S = 350 scenario would never happen, ever. But one could happen tomorrow, and to simply ignore this possibility would be negligent.
One way the board might find it useful to think about an S = 350 scenario involved assum-
ing 300, 500, 700, 900, or 1,100 fatalities associated with 0.800, 0.160, 0.032, 0.006, and
0.001 conditional probabilities, a discrete S value probability tree with five equal interval S
values representing a range of from 200 to 1,200 fatalities, with decreases for each successive
probability which were about a factor of five. This implied a conditional expected outcome
of 350. A simple continuous variable equivalent in the HAT framework involved assuming
the discrete values of 300, 500, . . . , 1,100 are class mark values for a rectangular histogram
conditional probability distribution. In continuous variable terms, 200 to 400 fatalities are
portrayed by a mid- range class mark of 300. Similarly, 400 to 600 fatalities were portrayed
by a class mark of 500 and so on (in practice, 200 to 399, 400 to 590, and so on).
The 350 needed rounding down to reflect the bias on the high side imparted by assum-
ing a rectangular histogram distribution and then rounding up again to reflect the effect of
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 503
possible values above 1,200. However, if this continuous variable distribution portrayal was
assumed, it was useful to recognise the need to deal with a monetary unit later anyway, and
the uncertainty levels associated with the TS transformations.
Of much more importance, the S = 350 scenario needed linking to a set of contributing ‘sub- scenarios’ which the safety and security team could put together when they explored
specific train layouts and the physical implications of what might happen in various kinds
of incidents. These sub- scenarios could range from worst to best case examples within the
catastrophe scenario range, with relative likelihoods, the shape of the curve as a whole, and
the understanding the safety and security team had already in different frameworks. This
would lead to adjusting the conditional distribution shape as well as the probability a colli-
sion involves 200 or more fatalities and the current nominal 200 and 350 values.
Her intermediate interpolation scenarios, S = 5, 20, and 70, used incremental increases
from 5 to 20 and from 20 to 70 which were not a consistent exact factor of five change.
This was to keep interpretation simple and credible, for reasons she would explain shortly.
The ‘S = 70 scenario’ was a ‘disaster’ scenario, nominally a conditional expected value of
70 in the 40– 199 fatalities range, embracing their recent 40 fatalities derailment example.
Although NER had never experienced as many as 70 fatalities in a single collision, they had
experienced an S = 70 scenario in her ‘disaster’ scenario sense in a derailment context.
The ‘S = 20 scenario’ was a ‘very serious incident’ scenario, nominally a conditional
expected value of 20 in the 10– 39 fatalities range, embracing their recent 30 fatalities
collision.
The ‘S = 5 scenario’ was a ‘serious incident’ scenario, nominally a conditional expected
value of 5 in the 2– 9 fatalities range, removing the ‘very’ from the S = 20 scenario, but
avoiding any language which played down the situation.
For illustrative purposes, to put the definitions of these five scenarios together and provide
some example PS values for S = 0, . . . , 350 in a deliberately simple tabular framework for
their ‘collision with fatalities’ example, they would use Table 9.1.
Table 9.1 was a plausible approximation to the assumptions needed by NER as a starting
point for safety and security decisions centred on collisions with fatalities. In addition to the
P0 and some of the CPS being contentious for reasons she did not want to discuss until they
had more developed unbiased expected value estimates, this table as a whole would need
revision and refinement by the safety and security team before it was used by the board to
make real decisions.
Table 9.1 Example five scenario portrayal of ‘collisions with fatalities’ incidents for L = 0.05, P0 = 0.9500 so (1 – P0) = 0.0500, and the assumed illustrative values for CPS shown.
Size scenarios
CPS for S greater than 0 and PS for scenarios S = 0 to 350
Contributions to F, the annual expected number of fatalities
S (nominal range) CPS (1 – P0) × CPS PS S × PS (as a %)
0 0.9500 0 × 0.9500 = 0.0000 (0%) 1 0.05 0.0500 × 0.05 = 0.0025 1 × 0.0025 = 0.0025 (0%) 5 (2– 9) 0.50 0.0500 × 0.50 = 0.0250 5 × 0.0250 = 0.1250 (15%) 20 (10– 39) 0.40 0.0500 × 0.40 = 0.0200 20 × 0.0200 = 0.4000 (47%) 70 (40– 199) 0.04 0.0500 × 0.04 = 0.0020 70 × 0.0020 = 0.1400 (17%) 350 (200– 1,200+) 0.01 0.0500 × 0.01 = 0.0005 350 × 0.0005 = 0.1750 (21%)
F = 0.8425 (100%)
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
504 Employing planning tools in practice
She needed to explain Table 9.1 to board members in some detail to ensure that they were
all comfortable with what it said, assuming that the numbers were reasonable for the illustra- tive purposes intended at this initial meeting. All board members needed to be comfortable with what Table 9.1 implied in this sense. She would build an example analysis on the impli-
cations of Table 9.1 to explore the implications of her approach as a whole, so understand-
ing its foundations clearly mattered.
On a column by column basis, starting with the left- hand column, the ‘size of incident’ was
portrayed by S = 0 plus the five ‘size scenarios’ S = 1, 5, 20, 70, and 350. As already explained,
S = 1 was the ‘lower benchmark’ scenario, S = 350 was the ‘upper benchmark’ scenario inter-
preted as a ‘catastrophe scenario’, and S = 5, 20, and 70 were the intermediate scenarios.
The ‘nominal range’ for each S in brackets was a convenient working assumption to assist
interpretation for scenarios other than S = 0 and 1. The crucial choice was the nominal range
for S = 350 of 200 to 1,200+ as already discussed. She had assumed that more than 200
fatalities involved a conditional probability of any particular S value in the region of 1,000
fatalities approaching zero fast enough to pull the expected value down to about 350 what-
ever the actual maximum number of fatalities involved might be. Her safety and security
team would have to explore the robustness of this assumption, and perhaps adjust the 350
currently defining the conditional expectation or the 200 currently defining the bottom end
of the range or the illustrative 1,200+ upper end. It was important that the S = 350 scenario,
revised as appropriate, provided an unbiased estimate of a catastrophe scenario defined by a
nominal range like ‘200 up to the physical maximum defined by the system accommodating
the possibility of more than one incident in any given year’. The exact shape of the curve
associated with this scenario range was important. How it linked to the rest of the curve for
all scenarios was also important. But a coherent and robust understanding of the ‘catastro-
phe scenario incident’ based on the knowledge of the physical systems involved and what
happened during an incident was essential, a deeper kind of understanding which she and
her team would have to ensure was robust. They would have to be able to convince the
board they had done this effectively for the board to do their job effectively, which is why
the board needed to start to understand what needed to be done that day.
Five size scenarios were judged to be a reasonable compromise between the unhelpful
simplicity of a smaller number and the unenlightening complexity of a larger number, but
four or six might prove to be preferable alternatives.
The example scenario S values she had used increased on a scale which was approximately
logarithmic – an exact factor of 5 increases from the lower benchmark value (from 1 to 5)
and to the upper benchmark value (from 70 to 350), slightly compressed in the middle
ranges (from 5 to 20 to 70). The size scenario ranges and spans also increase by approxi-
mately a factor of 5 each time. Approximate factor of five increases each time yielded the
overall range needed, given five size scenarios. A logarithmic approach provided a simple
way to capture the implications of the distribution’s extreme asymmetry (skew) which was
often used. In this case, Sophie thought from 1 to 5 was a useful start, but 5 to 25 followed
by 125 and then 625 made both the ‘catastrophe’ scenario of 625 and the ‘disaster’ scenario
of 125 too extreme, with a smaller probability for the 625 scenario than she wanted for her
example purposes.
Generalisation within this framework to S = 1, . . . , m could be based on fitting a con-
venient non- linear curve to the five scenario S values and an assumed zero probability upper
limit, then interpolating for all integer values of S.
The CPS represented conditional probabilities associated with the five S = 1, . . . , 350
size scenarios. She was assuming they would estimate these five values directly. They would,
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 505
of course, have to ensure that their CPS estimates sum to unity, as for her example values.
The Table 9.1 CPS values were plausible, but estimated for illustrative purposes only, to two
places of decimal.
For S = 0 the P0 = 0.9500 in the second of the two PS value columns was based on the Pois-
son distribution L = 0.05 assumption as discussed earlier. But for S = 1 to 350 the assumed
illustrative CPS values and the PS = (1 – P0) × CPS relationship was used to calculate the PS. Examining past NER collisions with fatalities, and then other railway system collisions
with fatalities, was obviously one starting place to build a story, if not a database in a direct
sense, to underlie Table 9.1. However, the ‘complete story’ associated with incidents involv-
ing fatalities had to look at all NER near misses, plus all the ways incidents might happen
which were suggested by a range of circumstances, from changes in operating practices to
changes in maintenance contract structures. Furthermore, some of the ‘missing data’ were
related to malicious incidents which did not involve railways.
For example, the 9/11 terrorist attacks in the US, and the 7/7 London bombings in the
UK, were two examples with important implications for NER. A heavy vehicle left on a level
crossing when two high speed trains were approaching from opposite directions was one
possibility, but there were many others. Such attacks would specifically target S = 350 and
70 scenarios, also increasing the chances of an S = 20 scenario. Depending on assumptions
about terrorist capability and intensions, the CPS increases would favour the higher S- value
scenarios to varying degrees.
All these PS values were contentious, but she was convinced that they were appropriate
for illustrative purposes, and NER would need to demonstrate why it was reasonable to use more favourable PS values if that was what the NER board believed. NER could not afford
to use any approach which implicitly assumed an optimistically biased variant of the implica-
tions of the S = 350, 70, 20, 5 and 1 scenarios portrayed by Table 9.1. She indicated she
had actually estimated the 0.01 probability associated with an S = 350 scenario first, having
adjusted her catastrophic scenario to get an answer of about 0.01 for illustrative conveni-
ence, and in practice a comparable approach based on suitable data could be useful.
The final columns showed contributions to F, with F defined as ‘the annual expected
number of fatalities’, equal to the sum of the S × PS. The sum yielded F = 0.8425, with the
relative size of each contribution expressed as a percentage in brackets. F = 0.8425 rounded
up to one implies that if these probabilities were of the right order of magnitude, NER
should expect about one fatality per annum from collisions ‘on average’.
The S × CPS might have been shown, and then summed to compute the expected number
of fatalities if a collision occurred, 16.85. In rounded terms this meant NER could expect
about 17 fatalities on average when incidents involving fatalities occurred, about 20 fatalities
rounded in a manner consistent with both one fatality per year on average and one incident
every 20 years on average.
She then emphatically emphasised a point of crucial importance, central to understanding
all safety and security analysis approaches and their implications – the ‘average’ was not even remotely representative of what was actually going to happen most of the time, and all the ‘aver- ages’ involved were only meaningful concepts over a VERY long period.
For all practical purposes an ‘average year’ involving one fatality caused by a collision was
exceedingly unlikely – the return period for this scenario was about 400 years. Furthermore,
an ‘average incident’ was also very unlikely – assuming an ‘average collision incident’ meant
17 fatalities or 20 if rounding was used. Even if 20 fatalities were interpreted in the S = 20
scenario sense, meaning 10 to 39 fatalities, this view of ‘an average collision incident’ had
only a 0.0200 probability, implying a 50 year return period, one every 50 years on average.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
506 Employing planning tools in practice
What was never going to happen was a regular pattern of collision incidents every 20 years,
including one S = 350 every 2,000 years, and one S = 1 every 400 years. No collision inci-
dent was the most likely outcome, with an assumed 0.9500 probability per annum.
Furthermore, when all incidents involving fatalities were considered later in her analysis,
they would have to formally recognise that most years involve no collisions and no derail-
ments but very large numbers of suicides (people taking their own lives using trains) and
small numbers of people killed accidentally in level- crossing incidents.
Random processes were notoriously difficult to visualise, even for people with extensive
relevant theoretical and practical background. Highly asymmetric distributions with one
very long tail made the implications of random processes even more difficult to visualise.
The portrayal of an asymmetric distribution with one long tail provided by Table 9.1 helped
to clarify some of the key issues driven by these very long tails in ways which were worth a
brief pause for reflection before building on the basis it provided.
The S = 350 scenario had only 1% of the conditional probability, a CPS = 0.01. But 21%
of the value of F was contributed by this scenario, as indicated by the percentage contribu-
tions shown in brackets in the last column. This was because the leverage applied by the long
distribution tail was crucial. The expected value was the point of balance of the probability
distribution in density form – technically defined as the first moment about the origin. It was
very important to understand that very small probabilities associated with very big numbers
of fatalities mattered greatly in expected value terms. This was in addition to obviously mat-
tering a very great deal if they occurred. In expected value terms, they mattered much more
than the very small probabilities might suggest – a counter- intuitive implication for most
people, including some experts.
The S = 70 scenario had 4% of the conditional probability, a CPS four times more than the
0.01 of the S = 350 scenario. However, 17% of F was contributed by the S = 70 scenario,
marginally less than the 21% contribution of the S = 350 scenario. This was because of the
way the logarithmic S value scale highlights the inherent implications of very long distribu-
tion tails.
The S = 20 scenario had 40% of the conditional probability, with a CPS ten times more
than the S = 70 scenario, 40 times more than the S = 350 scenario. But it contributed 47%
of F, only about three times as much as the S = 70 scenario, about twice as much as the
S = 350 scenario.
The S = 5 scenario had 50% of the conditional probability, but it contributed only 15% of
F, yet again because of the way the logarithmic scale highlights the importance of the long
tailed distribution effect.
The S = 1 scenario contribution to F was 0%, working to the nearest percentage, despite
the fact that it contributes 5% of the conditional probability.
The board needed to remember that NER had never had a ‘collision with fatalities’ inci-
dent greater than 30 before, their 40 fatalities ‘derailment’ was just into the S = 70 scenario
range of 40 to 199, and NER had never had an S = 350 size of incident of any kind before.
However, incidents involving more than 200 fatalities were clearly feasible, and if they hap-
pened the board would be directly accountable.
Table 9.1 indicates incidents involving 40 or more fatalities make a 17% + 21% = 38% con-
tribution to the expected number of fatalities per annum, despite being associated with only
5% of the conditional probability, so the probability of 40 or more fatalities mattered greatly
in expected outcome terms. Although the individual illustrative parameter estimates were
contentious, Table 9.1 was a plausible overall portrayal, and in broad terms her inferences
based on this table’s values were robust for that day’s discussion purposes.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 507
If NER used a biased approach to estimating PS and F values because of a lack of due atten- tion to what was possible but had not yet occurred, or there was a cultural tendency to ignore very low probability possibilities, NER was missing what really mattered. If this was the case, arguably serious negligence would be involved because NER experts should know what really mattered and the NER board members should know whether or not they could trust their NER experts to know what really mattered.
The board needed to remember that train carriages were much safer than they used to be,
and many significant improvements in other safety features for the system as a whole had
been achieved. NER had done a lot of very positive things to make rail travel safer. However,
trains were faster than they used to be, with still faster very high speed trains on the agenda.
The board also needed to reflect that recent terrorist concerns around the world, coupled
to a near- miss interpretation of the two major accidents in the past two years (both could
have been very much worse), suggested diligently maintaining a very healthy level of ongoing concern for ‘long tail incidents’ – learning to be afraid of potential extreme events.
Table 9.1 suggested they were unlucky to have had an S = 20 collision with fatalities inci-
dent (with 30 fatalities) in the last two years plus an S = 70 (with 40 fatalities) derailment,
but they were also lucky that neither was much worse. A further very serious incident in the near future would be widely interpreted as negligence if they could not convincingly dem-
onstrate why this was not the case, and at present they could not do so.
However, while this kind of inference was useful to familiarise the board with some basic
implications of long tailed distributions associated with low probability but high impact
incidents, the real purpose of Table 9.1 was starting to build a numerical model from the
bottom in a series of layers, which the board could use to understand the top- down perspec-
tive the board needed to make strategic decisions. As she had explained earlier, they had a
series of steps to take, and a broad top- down understanding of Table 9.1 was the first of
these steps.
There was an obvious need for the safety and security team to look at the conditional
probability values and their links back to L in terms of all relevant underlying models and
data. It would be crucial for her team and the board to preserve the clarity of the kind of
top- down view illustrated by Table 9.1, maintaining a focus on what seemed to matter most.
The evidence basis for the base- case model probabilities they would use in practice needed
to be as sound as possible, using all the lessons learned by an international community of
safety experts over many years, plus the accessible databases of all other relevant railways as
well as in- house NER data and expertise. However, synthesis of this evidence basis in an
NER context also had to rely on inherently subjective expert judgement, guided by a range
of models with inherent approximations built into their assumptions. One key issue was
‘To what extent, and for what reasons, would near future and longer term future possible
incident probabilities be different from those in the past?’ Another was ‘How should all relevant near- miss information and interpretations, including new security alerts, be linked
to a coherent overview which made full use of all available information in a robust manner?’
Sophie suggested that the Poisson model perspective was particularly useful when doing
this from a top- down perspective, and it would help them to integrate these concerns with
an approach to size scenarios which allowed them to avoid specific distribution assumptions
and make use of physical characteristics, like ‘the maximum number of passengers on the
trains currently in use or proposed’.
The interpretation of Table 9.1 which she had just explored in considerable detail empha-
sised that NER were dealing with a very simple basic model associated with a very complex
underlying reality, and it was important to understand in broad terms what the underlying
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
508 Employing planning tools in practice
complexities involved and implied. Table 9.1 could be given an even simpler format, using
a probability tree presentation without the range information, if this was useful, but the
underlying complexities always need to be understood and remembered.
While keeping the modelled portrayals simple for board level presentations, a rich inter-
pretation of significant underlying analysis would be their aim in practice. If useful as part
of this, a different scale or more size scenarios or more sophisticated underlying curve
assumptions could be used for additional clarity. NER could design an efficient approach
to increasing clarity with no bounds imposed by specific distribution assumptions or any
other assumptions that could not be tested for robustness, a defining characteristic of the
EP approach she was using.
A two part transformation from fatalities to fatality related cost, TS = V1 × ES
Sophie emphasised that most organisations involved in operations which could lead to fatali-
ties now used some variant of a ‘value of an avoided fatality’ concept to transform fatalities into
a cost equivalent. The avoided fatality concept was by no means new – she was aware of one
publication dating back to 1930, The Money Value of a Man (Dublin and Lotka, 1930). How- ever, the exact nature of the concept was still evolving, ambiguous, and highly contentious.
There were a number of very different ways people currently viewed the nature of ‘value
of avoided fatality’ concepts. They might be viewed on a spectrum, ranging from the 1970s
bad practice Ford Pinto misinterpretation of the NHTSA cost– benefit analysis approach
which she had just explained to a shadow price and revealed preference interpretation using
the kind of TS based approach she had used to interpret both the 1970s’ NHTSA approach
and Ford’s inappropriate use of the T = V = $200,000 NHTSA value. NER needed to
adopt an S dependent variant of this TS approach. It was important for NER to adopt the
most general (unrestricted) perspective available, make use of a shadow cost and revealed
preference interpretation with acknowledged ethical content, and chose numerical values
appropriate to the NER context which could be linked to other current practice in terms of
a grounded argument which all relevant parties could trust.
She observed that avoiding stealth assumptions was particularly difficult in a context
involving trust between multiple parties and the possible loss of lives. But this made it
especially important. A failure to address an effective shared understanding of all important
assumptions which clarified important differences in opinion, as well as what did not matter,
was arguably negligent in particularly serious terms.
As she had explained earlier, the simplest basic operational framework that NER could
adopt to meet the crucial generality and clarity concerns involved a two part value transfor-
mation of the form TS = V1 × ES.
V1 was a benchmark value of an avoided fatality at scenario level S = 1 relevant to NER, a suitable starting point or datum. ES was an everything else factor. TS = V1 × ES was the shadow cost and revealed preference transformation concept for all S = 1, . . . , m values,
assuming that ‘everything else’ (over and above the V1 value) which was relevant in terms of
‘costs related to fatalities’ was fully addressed by the ES. By ‘costs related to fatalities’ what
she meant in terms of the basic approach used earlier was ‘all relevant costs directly related to fatalities’. She now wanted to start to explore a generalisation which interpreted ‘costs
related to fatalities’ as ‘all relevant costs correlated with the number of fatalities occurring’. The five size scenarios were a working assumption approximation which would let them
focus on S = 1 and S = 350 scenarios initially and then interpolate for all relevant S values.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 509
When using either the basic or the generalised form of this conceptual framework in the
NER context, a T = V1 × E special case was a gross oversimplification to a degree which was
not acceptable in her view. But the board should know that the NER approach being used at present involved basic form approach with a $4 million value of an avoided fatality for all S values, so the NER current policy, implicitly if not explicitly approved by the board, was
equivalent to approving her V1 = $2 million suggestion plus ES = 2 for all S, a TS = $4 mil-
lion for all S.
This current T = $4 million value could be interpreted as ‘an NER board approved
revealed preference statement about the value of NER avoided fatalities for S = 1, . . . , m’.
T = $4 million was currently being used for all NER safety and security decisions with board approval. She appreciated that board members had probably not thought about it in these terms before. She believed they needed to start thinking about it this way immediately. Doing so was part of the process of beginning to understand why NER needed to remember to be afraid. It was also part of the process of beginning to understand what needed to be done about it and how they could develop a position which was defensible even if further serious incidents occurred.
The five scenarios approach would allow NER to treat fatalities associated with the
S = 350 scenario as more important per fatality than fatalities associated with the S = 1 scenario fatalities, with three intermediate levels of relative importance. This analysis could
then be used as a basis for a non- linear curve generalisation.
The controversial nature of the appropriate TS shadow prices and their revealed prefer-
ence basis would not go away, but being able to accommodate increasing ES values as S
increased would allow NER to alter the trade- offs deemed appropriate between expenditure
on making the system safer and tolerating fatalities as the value of S increased. Crucially, it would allow NER to explore and test all related working and framing assumptions to better understand what they were doing, facilitating trust which would be seen as well founded by all relevant parties, even if very serious incidents took place in the near future.
ES had to reflect NER board views about an appropriate uplift from V1 for S = 350 as well
as S = 1, addressing the common view that large numbers of people killed in one incident
was relatively more serious than the same number killed one at a time in separate incidents,
as noted earlier. But while S dependence was important for NER, overall ethical and related
trust concerns, including related tests for overall robustness, were much more important. ES values used to make NER safety and security decisions were a board level decision that
really mattered. Directly confronting the full implications of ES choices would help everyone to understand both what really mattered and why it really mattered.
Using ES = AS + IS + DS + RS decomposition to make analysis simpler and more effective
Given the need to address ES dependence on S to deal with fatalities, they could usefully
employ this S dependent framework to make the overall analysis simpler in several key
respects, as well as a more effective – achieving a higher level of clarity efficiency. They could
do this by decomposing her generalised interpretation ES involving cost related to fatalities
which included the implications of injuries plus other relevant concerns usually correlated
with fatalities into four component factors, each associated with three plausible options in
terms of approach. They could use different options for different purposes at different stages
in the analysis process. At first sight this might look even more complicated to a disturb-
ing extent, but each aspect of the complexity involved was constructive – designed to pay
significant dividends in terms of keeping it simple systematically at an overall analysis level.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
510 Employing planning tools in practice
As an initial overview, it involved decomposing ES into four components in order to embed
within ES:
1 the treatment of fatalities, plus 2 injuries which were correlated with fatalities, plus 3 physical damage to the system and directly related knock- on implications which were
correlated with fatalities, plus 4 any other issues which were correlated with the number of fatalities and not fully cap-
tured by the first three separate components, for ‘closure with completeness’.
While this might initially look unnecessarily complicated, in practice it actually involved a
form of local complication which greatly simplified overall management decision making in an enlightened manner, an important feature of the enlightened simplicity aspects of the EP approach she advocated.
Fully exploiting this multiple part decomposition of ES as a potential approach option for
all organisations would now be part of a generalisation and broadening of her interpreta-
tion of what all organisations could and should mean by ES and the complete range of roles
which ES could perform.
A central issue for the NER board that day was why this particular four part decomposi-
tion was not really an option for NER and why the board had to mandate the use of a four
component S dependent ES concept. The board needed to understand what had to be done,
which required an overview understanding of how to do it and why this complexity was
necessary.
The first component was the ‘avoided fatalities factor’ AS – a factor associated with adjust-
ing V1 for all the revealed preference implications of avoided fatalities on the NER railway
system as distinct from road traffic accident avoided fatalities. AS was focused on all NER
revealed preference aspects of fatalities which were not captured by V1.
‘Option A’ for this first component was further focused on the expected cost of legal
settlements imposed by the courts, out- of- court settlements, legal fees and fines for the S
level involved – the actual NER expected cash flow costs directly related to fatalities. Option
A would not attempt to assess common practice shadow cost or ethical and trust related
concerns which might go beyond these cash flow costs. This was her ‘minimum plausible’
option for the first component, which she believed was useful for some sensitivity analysis
purposes at a later stage in the analysis but not initially. They had to formally acknowledge
that AS was an uncertain parameter even for a given S value, and the associated option
A value was a useful plausible minimum for some purposes.
‘Option B’ would increase this plausible minimum to incorporate a fair and reasonable
interpretation of common practice shadow cost values of avoided fatality concerns for
the S level involved without going beyond a reasonable interpretation of ‘good practice’.
This was her ‘mid- range’ option, which she suggested they start with for initial analysis
purposes. She believed that a good practice option was a convenient starting position
working assumption and it was also a useful part of the basis for later discussion of best
practice choices.
‘Option C’ would further increase this mid- range value by adding a maximum plausible
ethics and trust provision to AS. This was her ‘maximum plausible’ upper bound option,
relevant to discussions of NER best practice AS choices.
She believed the board should eventually agree an AS estimate value in the option B to
option C range, using carefully considered and delivered advice and guidance from the NER
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 511
safety and security team, but without firm, specific recommendations from her team, because the choice involved was a board level judgement with crucial ethical dimensions.
As a summary for AS, her option A was a ‘plausible minimum’, option B added a good
practice uplift, option C added a further NER best practice driven ‘maximum plausible’
uplift, and she was suggesting starting initial analysis with the option B good practice uplift
with a view to the board eventually making a choice in the range between options B and C.
The second component was the ‘injuries factor’ IS – a factor associated with further
adjusting V1 in order to make a provision for injuries on a per fatality basis. At present,
NER were assessing the number of people injured to various levels of severity in each rel-
evant potential incident separately from their fatality estimates, making use of a number of
categories of injury seriousness. They then used a proportion of the T = $4 million value
of an avoided fatality for each of these categories of injury seriousness, plus the numbers of
people injured in incidents. This was a variant of the NHTSA and Ford approach, which was
still widespread common practice. She suggested NER could greatly simplify this approach
in a manner consistent with common practice norms and make it much more effective and
efficient at the same time. They could do so by looking at the expected number of injuries
in all relevant injury categories on a per fatality basis as S changed and focus NER attention
on the overall importance of these injuries relative to the associated number of fatalities because it was the joint importance of these correlated outcomes which really mattered.
They would obviously have to test the robustness of this correlation working assumption,
but she believed it would prove viable, given their focus on what really mattered.
She suggested they start with a mid- range option B working assumption for this IS com-
ponent. This was linked to the AS component option B choice, bearing in mind that esti-
mated option A and C alternatives for IS linked to the AS component could also be used for
the IS component. That is, for initial use they should estimate the expected value of avoiding
all the injuries which occurred in conjunction with S fatalities as an expected cost function of
S, using an approach consistent with the mid- range option B choice for AS and comparable
common-practice norms for different injury categories. Linked option A and C estimates
would also be available.
The third component was the ‘damage factor’ DS – a factor associated with further adjust-
ing V1 in order to make a provision for physical damage to the system plus all the associated
knock- on operating cost and lost revenue effects assumed to be correlated with fatalities.
Instead of estimating these direct and opportunity costs separately, NER should treat them
as an expected value function of S within the ES approach. In terms of the overall analysis
and its interpretation, this would be much simpler, a further useful improvement in clar-
ity efficiency. It was based on a strong and arguably crude assumption, which would need
testing for robustness and might involve significant variability about the expected outcomes
assumed, but she had no doubt that it would prove a viable working assumption given
their focus on what really mattered. She suggested NER begin with a mid- range option B
approach as for the first two components, but this time NER should recognise that the mini-
mum plausible option A approach would be a plausible minimum cost estimate of actual
direct costs and lost revenue, the maximum plausible option C was a plausible maximum
cost estimate of actual direct costs and lost revenue, and the NER mid- range estimate ought
to be an unbiased estimate of the expected outcome.
The fourth component was the ‘residual factor’ RS – a factor associated with a ‘residual
provision’ for any concerns not fully captured by the other three component factors of ES.
At its simplest this ‘residual provision’ would include a very simple rounding up to acknowl-
edge the importance of all relevant unresolved ambiguity. But more specifically, it would
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
512 Employing planning tools in practice
also include a crucial ‘prudence provision’, to be discussed later in the context of enlight-
ened prudence. In addition, RS as a whole should round up the TS relationship appropriately
to validate an NER board view that all relevant concerns had been addressed appropri- ately and grounded trust which was well founded was assured. This would ensure ‘closure
with completeness’. RS should explicitly address a residual to deal with any relevant issues not fully covered by the first three components, including all relevant trust contribution concerns not fully built into good practice approaches already embedded in the first three
components. In the RS case, the issues involved in the options A, B, and C approach would
have to reflect plausible minimum, maximum, and mid- range views about trust contribution
concerns given the option choices for the first three components. She suggested a mid- range
option B working assumption as a starting position, consistent with three earlier option B
choices, although the nature of the range involved was clearly much more complex, needing
elaboration best deferred until later.
At this stage she wanted board members to understand in broad overview terms how the
four components of ES played their collective roles, not the details, but as summary they
might find useful later, she was suggesting NER adopt her four option B assumptions and use
E = A + I + D + RS S S S S,
where:
1 AS was the ‘avoided fatalities factor’ associated with adjusting V1 for all the revealed
preference implications of avoided fatalities on the NER railway system as distinct from
the road traffic accident avoided fatalities addressed by the NHA. Assuming option B
would mean including a provision for all the S dependent implications of court judge-
ments, out- of- court settlements, legal costs, and fines, and it would also reflect good
practice approaches based on a shadow cost interpretation linked to earlier cost– benefit
analysis practices, but it would not go beyond current good practice norms for an
approach based on shadow cost and revealed- preference interpretation.
2 IS was the ‘injuries factor’ associated with further adjusting V1 in order to make a provi-
sion for avoided injuries on a per fatality basis. Injuries were assumed to be correlated
with fatalities. IS had to address revealed preference concerns comparable to those asso-
ciated with fatalities. But a range of degree of injury categories needed to be addressed,
and any other relevant differences between fatalities and injuries also needed consid-
eration, like how did the numbers and distribution of different types of injury change
as S changed. The approach to IS in terms of options A, B, or C obviously had to be
consistent with the approach to AS.
3 DS was the ‘damage factor’ associated with further adjusting V1 in order to make a pro-
vision for physical damage to the system plus all associated knock- on direct cost and an
opportunity cost portrayal of lost revenue effects assumed to be correlated with fatali-
ties. DS needed to reflect how all these costs on a per fatality basis changed as a function
of S, using an expected value estimate of actual cash flow costs and the opportunity cost
equivalent of lost revenues.
4 RS was the ‘residual factor’ associated with a ‘residual provision’ for any concerns not
fully captured by the other three ES component factors which rounded up the TS rela-
tionship appropriately to capture an NER board view of all the issues relevant to well- founded trust, including enlightened prudence, fully reflecting a board level test that all
relevant concerns had been addressed appropriately.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 513
Trust in NER board level decisions involving safety and security should be seen as directly
dependent upon the RS values given the other three ES factor values and the V1 value. But trust in board level decisions was also conditional on the rest of the NER approach to safety and security, including implementation as well as planning.
The choices of numerical values for AS, IS, DS, and RS which were preferred by each board
member might be very different. All board members needed a reasonable degree of collec-
tive agreement about plausible working assumptions for external discussion purposes. But
what really mattered was collective agreement about the board’s preferred overall TS values
because it was the overall TS values which defined the revealed preferences of the board
implied by decisions taken.
All board members needed TS values they could live with. TS values were a board approved
revealed preference statement by NER of how much NER was prepared to spend to reduce
by one the number of fatalities associated with each possible S value associated directly with
AS, together with all the correlated issues addressed via IS, DS and RS values. TS values were
shadow costs given ethical content based on a revealed- preference interpretation. RS was a
useful focal point for board level agreement about TS, given assumed values of V1, AS, IS, and
DS subjected to earlier scrutiny. But different board members could have different views about RS values and other TS components, so long as they agreed to TS values, and the broad nature of the case for the underlying four component values.
The additive ES components relationship was ‘nominal’ for two reasons, both worth
understanding. One reason was that a multiplicative alternative to the additive form used
earlier might be preferred, and this choice was a matter open to debate. The major reason,
of much more importance, was the relationship defining TS in terms of any four parameter structure was a ‘nominal’ relationship because if somebody believed the TS values were
about right overall, but one of these component factors should be much bigger, while any
of the other components should be much smaller, with no net overall change in TS, it was crucial that they and everyone else understood this would make no overall difference to NER decision outcomes. The NER board’s revealed preferences were defined by the choices the board made as a collective group of people acting as a group, with no direct relationship
to the underlying preferences of individual board members. This second reason meant that
it might not be worth arguing about the details of different component values, although
discussing the rationale of the component structure and the nature of the values used was,
in general, very important, and crucial presentation issues would also need attention.
Trust depended on understanding in broad terms the practical implications of the TS val-
ues used to make choices, plus the nature and motives of alternative perspectives and values,
which need not be common, but have to be judged reasonable or not reasonable, given due
accommodations for other peoples’ different concerns and perspectives.
Crucially, the board needed to agree on the net effect of the ES given V1 without any
compulsory need for agreement about the details of the component contributions within this ES structure. Agreement about the component values was highly desirable and should be addressed in an effective manner, but it was NOT compulsory. Furthermore, and perhaps even more crucially, NER needed to make it clear to all other interested parties why it could take a comparable view.
Some substantive differences in component values might suggest differences in perspec-
tives which do need extensive clarifying discussion and reconciliation, and where these dif-
ferences of opinion really mattered they might need resolution. But when differences of
opinion did not matter, or did not matter very much, everybody needed clarity about why
this was the case. Discussion needed to take place within a commonly understood and
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
514 Employing planning tools in practice
accepted framework so that reaching an agreement could focus on what mattered, using
shared framing assumptions to clearly articulate alternative perspectives on what mattered
most. Without shared framing assumptions effective dialogue would not be feasible. A key aspect of the rationale for the use of the TS = V1 × ES structure defined in terms
of four components of ES was recognising the value of separate treatment of industry
benchmarks, other relevant good practice approaches based on shadow cost and revealed-
preference interpretations, and common practice values which were relevant over a bad or
debatable practice to best practice range, plus the underlying role of legal settlements, out-
of- court settlements, and fines.
Also key was board level judgements about each of these categories of concern in an integrated framework without the need for any more parameters or any more effort or any more agreement about components than necessary.
The structure itself was designed to make the first three of the four ES components free
of direct ethical and trust- based adjustments beyond good practice norms if a mid- range option B was used for the first two components. The option A approach to AS and IS meant
that if the good practice increases used to define option B were moved to RS, the first three
components would all be direct cash cost measures, and separate decomposition of RS com-
ponents could be a focus for sensitivity analysis in a useful manner that she would discuss
later.
She believed that the four components approach just outlined was sound in overall terms,
and its particular four option B choice form was a good place to start for initial presentation
purposes. However, a key EP tool was a rich range of alternative presentation strategies to clarify the analysis interpretation from different perspectives, briefly explored later. As an
overview now, option A (minimum plausible) approaches to the first two components could
be useful for comparative purposes, provided that the shadow price based good practice uplifts to the AS and IS, which were removed when dropping option B in favour of option
A for these two components, were reinserted in the RS in a clearly identified manner. This
would mean that all the first three components were strictly limited to cash flow measures,
allowing discussion of a decomposed RS structure to focus on concerns going beyond issues
which could be measured in cash flow terms. Having both these two alternative presentation strategy choices available was important, but starting with four option B choices was useful for several reasons. One reason was the need to make it very clear that even with four option
B choices, a positive RS was essential for enlightened prudence reasons.
In her initial four option B form, the RS component was focused on trust contribution
issues which the board believed were important and go beyond what is addressed by the
first three components given option B choices. This approach reflected the recognition that
good practices involved in defining V1 and the AS and IS components of ES would involve
some trust- driven provisions. It was important to avoid missing out key trust contributions,
but it was also important to avoid double counting when assessing good practice option B
values for AS and IS. V1 might be assumed to embody implicit ethics content basis even if
NHA was not explicit about this issue.
The RS component also formally addressed a ‘rounding- up option’, having required
formal testing for completeness without double counting. As part of this it was useful to
formally recognise the very ambiguous nature of revealed preferences, the further complex-
ity introduced by a ‘prudence provision’, plus the futility of excessively detailed calculations based on less important issues when the really big issues that really mattered were not amenable to precise numerical evaluation. If estimating RS beyond any given level of precision was not feasible, spending a lot of effort and money estimating AS, IS, or DS to a significantly
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 515
higher level of precision was a serious waste of time and money – clarity inefficient. This was
wasted time and money which could be spent making the system safer. Put slightly differ-
ently, wasted effort on unproductive precision was a serious distraction when other activities
would be much more useful, and the opportunity inefficiency which was implied needed to
be avoided.
Part of the rationale for using an S dependent ES instead of a T = V1 × E approach
was to allow incidents involving large numbers of fatalities to be weighted more heavily
than incidents involving small numbers of fatalities via the AS uplift. This issue might have
been pursued via an ES = AS concept which did not address the other three component
factors. However, overall clarity efficiency was enhanced by the simpler four component ES approach, considering S dependence only once, and allowing the board and other key
interested parties to focus on agreed TS values without necessarily reaching full agreement
on component uplifts.
As an illustrative example, one director might see NER corporate reputation concerns as
the really central issue, a second director might see the ethics of fatalities as the key, a third
might see the full implications of life- changing injuries as the key – but it was the combined
effect that actually mattered and required their collective agreement. They could agree to
differ on their individual views about component parameter values, and even the nature of
an appropriate V1 and some of the underlying relationships, but if they disagreed about the
total implied by TS they were going to have to confront the implications and resolve their
differences in a way they could all live with in the face of scrutiny by other interested parties.
Furthermore, passengers, staff, regulators, the courts, and the public needed to judge the
board’s decisions at a TS level, no doubt with significant interest in exactly how they arrived
at whatever numbers they used, but recognising the nominal nature of the components.
Different terminology and alternative views could be used, and comparisons which might
become crucial in terms of disputes involving differences of opinion were complex. For
example, an internal NER document about the $4 million value cited a report to the UK
Rail Safety and Standards Board, T430: The Definition of VPF and the Impact of Societal Concerns by Oxford Risk Research and Analysis (Rail Safety and Standards Board, 2006). This report used the term value of a prevented fatality (VPF), a very limited view of ‘revealed preference’ approaches that explicitly rejected an S dependent approach to the AS compo-
nent, did not address an IS or DS component, did not explicitly address ethics and trust con-
cerns, and omitted all other aspects of an RS component. This approach was consistent with
current NER practice and most common practice, the reason it was cited. In effect, Sophie
being as highly critical of the Oxford Risk Research and Analysis approach as she was of the
current NER approach, with concerns that went beyond the constant T = $4 million value.
She suspected the Oxford Risk Research and Analysis authors and many others might disa-
gree with her approach. She believed that the enhanced and enlightened perspective that her
overall approach provided could be defended against all critics who preferred alternatives by
examining the stealth assumptions implicit in those alternatives from the general perspec-
tive which her EP approach provided. She accepted that a good case might be made for her
simpler T = V1 × E special case in many contexts which did not involve the very wide range
of fatalities which railways had to cope with, but these would be special cases, not exemplars
for a general rule which railways ought to use.
At an overview summary level, replacing the common practice S independent value of an
avoided fatality approach with the more general S dependent TS = V1 × ES approach using
four component factors for ES might not always be useful in direct operational terms with
all four component factors S dependent in all contexts beyond NER. But the associated
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
516 Employing planning tools in practice
robustness tests were important in all contexts, and in her view the very wide range of
potential fatalities involved in a single railway incident made it essential for NER. Further-
more, even if S dependence for AS was not useful in some contexts (when only a very narrow
range of fatalities was feasible for example), distinguishing TS, V1, and ES concepts together
with the four component factors for ES was important in all contexts for several reasons. Acknowledging trust and underlying ethical concerns was particularly crucial, linked to the
shadow cost and revealed preference basis for the TS transformation, and these trust con-
cerns involved complex issues like enlightened prudence which she would explore later, as
repeatedly indicated.
Within this four component ES framework, NER directors had currently approved RS = 0 for all S, AS = $4 million for all S, and an approach which dealt with all IS and DS issues separately at a higher level of decision- making cost but a lower level of clarity. These choices involved stealth assumptions needing immediate attention.
The next stage in that day’s overview of why this was the case would involve developing
plausible illustrative numerical values for all four of the TS components.
The avoided fatalities component V1 × AS for S = 1 and 350
To develop more clarity via illustrative numerical values for all four V1 × ES components of
TS prior to a detailed analysis of the issues and available data, a suitable starting point was AS
values for S = 1 and 350. Sophie had some feel for plausible ranges, and using these ranges
she would suggest purely illustrative mid- range example values to outline her recommended
approach. She was convinced that numerical examples would make it much easier for the
board to understand what was involved, but she was also very concerned that they avoid
wasting time on discussions about the validity of her particular illustrative example values for
their initial meeting discussion, or even her assumed plausible ranges of values.
She would use A1 = 1.1 and A350 = 5 so that the avoided fatalities component of TS was
V A $2m 1.1 $2.2m for S 1,
$2m 5 $10m for S 350. 1 S× × = =
= × = =
=
This implied that a V1 × AS component of TS varying over the range $2.2 million to $10 mil-
lion as S went from S = 1 to S = 350 would replace the current T = $4 million value for all S.
She suggested that A1 = 1.1 might have a plausible range like −50% to +100%, but an
uplift of 10% on the NHA value of V1 = $2 million to $2.2 million provided what she
thought was an appropriate illustrative initial stance when one NER fatality was involved. It
implied a higher NER value than the NHA value even when only one fatality was involved
because a lower value might be construed as unnecessarily controversial. But it was not
much higher because she did not want to overstate the AS values needed.
Her rationale for A350 = 5 was that no further uplift on A1 = 1.1 did not seem plausible,
and an uplift by a factor bigger than 10 seemed equally implausible. The approximate mid-
point of these two equally implausible extremes was 5, a round number which would serve
to illustrate what a number in the middle of what she saw as the plausible range did to the
analysis results.
It was not worth trying to be precise about either of these example estimates for AS or
any of the other TS components that day. She would later explain the basis of her views on
ranges and outline the kind of research and operational development of parameters they
would have to undertake.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 517
For S = 1 and for S = 350 the essential core of this avoided fatalities uplift component of
TS was an adjustment to reflect the difference between road travel on public highways and
rail travel on NER trains in measurable revealed preference terms, presented and explained
as clearly as possible, which all parties could trust.
She thought it would not be useful to attempt to embed trust or ethically driven adjust-
ments directly in AS today in terms of a possible option C approach or address option
A approaches and associated numerical differences.
The AS fatalities uplift factor component of ES had to reflect context specific concerns
about both the underlying legal liability concerns associated with option A, plus the other
revealed preference issues, measured using an option B approach in a plausibly neutral
‘objective’ manner. At this option B level, experts could argue about both values and meth-
odological issues without addressing the ethical implications directly, but they did have to
understand that ‘common’, ‘good’ and ‘best practice’ were not the same, and each meant
different things to different people.
As an example of relevant context specific concerns, she suggested most people expected
railway travel to be safer than road travel for a range of reasons. One reason was the loss
of any semblance of control over personal safety by the traveller – people travelling in their
own cars had a perception of control which they clearly lost when travelling by rail. She
thought being sued by the families of those killed was generally not an issue for highways
agencies, but NER might be sued even if only one fatality were involved, and action against
NER in the courts would be almost inevitable following an S = 350 scenario, especially if
any possibility of negligence was suspected. She did not think the private sector status of
NER was very significant at present because NER was providing the only rail travel service available in the country, they operated like a public sector railway in most respects, including
regulation, they were widely seen as a ‘national treasure’, and they had never been seen as
an exploitive profit- maximising organisation. However, private /public sector status could
become an issue, and she would return to this later. Even if only one fatality was involved
in a collision, likely to be a train driver because of their position at the front of the train, in
her view both staff and passengers would expect a private sector railway company to try to
be safer than their country’s road system and to be as safe or safer than other comparable
private and public sector railways.
She thought that when S = 1 was involved, an AS avoided fatality factor component of E1
as low as 1.1 was appropriate because the highways agency T = $2 million which she was
proposing as V1 presumably assumed two or more fatalities in a single road traffic accident
was common, while the V1 × A1 = $2.2 million explicitly applied to only one fatality in a
railway incident, linked to explicit provision for S dependent increases. This line of thinking
could lead to plausible suggestions of a factor like 0.9 or less, involving an A1 value lower
than the benchmark, but she thought this would not be helpful in terms of approval by all
relevant parties. There was a case for higher values than 1.1, but she believed they should
link the current NER- assumed value of T = $4 million to ‘an average S level railway incident
scenario’ involving significantly more than one fatality for comparability purposes, part of
her rational for settling on A1 = 1.1 for illustrative purposes. Although she did not want
to debate specific values today, it was worth observing that when she had the rest of her
numbers and underlying models in order, this would be a key focal point for board level
discussion.
When S = 350 was involved, the A350 = 5 and V1 × A350 = $10 million values had to reflect
the difference between travel on public roads and rail travel on NER trains when very seri-
ous airplane crash levels of fatalities were involved. The public probably saw 1 to 3 fatalities
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
518 Employing planning tools in practice
as the usual outcome of car accidents involving fatalities, but 100 to 300 fatalities as the
usual outcome of commercial airliner accidents involving fatalities, so an S = 350 scenario
was a major catastrophe even by air crash standards.
NER might start by thinking about a V350 alternative to V1 associated with its national
airline instead of its national highways agency. The country’s national airline had safety stand-
ards comparable to international airline safety standards which might suggest a V350 as much
as ten times that of the V1 for road traffic incidents. However, this ratio was highly contro-
versial for a range of reasons, including the nature of the relationship between safety levels
and levels of spending to avoid fatalities, what was measured, and the rationality of people
expecting air travel to be safer than road and rail travel per passenger kilometre or mile.
She did not want to court controversy that day by overstating the case for an A350 which
was bigger than A1 to a questionable extent. However, in her view, the role of the courts and
other aspects of revealed preference information available suggested that using an A350 = 5
value so that V1 × A350 = $10 million provided a plausible round number illustration. The
range for the value of V1 × AS from $2.2 million to $10 million as S went from 0 to 350
involved an average uplift by a factor of $10 million/$2.2 million = 4.5, which seemed
plausible to her. Furthermore, this range was big enough to allow her to demonstrate what
S dependence could imply but modest enough to avoid seriously overstating its importance.
The current V = $4 million used by NER was closer to the bottom- end V1 × A1 = $2.2 mil-
lion value than it was to the top- end V1 × A350 = $10 million value and below the arithmetic
mean of $6.1 million, but that was consistent with her understanding of the thinking behind
the T = $4 million in current use.
Sophie reminded the board that they were talking about a catastrophic incident when
they discussed the S = 350 scenario, involving from 200 to 1,200+ fatalities. NER had never
experienced a catastrophe like this, and her working assumption was catastrophes at this
level had a 2,000 year return period. But the S = 350 scenario clearly could happen, and the
NER board had both a moral duty and a legal obligation to address how both the assumed
impact and the assumed probability could be reduced, treating an S = 350 scenario with
concern proportionate to its importance.
Assuming A350 = 5 was clearly debatable, over a very wide plausible range (like 1– 10), but
she did not want to engage in this debate today. The key issue here was she did not think
that NER would be trusted, nor should NER be trusted, if they did not associate serious airplane crash levels of fatalities with air travel levels of safety as expressed through court
judgements and all other relevant forms of revealed preference information. She had simply
used what she saw as a plausible uplift from a road traffic situation basis to reflect this for
illustrative purposes. When she had the rest of her numbers and underlying models in order,
this would be another key focal point for board level dialogue.
The injuries component V1 × IS for S = 1 and 350
Sophie explained that she had outlined the rationale for her AS illustrative numbers with
some care because she believed the credibility of her approach with the board demanded
this care, but it was not worth a comparable level of detail about her other example TS
component illustrative values for that day’s presentation. Available data could be used in a
relatively straightforward manner to develop initial estimate suggestions about appropriate
operational values for IS for S = 1 or 350 based on injuries at various levels of S, but she
had not attempted anything more than purely illustrative numbers to illustrate her recom-
mended approach for this initial meeting.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 519
She would use I1 = 0.6 and I350 = 1.8, so the injuries component of TS was
V I $2m 0.6 $1.2m for S 1,
$2m 1.8 $3.6m for S 350. 1 S× ×
×
= = =
= = =
This implied that an IS component of TS varying over the range $1.6 million to $3.6 million
would replace the current values derived from separate relatively complex computations.
She indicated that I1 = 0.6 meant she was assuming that when S = 1 the total number of
people injured in all categories of injury levels required slightly more than half of the weight-
ing given to one fatality (0.6/1.1 = 0.55). She thought this was an appropriate illustrative
initial stance when one NER fatality was involved in the absence of appropriate data analy-
sis, to avoid playing down the importance of injuries without understating the importance
of fatalities. A plausible range might be −50% to +100%, and both data and current NER
equivalent fatality assessments for injuries would be used later.
NER currently used a proportion of their T = $4 million for injuries in a number of dif-
ferent injury categories, comparable to common practice since the 1970s NHTSA and Ford
Pinto approaches. In this way NER currently converted estimates of the numbers injured into
‘equivalent fatalities’, treating a small number of seriously injured people as equivalent to one
fatality, a larger number of moderately injured people as equivalent to one fatality, and so on,
separately estimating numbers of people in each of these categories. She was recommending
they avoided this kind of detail, saving effort as well as increasing clarity. They would be bas-
ing their injuries component of TS on an assumed expected number of injuries in different
categories of seriousness for each fatality plus an average V1 × AS equivalence, linked to sce-
nario examples to clarify the weighting behind these averages. Estimates of expected numbers
of people injured to different degrees and associated ranges could be provided whenever
injuries were an issue, and the kind of ‘equivalent fatalities’ metric currently used by NER
might provide particularly useful initial guidance, drawing on their approach when model-
ling incidents involving injuries but no fatalities (grade 2). However, in her view modelling
injuries within the ES factor via IS would save time currently spent on detail which was not
very productive. It would also ensure that the correlation between the numbers of people
killed and injured in incidents was dealt with effectively. And crucially, it would contribute to
integrating all cost related to fatalities in a useful manner which allowed different people to
weight injuries relative to fatalities in different ways within an agreed overall TS value.
In her view, V1 × I1 = £2 million × 0.6 = $1.2 million was a plausible order of magnitude
until they had done more work on available data, but the numbers were purely illustrative,
not intended to suggest recommendations. She suspected the board might want a very
careful and detailed discussion of these issues after she and her team had done a significant
background preparation. Testing the relationship between this greatly simplified IS factor
approach and current practice would be an essential part of initially estimating appropriate
IS values.
Her I350 = 1.8 rationale was simply a factor of three increase on the I1 = 0.6 value, to reflect
a lower level S effect than for fatalities, assuming the number of people injured in various
categories for every fatality did not change too massively as S increased. She currently had no
real understanding of how the scale and nature of injuries could be expected to change as S
increased, so her purely illustrative example numbers for the IS involved a particularly wide
uncertainty range associated with her very simple factor of three assumption. Some directors
might hold very different views about the relative importance of injuries versus fatalities,
making discussion crucial when effective prior data analysis could support it properly.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
520 Employing planning tools in practice
The damage component V1 × DS for S = 1 and 350
It was obviously important to consider directly associated knock- on operating costs and
opportunity costs associated with lost revenue in addition to any physical material damage
costs as part of the costs NER had to meet given an incident involving fatalities. In her view
a further TS component factor defined by V1 × DS for the cost of damage to rolling stock
and tracks plus all associated knock- on costs, including an opportunity cost portrayal of lost revenues, was both convenient and useful, instead of treating these costs separately, the cur-
rent norm. For clarity efficiency reasons, it was important to simultaneously address fatalities
plus correlated injuries, plus correlated cost implications of physical material damage to the
system, plus all costs associated with this damage. The rationale was directly linked to the
rationale for her separate IS approach, ensuring effective internally consistent decision mak-
ing with minimal effort and associated decision- making cost.
The board would no doubt want to see direct material damage expected cost estimates and
associated knock- on expected cost estimates for a range of sub- scenarios associated with each
of their five S value scenarios, and future practice would require modelling the relationship
between fatalities and material damage plus direct knock- on costs in fairly complex terms.
For illustrative purposes she would just use D1 = 0.2 and D350 = 0.4 so the damage com-
ponent of TS was
V D $2m 0.2 $0.4m for S 1,
$2m 0.4 $0.8m for S 350. 1 S× ×
×
= = =
= = =
This implied that a DS component of TS varying over the range $0.4 million to $0.8 million
would replace the current values derived from separate relatively complex computations.
Her D1 = 0.2 meant that when S = 1 she was assuming the average total material dam-
age plus knock- on cost was $400,000. She thought this was an appropriate illustrative ini-
tial stance when one NER fatality was involved in the absence of appropriate data analysis
to avoid playing down the importance of injuries and fatalities while acknowledging the
role of more direct material damage and knock- on costs. A plausible range might be −50%
to +300%, and she suspected knock- on cost issues addressing lost revenues fully explored
might prove much more important than her illustrative values would suggest.
Her D350 = 0.4 meant that when S = 350 the scale of the physical damage and knock- on
implications involved per fatality might increase by a factor of two, as an illustration of pos-
sible scale changes, but again this was purely for illustrative purposes.
The residual component V1 × RS for S = 1 and 350
It was crucially important to address the dependence between the number of fatalities and
a residual component to reflect anything not fully embedded in the other three TS com- ponents. In her view it was useful to see the role of RS as a very broadly defined ‘residual
provision’ in a ‘closure with completeness’ sense which explicitly included a full synthesis
of all three other TS components, plus the ‘prudence provision’ with a nature and role still
outstanding on their initial meeting agenda, plus an overall ‘rounding up when in doubt’.
For illustrative purposes, R1 = 0.1 and R350 = 2.8 would be used so that the ‘residual provi-
sion’ component of TS was
V R $2m 0.1 $0.2m for S 1,
$2m 2.8 $5.6m for S 350. 1 S× ×
×
= = =
= = =
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 521
This implied that an RS component of TS varying over the range $0.2 million to $5.6 million
would replace the current implicit NER values of zero for all S values.
She suggested they start to interpret these purely illustrative values in terms of the overall
role of the ‘residual provision’ over the S = 1 to S = 350 range.
If RS = 0 for all S was assumed, the implication was that expected costs as measured by the
other three ES component factors needed no further uplifts. In her view this would not be seen by all relevant parties as an appropriate position for the NER board to take, and there
was a reasonable case for saying that it should not be seen as an appropriate position. The obvious and basic reason was RS = 0 did not reflect a trust contribution based on the full
range of ethical considerations which most people would hold to be relevant. She did not
want to labour discussion of issues involved that day, but did want to emphasise the need
to accommodate all relevant trust- based considerations. It was also crucial to accommodate
a range of underlying technical concerns which she would touch on when they became
relevant.
The TS residual component R1 = $0.2 million implied that in an S = 1 context, NER was
prepared to spend this sum in addition to the £3.8 million associated with the other three
components of TS, a total of $4 million.
This would round up the $3.8 million illustrative value to $4 million. Relative to using
T1 = 3.8, it implied a marginally reduced expected number of fatalities in an S = 1 con-
text via additional spending on safety and security, attributing 5% of the total TS value to
an increased effort to reduce the probability of S = 1 scenario incidents because of NER
revealed preferences associated with seeking trust for their overall approach. But it would
also preserve consistency with the current NER use of T = $4 million, and it would avoid
suggesting that NER would reduce their concern for single fatality incidents, which she
thought was important.
What the TS residual component of R350 = $5.6 million implied was in an S = 350 context
NER was prepared to spend this sum in addition to the $14.4 million associated with the
sum of the other three components of TS for S = 350, a total of $20 million.
This would round up the $14.4 million illustrative value to $20 million, a significantly
bigger adjustment in proportionate terms than for S = 1. It should lead to increased focus on
reducing the expected number of fatalities in an S = 350 context relative to TS = $14.4 mil-
lion, attributing 23% of the total TS value to an increased effort to reduce the probability of
S = 350 scenarios occurring because of NER board approved revealed preferences associ-
ated with a desire for trust by all relevant parties which went beyond the implications of the
AS + IS + DS components and R1 as S increased. Part of the rationale for this bigger uplift as S
increases was based on the role of enlightened prudence provision, to be considered shortly.
Taking these two S value RS scenarios together, the numerical example had simply rounded
the $3.8 million up to $4 million and the £14.4 million up to £20 million. This provided a
simple factor of five illustration for the increase in TS over the range S = 1 to 350. It assumed
that a very modest residual provision was appropriate for S = 1, in conjunction with a more
significant residual provision for S = 350.
The absolute values would obviously matter in practice when a developed version of her
approach was implemented. But it was the implications of NER explicitly showing more
concern about the implications of very big S value incidents that she particularly wanted
to communicate at their initial meeting, with medium range increases in TS illustrating the
effects of an overall S dependent relationship.
In summary terms the implication of her example was that NER was prepared to make
only a nominal 5% uplift to go beyond the shadow cost of $3.8 million for the S = 1 scenario
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
522 Employing planning tools in practice
to avoid a reduction relative to the current use of $4 million purely for fatalities. However,
when S = 350 was involved, the NER residual provision was midway between fatalities and
injuries in importance, and much more important than physical damage to the system.
What she wanted to demonstrate to the board at their initial meeting was the implications if NER had a significant revealed preference for avoiding high S value scenarios which it was prepared to demonstrate to their passengers, their staff, their regulator, the courts, and the pub- lic at large, in terms of their willingness to spend additional money to reduce the risk of very low probability but very high impact incidents, going well beyond current ‘good practice’.
Why this might be reasonable would be explored in more detail later that day, but some
immediate indications of ‘anything else’ concerns associated with RS included a need to
regain the trust lost by two serious accidents in the last two years. This issue could become
crucial if these two incidents triggered a campaign to nationalise NER. It was certainly very
important to avoid complacency after long periods of no serious accidents, but significant
additional weight given to preventing serious accidents in the present circumstances might
be presented and interpreted as a prudent act of contrition, undertaken with a view to
regaining lost trust, an acknowledgement that they had to demonstrate a willingness to try
harder, and they had to back up their promises to try harder with cash expenditure.
There was also a case for the board making a generous S = 350 uplift via this fourth TS
component based on the grounds that for high S values it was better to err on the cautious
side given the uncertainty involved, a variant of the ‘precautionary principle’, and of ‘going
the second mile’. That is, in terms of seeking an optimal TS value, which is by nature very ambiguous and contentious, the ‘penalty function’ associated with getting it wrong was not
symmetric, so erring on the high side when catastrophes were involved was a sound strategy.
Formal and explicit acknowledgement of the need to round upwards to deal with ambiguity
to cope with what the board acknowledged NER did not know, in addition to the role of
asymmetric penalty costs, would not go amiss.
As a further somewhat separate issue, if the board wanted to develop its ‘national treasure’
status, as ‘a gift to the nation’ the NER board might want to go beyond the minimum needed
to regain and then preserve trust. They might interpret this further uplift as an explicit ‘gift
adjustment’. Their intentions here could be sincere philanthropy or a bid for more passengers
as a variant of Trevor’s ‘non- price advantages’ (as discussed in Chapter 6). A ‘bid for more pas-
sengers’ interpretation was comparable to Volvo promoting their reputation for cars designed
and built for safety starting many decades ago, when other car makers seemed less concerned
about safety. The board should consider this component very carefully at a later date for all S
values, and very significant uplifts might be appropriate for S = 350.
The NER board would have to address all relevant TS residual component issues associ-
ated with RS very carefully in the near future, including the role of enlightened prudence,
still on their ‘to do’ list. But for current purposes they might just assume that NER was scal-
ing up the AS, IS, and DS components so that an overall factor of five for increase for TS as S
went from 1 to 350 could be used for numerical illustration simplicity.
Sophie’s further reflections on the overall transformation factor TS for S = 1 and 350
Sophie began a reflective review by suggesting that the illustrative example values selected
meant that for S = 1 the overall value of TS was $4 million, while for S = 350 the overall value
of TS was $20 million, a simple factor of five increase which could be linked to the pattern
of component changes with a view to comparable illustrative simplicity.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 523
With S = 1, the fatalities contribution defined by AS dominated, providing 55% of the
total, followed by injuries defined by IS contributing 30%, followed by damage defined by
DS adding 10%, with the residual defined by RS contributing just 5%.
With S = 350, fatalities still contributed 50% of the total, with injuries contributing 18%,
damage adding only 4%, and the residual now a significant 28%.
She was sure they would find the implications and the full role of the residual provision
puzzling until they saw how it worked in practice. But she would explain the rationale of the
prudence provision shortly, and she believed they would soon begin to see the rationale for
her illustrative example numbers for the first three components, even if they thought big-
ger or smaller numbers more appropriate for these components. Furthermore, she believed
they would soon begin to see that what the RS value was doing was providing them with
the ability to increase the relative importance of RS as S increased towards 350 for reasons
related to the NER board’s revealed preferences which could go beyond current industry
good practice safety and security concerns if the board believed that was appropriate. RS = 0 for all S values was a feasible choice within her framework – but she would strongly advise
against this choice, and she believed they would see why shortly.
What was crucially important in her view was the development of a properly grounded
TS concept for operational NER use in the very near future which the board understood
and took responsibility for. She believed this would prove a key aspect of NER building and
retaining grounded trust by all relevant parties, even if tested by further serious incidents.
The acid test of trust which was well founded was the maintenance of trust when things
went wrong.
Their initial meeting was about the framework NER needed to build and maintain
grounded trust, and the component concepts needed, in broad what needs to be done
terms. The board should not get distracted by the example numbers used.
What the board had to do to when assessing the complete set of four components in the
light of analysis her team would have to complete as soon as possible was ask the question,
‘Were the TS values fair and reasonable, deserving the trust of our passengers, our staff,
and other interested parties, including shareholders, regulators and courts, given the way
they were defined, and given the NER overall approach to safety and security?’ If the board
accepted V1 but believed any TS was too low or too high, they should mandate NER staff to
adjust ES up or down by revising appropriate components until the board were comfortable.
Being comfortable with the basis of V1 was the obvious starting point.
Furthermore, when they were confident that their approach was the basis of well- founded
trust, NER needed to make the whole process public and transparent, in the same ‘full
disclosure’ manner which should be expected of any public sector organisation in order to
maintain NER ‘national treasure’ status. And NER needed to be ready to respond to reason-
able representations to increase or decrease ES or V1 or adjust other aspects of its approach.
Individual board members need not fully agree on individual components – a higher
AS with a compensating lower IS would not affect decisions. However, this flexibility did
not justify bias, especially wilful and amoral bias like deliberately underestimating DS in
order to inflate AS for public relations purposes. Legitimate differences of opinion should
be anticipated but tested for legitimacy during the process of reaching agreement on overall
TS values.
Selecting a four factor ES component set given V1 and an overall approach to safety and
security involved ethical aspects which should be related to trust concerns by all relevant
parties. The TS values selected could and should be judged as ‘reasonable’ or ‘not reason-
able’ by all relevant parties given all the assumptions used to determine those values.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
524 Employing planning tools in practice
‘Relevant parties’ had to include courts of law, and satisfying the legal interpretation
of ALARP in an NER context was clearly a very important issue. However, the overall
judgement by the board on behalf of NER was much broader. The board’s view had to
accommodate differences in personal opinions within the board, but even bigger differences
between shareholders, train drivers plus other ‘on- board’ NER staff, passengers, and the
public. A key role for the board was setting the tone for the NER approach to the ambiguity
implied by all these differences as judged by the board.
If boards of private sector organisations, or their public sector equivalents, were deemed untrustworthy in this role, some people would argue that governments had a clear duty to
punish the directors for any failures to meet reasonable standards and relieve them of their
responsibilities, in addition to punishing shareholders if a private sector organisation was
involved. Most people would agree that it was crucial to avoid punishing passengers via
crude fines which were simply passed on via ticket price increases.
Governments had a clear duty to regulate both private and public sector organisations
in an effective and efficient manner, and they should not wait for a series of catastrophic
accidents to happen – they should be proactive. If governments failed in this respect, railway
user associations and other comparable bodies might need to raise funds to use the courts
to sue relevant parts of the government, as well as NER, and apply other forms of leverage,
seeking support from the press.
Sophie acknowledged that some board members might see these statements as controver-
sial, and perhaps too boldly confrontational, but she thought the NER board should see all
of them as issues worth addressing before others decided to do so.
Defining TS for S = 5, 20 and 70
Moving on to ES for S = 5, 20, and 70, Sophie now assumed
E 3, so that T 2 3 6,
E 5, so that T 5 5
20 20
= = =
= =
×
22 5 10,
E 8, so that T 2 8 16.70 70
×
×
=
= = =
She suggested that in practice they would have to use the same four components decom-
position pattern to clarify what they were doing in detail for each relevant S value, but for
that day’s purposes she would simply use an illustrative interpolation approach without even
attempting to explore a rationale for the numbers. Her example assumed that the S = 5 con-
text was closer to the S = 1 context than the S = 20 context. Furthermore, she had assumed
that the S = 20 context was closer to the S = 1 context than the S = 350 context. But the
S = 70 context was assumed to be closer to S = 350 context than the S = 20 context.
Overall these three intermediate scenario values defined an asymmetric ‘S curve’ with an
asymptotic approach to both E1 and an E350 upper limit defined by the physical capacity of
the system. There was clearly room for a lot of debate about the best shape for the S curve
used, but much more work was required before this debate would be productive.
Any other non- linear relationship which the NER board wanted to use could be por-
trayed with a five scenarios approach, and more or fewer than five scenarios could be used.
One of the reasons five scenarios were used was the need to capture this level of flexibility to
provide a plausible basis for discussion.
In Sophie’s view an ES which could not be given a non- linear asymmetric S curve as the
parameter S increased was inappropriate, and the current NER use of T = V1 × E = $4 million
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 525
for all S = 1, . . . , m was an extremely limited special case which could not be defended. She
believed testing its implications in the more general framework she was proposing would
clearly demonstrate why it was not defensible. To fail to even consider such a test was unac-
ceptable from an EP perspective. What they had to avoid at this point was locking NER into
any particular restrictive approach.
The fatalities metric S transformed by TS to define ‘cost related to fatalities’
Sophie presented Table 9.2 next, building on Table 9.1 and the intermediate discussion plus
the earlier relationship definitions CS = S × TS = S × V1 × ES. She was now formally defin-
ing CS as ‘the expected cost ($m) related to fatalities if scenario S occurred’ and C as ‘the
expected cost related to fatalities for all S’.
Interpreting Table 9.2 involved understanding the notation definitions for CS and C just
provided plus the nature of the concepts and the underlying relationships.
CS = S × TS was ‘the expected cost ($m) related to fatalities if scenario S occurred’, a con-
ditional expectation associated with the complete set of S values in the range portrayed by
the scenario. The TS transformed fatalities as measured by S into $m, using the two compo-
nent transformations factors defined by TS = V1 × ES to obtain the CS values.
C was ‘the expected cost ($m) related to fatalities for all scenarios per annum’, an uncon-
ditional expectation computed using the CS and the associated PS. The PS weighted the CS
values to compute the unconditional expected value C.
The first column of Table 9.2 replicated Table 9.1 apart from dropping the size scenario
ranges in brackets to keep Table 9.2 as simple as possible.
The second column simply repeated the PS values of Table 9.1.
The third column showed the computation of the CS values using S × V1 × ES for each S
value. The first row’s entry accommodated the S = 0 special case. The second showed the
calculations for S = 1 assuming V1 = 2 and ES = 2 to produce TS = 4 ($m). The third and
subsequent entries showed ES = 3 for S = 5 rising to ES = 10 for S = 350, using the illustra-
tive non- linear interpolation suggested in the last subsection. In addition to wanting to be
able to deal with non- linear asymmetric ‘S curves’, Sophie wanted to provide a plausible
range of variations in her set of everything else factors to illustrate how this transformation
fed through to C.
Table 9.2 Cost related to fatalities table, based on Table 9.1.
Size scenario Unconditional probability
CS, expected cost ($m) related to fatalities if scenario S occurred
Contributions to C, expected cost ($m) related to fatalities for all S per annum
S PS S × V1 × ES = CS PS × CS (as a %)
0 0.9500 0 0 (0%) 1 0.0025 1 × 2 × 2 = 4 0.0025 × 4 = 0.010 (0%) 5 0.0250 5 × 2 × 3 = 30 0.0250 × 30 = 0.750 (8%)
20 0.0200 20 × 2 × 5 = 200 0.0200 × 200 = 4.000 (38%) 70 0.0020 70 × 2 × 8 = 1,120 0.0020 × 1120 = 2.240 (21%)
350 0.0005 350 × 2 × 10 = 7,000 0.0005 × 7000 = 3.500 (33%) C = 10.500 (100%)
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
526 Employing planning tools in practice
C was the expected cost related to fatalities for all S scenarios per annum, an expected
value defined by the sum over S = 1, . . . , 350 of the PS × CS products, a key summary
parameter.
The fourth column showed the contributions to C for each S value. C was the $m cost
of annual expected number of fatalities (F in Table 9.1) transformed into a $m equivalent.
The percentage contribution to C made by each size scenario was shown in brackets, indi-
cating that the significant proportionate contribution to F made by the higher S values in
Table 9.1 was further increased here, as might be expected given the S dependent TS effect.
However, it was crucial to understand that the further increases driven by the S dependent
TS were not as big as some people might expect, perhaps counter- intuitive, and of consider-
able importance in terms of understanding what really mattered in broad overview terms.
To begin to elaborate the relative importance of the key drivers and their overall effect,
the F and C percentage contributions for each S size scenario were
S equal to 1 5 20 70 350
F % contributions from Tablle 9.1 0 15 47 17 21
C % contributions from Table 99.2 0 8 38 21 33
What a simple starting point interpretation of C = 10.500 suggested was about $10 million
per annum was the annual expected value of costs related to fatalities using all the underlying
illustrative example parameters.
Linking this to the basic parameters L and F, in annual expected value terms using these
example values, NER were dealing with
L = 0.05,
F = 0.84,
C = $10m,
working to comparable levels of precision to emphasise these were order- of- magnitude
example numbers.
The earlier breakdowns suggested that at least half of this C value was associated with
fatalities, but some of it was associated with related injuries, physical damage to the system,
associated knock- on costs, and a residual provision.
The Tables 9.1 and 9.2 contribution percentages implied that the S = 20 scenario con-
tributed the most to both F and C, but the S = 70 and 350 scenarios played an even more
significant role in C than they do in F.
On average less than one fatality per annum was expected, but Table 9.1 suggested that
85% (47 + 17 + 21) of this expected value was associated with S scenario values of 20 or
more, and Table 9.2 suggested that 92% (38 + 21 + 33) of the expected transformed value
of $10.5 million per annum was associated with S scenario values of 20 or more. What mattered crucially in terms of the average outcome happened relatively infrequently, primarily because of the effect of the very long distribution tail. However, a factor of five increase in TS as S increased from S = 1 to S = 350 did amplify the highly asymmetric distribution effect driven by one very long tail.
The weighted mean value of TS was $10.5 million/0.8425 = $12.5 million, approximately
the same as the simple mean of ($4 million + $20 million)/2 = $12 million, with about
$6 million of this associated with fatalities. A mean of about $6 million associated with the
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 527
fatalities component of TS equated to a 50% increase on the T = $4 million currently used,
weighting the increase by relevant PS. In effect, when an incident happened, S = 1 and S = 5 scenarios were not very important in either F or C terms. It was the S = 20 or more scenarios
which really mattered. Furthermore, the S dependence in the TS amplified the long tail dis-
tribution effect portrayed by Table 9.1 in terms of F, but the long tail distribution effect on F was crucial, arguably at least as important overall as the S dependence of ES issue.
Sophie emphasised that she had deliberately erred on the modest side for the illustra-
tive effects of S dependent ES changes as S increases from 1 to 350, and she had similarly
erred on the low side when estimating CPS values for high S values. She wanted everyone
to concentrate on points of principle without feeling unduly threatened by the illustrative
numbers. She suspected that the full implications of using real numbers would be much
more concerning than her illustrative numbers, but she had deliberately avoided unneces-
sarily overstating any of the suspicions she had about possible concerns with big impacts, in
case these suspicions proved incorrect.
The key thing she wanted the board to begin to understand from these two tables and
their use so far was the need for NER, operating as a ‘big team’, led by their board, to learn to be very afraid of very big S value scenarios, and never forget this fear. The low probability of big S value scenarios occurring was never a valid excuse for complacency.
The key follow- on issue was the need to vigorously pursue opportunity efficiency in a very
optimistic and positive spirit, always striving for zero injuries and fatalities.
This combination of well- founded fear and an enterprising and creative approach to remain- ing optimistic, grounded on a clear understanding of ambitious targets and opportunity effi- ciency, should be the key overall NER goal, and it could not be achieved without board leadership.
Building and using the rest of the base- case model set
Sophie indicated that practical use of the holistic base- case model set would require an
extension of Tables 9.1 and 9.2 in terms of the incident sets addressed. These tables just
addressed collisions with fatalities, and they would have to extend the range of incidents
covered by such tables before they could use them directly to make decisions.
To address all aspects of collisions they would have to add consideration of collision
incidents limited to injuries or just material damage and knock- on costs. Furthermore, they
would have to consider ‘derailment’ incidents (not leading to collision), ‘level-crossing’
incidents (not leading to derailment), and ‘any other incident’ component models. Fur-
ther still, they would have to use the level- crossing models to build a component of the
derailment models which dealt with the subset of these incidents initiated by level- crossing
incidents, and a component of the collisions models which dealt with the subset of these
incidents initiated by derailments. Still further, they would have to decompose all these
models to provide greater granularity for the NER safety and security team to understand
failure modes and effects in the various ways which safety engineers have found useful for
addressing the root causes of accidents for many decades. This might include, as an example,
detailed studies into the implications of alternative particular designs for the new generation
of very high- speed trains recently suggested.
The board did not need to explore what this involved until it was accomplished, but it
was important to understand the need for a critical review of all current NER safety models,
fitting them into the top- down structure just outlined. This review would have to modify
the existing models as needed for compatibility and consistency, generalise their capabilities,
and fill in any important gaps. The board should now assume this would be done as soon as
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
528 Employing planning tools in practice
possible and consider in outline how this extended base- case model set with its more com-
prehensive F and C values could be used to help NER make opportunity efficient choices.
For numerical example purposes in this first meeting she would just build on Table 9.2
and all the underlying parameter assumptions, assuming they incorporated all relevant inci-
dent sets, to keep her presentation simple. This implied a degree of optimism which she and
her team would strive to achieve, but it might be useful to see her example results today as
an aspirational stretch target. That is, she would, in practice, try to ensure that the numerical
equivalent for NER addressing all types and categories of incidents was no more pessimistic
than Table 9.2, but they should not be surprised if later estimates for the complete model set
were significantly less optimistic. They should certainly not see her current numerical exam- ples as pessimistic illustrations designed to frighten them. She wanted the board to have
realistic expectations, but stay optimistic about what could be achieved by a well- grounded
opportunity management approach. She would now begin to give them some hints about
what fully developed opportunity management instead of traditional safety management
could mean in a NER safety and security planning context.
A new generation signalling system as a simple ‘physical system’ example
Sophie suggested they start with a very simple analysis of a relatively simple ‘physical system’
example. Say they were considering investing in a new generation signalling system, and say
they made the very strong working assumption that the new signalling system would reduce
the (1 – P0) = 0.0500 underlying Tables 9.1 and 9.2 by 10% without affecting the relative
size of the conditional probabilities or anything else. The mechanics of the Table 9.1 and
9.2 calculations meant a 10% reduction in C, $10.500 million × 10/100 = $1.05 million.
If the ‘new system cost’ per annum as measured by the amortised annual cost of the new
system was less than $1.05 million, and all associated working assumptions were reasonable,
the new generation signalling system looked worthwhile, in the sense that the ‘flip- point’
value for the new system cost was $1.05 million per annum. For example, if the new system
cost per annum estimate was of the order of $0.5 million, there would seem to be a strong
case for agreeing to the new system in terms of an annual expected value reduction. How-
ever, if the new system cost per annum estimate was of the order of $2 million, there would
seem to be a strong case for rejecting the new system in terms of an annual expected value
cost increase.
Furthermore, if the new system cost per annum estimate was of the order of $1 million,
the analysis would suggest ‘no clear advantage either way in terms of C plus the new system
cost’. This would be very useful guidance because all other considerations not measured by
C or the new system cost can now be the focus of attention. Indeed, once it was clear how
useful it could be, it became obvious that what was not measured directly always needed careful consideration.
Valuing the marginal changes in expected outcomes using the assumed costs defined by
a benchmark avoided fatality parameter V1 and an ‘everything else factor ES’, the board had
clear guidance whatever the outcome, provided all the working assumptions were deemed reasonable, the board was clear about what TS did not measure as well as what TS did meas- ure, and annual expected cost values could be used without considering departures from these expectations – a particularly big assumption which enlightened prudence would address.
Apart from board level inferences, Sophie’s team could use this kind of framework for
developing proposals, in conjunction with any NER group which wanted to develop pro-
posals with safety and security implications.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 529
Of critical importance, leaving aside working assumptions about the TS parameters for
the moment, the board and all NER groups bringing proposals to the board have a clear
basis for testing and exploring the implications of all working assumptions within an expected outcomes framework, addressing qualitative assumptions as well as quantitative assumptions. Furthermore, we can later systematically generalise this framework, including addressing enlightened prudence in a much better framework for understanding what ‘risk’ ought to mean in a safety and security context.
One of several crucial inferences which can be made building on this very simple example
which Sophie wanted to emphasise was that safety and security planning needed to pay spe-
cial attention to reducing the conditional probabilities associated with the high S values, and
the implications of departures from expectations involving these scenarios. For example, if
the assumed S = 350 scenario unconditional probability of = 0.0005 could be driven down
by 30%, the reduction in C implied by Table 9.2 of $3.500 million × 30/100 = $1.05 mil-
lion implied it would be worth about $1.05m per annum in expected value terms before
considering any of the knock- on changes to other parameters, of the same order of magni-
tude as a 10% reduction in (1 – P0). The reason was the S = 350 scenario contributed about
a third of the value of C in Table 9.2. In addition, it made the S = 350 scenario much less
threatening in terms of potential departures from expectations, a risk issue going beyond
expected outcomes which needed exploration later.
More generally, reducing the seriousness of incidents if they happen may be much more
effective and efficient than reducing the probability of incidents happening. The base- case
model set provided a comprehensive and coherent basis for exploring these central trade- off
issues. Trade- offs between prevention and mitigation were a crucial issue to understand and
deal with effectively.
In part this was why ‘staying very afraid of high S value incidents’ was a key part of the cor- porate culture NER needed, bearing in mind that the very low probability of such incidents
made it exceedingly difficult for people to sustain focus on safety when immediate pressures
on profit or cost were difficult to resist, and the practical difficulties associated with estimat-
ing very small probability values.
Sophie then very briefly considered some of the complications which would need address-
ing in practice.
A new generation signalling system might reduce some of the conditional probabilities
more than other considerations. NER would need to model the underlying failure mecha-
nisms to understand this, as well as linked impacts on L via P0. Furthermore, the new-
generation signalling system might have operational benefits which had significant value
beyond safety, like allowing a higher traffic volume with higher revenue implications. They
would need to ensure that this benefit was deducted from the new system cost or considered
in conjunction with the new system benefits in relation to the total system cost changes.
The latter approach would usually be simpler in practice. Sophie noted that there was no
need to separate new system costs and benefits for different S values, and it was important
to understand that in practice additional costs for decreasing expected fatalities could be dif-
ficult to separate from additional costs for other benefits. Benefits might come as bundles,
roughly comparable to a generalisation of the marketing concept ‘buy one, get one free’,
but potentially much more complex in a variety of important ways which needed context
specific understanding.
Indeed, separating signalling from other aspects of the ‘operations information system’
might be very unwise if they wanted to develop the kind of redundancy, resilience, and
robustness generally advised. For example, if a signalling system failure meant that a train
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
530 Employing planning tools in practice
driver was not advised that there was a stationary train on the track two kilometres ahead,
a key question was, ‘What other ways could a train driver get this information?’ Sophie
indicated she was not a railways expert, but on the train to their meeting that morning, she
wondered if large screen TVs in train driver’s cabs linked to a telescopic lens night- vision
capability CCTV camera on the front of each train, perhaps computer adjusted for track cur-
vature with driver override, might serve as backup to signal failures. But the point she really
wanted to make was this kind of capability might also help to avoid or reduce the impact of
full speed collisions with derailed trains and vehicles on level crossings – including vehicles
maliciously placed there just before a train was due. If train drivers plus all other experts who
understood the issues and the technology possibilities worked together, they might come up
with a concept like that illustrated by her camera speculation that was novel, effective, and
efficient. Furthermore, in conjunction with good controller- to- driver radio communication
systems, adding this kind of feature to the system might be better value for money than new
signalling system hardware, but neither or both may be worthwhile. Sophie emphasised that
she had no idea whether CCTV cameras on the front of trains made any sense, but she was
trying to illustrate the kind of creative thinking required of all those within NER, or availa-
ble to NER, who might contribute to a creative search for safety and security improvements.
Furthermore, an even wider view of the physical system was important. As an example, the
recent derailment interpreted as a collision near miss, not to mention recent terrorist tactics,
meant that NER had to think about how to lower the fatalities and injuries implications of
high impact incidents via creative design and operations changes which might be radical. The
new very high- speed railway lines they were planning at present might benefit significantly
from much greater separation of the tracks than currently envisaged, especially at vulnerable
locations, perhaps also using a deflection barrier comparable in function if not in appear-
ance to those used on separated motorway lanes. A related alternative which might be worth
considering was a three or four track norm for all high speed lines, with only the outer tracks
normally used unless maintenance was required. Driverless trains might be considered, with a
view to large trains being replaced by sets of small trains, decreasing the time interval between
trains for the convenience of passengers in terms of more frequent service as well as addressing
catastrophe scenario safety concerns, just to illustrate the scope for flexible creative thinking.
Further still, a very effective approach not requiring driverless train technology might be
just reducing train sizes by 25% but running more services at peak times, using safer car-
riages which were also more comfortable, and charging more for tickets for a better service,
phasing in these components of a new overall strategy as soon as each component was fea-
sible. This might let NER increase total passenger numbers, revenue, and profit simultane-
ously while also significantly increasing safety and security as measured by C.
What the board needed was a top- down view of potential physical system improvement
options, with their safety and all other operational benefits fully integrated with their costs
and revenues. What NER as a whole needed was extensive use of this perspective by the
NER safety and security team, plus all other relevant NER players, to develop well- polished sets of options to recommend to the board. The board would need confidence that the
safety and security team, plus all other NER staff with contributions to make, had explored
a much wider set of combinations and permutations of changes in an internally consistent
manner to eliminate dominated options.
The recommended proposals for board consideration would involve important trade- offs,
leaving the board, with suitable support, to consider the trade- offs that were the board’s
responsibility.
The recommended changes might have been initiated by the perception that a new gen-
eration of signalling technology was now available, and the working assumption that it was
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 531
worth a good look by NER. However, NER needed to ensure that everyone with useful
input and creative ideas was able to contribute to the ideas ultimately tested before propos-
als came forward to the board. Train drivers who had experienced near misses or observed
features of the current system which could lead to problems had knowledge which needed
systematic gathering and integration. Other on- board NER staff, signalling staff, passenger
associations, and railway safety consultants also had knowledge which might be useful.
Track maintenance contracting as a ‘procedural systems’ example
Sophie suggested they now consider looking at system changes which were initiated from
a somewhat different perspective, using track maintenance contracting as a ‘procedural sys-
tems’ example. She did not wish to get into recriminations about what made the 40 fatality
incident recently experienced ‘an accident waiting to happen’, as reported in the press. What
she did want to explore briefly was a framework for considering procedural systems from a
‘designing desirable futures’ perspective.
What she meant by this was starting with the procedural system features and associated
implications which they wanted, a luxury they could not afford in terms of most aspects of
NER physical systems because of the ‘heritage’ issues. In the context of a procedural system they could afford to scrape the existing framework completely, after designing and appro-
priately testing a completely new procedural system to achieve as much of their ‘desirable
future’ as they could. They could begin the design process by focusing on effectiveness in
a ‘creative thinking mode’, without too much concern for efficiency issues. This could be
followed by testing and shaping the proposed procedural systems for efficiency to get a prac-
tical and robust approach. When they were confident that the new approach could work in
practice, they could implement it, perhaps initially in a prototype form in selected regions,
gradually developing efficiency as well as effectiveness via practice focused on continuous
improvement.
For example:
1 The current use of a very small number of big contractors might be convenient for
some NER staff, in particular, those responsible for letting the contracts. However, a
much larger number of contractors, some relatively small organisations, might allow
much better response rates when urgent unscheduled maintenance work was needed.
This approach might also help to keep cost rates more competitive for all scheduled
and unscheduled maintenance work. It might also increase opportunities for new and
creative responses, including new technologies. It would require more sophisticated
NER inspection and management approaches, and it would be more challenging for
all of Oscar’s operations staff (Sophie was looking at Oscar when she said this, but he
was smiling because he knew this comment was coming). She believed there was a
very strong case for a more sophisticated approach which provided both less safety and
security risk and lower overall cost, and she was determined to explore the possibilities
effectively.
2 The current budgeting constraints and operations constraints on work deemed urgent
by inspectors had to be eliminated, as soon as possible. Maintenance should be planned
and associated budgets controlled as far as possible, closely integrated with operational
needs. But a much enhanced and empowered safety and security inspection team who
were always fully informed about the state of the system and empowered to act imme-
diately without inappropriate financial constraints was a clear necessity in her view (she
was looking at Oscar and the Finance Director alternately when she said this, but Oscar
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
532 Employing planning tools in practice
had primed the Finance Director, and he was comfortable with the implications as he
understood them thus far).
3 The kind of safety and security inspection team Sophie had in mind would have to work
closely with the scheduled maintenance planners. But in Sophie’s view they also needed
to work closely with all their maintenance contractors and associated subcontractors, in
‘coaching mode’ as well as ‘critic or examiner mode’. Furthermore, they needed to act
as a conduit for feedback from staff working for NER contractors and their subcontrac-
tors on all relevant concerns. People working for contractors and subcontractors should
feel they were an important part of the ‘NER big team’, responding as good team
members in a carefully designed partnership relationship. If they could not be trusted
to act as trusted partners in a big team they should be sacked as quickly as possible. If
the NER staff managing them or designing the partnership arrangements could not be
trusted to do so effectively, they too should be sacked or transferred to other jobs as
soon as possible.
4 Having developed what Sophie and the NER ‘core team’ thought was the most effec-
tive contracting approach, more efficient regimes, in terms of reduced overall expected
cost and possible departures from expectations, could be tested. They could try to
ensure both fatality efficiency and appropriate trade- offs between the overall expected
cost, the expected level of fatalities, and the risk of very high levels of fatalities like an
incident portrayed by the S = 350 scenario.
A change in culture or ethos
Sophie was convinced that at a more general level, common to both physical and procedural
examples, NER needed a significant culture change led by the board. This culture change had to be coupled to and facilitated by the other safety and security planning changes she
was proposing. All the changes involved might be characterised as starting with a ‘fail- safe
culture’, both driving and driven by a ‘fail- safe planning’ approach. She made these points
as a starting position for a relatively ‘soft issues’ insertion in her otherwise predominantly
numerical model focused presentation.
She had a personal preference for the word ethos instead of culture, to signal the trust and underlying ethical issues involved, but she would stick to the more common ‘culture’ word
that day. They could debate terminology later.
They could discuss and change the labels and revise the features she was suggesting, but
what she had in mind in broad terms was building on the idea that the Westinghouse fail-
safe braking system for trains first introduced in the US many years ago was one of the big-
gest transformational steps forward in railway safety history. The basic idea behind ‘fail- safe’
brakes was a train would not be able to move if the brakes were not functioning properly.
She did not want to suggest NER should not be able to move if the NER approach to
safety and security was not functioning properly, but enlightened use of the fail- safe prin-
ciple might have significant useful implications which went well beyond the physical and
procedural systems and needed careful development.
Building on this starting point, ‘an enlightened culture’ for NER in terms of safety and secu-
rity might be encouraged if the overall strategy mandated by the board adopted policies like:
1 all changes to the system which demonstrated a reduction in the value of C were always
implemented as soon as possible unless there was a very clear case against for other cost or further reasons,
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 533
2 all changes to the system which demonstrated an increase in the value of C would always be
rejected unless there was an extremely clear case in favour of other cost or further reasons, 3 NER at all levels acknowledged that changes in C should always be an issue, 4 the search for reductions in C and improvements in other aspects of operations plan-
ning were always fully integrated, and
5 NER remained open to the value of any further ‘culture’ improvements which could be
engineered.
To achieve the implementation of these policies in terms of some illustrative example fea-
tures for engineering an enlightened culture:
1 All choices with safety and security implications would be assessed using an approach
based on the set of V1 and ES values developed for and approved by the board. These
parameter choices would drive safety and security related spending, not the other way
around, as soon as NER had a set of values for these parameters the board members believed were appropriate. Initially significant iteration should be anticipated, testing assumed values of RS and other ES values until the safety and security team and the board were
confident about the values used, with effective plans to manage the initial iterative
process in place before iteration started. In the longer term, further iteration would
be needed to revise the process as circumstances changed, new issues were identified,
and they ‘learned by experience’. A key aspect of learning by experience would be the
‘enlightened prudence’ role of RS, to be clarified shortly.
2 The benefit of the doubt when making choices would always favour the safer and more
secure approach, at the board level, at the level of a member of the safety and security
inspection team asking for urgent unscheduled maintenance, at the level of adding or
deleting approved contractors or changing the terms of contracts, at the level of manag-
ing staff culture and morale issues with safety implications, and at the level of decisions
by individuals like train drivers or controllers.
3 The small ‘core team’ of safety and security staff managed by Sophie would be ‘safety
planners’. They would be responsible for developing safety and security plans and involved in integrating those plans with all other plans. They would emphatically not be ‘safety managers’.
4 Responsibility for managing safety would rest with the ‘NER big team’ in terms of
all relevant operations decisions. So far as possible, an inclusive and flexible ‘NER big
team’ approach to safety planning and management would be adopted. Contractors’
staff and subcontractors’ staff, as well as NER staff, would be seen as contributors to
corporate knowledge relevant to safety and security planning and responsible for man-
aging their contribution to the system’s safety and security.
5 A team of safety and security inspectors would audit all safety and security related per-
formance, approaching this task with a supportive perspective, ‘helping everyone to
do better’ rather than ‘imposing rules which interfere with performance’. They would
focus on ‘enhancement’ instead of ‘compliance’, ‘enabling’ rather than ‘blocking’.
6 So far as possible, all the desirable features of the current approaches to safety and secu-
rity would be preserved. But all the undesirable features of the current ‘box- ticking’ and
‘back- side protection’ health and safety management culture would be designed out,
linking these mindset changes to toolset and skill set changes in some areas.
7 The wider culture change implications would focus on identifying ‘bad behaviours’ and
replacing them with ‘good behaviours’.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
534 Employing planning tools in practice
8 Staff morale and its role in a healthy ‘fail- safe culture’ would receive direct and focused
attention. Formal recognition that best estimates of PS values were always a function
of morale and forms of contract, as well as physical systems, would be central to this.
Eliminating resistance to acknowledging this dependence, driven by unenlightened
prejudices about the role of subjective probabilities, would need early attention.
9 Current NER staff who were unwilling to change quickly enough, especially those who
made it difficult for others to change, would be removed as quickly as possible, what-
ever their function and level in the management hierarchy.
10 NER could rely on Sophie to lead the planning of a culture change programme for
NER, but all NER staff had to work collaboratively across all corporate boundaries
as well as within their own areas. They would have to collectively ‘engineer’ a culture
change, seeking to creatively synthesise ends and means in the sense Sophie understood
was the driving culture of the engineering profession. She suggested there was a large
literature on culture which was relevant but little meaningful agreement on what corpo-
rate ‘culture’ was, never mind what to do about it. In her view, culture could be given a simple working definition like ‘the way we do things around here’. This was a per-
spective widely adopted and useful as a starting point. However, once NER started to
approach how to change culture, the issues would become very complex very quickly.
Changing culture takes all organisations firmly into the realm of values, leadership,
teamwork, trust, and many other relevant bodies of knowledge which are not easy to
disentangle. It also takes organisations into what Hyde (2006) referred to as ‘mysteries’
when addressing what makes people creative. In Sophie’s view the safety- first enlight-
ened culture which NER needed included a team spirit that was creative, supportive
of others, optimistic in a constructive sense without forgetting what to be frightened
about, professional in the vocation sense, ethically sound, and unambiguously intoler-
ant of a lack of integrity in others. There was no room for ‘faking concern’ or using a
public relations approach to selling a sense of values that was not genuine. Leadership
had to be ‘from the front’, but that might mean ‘from the side’ or ‘from the bottom’
as well as ‘from the top’, and all the leaders involved needed to be open to constructive
criticism and willing to learn from all relevant sources.
Sophie urged the board to be cautious about how quickly they could move with NER cul-
ture changes, but not at all cautious when estimating how wide and how crucially important
the implications of an approach driven by this kind of opportunity management might be.
Using risk efficiency as a basis for option choices
For the NER board to understand the use of risk efficiency incorporating fatality efficiency,
including the use of decision diagrams with underlying sensitivity diagrams and risk– reward
trade- off concerns, Sophie needed to explain some of the material in earlier chapters of this book.
This chapter simply assumes that she did so now, if not at an earlier point in the presentation.
Sophie had discussed the role of C as an expected value measure of the cost related to
fatalities including correlated injuries and other consequences associated with S = 1, . . . ,
m, transformed via TS in scenario terms into a monetary metric. This extended the reach of
the analysis beyond fatalities to other concerns which were correlated in her scenario frame-
work, but it was consistent with conventional safety practice thinking within a generalised
EP framework, in the sense that it limited its consideration of risk to the expected value of
C. That is, the safety management common practice employed by NER used an expected
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 535
value metric for risk based on a definition of risk of the form ‘risk = probability × impact’, with ‘impact’ measured in ‘value of avoided fatality’ terms calculated for each ‘risk’, and
summed over all risks.
What she had delayed considering until now was further generalising this approach to
an extended view of risk which fully embraced the implications of potential departures
from expectations associated with all relevant costs, the role of risk efficiency, the need to
address risk– reward trade- offs on a risk efficient frontier, plus uncertainty about all model-
ling assumptions which did not involve quantification – the ‘conditions’ any probabilistic
analysis rested on. She had often hinted at the need to do so, frequently mentioned enlight-
ened prudence and a prudence provision, and sometimes linked this to RS, but she had
not developed any of the underlying concepts. The reason for this delay was she believed
the illustrative numerical examples she had developed so far, plus the softer culture change
issues just explored, would help them to better understand the key aspects of the complexity
involved in what needs to be done terms.
Sophie now wanted the board to begin to understand why risk in terms of potential
departures from expectations always needed to be addressed when considering options involving all possible changes to the railway system with strategic implications.
This would involve thinking about risk in terms of a much broader perspective than
the common practice way of thinking about risk in a safety context. Although the much
broader view of risk embracing departures from expectations was directly comparable to
common practice approaches which had been used in security investment contexts for a very
long time, with a growing role in best practice in many other contexts, common practice
safety analysis had not yet adapted a direct and explicit formal planning role for risk metrics
concerned with departures from expectations. In her view this was a very serious and fun-
damental flaw. Risk efficiency and all the more sophisticated ideas based on risk efficiency
explored earlier in this book were simply ‘off the radar’ for common practice safety analysis
approaches, and changing this position was long overdue.
NER had to take an approach to safety and security which avoided this significant common
practice shortcoming. Doing so involved thinking about risk– reward trade- offs between
expected outcomes and risk in a risk efficiency framework, working towards making deci-
sions at the board level about which strategic changes should or should not be made in a
manner directly comparable to BP board use of decision diagrams, but extended to address
a generalisation of the enlightened prudence concept discussed earlier in this book.
To illustrate what was involved, Sophie suggested they start by associating Table 9.2 with
the annual ‘cost related to fatalities’ for the current system – a ‘no change’ situation she
would refer to as the ‘old’ system option.
An associated ‘new’ system option might be the best overall set of physical and procedural
changes for the old system that the safety and security team could devise, working with all
relevant NER employees, consultants, and contractors, developed after eliminating all domi-
nated alternatives, incorporating systems changes proposed for reasons other than safety by
all relevant NER players, incorporating current views about corporate capability- culture
issues. This new system might involve an extensive portfolio of changes, like new signal-
ling plus new communications equipment, coupled to new operations procedures and sup-
portive training programmes, as well as new contracting approaches. Alternatively, this new
system might just involve smaller trains run more frequently, any of the other approaches
suggested earlier on its own, or a very different new one.
Sophie then suggested they use Table 9.3 to compare this proposed new system option
with the old system option.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
536 Employing planning tools in practice
Table 9.3 Table of cost related to fatalities for two options, based on Table 9.2 for the old system.
Size scenario
Old and new system PS and percentage change
Expected cost ($m) if scenario S occurred
Contributions to old and new C ($m)
Cumulative old and new PS
S old PS new PS % CS old new old new
0 0.9500 0.96230 +1.3 0 0 0 0.9500 0.96230 1 0.0025 0.00225 −10 4 0.010 0.009 0.9525 0.96455 5 0.0250 0.02000 −20 30 0.750 0.600 0.9775 0.98455
20 0.0200 0.01400 −30 200 4.000 2.800 0.9975 0.99855 70 0.0020 0.00120 −40 1120 2.240 1.344 0.9995 0.99975
350 0.0005 0.00025 −50 7000 3.500 1.750 1.0000 1.00000 old & new C 10.500 6.503
The first and second column S values and PS values were taken directly from Table 9.2.
She would now explain the other columns from left to right, beginning with the third col-
umn ‘new system’ PS values.
Sophie’s discussion of the PS for the new system began with her explaining that she wanted
a numerical example involving just two options which portrayed the key messages the board
needed clarity about and was as easy to understand as possible. For simplicity, the old system
PS values for S = 1, . . . , 350 scenarios had been reduced by 10%, 20%, 30%, 40%, and 50%,
as shown in the third column. The rationale for her use of these percentages was simply
illustrating the effect of prioritising high S value PS reductions – the focus which she believed
was essential for NER. The associated percentage changes were noted in the fourth column.
The sum of these new PS values subtracted from unity defined the new system P0 value,
with an entry of 1 – 0.03770 = 0.96230 for the new system. This implied a 1.3% increase in
the probability of no incidents in any given year, as indicated. It also implied an associated
decrease in the value of (1 – P0) of 0.0500 – 0.03770 = 0.0123, a 25% decrease, arguably a
more useful way to look at the same change.
This 25% decrease in (1 – P0) was a weighted average of the assumed set of five PS reduc-
tions for S = 1, . . . , 350. It was smaller than a simple arithmetic mean of (10 + 20 + 30 + 40
+ 50)/5 = 30% because high S value PS were relatively small and the PS weighting mattered.
To start to explore some immediate implications, Sophie observed that in practice her
team would have to start with the CPS conditional probabilities for S = 1, . . . , 350, and compute the unconditional probabilities, proceeding as illustrated in Table 9.1. The more
technically minded directors might be interested in the implicit Table 9.3 CPS values as well
as the PS, and the implied CPS could be inferred from the PS and (1 − P0) values provided. To
demonstrate how this worked for those interested, the assumed conditional probability of an S = 350 scenario for the old system as indicated directly in Table 9.1 could be computed
using the unconditional P350 value of 0.0005 and (1 – P0) of 0.0500. In a sense this involved
working backwards to obtain CP350 = 0.0005/0.0500 = 0.01. Using an identical approach
for Table 9.3 and the new system, the conditional probability CP350 = 0.00025/0.0377 =
0.006631 was significantly lower than the 0.01 value for the old system, a 34% reduction. Using
this same approach again, the conditional probability CP1 = 0.00225/0.0377 = 0.059682
for the new system was slightly higher than the 0.05 for the old system, a 19% increase.
In other words, for the new system there was a 25% reduction in the probability of any
kind of incident, and if an incident did happen, it was slightly more likely to be an S = 1,
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 537
significantly less likely to be an S = 350, as might be expected given the unconditional prob-
ability changes.
The fifth column transformed fatalities metric CS value entries are the same as Table 9.2.
The sixth column old system C contributions were the same as those of Table 9.2, and the
seventh column new system C contributions were obtained via column products in the same
way. The old system C = $10.5 million was reduced to $6.5 million for the new system, a
$4 million or 38% decrease.
The final two columns cumulative probability entries just accumulated the second and
third column’s PS values for the old and new systems as a basis for Figure 9.1. Figure 9.1
used these cumulative probabilities to plot the basis for several operational ‘decision dia-
gram’ examples. She would consider these examples in terms of their implications for deci-
sion making as soon as she had explained how Figure 9.1 was related to Table 9.3, and what
Figure 9.1 implied.
The y- scale for Figure 9.1 measured cumulative probability, as for all earlier decision
diagrams in this book. The only difference was the use of a ‘squiggle’ below the cumulative
probability of 0.95 to allow a focus on the region that mattered, from 0.95 to 1.00.
The first of two x- scales measured fatalities on the scale 1, . . . , m, using S size scenario
values taken from the first column of Table 9.3. The second x- scale measured cost related
to fatalities. It used the size scenario values taken from the CS values in the fifth column of
Table 9.3.
Both the S scale and the CS based cost related to fatalities scale were approximately loga-
rithmic (base five). They were directly linked via the TS transformation.
The Figure 9.1 piecewise linear curves involved working assumptions linked to Table 9.3
which needed a high level of clarity for everyone involved, including the board. The board
needed to be well aware of the approximations involved, but entirely comfortable with the
robustness of what this kind of diagram was saying. Sophie wanted board members to begin
by considering the general statement made by Figure 9.1 in terms of the situation being
fatalities using size scenarios and the logarithmic S scale of Table 9.3
0.95
1.00cumulative
probability
0
expected cost outcomes for the two options indicated
(old system $10.5m, new system $6.5m)
new system
old system
0 1 5 20 70 350
cost related to fatalities using the logarithmic $m scale of Table 9.3
Figure 9.1 The basis for operational decision diagram examples provided by Table 9.3.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
538 Employing planning tools in practice
addressed, then the implications in terms of the old system option, and then the implications
in terms of the new system option.
Figure 9.1 and its relationship with Table 9.3 used a variant of the rectangular histogram
HAT based approach illustrated earlier, but this time the discrete value probability distribu-
tion for S being portrayed by a continuous variable distribution involved an approximation,
because S really was a discrete value variable. This could be slightly confusing at the bottom
end of the S value range if the approximation involved was not understood. The S = 0 value
was not a problem, but S = 1 involved a particularly awkward approximation in presenta-
tional terms in Figure 9.1. In discrete value terms, S = 1 interpreted in terms of discrete
numbers of fatalities had to involve a zero range, but in terms of a continuous variable HAT
approach the linear curve segment had to associate S = 1 with a uniform probability distribu-
tion range from 0 to 2 fatalities, with a conditional mean of 1. In discrete value terms, S = 5
was nominally associated with 2, 3, . . . 9 fatalities, but in terms of a continuous variable
HAT approach the linear curve segment associated S = 5 with the continuous variable range
from 2 to 10 fatalities. There were similar interpretations associated with the S = 20, 70 and
350 scenarios, to make the piecewise linear curve as a whole continuous.
Considering the implications of these observations in relation to the old system option, if
the S scale of Figure 9.1 was used to interpret the curve, the integer value S = 1 was associ-
ated with the midpoint class mark value of a continuous variable in the range 0– 2 fatalities,
so the approximation was particularly obvious in visual terms, but no errors were involved
in terms of expected value calculations or other aspects of the portrayals.
If the cost related to fatalities scale from $0 million to about $10 million is used to
interpret the 0– 2 fatalities segment of the curve, the presentational implications of the
approximation ceases to be a problem, because the IS transformation factor component of TS
actually operates in terms of ‘equivalent fatalities’, transforming S into a continuous variable,
and the uncertainty about all four TS transformation components also mean cost related to
fatalities is a continuous variable. In either case, the old system linear segment for this range
rising from a cumulative probability of 0.9500 (defined by P0) to 0.9525 (defined by P0 +
P1) provides a robust linear approximation to the underlying situation.
The integer value range for S = 2, . . . , 9 associated with the S = 5 scenario was portrayed
by a continuous variable range S = 2 to 10, with a corresponding CS range from $10 million
to $100 million. Whether or not a discrete variable interpretation of S was of interest, any
approximation involved was much less obvious and hardly worth a mention apart from the
need to maintain confidence in the validity of what the portrayal of Figure 9.1 was saying.
The old system linear segment for this range rises from a cumulative probability value of P0 +
P1 = 0.9525 to a value of P0 + P1 + P5 = 0.9775, a much steeper rise than the first segment.
The other piecewise linear segments were added in the same way, using the same HAT-
based approach employed earlier in a form which accommodates both a discrete value inter-
pretation for S or a continuous variable interpretation for cost related to fatalities.
The same working assumptions applied to the new system curve, and its construction
followed the same pattern.
Because a continuous variable approximation to a discrete fatalities scale is involved, if an
integer value interpretation of S is used, interpreting the two curves defined by Table 9.3 in
the S = 0 to 2 range is intuitively awkward using the S scale. However, in practice, because
injuries require consideration in equivalent fatalities terms and cost transformations are
essential, sticking to the $m monetary scale means interpretation becomes straightforward.
This bottom- end segment of the curve matters in presentational terms if the S metric is of
direct interest, accentuated by the expected value being located in this region, but it was
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 539
purely a visual presentation issue, avoided in practice by working with just the $m monetary
scale, the only x- scale used for the operational examples to follow.
Comparing the two curves in Figure 9.1 using either scale, the new system was clearly the
only risk efficient choice. Stochastic dominance was involved. The probability of exceeding
any given level of fatalities or cost was lower for the new system than for the old system, and
in cost terms the expected outcome was $4 million less if the new system option is taken.
On Figure 9.1 the $7,000 million segment of the new system curve almost overlaid the
old system curve because both cumulative curves were so close to a cumulative probability
of 1.0000. However, Table 9.3 made it clear that the new system probability of an outcome
value in this range was only half that of the old system, the differences for the P70 and P20
values were also important, and it was very important to understand this in order to appreci- ate the full picture portrayed by Figure 9.1 plus the underlying Table 9.3 values. Indeed, this
difference in the curves at the top end was a matter of substance and importance. It must
not be overshadowed by the bottom- end presentational issues, and it was part of the case for
preserving a table plus curve approach based on a HAT framework in operational practice. The more commonly used Monte Carlo simulation approach could provide a comparable
table but rarely did. This, in turn, was part of the broader case for ensuring that someone
involved in implementing practical approaches clearly understood what really mattered, and
what was relatively superficial. Sophie had risked boring the board with the bottom end of
the curve explanation to make sure they appreciated the approximations involved did not
matter because she wanted the board to be confident about all the expert advice they used in
terms of what really mattered and what did not matter enough to need ongoing attention. The expected cost values were plotted on the cumulative probability distribution curves,
with expected cost values also noted in the text, to help the board deal with the com-
bined effect of the piecewise linear histogram approximation and the logarithmic scale. The
expected value of $10.5 million for the old system was plotted on the cumulative probability
curve near the beginning of the second segment, while the $6.5 million value for the new
system was plotted on the first segment, as illustrated.
Smoother curves could be used in practice, and board members would probably prefer
smooth curves with good reason, but they needed to understand what underlay smooth
curves in terms of all key working assumption. Indeed, more scenarios would obviously
refine the precision in the definition of the probability distributions. But always basing inter-
pretation on cost related to fatalities with a non- linear smooth curve fitted to S = 0 and
the other five scenario values would probably be much more clarity efficient, and using m
scenarios with a very large m would definitely not be clarity efficient.
As soon as Sophie was sure that board members were happy with the relationship between
Figure 9.1 and Table 9.3, she explained that if the annual cost related to fatalities as measured
and portrayed both in expected value (C) and associated CS variability terms in Table 9.3 and
Figure 9.1 was their only concern, then the new system was clearly better than the old system.
However, to make decisions of strategic importance in any realistic operational context,
NER had to consider all cost, revenue, and other non- measurable issues which were rel- evant to making decision choices in a fully integrated manner. She now wanted to use Figure 9.1 as the basis for operational forms of decision diagrams with examples which
involved considering a fully comprehensive annual ‘cost relevant to system choices’, defined
as the sum of two items:
1 the ‘cost related to fatalities’ which they had just addressed and
2 the ‘cost not related to fatalities’.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
540 Employing planning tools in practice
‘Cost not related to fatalities’ was defined to include the amortised (annual) capital costs for
the new system (net of any scrap value associated with the old system) plus any associated differences in annual operating costs for both systems (net of any revenue benefits in the
case of the new system).
Instead of subtracting increased revenue associated with the new system from the new
system costs, in some cases it might be more convenient to treat increased revenue for the
new system as an opportunity cost for the old system or address profit directly. But this was
a technical detail they could address later.
For ‘new system A1’, the first of two ‘new system’ examples she would use, Sophie sug-
gested they assume that the new system’s ‘cost relevant to system choices’ which was cost
not related to fatalities was estimated at $50 million per annum, and the corresponding old
system ‘cost relevant to system choices’ which was cost not related to fatalities, was also
estimated at $50 million per annum, in both cases with too little uncertainty to matter.
This meant that Figure 9.1 could be transformed into Figure 9.2 by just changing what the
second x- scale measures to the cost relevant to system choices and adding $50 million to
the scale, as indicated. Dropping the S value x- scale was a convenient option which she had
adopted, so they could consider fatalities in the broader context of cost relevant to system
choices.
The x- scale $0m in Figure 9.1 had become $50 million in Figure 9.2, with other class
mark values also increasing by $50 million: $4 million became $54 million, $30 million
became $80 million, and so on.
Assuming that board members were all comfortable with an EP perspective on risk effi-
ciency, the Figure 9.2 decision diagram clearly indicates that:
1 for new system A1 the risk of any outcome greater than $50 million was significantly
less than it was for the old system, and
2 new system A1 had a $60.5 million – $56.5 million = $4 million expected cost advantage.
The expected cost of the old system was $60.5 million ($50 million + $10.5 million) in
comparison to $56.5 million for the new system ($50 million + $6.5 million), and the new
cost relevant to system choices using the logarithmic $m scale of Table 9.3 + $50m
0.95
1.00cumulative
probability
0
expected cost for the two options indicated (old system $60.5m, new system A1 $56.5m)
new system A1
old system
50 54 80 250 1170 7050
Figure 9.2 Decision diagram for the new system A1 option.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 541
system curve was higher over the whole $m x- scale range of both curves, as indicated in Fig-
ure 9.2. Stochastic dominance was involved. New system A1 was the only risk efficient choice. Assuming the analysis underlying Figure 9.2 was complete and unbiased, the new system
was clearly preferable to the old system in terms of both expected cost and risk.
Sophie ensured all board members were comfortable with this interpretation of Fig-
ure 9.2. She explained that some uncertainty associated with relevant system costs other
than costs associated with fatalities as considered in Table 9.3 would imply that the rise in
the cumulative probability curves to the 0.95000 and 0.96230 levels would be less than ver-
tical. However, this uncertainty could be associated with bearable risk which was just ‘noise’
relative to the hugely significant uncertainty and associated risk which Figure 9.2 captured.
For those board members who might be focused on what usually happens, she empha-
sised that it might seem important that there would be no observable outcome change 95%
of the time. That is, 95% of the time the old system would enjoy a $50 million ‘all relevant
systems costs’ outcome associated with no incidents, while the new system would involve
a $50 million cost associated with no incidents 96.23% of the time, a very modest looking
change, the +1.3% change noted in Table 9.3.
For those board members who might be particularly concerned about how her EP
approach related to common practice risk management based on expected values, she
emphasised that the $4 million expected value cost advantage would be picked up by a com-
mon practice approach if the same numbers were used, but the associated variability issues
would not be part of the analysis framework, because associated variability was not part of
the risk concept.
The numerical values underlying this illustrative example had been developed with a focus
on 1– 1 scenarios – collision incidents involving fatalities. If the limited 1– 1 incident nature
of Tables 9.1 through 9.3 was extended to cover all type– grade scenarios involving relevant
incidents, the sharp transition from mere noise to crucial distribution tails would be sof-
tened, but overall expected values and the very long probability distribution tails would still
be the central concerns, and most of the softening of the sharp changes in the curves around
the $50 million point would be in the $50 million to $60 million range. Changes in this
region might matter in terms of expected outcomes, but they would not matter very much
in terms of risk associated with departures from expectations.
Sophie emphasised that the NER current view of risk for safety and security decision-
making purposes was limited to expected outcomes, completely ignoring risk in terms of
departures from expectations, and this had to change immediately.
All board members now needed to begin to understand the implications as her examples
became more sophisticated, beginning with the next example.
For ‘new system A2’, the second of two examples she would use, Sophie suggested that
they assume that new system A2 ‘cost relevant to system choices’ which was ‘cost not related
to fatalities’ was estimated at $55 million per annum, and the comparable old- system cost
relevant to system choices which was cost not related to fatalities was still estimated at
$50 million per annum, in both cases with too little uncertainty to matter.
Then Figure 9.2 could be transformed into Figure 9.3.
The old system curves of Figures 9.2 and 9.3 were the same, with the same $50 million
added to the x- scale. The difference between the new system A1 and A2 curves of Fig-
ures 9.2 and 9.3 was of crucial importance, as were the implications.
The new system A2 curve of Figure 9.3 involved accommodating the same x- scale by
shifting the new system A2 curve to the right by $5 million, with a vertical segment at
$55 million rising from zero cumulative probability to 0.96230 as a result of the $5 million
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
542 Employing planning tools in practice
shift and the expected value becoming $61.5 million. There was a very noticeable shift in
the new system A2 curve on the left- hand side. But the logarithmic nature of the x- scale
meant that the visible differences in the relative positions of the curves rapidly diminished
as the curves moved towards the right- hand side, with no real change in the region of the
$7,050 million scenario.
What the Figure 9.3 decision diagram portrayed was 95% of the time the old system would
involve the $50 million cost outcome, and the new system A2 would be $5 million more
expensive, at $55 million. However, the risk of any outcome more than $55 million would be less for new system A2, with a massive 50% reduction in the chance of a $7,050 million
scenario, a 40% reduction in the chance of an $1,170 million scenario, and so on down to a
10% reduction in the chance of a $54 million scenario.
Furthermore, new system A2 would cost $5 million more 95% of the time, but the expected cost disadvantage of new system A2 was only $61.5 million – $60.5 million = $1 million.
In this new system example, new system A2 did not dominate the old system in risk-
efficiency terms because it increased the expected cost by $1 million. Risk– reward trade- offs
were involved.
But new system A2 did significantly reduce risk in the general EP sense which considers potential departures from expectations as well as expected outcomes. For a relatively small
$1 million increase in expected cost, there was a massive decrease in risk in the departures
from an expectations sense, including a 50% decrease in the S = 350 scenario context, and
risk in the departures from expectations sense was actually the aspect of risk that really mattered here, arguably of MUCH more importance than a $1 million difference in expected outcomes.
Furthermore, for anyone who understood risk efficiency and used it as a basic conceptual tool, restricting the interpretation of risk to expected outcomes in the context of this example was clearly seriously myopic. What was of crucial importance in practice was the risk– reward trade- offs involved when making choices, with risk FOCUSED on departures from expectations, not defined to EXCLUDE departures from expectations.
New system A2 delivered a massive reduction in risk for an expected annual cost increase
of only $1 million, and rail passengers might be very pleased to pay marginally increased
cost relevant to system choices using the logarithmic $m scale of Table 9.3 + $50m
0.95
1.00cumulative
probability
0
expected cost for the two options indicated
(old system $60.5m, new system A2 $61.5m)
new system A2
old system
50 54 80 250 1170 7050
Figure 9.3 Decision diagram for the new system A2 option.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 543
fares to enjoy this increase in safety, so even a $1 million decrease in reward might not be
necessary for NER. It was actually crucially important to understand that safety and rev- enue driven by ticket prices were not independent, and the interdependencies should NEVER be ignored.
Some uncertainty associated with relevant system cost components other than cost related
to fatalities would imply the rise in the cumulative probability curves to the 0.95000 and
0.96230 levels would be less than vertical. As in the Figure 9.2 context, this uncertainty
could be associated with bearable risk – mere noise relative to the uncertainty and associated
risk which Figure 9.3 captures. But in practice it could confuse decision makers if they were not very focused on what really mattered, so it was important somebody kept everybody focused on what really mattered.
Furthermore, in the Figure 9.3 context, even modest uncertainty of this kind poses the
question, ‘Could this kind of uncertainty flip the $1m expected cost disadvantage for the
new system into an advantage of $1m or more?’ There might be a significant chance that
new system A2 could have both a lower expected cost and less risk, an aspect of higher- order uncertainty not considered directly by Figure 9.3, but important to understand when
interpreting Figure 9.3 and more complex operational equivalents.
At this point Sophie believed that it was worth pointing out that the conditional nature
of these curves and associated uncertainty mattered a great deal. For example, the P0 and
CPS estimates might be biased, and even if they were not, associated variability uncertainty
might matter. Possibly even more important, the TS values might be too low or too high.
Uncertain parameters central to the analysis inferences involved a higher order of uncer-
tainty and risk which the decision diagrams could not capture, but which might really matter
in a context like this.
The $1m expected cost advantage of the old system was not a sure thing – and it was ambig- uous for an extremely complex set of reasons. In terms of a simple example, comparatively speaking the massive 50% reduction in S = 350 risk which was assumed might well be a
sure thing, even if the associated unconditional probabilities were uncertain, with the same
comment applying to the other high S value risk reductions. That is, NER might be able to
estimate the probable reduction in fatalities given an incident of known physical characteris-
tics occurred quite accurately relative to estimating the probability of that incident happen-
ing. Ambiguity was involved for both concerns, but not comparable ambiguity. The difference might matter a very great deal, the reason clarification of ambiguity can matter a great deal in practice.
In practice, the limited incident nature (1– 1 type– grade) of Tables 9.1 through 9.3 had
to be extended by NER to cover all relevant incidents. This meant that the issue of possible
errors and bias in the estimates had a direct bearing on the question, ‘Could uncertainty
about costs which did not involve risk in a direct sense flip a computed expected “cost rel-
evant to system choices” disadvantage for a new system into an advantage?’
Addressing this question together with the closely coupled issues associated with the new system risk reduction associated with high S scenarios took the discussion into enlightened prudence ter- ritory in a safety and security context, which was next on Sophie’s agenda.
To start to consider the remaining aspects of the conceptual framework needed to
explain enlightened prudence at a board level, Sophie posed a broad and fundamental
question in an initial form: ‘Assuming for the moment that the analysis underlying Fig-
ure 9.3 was complete and unbiased, but it incorporated sizable non- systematic errors, and
there were no options available other the new system A2 or the old system, what choice
should NER make?’
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
544 Employing planning tools in practice
She said the board should find it useful to use Figure 9.4 to visualise how they might start
to answer the initial form of this question and its generalisation.
You might begin to understand Figure 9.4 by observing that it is a variant of Figure 8.8
used in Chapter 8 and Figure 3.8 used in Chapter 3. The key modification is risk measured
on the x- axis is focused on risk related to fatalities, ignoring other corporate risk because it
was not immediately relevant to the NER board – it was ‘noise’, or it was associated with
other kinds of strategic choices beyond current concerns.
Sophie used Figure 9.4 to provide the NER board with a thorough briefing on the role
of risk efficiency and associated risk– reward trade- offs in the NER context, building on her
earlier introduction to these concepts as addressed earlier in this book. You might like to
recall the discussions in Chapters 3, 6 and 8, and assume that the NER board acquired a
comparable understanding to yours in BP, IBM, Astro, Canpower, NER, and related con-
texts as a result of this discussion, so we do not need to revisit earlier material.
Building on this broad overview understanding, Sophie then explained that currently
NER used an approach to safety based on minimising the expected cost associated with
fatalities using a ‘risk = probability × impact’ perspective. This perspective implied that NER
would currently choose the old system option, declining the new system A2 option because
it increased risk as currently measured in expected value terms by $1 million. This perspec-
tive implied that the old system choice portrayed in Figure 9.3, which minimised expected
cost, might be visualised as point ‘a’ in Figure 9.4, assuming that it was risk efficient and that
risk associated with cost related to fatalities was the concern that really mattered for NER in
terms of the current discussion. Prior to Sophie’s arrival, NER moving away from point ‘a’
towards points b1 or b2 would not be addressed for reasons associated with safety and security
risk in terms of NER reward which fails to meet
expectations because of risk related to fatalities
NER expected
reward in terms of expected profit incorporating the expected value C In the expected value of cost relevant to system choices
a–c efficient boundary target, with
a maximum expected reward approach
c minimum tolerable expected reward
bi intermediate expected reward approaches
ei inefficient approaches
c–d inappropriate reward levels
key:
a
b3
c
d
competent management area (usefully interpreted as an opportunity management area)
e1
non-feasible area
b2
b1
e2 e3
incompetent management area
Figure 9.4 NER expected reward and risk immediately relevant in an efficient- frontier portrayal.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 545
because formal assessment of safety was currently addressed in expected value terms without considering departures from expectations.
In current NER formal planning terms, risk associated with departures from expectations was entirely ‘off the radar’. This was a fundamental and wholly unacceptable failing which needed immediate resolution, a huge blind spot in NER planning. It was a blind spot which was not unique to NER, but that did not make it tolerable.
Assuming that new system A2 was also risk efficient, choosing new system A2 implied a
movement towards point b1, with a very modest decline in expected reward (a $1 million
increase in expected cost) but a very significant decrease in risk (measured in terms of the
relative likelihood of high $m- value size scenarios). Indeed, if train passengers were pre-
pared to pay higher fares for a much safer service, reward might actually increase. Even if
this opportunity to shift the boundary is ignored, assuming rail fare ticket prices could not
change, the efficient frontier curve in the region of point ‘a’ was very flat, and NER could
and should take advantage of this massive decrease in risk for a modest increase in expected cost to meet their ALARP concerns. To have simply ignored this trade- off in the past was a
symptom of a deeply flawed approach. There should be no argument about the need to consider it. The only debatable issues were how to do so, and how far to go when moving towards b1 and then b2 and perhaps b3.
To address the trade- offs involved the board might start by looking at the S = 350 sce-
nario as the $7,050 million scenario, perhaps usefully rounded to a $7,000 million scenario.
The old system had a 2,000 year return period, compared to the 4,000 year return period
for new system A2. That is, halving P350 for the new system S = 350 entry on Table 9.3
implied doubling the return period.
The board might observe that the S = 70 scenario, viewed as a $1,170 million scenario,
perhaps rounded to a $1,200 million scenario, needed joint consideration too. The S = 20
scenario, viewed as a $250 million scenario was also important, and even the S = 5 and 1
scenarios had related relevance.
But before they took this exploration too far, it would be crucially important to examine
some underlying assumptions about NER risk-taking capability as well as NER risk– reward preferences. ‘Risk-taking capability’ as a supplement to ‘risk appetite’ defined in terms of prefer- ences was crucially important.
As soon as the risk-taking capability issue was raised the board might observe that
NER would be bankrupted if it had to pay out a figure of the order of $7,000 million, or
$1,200 million, which immediately raised five key questions:
1 How much of the CS would NER be legally required to pay in cash terms?
2 Who would bear the CS costs in cash terms which NER was legally required to pay but
not capable of paying?
3 What parties would share the further implications, and to what extent?
4 What about insurance?
5 What about ALARP from the perspective of the regulator and the courts?
The answer to the first question was NER would have to pay the AS and IS components
associated with their option A approaches (full compensation costs for fatalities and injuries
set out by the courts or associated out- of- court settlements, plus associated fines), plus
actual outcomes associated with the DS component. NER would not have to pay the full
option B or C amounts associated with the AS and IS component amounts (assuming that
these option B or C figures were higher, which was a reasonable working assumption if
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
546 Employing planning tools in practice
their estimates were unbiased). If RS made an accurate provision for knock- on implications
of loss- of- reputation issues and all other associated effects not covered by the other com-
ponents, these opportunity costs and more direct cash costs would also have to be paid by
NER.
The answer to the second question was the passengers and on- board NER staff involved
in the incident, plus their dependents, would have to bear the consequences beyond the
compensation amounts actually paid. Shortfalls in compensation payments might well be
the least of the concerns of those who lost loved ones, but they would ‘add insult to injury’.
The government might share some of these losses if the government stepped in as an insurer
of last resort in terms of fatality and injury compensations which were awarded by the courts
but beyond NER resources. But this might not happen, and it was particularly unlikely given
the private sector status of NER.
The answer to the third question was passengers and on- board NER staff involved in
the incident, plus their dependents, would inevitably share the major portion of a failure to
pay full compensation for fatalities and injuries as measured by good practice or best prac-
tice shadow price value of an avoided fatality concepts. The government could in principle
share some of these losses if the government stepped in as an insurer of last resort in terms
of fatalities and injuries and did not limit itself to court- based decisions if shadow costs
were higher. But Sophie had never heard of any governments even thinking about doing
this. The government would also share the loss in terms of all the other costs involved in
rescuing a bankrupt railway which had a catastrophic incident to deal with, whether or not
the government saw this in ‘insurer of last resort’ terms. Many further parties might also
suffer knock- on effects, as for any bankruptcy, including all the shareholders and creditors
of NER. Some of the creditors may be small local businesses which become bankrupt as a
direct consequence.
The answer to the fourth question was complex. However, any commercial third party
insurer with access to Table 9.3 and Figure 9.3 would probably want a good deal more than
$1 million extra per annum in premiums to provide full cover for the old system instead of
the new system, unless NER was planning to go bankrupt if a major claim arose and full
compensation at shadow cost levels was unlikely to even be considered. At present NER
did not have full insurance cover and bankruptcy was probably inevitable. This presumably
implied that passengers and NER staff were taking a significant uninsured risk, as were other
parties, and the government as an insurer of last resort was a possibility which could not be
relied on.
The answer to the fifth question was more complex still, and perhaps even more conten-
tious. At a very high overview level, if the regulator acting on behalf of the passengers,
employees, and the government was given access to Table 9.3 and Figure 9.3, he or she
should argue that the ALARP principle required NER to opt for new system A2, even if the expected cost difference was MUCH more than the $1m assumed for Figure 9.3. Furthermore, if NER did not disclose this information, and it was revealed post- incident, arguably NER
would and should face a variant of the Ford Pinto controversy, intensified by a number of factors, with knock- on implications for the regulator and the government.
More generally, simply assuming that point ‘a’ in Figure 9.4 was acceptable without
explicit and careful consideration of the risk of possible very high S and $m value scenarios
was not acceptable from an EP perspective – but this could be inferred to be the current
NER position, without its regulator or the government being aware of this situation. The current NER revealed preference was a definition of risk which did not address any of the risk efficiency concerns associated with Figures 9.1 through 9.4, and the current situation
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 547
seemed to involve their regulator and the government not recognising that this approach
posed any serious problems.
Current NER practice meant they were not explicitly thinking through the implications
of high S value scenarios arising in a risk efficiency framework. Sophie believed that the NER board members needed to fully understand the implications and make decisions accordingly, beginning that day, with a view to eventually sharing their new insights with all other interested parties.
A key limitation of the current NER framework was that NER did not address risk– reward
trade- offs on an efficient frontier in the Figure 9.4 sense. As a direct consequence, NER
failed to address the trade- offs between risk measured in terms of expected outcomes and
risk portrayed in terms of possible downside departures from expectations in the way she
had illustrated with Figure 9.3.
Sophie indicated that NER was by no means the only organisation doing this, but in her
view the very widespread nature of this common practice was no longer a viable defence for
NER or any other comparable organisation. She believed she had a professional duty to act
on this view, and she would be doing so, whether or not she continued with NER. That is, any organisation which limits its definition of risk to ‘impact × probability’ was
defining risk in ‘expected outcome value’ terms. If this was done, what she would call
enlightened prudence was always ignored by the formal planning framework because the
enlightened prudence concept along with its risk– reward trade- offs based on a risk efficiency
perspective were outside the conceptual framework being used. Those organisations which
did not use risk efficiency and enlightened prudence concepts were using a crucial ‘stealth
assumption’ – that ‘enlightened prudence did not matter’, nor did the difference between
enlightened prudence and enlightened gambles or enlightened caution. What mattered a great deal was beyond the framing assumptions being used.
It was vital for the NER board to understand that day why the search for a point like b1 was important, why b2 or b3 might be more appropriate, why all relevant risk needed to be
addressed, and why full consideration of risk generated by the significant uncertainty associ-
ated with high S and $m size scenarios might lead them to a very different overall corporate strategy.
Her view was it would not be helpful to try to take formal analysis of these issues further
that day in terms of more sophisticated numerical examples and decision choices. They
needed feasible real options requiring actual decisions, with realistic estimates which sepa-
rated out the components of her transformed values via the TS so that the board mem-
bers could understand the complex nature of the expected values and the associated risk of
downside departures in a specific real context. They also needed to deal with uncertainty
and associated risk which she had not attempted to embed in Figures 9.3 and 9.4 which
might also matter.
However, she would now try to clarify enlightened prudence in relation to enlightened
gambles when ‘clarified trade- offs’ were involved, in addition to explaining ‘clarified domi-
nance’ including enlightened caution, using simple variants of her Figure 9.3 example. She
would also link these concepts to the ‘fail- safe culture’ which she believed NER needed.
That is, she would look at ‘what has to be done’ in overview outline terms at a strategic
level, without getting into the tactical level details of ‘what has to be done’, never mind
the how to do it concerns, further drawing on our Chapters 3 and 6 discussion in an NER
context.
Again, the next section focuses on the new issues for you, avoiding repetition of earlier
material not essential for the new context.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
548 Employing planning tools in practice
Enlightened prudence and the board’s leadership role
Sophie suggested the board reconsider Figure 9.3 in terms of a series of adjustments to the
location of the new system curve, initially moving it to the left. These shifts might be associ-
ated with further revisions of the new system A2, designated A3, A4 and so on.
If new system A3’s ‘cost relevant to system choices’ which was ‘cost not related to fatali-
ties’ was $53 million, $2 million less than assumed for Figure 9.3, new system A3 would
be more expensive 95% of the time, but $1 million cheaper on average. New system A3
would be the only risk efficient choice. In this case choosing the new system would involve
an enlightened caution form of ‘clarified dominance’, in the sense demonstrated by BP’s
use of decision diagrams as discussed in Chapter 3 and Astro’s use of decision diagrams as
discussed in Chapter 6.
If the new system’s cost relevant to system choices, which was cost not related to fatalities
was less than $50 million, enlightened caution was not needed because the new system’s
cumulative probability curve was entirely to the left of the old system curve. Stochastic
dominance made the new system the only obvious choice – clarified dominance in the sense
discussed in terms of Astro’s use of decision diagrams in Chapter 6. The decision diagram
made the stochastic dominance involved clear and easy to explain so that enlightened cau-
tion was not needed, and sophisticated non- linear cumulative distribution curves were not
needed. They were beyond the point illustrated by Figure 9.2 in terms of the new system
curve moving to the left.
In summary, so long as the new system’s cost relevant to system choices which was cost
not related to fatalities was less than $54 million, the new system was the only risk efficient
choice. Enlightened caution could be associated with choices involving a new system in the
$50– $54 million cost not related to fatalities range because most of the time the new system
would cost a small amount more but the expected cost was lower. Below $50 million, the
new system was always cheaper, which was obvious in terms of clarified dominance without
complex curves being required.
Sophie then asked the board to consider moving the new system curve to the right to
define new system A4. Its position in Figure 9.3 associated with A2 had already taken it
through the $54 million flip point, so further movement to the right meant that the $1 mil-
lion expected cost relevant to system choices disadvantage would further increase. However,
the risk reduction obtained for the expected $1 million cost relevant to system choices dis-
advantage was so huge it should be clear that an enlightened gamble in the sense discussed
in relation to IBM in Chapter 3 and Astro in Chapter 6 was not involved if the old system
was chosen. Making this expected cost disadvantage for new system A4 rise to $2 million or
$5 million or even $10 million was not going to change this. Enlightened prudence, in the
sense developed in Chapters 3 and 6, was involved if the NER preference remained the new
system as the cost relevant to system choices disadvantage increased over the range $0 mil-
lion to $5 million or $10 million and beyond.
The very long probability distribution tails involved implied that the risk in terms of
departures from expectations was a very serious concern, and the importance of this concern
made the new system the preferred choice even if the expected cost was significantly higher.
The enlightened prudence this involves was based on the very rational preference for a small
expected cost increase to achieve a massive risk decrease. The 50% reduction in the risk of a
$7,000 million scenario was particularly attractive, but the smaller reductions in the risk of
smaller $m value scenarios were also important.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 549
In classic insurance terminology, in return for a very nominal insurance premium, a very
big risk was avoided – comparable to a private individual insuring his or her house against
fire and comparable risks for a premium larger than the expected cost per annum of fire
damage to the house. The new system’s failure to completely eliminate catastrophe risk was
comparable to house insurance which covered fire or flood but not riots or warfare.
If new system A4’s cost relevant to system choices, which was cost not related to fatali-
ties was $150 million, it might be rational to argue that an additional $100 million was too
much to add to the expected cost. This would certainly be a credible position if passengers
were explicitly asked on a well- informed basis if they would pay for the extra $100 million
through ticket price increases and a significant majority declined. Then NER would have a
variant of the enlightened gamble situation discussed in Chapters 3 and 6 contexts, which
Sophie explained. Certainly, an extra $1000 million suggested this. But the flip point was
not $0 difference in the expected costs, and it was probably much more than $10 million
given the risk differences portrayed by Table 9.3 and Figure 9.3.
Sophie then asked a question: ‘Say the board had to choose between the old system and
new system A4, the additional expected cost was significantly in excess of $10 million, and
the board believed the expected additional cost was worthwhile. Was there anything else the
board should do before deciding?’
She then suggested the EP response was ‘the board should mandate Sophie’s safety and
security team to study a pattern of RS adjustments over the S = 1 to 350 range which would
allow the expected cost C to increase for further reductions in risk by moving from new
system A4 to new system B1, B2, B3, and so on, increasing the relative importance of RS
via “prudence provisions”, because it was important that NER explore the opportunity for
doing better than new system A4 using an operational approach to “enlightened prudence”
which she would now just outline in overview terms’.
In brief, what she had in mind was starting the development of a further full analysis of
a new system B1 using a starting position set of TS values comparable to her examples as
discussed so far. She would have her team increase the RS component for S = 1 to 350 so
that it was bigger by a systematic set of increments, adding a prudence provision to RS which
would lead to new systems B1, B2, B3, and so on. What she would be doing is looking for
points like b1, b2 and b3 on the risk efficient boundary portrayed by Figure 9.4, assuming
that the only operational way to do this was finding an optimal profile of prudence provi-
sions to embed in RS for S = 1 to 350 for points b1, b2, b3, and so on.
It should be reasonably clear from the earlier discussion that if NER attached additional
shadow cost weight to higher S value outcomes using a larger prudence provision in RS in a
systematic manner, expected cost increases should result in decreases in risk in the EP sense
associated with Figure 9.4. It should also be reasonably clear that ultimately what really
mattered was the complete set of RS values over S = 1 to 350 given everything else, and the
framing assumptions being used allow the RS values to depend on all relevant considera-
tions. Furthermore, this approach was also required if new system A1, A2, or A3 was the
alternative to the old system, and the old system needed evaluation for comparative pur-
poses using the same set of RS values.
Sophie suggested that she and her team would have to begin using extensive sensitivity
analysis associated with a systematic identification of points b1, b2, b3, and so on for a real
context, then share what was learned with the board, and then help the board to choose a
target RS profile for future analysis. If this was comparable to point b2, it could be associated
with plausible minimum and maximum prudence provision bounds. Future analysis of real
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
550 Employing planning tools in practice
system decisions could use this target profile and associated bounds, gradually refining the
approach as experience was gained.
The crucial issues which the board needed to confront when deciding how big the RS
should be for S = 1 to 350 included an ethical approach to the interests of the other parties
involved in a way which reflected the validity of the case made for the new and old systems
in quantitative terms, bearing in mind the limitations of the numbers used. These limitations
obviously included NER competence in estimating probabilities like the PS. But they also
included NER competence in estimating parameters like the TS, arguably MUCH MORE demanding because of the inherent ambiguity and subjectivity.
For example, as NER Operations Director, Oscar had to judge the extent to which the
estimates used to make decisions were fair and frank views of all the issues involved, with
any suspected bias on the optimistic side requiring a pessimist adjustment to the appropri-
ate boundary between enlightened prudence and enlightened gambles. If other directors
did not have Oscar’s grasp of the details, they still needed to satisfy themselves that sound
decisions were being made in a competent manner in terms of a sufficient understanding of
what really mattered.
In Sophie’s view, the board’s leadership of an enlightened culture which included a ‘fail-
safe culture’ transformation had to start with a robust choice for V1. This had to be followed
by a robust and ethically sound set of choices for the ES values conditioned by the V1 choice.
This had to include a prudence provision contribution built into the ‘residual provision’
associated with RS. In addition, this had to be followed by a competent and ethically sound
choice for the boundary between enlightened prudence and enlightened gambles in the
context of decision diagrams and full background information for specific real cases, under-
standing the relationship between a prudence provision built into the residual provision and
the boundary between enlightened prudence and enlightened gambles addressed in terms
of decision diagrams plus underlying sensitivity diagrams.
An aspirational target for some directors might be to eventually fully embed the needed
enlightened prudence into the TS transformation as part of the prudence provision RS com-
ponent so that there was no need to keep judging a suitable $m value boundary between
enlightened prudence and enlightened gambles using decision diagrams. But fully building
all these ethically based concerns into the quantitative analysis was not an immediate pros-
pect in her view. For the time being, they should limit themselves to RS residual provisions
including a prudence provision which the board saw as a clearly required target over and
above current industry norms, adjusting this target prudence provision RS if it remained
clear that all the decisions they were making required a further implicit modification to
reflect the preferred flip point from enlightened prudence to enlightened gambles. An RS
which was clearly ‘too small’ or ‘too big’ would damage the credibility of the approach in
her view, and they did not want to rush towards simplifications which obscured the signifi-
cant and difficult complexities they had to address. Her current perception was NER should
make a significant movement towards a position where RS embodied a board approved
prudence provision which reflected NER willingness to reduce the probability of high S
value size scenarios in any current NER context. But they should commit to continuing to
test all key decisions for enlightened prudence to enlightened gamble flip points via decision
diagrams plus supporting sensitivity analysis, communicating board preferences to all other
relevant parties in terms of specific key choices with a comprehensive explanation of their
overall rationale as well as provisions embedded in RS and the other TS assumptions. And
they should anticipate RS changes over time for a range of reasons.
All other NER staff could build on this board led initiative, but in Sophie’s view, building
a new enlightened culture had to start at the top. In her view any attempt to minimise the
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 551
ethical issues by the board would compromise the NER culture as a whole. Sophie did not
say so for obvious reasons, but she saw the board’s view of this as a key determinant of her
wish to stay with NER or not. She took the view that the necessary culture change was not
going to happen if the board members could not be convinced they needed to lead from
the front on this issue.
The initial operational value of the prudence provision component of the residual provi-
sion defined by RS was a target value – an initial best guess which needed to be revised via
extensive consideration of enlightened prudence based on the use of decision diagrams plus
sensitivity analysis to test the implications. Testing the implications could include compar-
ing the decision diagram implications of different levels of prudence provision via RS in real
contexts.
The use of decision diagrams like Figure 9.3 with extensive sensitivity analysis backup
including clarification of actual cash costs as well as shadow costs could then be seen as a
robustness test, a confirmation that the NER point of transition from enlightened prudence
to enlightened gambles was ethically sound. All interested parties should see the basis of the
decision- making process being used before as well as after catastrophes occur. Alternative ways of presenting choices which Sophie thought most directors might find
informative included reducing the AS and IS components of TS to the minimum plausible
‘option A’ values and adding what was taken out back to RS in various ways so that cost
relevant to system choices was actual cash costs in terms of the first three component fac-
tors. This might provide a useful basis for alternative good and best practice provision
testing, using sensitivity analysis in a way which would help them to understand what was
involved.
Despite the complexity the board would have to engage with, if board members had
decision diagrams for real choices they had to make with carefully designed example TS val-
ues and appropriate underlying sensitivity diagrams to clarify what was involved, the board
could begin to make appropriate choices in a practical context. They could start to develop
a feel for ethical and defensible decisions if RS took on various possible values given suitable
working assumptions for the other TS component values, gradually working towards values
of RS they would generally be comfortable with.
This implied a set of real decisions with trade- offs requiring extended discussions looking
at complete CS and FS distributions as well as C and F values under a range of assumptions
involving debatable parameters. In the early stages this would be a deep learning process
for everyone involved, a non- trivial investment in corporate learning. But once they had
acquired the mindsets and skill sets needed, decision making could become efficient as well
as effective, with multiple order- of- magnitude reductions in the effort expended.
As a roughly comparable example, in the late 1970s BP spent about a year working on a
prototype SCERT methodology design study, then six months on their first SCERT analy-
sis, but within a year of making this approach mandatory worldwide, comparable analysis
was taking about six weeks. It was worth bearing in mind BP invested this effort to reduce
overall cost and increase overall profit via risk efficiency – to manage opportunity, and they
were convinced they made a very large rate of return on this investment. Some aspects of
NER’s issues were much more complex, and in some respects much more new transforma-
tive learning was involved, implying an even bigger need to progress slowly and carefully
with a focus on effectiveness first, later seeking massive efficiency improvements once they
understood what really mattered. But the pay- off for NER was not just a better under-
standing of its safety and security concerns. It was an improvement in very broadly defined
opportunity efficiency terms. It ought to deliver a wide range of improvements in corporate
performance, not just improvements associated with safety and security.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
552 Employing planning tools in practice
Sensitivity diagram portrayal of option costs
Sophie suggested that whenever the board wanted to choose between two options, like no
new system or a new generation signalling system coupled to other fully integrated changes,
it would be crucial to understand the roles of expected values and associated variability for
all key components of cost relevant to system choices plus underlying cost related to fatali-
ties components. She indicated that sensitivity diagrams were the best way to portray this
kind of information which she was aware of, and she believed that NER would find sensitiv-
ity diagrams indispensable aspects of sensitivity analysis exercises at the board level, as well
as essential for her team plus all other NER staff involved in related bottom- up analysis and
decision making. She did not have a developed example directly geared to their safety and
security concerns, but she could show them a project planning example used by BP drawing
on Chapman and Ward (2011), and she could explain how NER safety and security con-
cerns could be addressed in this framework. She then showed the board Figure 9.5.
Sophie started by explaining how BP used this kind of sensitivity diagram as discussed
in Chapter 4 using Figure 4.1, the basis of Figure 9.5. You might recall this explanation
Mar NovDec Jan Feb Apr May Jun Jul Aug Sep Oct
probability jacket fabrication will be complete by the dates indicated (probability of cost within value indicated)
0.2
0.4
0.6
0.8
1
0
0.3
0.5
0.7
0.9
0.1
base plan completion date (target for cost relevant to systems choices)
1 2
3 4
5
6
probability curves show the cumulative effect of the following sources of uncertainty:
dates with relevant years indicated (cost [$m] relevant to system decisions)
1. yard not available, or mobilisation delays (annualised capital cost of new system)
2. construction problems/adverse weather (adjusted operating costs of new system)
3. subcontracted nodes delivery delays (physical system damage cost)
notes: 1. the curves assume a minimum fabrication period of 20 months (current legal climate) 2. no work is transferred offsite to improve progress (trial residual provision approach) 3. no major fire, explosion, or other damage (no incident sources not currently assessed)
6. delayed award of fabrication contract (residual provision – decomposed on
a lower-level sensitivity diagram)
4. material delivery delays (compensation for fatalities)
5. industrial disputes (compensation for injuries)
Figure 9.5 Sensitivity diagram: North Sea oil project platform jacket fabrication example (with NER new system reinterpretations in brackets assuming no curve shifts).
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 553
looking at Figure 9.5, perhaps revisiting Chapter 4 if you do not remember the basic ideas.
The use of sensitivity diagrams was also discussed in Chapter 6 in relation to Figure 6.11,
and again in Chapter 7 in relation to Figure 7.5, which you might revisit later for a fuller
understanding if necessary. This kind of sensitivity diagram was first used by BP in the
1970s, and in the following decades it was widely adopted by many other organisations for
many other purposes. In the NER context of current interest Sophie suggested the board
might re- interpret it as indicated using brackets, and she would take them through this now.
Beginning with the Figure 9.5 title in brackets – the NER ‘new system’ reinterpretations
in brackets assumed no curve shifts to keep this example explanation simple, but in practice
the curve shapes would change, radically in some cases. She would comment on this issue as
she discussed each curve, after considering the axes.
Reinterpreting the x- scale as ‘cost ($m) relevant to system decisions’ meant that curve 6
in Figure 9.5 might be interpreted as the new system curve in Figure 9.3 in a smooth curve
form, with the gaps between the six curves reinterpreted as a six component breakdown
which used the minimum plausible option A approach to AS and IS, embedding the uplifts
to option B good practice plus the option C maximum plausible best practice approaches
in RS. This might imply three versions of Figure 9.5 being required for considering three
alternative approaches to RS: a zero RS version considering only cash flow costs, a mid- range
good practice version, and a ‘maximum plausible NER best practice’ version based on care-
fully identified working assumptions about uplifts added to the good practice version. The
key working assumptions involved would be specified in ‘note 2 – trial residual provision
approach’. The other two notes were self- explanatory. The board would presumably be
seeking agreement to a position between the second two and might also want trial versions
of working hypotheses about this.
The ‘base plan completion date’ might be reinterpreted as the ‘target for cost relevant to
systems choices’ (as indicated in brackets) for the new system, with a cumulative probability
of being within any cost value on the x- scale being indicated on the y- axis. The ‘target’
used might be a stretch target equated to the expected cost of the old system which the
new system would replace. If this were the case, a chance of achieving it of about 15% for
the assumed ‘maximum plausible best practice’ residual provision choice might be deemed
reasonable in a context like Figure 9.3, with some expected cost increase being tolerated for
a massive reduction in risk.
Each of the uncertainty sources might then be reinterpreted as indicated in brackets, and
Sophie explained them as follows.
Source 1 plotted on the revised NER version of Figure 9.5 might be the annualised
(amortised) capital cost of the new system per annum, net of any relevant scrap value adjust-
ments associated with the old system. If relevant NER estimates were used, curve 1 would
leave the x- axis at a plausible minimum value defining the left- hand end of the x- scale, and it
might rise with the same overall ‘S’ shape as the curve shown in Figure 9.5 but a significantly
steeper climb rate throughout.
Source 2 might be the annual operating cost of the new system adjusted for increases in
revenue. Again, if relevant NER estimates were used, curve 2 would leave the x- axis at a
plausible minimum for the first two components, it might rise reasonably parallel to curve
1 if little anticipated additional uncertainty was added, and it might have the same shape as
the curve shown in Figure 9.5 but a steeper rate of climb throughout.
Source 3 might be the cost of physical damage to the system in terms of damaged trains and
tracks as a consequence of all incidents relevant to the system changes being considered, plus
direct knock- on operating costs and an opportunity cost view of lost revenues. Curve 3 might
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
554 Employing planning tools in practice
overlay curve 2 near the bottom and over the mid- range, with a significant chance of no cost
in this range, and then move sharply to the right close to the top, with a long tail reflecting
the very low probability of a catastrophic incident. Expected values plotted on all the curves
(as in Figures 6.11 and 9.1– 9.3) would probably be very useful to help to clarify how both the
cumulative expected values and the ranges changed as more sources were considered.
Source 4 might be the cost of compensation for fatalities, including legal fees, fines,
and payments to relatives in out- of- court settlements, as well as court- imposed settlements,
separating out the components of the cost of an avoided fatality which would have to be
actually paid by NER in cash terms from the good practice uplift incorporated in the residual
provision of source 6. Curve 4 might overlay curve 3 for its lower and mid- range portion,
plus its early dramatic movement to the right. But it would then move farther to the right
very sharply to a massive extent, with an overall increase in importance which was signifi-
cantly bigger than that shown in Figure 9.5.
Source 5 might be the cost of compensation for injuries. Like source 4, this might include
legal fees, fines, and payments to relatives as well as payments to the injured parties them-
selves, separating this from good and best practice provisions incorporated in the residual
provision of source 6. The overall shape might be similar to curve 4, almost overlying curve
4 over the lower portions, then moving off dramatically to the right. As for source 4, the
relative increase in expected value and variability might be huge, very much bigger than that
associated with Figure 9.5.
The shapes of curves 3, 4 and 5 would all be very different from that shown, with a sig-
nificant bulge out to the right near the tops of the curves to reflect serious to catastrophic
incidents.
Source 6 might be the residual provision, initially using the three alternative approaches
already noted: a cash flow costs only minimum plausible approach (curves 5 and 6 are the
same), a good practice approach (shifting curve 6 to the right), and a NER maximum
plausible best practice approach (further shifting curve 6 to the right). A breakdown of the
components of this sixth component could obviously be provided using the same Figure 9.5
sensitivity diagram framework, to clarify the relative importance and absolute size of all the
key components of RS, further discussed in a moment. Later tentative board choices which
interpolated between the initial bounds could be explored in the same framework.
Separating source 6 from sources 4 and 5 in this way in effect separated the actual cash
flow costs from the other revealed preference contributions to sources 4 and 5 in a way the
board would probably find informative. All the relevant actual cost estimates could be por- trayed using five components via sources 1 to 5 using this approach. Sophie was convinced
that the board would want to see these components separated.
It would also be helpful to see good practice residual provisions, maximum plausible best
practice residual provisions, and later proposed intermediate positions compared, in what-
ever structure the board preferred. This might include perhaps putting the good practice
provisions back in AS and IS components and experimenting with associated best practice
provisions in AS and IS, as well as a separate focus on enlightened prudence in RS.
The overall structuring being suggested was slightly different from the structuring used to
arrive at the TS values at the outset of the analysis. However, it used the presentation options
A, B, and C identified earlier, and NER was free to use whatever structure was most appro-
priate to the case being made and the story being told. This might change as the analysis
progressed and NER learned what really mattered.
At a much later date, regulators, courts, and passenger representative groups might also
find this kind of decomposition useful, with structures tailored to their interests.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 555
Curves 1 to 5 as well as 6 could be decomposed in a nested structure with as many
levels as the board or other users wanted, built from the bottom up by the safety and secu-
rity team, explained from the top down to the level of decomposition currently available
according to board concerns, or the concerns of other parties. Further diagrams portray-
ing the composition of curve 6 would clearly be needed, but because any of the other
curves would have to be built from the bottom up by her team, they could be explained
from the top down using the initial structure or alternatives which the board or other
users preferred.
As in their use in the contexts of Chapters 4, 6 and 7, the safety and security team could
use nested sensitivity curves to develop their understanding of all relevant decisions, includ-
ing process decisions about the need to do more analysis before making recommendations
to the board. They could then use final forms to explain their recommendations after feed-
back on proposed formats.
In some cases which might arise in practice, the curve 2 sources for the new system might
involve significant revenue benefits as well as reduced operating costs relative to the old
system. This might be of special interest, perhaps actually driving the decision to change to
the new system. Extensive decomposition of revenue benefits, as well as cost savings, might
be of great interest to the board, and it might require careful attention to the effective
portrayal.
BP’s concerns in a North Sea project context were clearly very different from those NER
had to address, but the opportunity efficiency approach Sophie was advocating shared more
than just the use of decision diagrams plus underlying sensitivity diagrams, and its use in a
wide range of contexts beyond BP was very extensive. One important common aspect was
the extensive use of a nested analysis structure. The NER safety and security team could use
this toolset with other NER staff to seek opportunity efficiency while building a bottom- up
understanding of complex interdependent uncertainties with appropriate specific and gen-
eral responses which the board could view from the top down when approving important
recommendations in the NER gateway processes. Their goal would be opportunity effi-
ciency in very broad terms, not just a safer and more secure railway system.
A process for clarifying NER revealed preferences
Sophie indicated that the TS values adopted would define the basis of NER revealed pref-
erences for decision- making purposes, with the RS playing a pivotal role. At present NER
implicitly set the RS to zero and did not acknowledge the role of an S dependent TS con-
cept, but that was only part of the differences between current practice and required new
approaches.
She emphasised that it would not be easy to develop and agree on initial RS and other
TS operational values, because they would have to define NER revealed preferences for
decision- making purposes in a way which reflected an ethical approach to interpreting the
revealed preferences of their passengers and staff plus their dependents, plus other parties
like their shareholders. Views about all these issues might not be shared by different groups,
and opinions could vary significantly within groups.
She suggested trade- offs like those involved in RS, if they assumed that treatment of
enlightened prudence ought to be as fully embedded as possible, would be more difficult
than those associated with AS and IS, and even DS would not be straightforward. However,
they could leave some of the particularly difficult issues until they needed to confront them
in particular cases, dealing with all the issues involved on a case- by- case basis.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
556 Employing planning tools in practice
To address all these issues they could develop a process for clarifying NER revealed prefer-
ences. This process would be designed to make it as easy as possible to use the TS approach
in practice. She thought the following steps might serve as an outline of this process:
1 Safety and security staff could undertake background research on V1 values, related
potentially relevant V350 values, and TS equivalent values used by other organisations,
especially those used by relevant road traffic authorities, air safety authorities, and rail-
way safety authorities. They would need to pay special attention to the explicit and
implicit assumptions used by these other organisations which were relevant to the set of
V1 and ES to be adopted by NER.
2 Safety and security staff could undertake related research to clarify the relationship
between numbers of fatalities in railway incidents, correlated numbers of people injured
in various categories, and the $m costs to be embedded in the DS associated with related
physical damage to facilities plus knock- on cost and lost revenue implications.
3 Safety and security staff could propose initial sets of V1 and ES values for NER based
on the two sets of research outlined in the first two steps and both ‘good practice’ and
‘maximum plausible best practice’ proposed bounds, as well as an estimated ‘cash costs
only’ lower bound.
4 Safety and security staff and other relevant NER staff could build on the first three steps
using an extensive exercise applying these values tentatively to a range of relevant real deci-
sions, picking a rich mix of decision types, including any currently contentious choices.
5 An extensive set of tentative decision recommendations and associated assumptions
from step 4 could be used for board level discussions about appropriate modifications
to the sets of V1 and ES values suggested by the safety and security team when the
budget, F and C implications became clear. The budget, F, and C implications would
all be driven by the V1 and ES choices, along with the probability estimates plus NER
creativity in terms of the best possible ways to approach all planning.
6 Extensive and open debate at the board level about NER creativity in terms of the best
way to approach safety and security planning, and its integration with all other plan-
ning, could then consolidate what was learned from the first five steps.
7 Formal implementation could begin, initially using the tentative approaches developed
in the first six steps to make definitive specific decisions, with constant review to develop
better approaches as experience was gained.
8 Using these initial definitive decisions in particular contexts to start to explore NER
revealed preferences in more general terms could be a further level of purely internal
NER learning.
9 When NER had developed a feel for what was involved, which was understood and
trusted at an appropriate level by all the NER staff involved, a process of opening the
discussion to the regulator and passenger representatives could begin. The goal would
be a completely open relationship with a view to establishing and maintaining trust,
even if a serious incident occurred.
The core revealed preference issues for NER
Sophie suggested that what the board ultimately agreed about the operational form of NER
revealed preferences for decision- making purposes had to be seen as reasonable by the regu-
lator to gain the regulator’s approval and reasonable by passengers to achieve customer
trust, in addition to being acceptable to NER staff and shareholders, unless NER chose to
hide this information or deny that it was relevant.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 557
NER revealed preferences had to be defined in terms of an NER view of what constituted
a reasonable and fair compromise between the potentially conflicting views of these different
parties, given all the ambiguity involved.
Current NER practices meant that none of the choices in her examples would be taken to
the board level. Choices were currently being made on technical grounds without the care-
fully structured board input needed – equivalent to assuming T = V1 × E = $4 million for all
S with no concern for risk efficiency and related concepts like enlightened prudence plus a
range of other stealth assumptions, with the board not even understanding that this was the
case. She argued that the basis of V1 should be of interest to the board, and the role in TS
of an S dependent ES factor bigger than unity ought to be a very big concern for the board
because all associated issues were unequivocally a board responsibility. The new framework they would use to avoid ignoring the enlightened caution and
enlightened prudence issues, which she had raised in relation to Figure 9.3 and the residual
provision captured by RS, was ultimately going to be of huge importance. Although she
could not effectively demonstrate the full implications of all these issues at this initial meet-
ing, she hoped they now had some feel for what was involved.
Sophie suggested that part of the reason NER had been taken by surprise by the public
reaction to their 40 fatality derailment was that they had failed to understand the need to
discriminate between expected fatalities associated with small numbers of people killed on
a regular basis (as on the roads) and large numbers of people killed on an infrequent basis
(as in serious airplane crashes). They had failed to see how elements of both had to be con-
fronted by NER, and they had failed to undertake effective contingency planning associated
with both prior preventative measures, and post- event mitigation measures, for high S level
incidents. This could be linked to the use of a TS concept which did not recognise the S
dependency. It could also be linked to a failure to use any effective form of ‘everything else’
factor and a failure to understand both risk efficiency and enlightened caution. Further-
more, it could be linked to a failure by NER safety experts to understand that risk could
not be measured just in terms of expected outcomes, and addressing possible departures from expected outcomes was crucial to recognising the role of enlightened prudence within an effec- tive framework for managing risk and opportunity. Any formal risk analysis framework which ignored possible departures from expected outcomes was untenable in her view.
The unconditional and conditional expected implications of catastrophes mattered, but so did the full range of potential departures from expected outcomes if an incident occurred,
and the implications of any associated ambiguity. From an EP perspective, it was wholly
unacceptable to argue otherwise.
These were all technical safety and security management shortcomings which needed
to be addressed. They were also clear evidence of why NER had not yet learned why they
needed to be afraid. NER needed to develop and then use this fear productively to creatively
address the issues associated with very low probability but very high impact incidents, plus
all related intermediate level incidents. Exactly how they would address these concerns was
still an open question, but the basic issues associated with the need to do so, and what had to
be done, were clear to her, and she hoped they were now becoming clear to board members.
Communication and capability issues for NER
Sophie suggested that underlying what might be seen as technical problems associated with core revealed preference issues for NER, there was a set of communication prob- lems and linked capability concerns which had to be addressed. NER needed an approach to these communication problems in terms of appropriate forms of dialogue or narrative
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
558 Employing planning tools in practice
fully embedded in the NER approach to corporate capability. Requisite capability included
understanding how to synthesis goals for safety and security which were focused on effec-
tively managing a zero fatalities aspirational target coupled to a sophisticated understand-
ing of the difference between aspirational targets and other targets. An associated nuanced
understanding of what should happen on average over an extended period also needed
to be coupled to commitments which accommodated low probability but high impact
incidents over a very wide range of possibilities. For example, the board might approve
publication of a statement like
NER is dedicated to always seeking a target of no injuries and no fatalities, and all NER
staff will be ceaseless in their efforts to achieve this goal. But everyone involved has to
live with the fact that even ‘catastrophic’ incidents are possible. Although a catastrophe
like 350 fatalities in one incident is possible, NER believe that the current estimate of
a return period of more than 4000 years for incidents of this severity involves a level of
risk that is both prudent and acceptable in terms of a compromise between the interests
of all the parties involved, having consulted with and listened to those parties, including
railway passenger groups and the railway industry regulator.
The board might prefer a very different form of words, but they needed a tactful way of
explaining clearly what was involved to everyone who was interested.
There was a technical aspect to this communication issue, associated with the need to
explicitly and formally distinguish between aspirational targets, what usually happened,
expected outcomes, and commitment targets defined in terms of target low probabilities for
large S value scenarios. NER needed to generalise the ‘ABCs of targets’ initially discussed in
Chapter 3, illustrated in later chapters. NER needed an ‘XYZ of targets’ addressing a much
more sophisticated view of commitments and balanced target concepts. But there were also
deep communication and linked cultural implications in terms of what they were trying to
achieve, as well as what needed to be done, how they ought to set about doing it and how
they ought to explain their position in terms of variations from both aspirational targets and
expected outcomes associated with probability distributions with very long tails for possible
injuries and fatalities which the industry had to live with.
Arguably a still deeper set of issues NER had to address was the current lack of an enlight-
ened culture driven by an enlightened generalisation of a fail- safe culture which always bal-
anced the need to remain optimistic about NER ability to improve risk efficiency related
concerns while remaining alert to all possible adverse incidents, avoiding any complacency.
Part of the solution to these concerns was first understanding and then managing some of
the complex insurance and risk sharing issues which were involved. This needed further
attention immediately.
An S = 350 scenario would bankrupt NER given current insurance arrangements as
Sophie understood them, and much smaller S value incidents would risk NER bank-
ruptcy, unless new insurance arrangements could be put in place. She understood that
the majority of NER shareholders were domestic pension funds. She suspected that they
were unaware of this risk, and would probably be very unhappy about it when they were
made aware of it. She suspected that both the regulator and the government were also
unaware of the scope and nature of NER bankruptcy risk, not to mention passengers and
passenger associations. She believed that the NER board and their shareholders should
seek insurance to avoid this risk, and by being open about the need for it with the regu-
lator, they might be able to encourage the regulator to provide appropriate insurance
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 559
via the government in a cost- effective way that would be a collaborative win– win for
everyone.
She would now briefly consider what might be involved.
Enlightened regulation and risk sharing
Sophie emphasised that NER very clearly needed to develop a good working understand-
ing of the approach she was proposing before approaching the regulator, but it might be useful if she now briefly sketched out what she called a ‘new deal’ scenario which the board
might develop to take to the regulator, once they had some specific practical propositions
to negotiate with.
They might start thinking about the issues from the perspective of an enlightened regula-
tor. If she was planning the kind of meeting she thought NER needed with an enlightened
regulator, there were a series of questions she might ask the regulator, with related observa-
tions. She would discuss these questions and observations in detail with the board well in
advance of such a meeting, based on what they had learned in the meantime. But she would
now outline her current view of some of the key questions to give the board a flavour of
what might be involved.
The first question, after explaining NER’s position, was, ‘If NER chooses b3 or b1 instead
of b2 in Figure 9.4, would you like to be consulted and see the trade- offs involved linked to
other aspects of your role as regulator, including rail fare levels and the structure and level of
subsidies?’ Sophie observed that choosing b1 or b2 or b3, plus the V1 and ES choices made by
NER, would seem to be core issues for regulator discussion with NER. NER proposing this
degree of open sharing could be the basis of a ‘new deal’. This new deal would have to fully
acknowledge the need to preserve the independent stance the regulator needed to protect
passenger, employee, and government interests. However, it could be based on assuming
that transparency and trust were important for both NER and the regulator to a degree
not fully appreciated or effectively developed earlier. The reason was neither NER nor the
regulator previously understood the issues in a framework which made an open relationship
an obvious win– win possibility for all parties involved – which all parties ought to insist upon. The second question was, ‘The government has made it an offence to drive motorised
vehicles on public roads without liability insurance, for very obvious good reasons, so why
do you not insist that NER had liability insurance in place to cover the liability aspects of
$7000m incidents or smaller incidents which could lead to NER bankruptcy?’ This ques-
tion highlighted an obvious weakness in the regulator’s current role if NER had another
serious accident and became bankrupt – and someone might raise this question, even if no
further accidents happened. The spirit of the question was revealing an uncomfortable truth
for both NER and the regulator which they needed to deal with jointly. At this point in the dialogue it would be important to make it clear that NER profitability was so modest in gen-
eral and so damaged by the two recent major accidents that there was no scope for further
reducing it without forcing NER into bankruptcy.
The third question was, ‘Private sector commercially provided liability insurance would
clearly be very expensive, and its cost would have to be added to either the current fares
passengers pay or current government subsidies for rail travel, so why not consider making
this mandatory liability insurance part of your rail regulation remit, providing it as a govern-
ment underwritten cost- effective part of your oversight role and extending the role of this
insurance to ensure that massive physical damage to the system leading to NER bankruptcy
was also avoided?’ This would save money for passengers and the government by avoiding
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
560 Employing planning tools in practice
private sector commercial insurance companies duplicating work which the regulator argu-
ably needed to do anyway. If it included prescribed payment guidelines for fatalities and
injuries, it might be a popular way of avoiding much of the legal costs surrounding the
ongoing disputes following the recent NER accidents. It would allow the regulator much
closer supervision of NER safety and security practice. It would also facilitate incentives for
safety and security which were directly linked to insurance premium levels.
The fourth question was, ‘would the regulator like to take a proactive pioneering role as
a partner with NER in the development of a “new deal” approach to regulation which both
parties are fully comfortable with?’ If so, the railway regulator might generate a model for
much wider use, beyond the railway industry, and take the credit for doing so.
The fifth question was, ‘Would the regulator like to involve passenger association repre-
sentatives, and perhaps even government minister’s representatives, in early discussions, as
part of showing regulator leadership when fashioning a “new deal” for all involved parties?’
Sophie did not want to oversell the insurance issues, or any other particular aspect of
the ‘new deal’ approach which she thought might be necessary, or take more board time
that day. However, she did want to make it clear to the board that complex issues like self-
insurance, partial insurance and risk sharing, in addition to economic efficiency, needed
consideration. In addition, very important ethical issues and ‘social efficiency’ issues were
also involved, because the trade- offs that needed making involved the interests of all other
relevant parties. Furthermore, if NER wanted to rebuild trust with all these relevant parties,
NER should show these parties enlightened plans which clarified who was taking what risks
for what rewards and how an open and collaborative approach could prove a winner for all
relevant parties, with the possible exception of some lawyers.
NER needed to show initiative and leadership but let the regulator, as well as rail pas-
sengers and NER staff on the trains, share the benefits from the change process with NER
shareholders, provided the regulator was sufficiently enlightened to be open to this kind of
approach. A hostile or uncooperative regulator would require a different plan.
Ambiguity which may be resolvable at least in part
Sophie indicated that she was about to draw her contribution to their current meeting to a
close, but it was important to stress that the need for ethically grounded cost transforma-
tions associated with fatalities and correlated levels of injuries which would generate trust
was beyond dispute. Furthermore, the need for appropriate risk efficient trade- offs involv-
ing prudence provisions was also beyond dispute. Despite the ambiguity about the relative
merits and drawbacks of common practice TS values and associated analysis frameworks
versus her proposed approach, the basic ideas behind risk efficiency, cost related to fatali-
ties transformations, and associated revealed preferences, had been implemented in practice
across the world in a variety of forms at various levels of sophistication, and they would make
intuitive sense to everyone interested in NER safety and security. For example, people did
not have to understand multiple criteria mathematical programming and associated goal
programming concepts to intuitively understand the idea that if you wanted to minimise
expected fatalities and you also wanted to minimise associated cost, there must be trade- offs involved. When key decision makers commit to a course of action, they obviously reveal
their preferences about appropriate trade- offs, and these trade- offs have ethical implications,
whether or not the decision makers understand and accept this is the case. The NHTSA
understood most of these issues in terms of its TS = T or V concept variant many decades
ago, as did most other bodies responsible for trade- offs of this kind. In addition, there were
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 561
important trade- offs between expected outcomes and possible downside departures from
expectations which most areas of risk management had confronted to some extent. The
existence of trade- offs always implied shadow costs if optimality was a concern. To some
extent these ideas had always been implicit in the ALARP concept, although significant
ambiguity of interpretation was also involved.
The ambiguity about common practice TS = T or V based approaches versus Sophie’s
framework was resolvable, as was ambiguity about how best to adjust and develop the opera-
tional form of her proposed approach, including the risk efficiency and enlightened pru-
dence discussion associated with Figures 9.1 through 9.5. What was not resolvable was the
ambiguity inherent in revealed preference interpretations of the preferences of other people
or the inherent ambiguity in the ethical aspects of the residual provision in RS or the inher-
ent ambiguity associated with the boundary between enlightened prudence and enlightened
gambles in the context of specific decision- making choices. NER had to confront all this
ambiguity as best they could or admit defeat – and admitting defeat was not a tenable posi-
tion for members of the board of NER.
More generally, difficult problems we are not sure how to solve are never resolved or dis-
solved by pretending that they do not exist. Difficult problems that are ignored simply become
invisible off the radar sources of lost opportunity risk. In the NER context, this really mattered
because hundreds of people might lose their lives or be critically injured when relatively small
increases in expected cost spent in the most effective way could avoid or significantly reduce
the chance of this possibility, perhaps simultaneously also improving the quality of service
provided in other ways. Everyone involved should prefer an operational form of enlightened prudence, embedded in opportunity efficiency, to failing to even consider resolvable ambigu-
ity because resolvable ambiguity was off their radar and beyond NER competence.
It was imperative that NER seek approximately correct answers to the right questions,
avoiding wasting effort and misleading everyone by answering the wrong questions.
Sophie did not say so for obvious reasons, but in her view if the NER board failed to rise
to this challenge, all other interested parties needed to understand why this was not acceptable and do something effective about it.
Sophie’s closing remarks
Sophie said she appreciated that her presentation was technical, intense, and lengthy, but
the problems addressed were difficult, complex, and extremely important. She apologised
for not explaining any points as clearly as members of the board might have expected, but
promised to spend more time whenever this was requested. She also apologised for any can-
dour they might have found offensive, but suggested that the topic was much too important
to risk failing to make the need for key changes clear. She finished by saying she hoped that
the board understood the flavour of what she had been saying in broad terms, sufficient to
see whether or not further dialogue would be fruitful.
She might have addressed the issues considered in the next section by way of a summary.
She refrained, but they are a useful summary of some core aspects of what this chapter has
been about from your perspective.
Biases, blind spots and bounds on formal analysis
From the board level down, all organisations can benefit from an inherently optimistic
opportunity management perspective on safety and security, looking for creative ways to
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
562 Employing planning tools in practice
reduce expected fatalities or comparable issues in positive terms. ‘Accentuate the positive’ is
a good general rule. However, ‘eliminate the negative’ is a relevant related goal providing
constructive tension. At a 100% level of achievement, eliminating the negative is usually not
feasible, no matter how positive the approach.
It is important to maintain a clear view of both the expected cost of failure and a plausi-
ble maximum realised cost if failure occurs, as well as a spectrum of intermediate scenarios.
Always being aware of the full implications of catastrophic scenarios if they occur, and the
full range of reasons they might occur, is part of avoiding complacency, and part of not
allowing short- term financial considerations undue weight.
For example, NER board members and some other NER staff members needed routine
reminders that the S = 350 scenario – if it happened – carried a CS = S × V1 × E350 ‘price tag’
of the order of $7,000 million. When appropriate, they also needed to be reminded that
the return period associated with this level of catastrophe would be directly influenced by a
wide variety of procedural systems and cultural changes which only a ‘fail- safe’ culture could
address in an effective and efficient way. They needed to set the cultural tone based on this
understanding and help to ensure that it reached ‘the big team’ as a whole. An effective and
efficient approach to safety and security risk, driven by an optimistic search for less risk in
conjunction with lower costs for better service, might transform a 2,000 year return period
estimate into a 4,000 or 8,000 year return period estimate, but eliminating any possibility of
the maximum plausible number of fatalities was rarely an option. Staying vigilant was crucial.
A complacent approach to outsourcing some safety critical issues might change a 2,000
year return period to 200 years or even 20 years, especially if the way their contractors made
use of subcontractors was not very carefully controlled. NER had to remain fully responsible for all outsourced components of their operations.
Enlightened optimism tempered by enlightened pessimism might be the way to present a
well- balanced approach to ‘do not forget to be afraid’ with a view to everyone playing their
role in responding to this key message.
A number of issues underlie the importance of this message and the overall approach of
this chapter, briefly reviewed now under the headings ‘biases’, ‘blind spots’, and ‘bounds
on formal analysis’.
Biases
Bias is an endemic estimation problem. At its simplest, if we ask how long an activity will take
or how much it will cost, assuming that conscious bias is not a concern, unconscious biases
like ‘anchoring’ need addressing. During the discussion of a minimum clarity approach to
estimation in Chapter 3 it was pointed out that asking for a P90 maximum plausible estimate
first, a P10 minimum plausible estimate second, tends to reduce unconscious anchoring bias
and offset some of the usual inherent bias associated with failing to understand the full scope
of uncertainty. Identifying implicit conditions like ‘this expected outcome estimate assumes
the project will not be cancelled’ pushed the sophistication of low clarity approaches to
unconscious bias a little further. The extensive decomposition of sources of uncertainty prior
to deciding what to quantify and what to treat as a condition, illustrated by the BP SCERT
approach discussed later in Chapter 3, pushed the sophistication of higher clarity approaches
further still. The sensitivity diagram discussion a few sections ago, and other links between
best practice project planning and operations planning, should make it clear that this level
of sophistication can be embedded in safety and security planning. There is a strong case for
arguing it should be formally required as part of seeking an ALARP position.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 563
One purpose served by emphasising the need for an early focus on a maximum plausible
impact scenario like Sophie’s S = 350 was minimising bias in this sense, using all available
expertise from safety and security experts, as well as the more general literature on uncon-
scious bias.
A closely coupled further purpose was providing a basis for explicitly addressing con-
scious bias. We know from widely reported 1960s PERT experience that forcing people
to estimate ranges as a basis for providing expected value estimates reduces conscious as
well as unconscious bias. It follows that forcing people to start with a plausible upper
bound on the size and probability of a maximum plausible catastrophe scenario makes
it relatively difficult to conceal the scale and likelihood of both potential catastrophic
outcomes and less serious related outcomes. Making understanding the full implications
mandatory at the board level, followed by regulatory approvals processes exposed to pas-
senger, on- board staff, and more public scrutiny, can be seen as part of this attack on bias.
Current common practice does not frame concerns about bias in this way, and doing so
has important advantages.
Blind spots – expected outcomes, what usually happens, and the XYZ of targets
Safety and security planning, and managing the implementation of plans, needs clarity about
expected outcomes, what usually happens, and the XYZ of targets. For example, NER needs
staff and passengers to focus on zero fatalities and injuries as the aspirational target, under-
standing that zero fatalities and injuries are only going to happen if they are all both diligent and lucky. Each and every incident involving fatalities and injuries needs a proportionate assessment to ascertain whether bad luck or bad management was involved, dealing appro-
priately with root causes and responsible parties. This is part of managing safety and security
in relation to aspirational targets.
However, targets which safety and security regulators control to, publish, and reward per-
formance or punish in relation to should not be confused with aspirational targets, expected
values used to make decisions, other balanced targets, or what usually happens. Targets used
by safety and security regulators should be commitment targets with identified probabili-
ties of exceeding a range of incident scenario values like the S = 5, 20, 70, and 350 values
used here for specified categories of incidents like collisions with fatalities, derailments with
fatalities, and level- crossing incidents with fatalities, with issues like terrorist incidents kept
separate for some purposes, and suicides involving trains kept separate for most purposes.
Commitment targets based on aggregate numbers of fatalities per annum, interpreted as a
number not to be exceeded, are not appropriate. Any high S value incident should neces-
sitate in- depth but proportionate independent audit of root causes and responsible par-
ties which will be made public. This is part of managing safety and security in relation to
commitment targets defined in probability of exceedance terms, comparable to using 100 ,
1,000 , or 10,000 year return period specifications for the risk of structural failure for build-
ings and other facilities due to earthquakes, which traditionally distinguish between very
different structures with different failure implications, like houses, high- rise apartments,
large dams, and nuclear power stations.
There is a strong case for arguing that the XYZ of targets in this context must include
full disclosure and effective dialogue between all interested parties. From an EP perspective,
there is no plausible case for not making it legally required practice, and treating a failure
to meet this required practice as a very serious matter involving custodial sentences for the
senior level decision makers who should have understood the issues.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
564 Employing planning tools in practice
Blind spots – the ethical interpretation of shadow cost and revealed preference aspects of transformations from a fatalities metric S to a cost metric CS via TS
Safety and security planning and associated management by both public and private sector
organisations, plus associated regulation, and ALARP discussions in a legal context, need
clarity about the nature of the transformation from a fatality metric S to a CS cost concept
involving the value of avoided fatalities.
One aspect is the shadow cost nature of the transformation which dictates an explicit
value for consistent decision making. If TS values are not made explicit, both lives and
money will be wasted, because different values will be used implicitly in different contexts
in an inconsistent manner, sometimes spending too little, sometimes spending too much.
Explicit values used consistently are required for fatality efficiency to minimise the number
of fatalities for any given level of expenditure with safety and security implications.
A second aspect is the revealed preference nature of the shadow cost employed, dictating
an interpretation which requires consideration of ethical issues. Whether or not the money
being spent and the lives and health at stake are associated with the same set of people, all of
them ought to have a say in the process of determining appropriate trade- offs, and there will
be a range of views which cannot be averaged without ethical implications. Addressing these
implications requires ethical judgements consistent with the ALARP principle, plus notions
of fairness with respect to winners and losers. There are important implications which need
considerable care.
In my experience, common practice is neither unambiguous nor consistent about these
issues. From an EP perspective, there is no plausible case for not making a consistently
high level of clarity associated with all relevant practice both an acknowledged ethical duty
for all relevant decision makers, and a legal requirement enforceable at the level of suitable
accountable individuals, with no acceptable avenues for organisations to avoid having appro-
priate accountable individuals.
Blind spots – the ethical interpretation of TS values which are not S dependent
If the ethical nature of any TS concept is clear, then making TS independent of S in a con-
text like road travel or aircraft travel may be a reasonable judgement if all those involved see
no advantages in the added complexity of S dependent analysis that override the simplicity
advantages of a TS = T approach. But being able to choose may be important. It is always
an asset to have the ability to test the implications of possible simplification, and arguably
it is essential to use framing assumptions which allow testing any simplifying assumptions which may really matter. The risk of seriously inappropriate simplifying assumptions is
important.
In a railway travel context, insisting that 1,000, 100, or 10 fatalities in the same incident
all have the same implications per fatality as single fatality incidents seems an inappropri-
ate framing assumption at present, for a range of reasons, the need to consider enlight-
ened prudence associated with departures from expected outcomes perhaps being the most
fundamental. Although enlightened prudence was considered very late in this chapter for
expository reasons, it matters greatly. If government insurance paid full compensation for
fatalities and injuries at standard rates, and all the parties involved understood all the issues,
then all parties might agree that TS = T was appropriate. However, at present it seems plau-
sible to assume that most people involved would prefer additional concern for high S value
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 565
incidents relative to low S value incidents for a very complex but entirely rational set of
reasons. This should include NER shareholders and NER board members. It should also
include rail regulators and governments.
In my experience TS = T is a common practice working assumption given framing assump-
tion status, with closely coupled implications directly dependent on ignoring risk associated
with departures from expectations. From an EP perspective, there is no plausible case for
not making S dependence an option to be tested and used whenever appropriate.
Blind spots – the ethical interpretation of enlightened prudence in a CS context
When a CS concept is involved, enlightened prudence has an ethical aspect which goes
beyond financial prudence. An NER S = 350 implying a CS of the order of $7,000 mil-
lion should frighten everyone involved enough to think very seriously about what can be
done to reduce the probability and mitigate the impact as far as possible, however they
chose to weight the relative importance of the fatalities, the injuries, and all the knock- on
consequences.
In my experience, common practice does not address enlightened prudence in a safety
context because risk efficiency and associated risk– reward trade- offs are not addressed
explicitly. In effect, there is no framework for considering enlightened caution, enlight-
ened gambles or enlightened prudence. This means there is no framework for thinking
about distinguishing between good management and good luck, bad management and bad
luck. This, in turn, means that motivating all staff to behave appropriately, including board
members, is severely handicapped. Furthermore, an organisation’s ability to retain good
managers and dispense with bad managers is handicapped. From an EP perspective, this is
inappropriate and unnecessary, and it is a board level failure, along with most of the other
blind spots, especially the next.
Blind spots – the need for ‘add- in’ strategic consideration
If all the earlier blind spots are understood and dealt with, the need for board level top-
down strategic management in a holistic framework addressing all related bottom- up plan-
ning becomes obvious, arguably ‘blindingly obvious’. Boards clearly cannot get directly
involved in tactical planning issues, but it is a board’s responsibility to guide and deal with
the strategic implications of lower level tactical choices. The implications of possible large
numbers of fatalities and injuries require consideration of crucial issues that need to be
explicitly built into board level strategic choices via a clear formal recognition of the role of revealed preferences about trade- offs among profit, fatalities, and injuries.
In my experience this is currently not common practice.
Blind spots – the clarity efficiency of an ES which goes beyond fatalities
Whether or not an S dependent TS concept is employed, being able to treat injuries and cost
associated with damage to the system at the same time as fatalities within the ES concept is
an important clarity efficiency issue. It means that injuries and physical damage to the system
plus knock- on cost and lost revenue concerns which are assumed to be correlated with fatali-
ties can be addressed by the use of one common overall shadow cost and revealed preference
transformation factor. This, in turn, means that agreement involving different directors with
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
566 Employing planning tools in practice
different views is facilitated. More important, different parties with different perspectives,
like passengers and NER shareholders and regulators, can focus on what really matters.
This is an opportunity to improve the clarity efficiency of a general approach which was
identified as a consequence of the need to consider the added complexity of an S dependent
AS in a particular context. It can be widely exploited, even if an S dependent AS concept is
not required. If it can also be made use of in other contexts which do not need the complex-
ity triggering the insight, then it may provide further added value in a manner that needs
exploration and exploitation. For example, it might prove useful in a road safety context, an
air transport safety and security context, or many other very different contexts which have
nothing to do with transportation.
Bounds on formal analysis
The generalisation of risk, clarity and opportunity efficiency as explored by Sophie for NER
takes us to the limits of sophistication for EP which I am currently comfortable with. You
may need to take things much further or in very different directions. But there are bounds
on the levels of complexity which formal analysis can cope with, and it may not be straight-
forward to deal effectively with what we do not really understand or areas where we collec-
tively understand the issues but do not agree about their relative importance.
‘Understanding the scope of relevant ignorance is the beginning of wisdom’, stated in
numerous variants, is often quoted common wisdom which might be seen as a crucial com-
ponent of ‘remembering to be afraid’. It is also worth remembering in more general terms
in the context of this chapter. More generally still, it is a useful mantra for the enlightened
planning approach as a whole, which can be worth remembering even in comparatively
simple contexts.
Earlier examples of related safety and security analysis
Sophie’s approach to safety and security is directly grounded on 2010– 2013 experience
working in a military context. The central importance to the tale of this experience means
that it is not feasible to further develop the tale of this chapter in a way that is grounded
in empirical tests which can be discussed. However, the framework of Sophie’s overall
approach, and variants of the specifics of her approach to implementation, seemed to be
making progress in a military context when feedback was last provided.
An earlier real case involving Railtrack, which Sophie’s tale is also directly based on, tried
in a different way to implement a somewhat different structuring of some of the basic ideas,
including a prototype S dependent TS shadow cost concept. My recommended approach
failed to convince the key decision makers, no doubt at least in part because the four com-
ponent ES concept was missing, and at least in part because the implications of high S values
and risk in terms of departures from expectations could not be explained with sufficient
clarity to motivate a desire to ‘learn to be afraid’ in the sense this chapter addresses. Other
relevant issues may have included a natural desire by Railtrack to keep the approach used
simple in a manner which was in line with conventional expert thinking, unless there was
overwhelming evidence of a need for change, and the lack of a clear case for specific reason-
ably simple defined changes.
Two subsequent very serious rail crashes were instrumental in Railtrack becoming
bankrupt as much of the press interpreted it at the time, although some later explana-
tions offer other reasons for Railtrack’s bankruptcy. As a result, Railtrack was replaced
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 567
by Network Rail and the current associated structure. This might be interpreted as evi-
dence of the need for change in any organisation in a context comparable to NER, using
approaches comparable to those attributed to NER which have yet to experience a crisis,
in the sense that they may have been lucky so far, but sooner or later their luck may run
out. I would strongly encourage this view for any organisation currently using a TS = T or
V concept with no S dependence capability, no multiple part ES concept, and no enlight-
ened prudence concept linked to risk efficiency plus a more general opportunity- efficiency
concept. This view may irritate some safety experts, but relevant board members may
need to confront their experts. Additionally, passengers, on- board railway staff, regula-
tors, and their political masters may need to take effective action contrary to relevant
current expert advice.
An even earlier real case, that triggered the MoD study and provided some insights you
may find useful, was undertaken for a public sector water authority. It involved addressing
the risk of terrorist bomb attacks on water mains, when working with Sir William Halcrow
and Partners over the period from 1986 to 1988 for the Severn Trent Water Authority on an
overall review of their water supply systems. The analysis of a potential terrorist attack on a
major water main supplying the city of Birmingham was highly successful at the board level.
It resulted in a decision not to spend money on protecting the water main from potential
terrorist attacks, so far justified empirically by good luck and political background changes,
but it was explicitly designed to protect the board should bad luck occur in addition to mak-
ing an appropriate decision. It triggered the military study initiated in 2010 by a very senior
MoD manager who remembered my using the Halcrow example decades earlier for an MoD
in- house course which he had been responsible for running.
The Halcrow potential terrorist attack approach adopted was very simple in comparison
to the relatively high clarity approaches demanded by the MoD and NER contexts, but this
makes it a useful example to explore briefly now.
The Severn Trent Water Authority board’s concern was initiated by an unsuccessful
attempt to blow up this particular water main by a Welsh extremist who objected to Welsh
water going to England. This unsuccessful attack made the board aware that all signifi-
cant water mains were exposed to possible successful sabotage attempts using explosives to
destroy them and that the board would be exposed to reasonable charges of negligence if it
did not formally analyse an appropriate response to this clearly demonstrated threat.
The starting point for analysis was the client’s knowledge that the pipeline had been in
place for a very long time with no successful attacks, a subsequent quick review which sug-
gested that there was no other relevant data to assess the probability of a successful attack
in terms of its expected value or an associated plausible range, and the resulting working
assumption that a parametric approach centred on this probability expected value and its
plausible range might be a good place to begin.
A review of the nature of the pipeline over its full length revealed that the obviously vul-
nerable points for terrorist attacks involving explosives were river crossings where the pipe-
line came out of the ground and crossed the river on its own trestle or bridge. The expected
cost C1 and a plausible associated cost range to bury the pipeline at all river crossings was
assessed. This expenditure would not eliminate the risk of successful terrorist attacks, but
it was the obvious cost- effective first step, and it involved reasonably straightforward civil
engineering project cost estimates.
The expected cost C2 and a plausible associated cost range to deal with all the direct and
indirect cost implications of a successful attack was then assessed. Part of this was concerned
with repairing physical damage, but the impact on domestic water consumers and industry
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
568 Employing planning tools in practice
was crucial, and there was particular concern for the complex economic implications of hav-
ing to shut down water- dependent industrial operations.
The prior view when choosing a parametric approach was that if the C1/C2 ratio was of
the order of 1/10,000 or less, it would probably be prudent to spend the money burying
the pipeline without further analysis because even very small probabilities of a successful
attack made this a risk efficient choice, and nobody could make a convincing case for a prob-
ability lower than this parametric analysis ‘flip point’. However, if the C1/C2 ratio was of the
order of 1/10 or more, most customers of what was then a publicly owned water company
would probably support the view that it was not worth spending the money required to
bury the pipeline or pursue further analysis because a successful attack every ten years on
average was implausibly pessimistic. A C1/C2 flip point in the 1/10 to 1/10,000 range
would require a more carefully considered decision, considering the ranges associated with
C1 and C2 expected values, as well as any important issues not measured by these costs, like
reputational concerns. A flip point in the middle might suggest that the choice was reason-
ably neutral in measured cost terms, and the choice could perhaps be made with an explicit
focus on what had not been measured. The flip point measured, and the explanation of its
implications for the board, let them make decisions they were comfortable with and could
defend whether or not a successful attack took place, which they saw as based on very sound
and effective clarity efficient analysis.
A key general message to take away from this example is that massive uncertainty about
only one key parameter, like the probability of a successful attack on the water main, can
be addressed using a parametric approach, a special kind of sensitivity analysis, linked to an
iterative multiple pass approach. More than one key parameter involving massive uncertainty
means sensitivity analysis remains crucial, but we have to simplify the key parameter relation-
ships in a systematic manner, and test sensitivities in a systematic, multiple pass manner. That
is at the core of the EP approach to the NER context.
The starting point for the differences between this Halcrow study and the MoD and NER
contexts was the need for an explicit shadow price concept related to fatalities, and the value
of a two component T = V1 × E structure to link the transformation of a ‘fatalities’ metric
to a ‘cost’ metric. Making TS = V1 × ES dependent on the value of S was a further key new
complexity, but getting all relevant parties to understand the way the four component ES
factor kept the number of parameters which drove decisions to a minimum was a key new
simplification, by embedding multiple ‘injuries’ metrics, plus the cost of physical damage,
plus a residual anything else factor which also addressed enlightened prudence to help cap-
ture concerns about departures from expected outcomes.
You might find it useful to see this as a demonstration of an elaboration of ‘keep it simple
systematically’ to ‘keep it simple systematically employing carefully considered complexity
composition’. The pay- off from finding an appropriate structure for decomposing the TS
concerns is this chapter’s illustration of why seeking carefully considered decomposition and
composition, whenever it really matters, is worth the effort, building on simpler illustrations
in earlier chapters.
In summary, Sophie’s approach is flexible, sound, and robust. Significant relevant experi-
ence is the basis of this judgement. However, I emphatically do not wish to overstate the
empirical basis of some details of its features or underplay the contentious nature of the very
difficult and complex issues involved.
The credibility issues associated with this chapter’s tale are again somewhat different from
those associated with earlier chapters, but there are some common aspects, including public /
private sector issues with regulator implications.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 569
Sophie’s views on politics and personal ethics
Sophie was politically neutral, in that she was not ideological in a political party or political
movement sense to a degree which mattered for current purposes, and she resisted doctri-
naire views on all related topics. She saw serious flaws as well as useful strengths in most of
the approaches to politics and issues like public or private sector status for industries like
railway travel which she was familiar with, acknowledging that the scope of her knowledge
was limited. She had what most people would see as a clear moral compass, in terms of
treating other peoples’ interests and concerns empathetically, and taking full responsibility
for her own actions.
Sophie had developed an approach for NER based on what she thought were reasonable
beliefs. She was reasonably sure the NER board would be open to accepting leadership in
the cultural and ethical terms she proposed, and while her hope that the regulator would
also prove enlightened was more tentative, she thought it worthwhile starting with an opti-
mistic premise and an ambitious stretch target. She saw these two working assumptions
as a fundamental aspect of applying a UP with minimal constraints to ‘design a desirable
future’ – to be as creative as possible in the time available. Part of her associated robustness
testing was testing her working assumptions about the regulator and the board.
For example, the board would have to approve any position which she and Oscar negoti-
ated with the regulator, and an approach which was as open as that proposed would need
clear board approval in advance. If the board was convinced that the regulator would never
be able to provide insurance, and nobody working for NER should explicitly encourage
speculation about large S value incidents in the way that using an S = 350 scenario in her
illustrative analysis did, she would have to respect that judgement to continue working for
NER. It would marginally compromise her ethics to suggest an approach like reducing her
five size scenarios just four defined by S = 0, 5, 25, and 125, merging the earlier S = 70 and
325 into a single, less frightening S = 125 scenario which still covered fatality levels up to the
feasible maximum with the same underlying probability distribution, increasing the S = 20
to 25. But apart from presentational issues, these four size scenarios now use a simpler strict
log- base five scale, which would on the surface look like a benefit. The associated downside
was the numbers she had used suggested that the S = 125 scenario would now be hugely
important, with no framework for reassurance about what might be seen as an S = 625 sce-
nario not addressed. She understood what she referred to with close associates as ‘frighten-
ing the horses’ risk. ‘Frightening the horses’ was a euphemism she frequently encountered
for revealing unwelcome truths in a way which upset those you did not want to upset. In her
first meeting with the board she wanted the discussion to reveal that she would compromise
in this way without any problems if that was the board’s wish because she had no wish to
frighten the horses.
However, a crucial implication of this kind of compromise was the potential creation
of an associated risk of what she and many other people referred to as ‘an elephant in the
room’. She and all other relevant players would have to manage this elephant in the room
risk in a way which did not compromise her moral compass too far by obscuring what really
mattered, including issues like a possible S = 625 scenario needing attention. Making sure
that revealed preferences displayed by NER reflected any key NER risks, and appropriate
enlightened prudence, would be crucially important to Sophie.
As a practical illustration of the way forward, if these issues came up and her numbers
were realistic, she might propose S = 1, 10, 100, and 1,000 size scenarios, based on the
hypothesis the 100 fatalities scenario would prove the dominant concern, the 1,000 fatalities
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
570 Employing planning tools in practice
scenario demonstrably improbable. In terms of working assumptions and issues like what S
scenario size structure would work best, Sophie was flexible and creative.
In terms of ethical issues, Sophie’s ‘bottom line’ was that she needed to trust that the
board would engage with her basic TS structure ideas and risk efficiency based conceptual
framework and accept the need to work towards an ethically sound position to make much better the NER organisation which she believed was a genuine ‘national treasure’. If that did not seem feasible, she would prefer working elsewhere, perhaps for a regulator, perhaps
eventually becoming the head of a regulator organisation, as a different route to meeting
her professional goals.
Public /private sector issues with regulator implications
Sophie assumed that NER directors would want to take an ethical approach to TS deter-
mination and dealing with enlightened prudence, including addressing liability insurance
for large S value incidents with the regulator in an open manner. She based these assump-
tions on the ‘national treasure’ nature of NER, driven by shareholders and board direc-
tors whose ethos included the belief that shareholders and employees of NER should
all be fairly rewarded for providing the best service feasible, but there was no place for
sharp practice or unfair exploitation of their monopoly position. Given these assumptions,
Sophie was convinced that the current private sector ownership of NER was ‘socially
optimal’.
One of Sophie’s central concerns, which she linked to her personal variant of the UP of
Figure 2.1, and an EP perspective in general, was the concept ‘the good must be promoted
and protected, the bad contested and constrained’; otherwise, ‘the bad will drive out the
good’. She fully appreciated that an important corollary was ‘recognise that different parties
will have different views about what is “bad” and what is “good”, and some parties will be
driven by views entirely focused on what is good for them, including a wish for power or
material benefit’. She had come across the basis of this idea in market behaviour contexts,
discussed earlier in several chapters. For example, the winner in a bidding context is some-
times the lowest price bidder who least understands the problem or is the most prepared to
engage in ‘claims engineering’ to drive up the actual cost, which makes life very difficult for
more competent and honest bidders, as well as for all the customers/clients involved.
Sophie’s current working assumption was that the regulator would probably have con-
cerns about NER competence but would understand and want to preserve the ethos under-
lying the widespread ‘national treasure’ perception of NER. She hoped the regulator would
share her view that NER’s private sector status was socially optimal and assumed this was the
case when proposing her approach to insurance.
Sophie was aware that the modest profits earned by NER shareholders, damaged by the
two recent incidents along with the reputation of NER directors, put NER at risk of a hos-
tile takeover by international players of several different kinds – for example, a private equity
firm or a utility based outside NER’s country. She took the view this was an obvious risk
for NER directors, shareholders, and staff, but it was particularly important to her because
it was also a critical risk for passengers and the government, and it needed attention by the
regulator and possibly by other government agencies.
A predatory takeover by an organisation with a focus on profits remitted to another
country would be a social catastrophe in Sophie’s view. This kind of organisation might pass
ownership of NER assets to low tax rate countries to evade taxes and avoid loss of owner-
ship of the assets if NER went bankrupt, and having minimised the impact of bankruptcy,
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 571
they might minimise their concerns about service as well as safety and security, disguising
their short- term profit- driven stance with wholly unethical Machiavellian guile and cunning.
Sophie knew this issue was beyond her terms of reference, but finding the right way to
make it a concern for the regulator could be an important part of the NER dialogue with the
regulator. She wanted the regulator to see the need to make public a deal with NER which
would make a predatory takeover of NER very unattractive for any organisation which
would be prepared to exploit its position – to discourage any potential badness before any
bad parties gave serious thought to driving out the good as a carefully planned pre- emptive
strike. In Sophie’s view, if this was not done, and an organisation which the regulator and
everybody else could not trust took over NER assets, the situation would be exceedingly dif-
ficult to retrieve. A disastrous accident taking place against a background of other concerns,
followed by nationalisation handled by a competent government with equally Machiavel-
lian cunning, might be the only feasible way forward. This would clearly raise a number of
immediate and ongoing political difficulties and other complications best avoided. To use
another one of Sophie’s favourite mantras – ‘an ounce of prevention is worth a pound of
cure’. In Sophie’s view a ‘safety- first’ culture was needed by an enlightened regulator as well
as NER. Waiting for a massive train accident before dealing with the preventable loss of a
national treasure would be highly unenlightened.
The private sector status of NER should sharpen the focus of a regulator on safety and
security relative to the position if NER were state owned and intensify the need for clar-
ity about risk sharing and insurance issues. However, state ownership of a railway system
instead of a private sector organisation operated in a holistic manner like NER should not
change the nature of an enlightened approach to safety and security management in any
fundamental ways.
Public /private sector issues as very briefly explored in this section for NER are complex.
But they are very simple relative to other contexts. If a government has more than one
organisation to deal with, competition between these organisations is involved, and some
collusion may be a concern, the issues can become much more difficult.
Integration of safety planning and other corporate planning
The need to integrate safety planning with security planning and corporate planning as a
whole at a strategic level has been central to the EP perspective developed in this chapter.
This implies that integrating the approaches discussed in this chapter with the approaches
explored in Chapter 8 is essential.
A starting point for achieving this integration involves working within an agreed NER
version of the five categories of corporate planning framework summarised in Figure 8.1:
1 goals planning,
2 futures planning,
3 long- term planning,
4 medium- term planning, and
5 short- term planning.
Safety and security planning for NER also needs integration with some other aspects of plan-
ning, worth brief consideration now.
Oscar and Sophie both viewed safety and security planning as a special case of using an
EP approach to dealing with uncertainty and complexity. This meant Sophie could use the
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
572 Employing planning tools in practice
EP perspective developed in Chapters 1 through 8 plus all the ideas explored in this chap-
ter, but also draw on all useful aspects of the considerable literature and experience behind
safety and security management in general, and that related to railways in particular. It
meant that Oscar could use the same EP perspective, but in addition he could draw on all
useful aspects of the considerable literature and experience behind operations management
in general, and that related to railways in particular. Furthermore, it meant they could both
use these compatible frameworks to effectively integrate safety and security planning with
other aspects of planning within Oscar’s remit if the board approved Sophie’s proposed
approach.
An effective common perspective is important for full integration. It is also important to
appreciate that this full integration means that whatever both Oscar and Sophie add to their
EP frameworks needs internal coherence – any internal inconsistencies need resolution.
Oscar also needed a relationship with the NER Projects Director comparable to that
between Ollie and Paul in Chapter 7. This in turn meant that Oscar, the NER Projects
Director, and everybody else with NER interfacing with projects, needed to work within a
common NER version of the four Fs:
1 a project (asset) lifecycle framework,
2 a seven Ws framework,
3 a goals– plans relationships framework, and
4 a planning process framework.
Both an operations planning perspective and a project planning perspective can and should
use a common lifecycle structure and a common seven Ws structure to integrate their roles
in any organisation, as discussed in Chapter 7.
A goals– plans relationships framework for all planning purposes needs to be built on these
two frameworks, within the five categories of corporate planning discussed in Chapter 8.
The rationale is comparable to that discussed earlier, in corporate planning, project planning
and operations planning contexts.
Finally, a coherent approach to using processes in all contexts needs clarification for NER
to make full use of an EP perspective.
Applications in other contexts when well- founded trust matters
Sophie’s overview of what needed to be done to address safety and security planning in a
railway context clearly needs generalising to accommodate other industries with different
but comparable issues. This section briefly explores some of the issues which may need
attention, assuming that well- founded trust needing explicit formal planning when dealing
with important complexities remains a key concern. ‘Catastrophes’ in the low probability
and high impact sense may or may not be an issue, but a level of clarity which is proportion-
ate to what is at stake is always a clarity efficiency concern, and if a lot is at stake, some of this
chapter’s approaches may have useful formal roles in many different contexts.
The Shell HSSE (Health, Safety, Security and Environment) Award of the Association
of Project Managers illustrates one broader perspective – generalising safety and security to
also embrace environmental concerns plus associated long- term health implications. Adopt-
ing this perspective, by way of a specific example, what should Sophie say to an organisa-
tion involved in offshore drilling for oil or gas if they asked for her advice about preventing
the kinds of problems BP ran into over the Deepwater Horizon incident in 2010, or a
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 573
government concerned with regulating this kind of activity? Assume Sophie began by
explaining in outline the ideas developed in this book thus far, as a starting point. How
could she build on this?
The 11 fatalities involved in the Deepwater Horizon incident (people working on the oil
well rig when the blowout occurred) clearly mattered a great deal, and any injuries associ-
ated with this kind of incident clearly matter too. Fatalities and injuries plus damage and
knock- on implications comparable to those discussed by Sophie could be addressed using
the framework outlined earlier in this chapter. But it was the scale of the oil spill triggered
by the Macondo well blowout, and its ongoing pollution implications, which triggered US
government and other reactions, and it was the net effect of all reactions which crippled BP.
As of 2017, the direct cash flow cost estimate for BP exceeded $60 billion. Environmental damage with long- term as well as short- term implications, including ongoing health hazard
concerns, raises some further very different and exceedingly complex issues relative to those
addressed for NER.
This suggests a need to replace the NER fatalities metric with a generalised size of inci-
dent scale which is certainly more complex and might be multi- dimensional. A two dimen-
sional scale might use S1 = 1, . . . , m1 and S2 = 1, . . . , m2, where S1 measures incident
fatalities in a manner directly comparable to Sophie’s S, and S2 measures oil spill volume.
S2 would be associated with non- financial environmental implications, as well as long- term
health implications and economic losses associated with concerns like fishing, plus both
short- term and long- term clean- up costs. But three or more dimensions might be needed in
some cases. Assuming two dimensions for the moment, the ES1 S2 needed to transform S1- S2
combinations into a T measure generalisation of TS would have to be a two dimensional
matrix instead of a one dimensional vector.
Strong correlation or the dominance of one dimension might make it clarity efficient to
collapse this multiple dimension approach into a single dimension S concept which uses com-
posite scenario portrayals directly. For example, the S = 1 scenario might involve a minimum
plausible oil spill volume blowout, with no fatalities or serious injuries, no ongoing health
issues, and no fire or explosion. The S = m scenario might define the maximum plausible oil
spill blowout oil volume in conjunction with the maximum plausible number of fatalities in
total, the maximum plausible number of immediate injuries plus ongoing health impairments,
and all other relevant maximum plausible negative outcomes. The S = 2 scenario might rep-
resent a much bigger plausible oil spill blowout than S = 1, involving one fatality and several
injury cases. Fatalities and injuries plus a much bigger oil spill than S = 2 might be associ-
ated with S = 3, . . . , m − 1, portraying escalating disaster levels in terms of levels of severity
measured on several scales, with scale alignments based on assumed levels of correlation. By
S = m/2 everybody on the drill rig might be assumed to be killed, further scenarios reflecting
bigger and bigger oil spills, perhaps with fatalities associated with those trying to contain it
or living nearby. The S = m scenario might involve implications significantly exceeding all the
BP Deepwater Horizon consequences, with direct cash flow costs of $100 billion or more,
just in terms of fines and court settlements and out- of- court settlements.
The cost– benefit approach discussed earlier, plus a legal liability approach, could be used
to estimate plausible V1 and E1 values, then the first TS involving both fatalities and oil spills,
and so on, with TS values over the S = 1, . . . , n range addressing oil spill environmental
damage issues and knock- on health concerns, as well as incident fatalities and injuries. Some
variant of this kind of approach would seem to be an obvious and appropriate early step in
an analysis which needs to take quantification of what really matters to its limits, but it would be crucial to remain very clear about the limitations of all aspects of this kind of valuation.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
574 Employing planning tools in practice
Issues of trust and their ethical basis clearly need direct and very careful treatment over
the full range of relevant S values. No V1 or Vm or intermediate interpretations addressing oil
spill environmental damage issues and knock- on health concerns comparable to a ‘revealed
preference’ interpretation of the value of avoided fatalities used for decades by government
bodies responsible for travel safety are available which I am aware of. Although possible
starting places include a traditional cost– benefit analysis approach or implicit revealed pref-
erence values assessed via reviews of past incidents, these are very different starting places,
and both involve serious difficulties.
As part of processes associated with granting permissions to drill and operate by all
directly involved governments, the use of some variant of Sophie’s approach to TS values
as a starting point for an enlightened approach to planning might be required. Or it could
be volunteered by the oil companies. This might provide a useful basis for the regulation of
such development, and a useful way for oil companies to build trust with their regulators.
However, to be effective, it would probably need auditing by external bodies who could be
trusted by all parties, and auditing the procedural system and culture concerns, as well as the
physical systems, would be crucial.
Whether or not governments get directly involved in this kind of enlightened planning,
to be effective the risk of blowout addressed in this way has to be linked to all other threats
which might share related initiating events or outcomes plus preventative, reactive, or miti-
gating responses. For example, if blowout during oil well development, and blowout as a
consequence of an operating well being hit by a ship or an iceberg or a terrorist attack have
comparable implications and ways of preventing or mitigating such events have any underly-
ing commonality, a clear understanding of these relationships has to be a central part of the
analysis. A prototype EP approach to this kind of situation is the basis of a paper which you
might find useful (Chapman, Cooper, Debelius, and Pecora, 1985), but significant develop-
ment of its approach would be essential.
Some organisations may take very large risks, like that of a blowout, and go bankrupt
if these risks are realised because internal performance reward systems promote this. The
shareholders and boards may not appreciate this risk, but some may explicitly set themselves
up to profit from taking risks they cannot cover. In extreme situations they may operate like
the organisation Sophie associated with ‘hostile takeover risk’, setting themselves up in care-
fully planned structures to minimise taxes as well as other liabilities. If governments are con-
cerned about having to pay for the costs of clean- up because the responsible party has gone
bankrupt or has all its assets offshore protected by governments they cannot confront (or
will not prevail over if it do confront the responsible party), one obvious response is using
license fees and taxation systems to charge an insurance premium appropriate to the risk born
by the government and its citizens, or simply refuse permission in some cases. On- board
ongoing inspection to ensure that insurance conditions are fully observed might be required
by law. Clearly stated jail sentences might be the penalty for any compliance failures, with all
appropriate directors in the frame as well as their operators. If up- front ‘insurance’ payments
to cover the expected cost of all incidents of this kind are used explicitly, governments will
clearly need to audit all aspects of safety and security. To set appropriate ‘premiums’, this will
have to include physical and procedural systems, plus other capability- culture concerns, con-
ditioned by all other relevant issues. The implicit ‘uninsured risk’ involved becomes more
manageable because it is made visible when transforming it into an explicitly insured risk
with this insurance mandatory. In turn, the ‘dysfunctional incentives’ driving organisations
to take inappropriate risks become much easier to confront – mandatory insurance involving
prohibitive insurance premiums might stop organisations taking risks which can be reduced
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 575
effectively and efficiently, or stop them attempting this kind of project, providing wilful
dishonesty by any key players is viewed as a criminal offence, law enforcement is viewed as
non- negotiable, and the legal system is free of undue political influences. These provisos
may be non- feasible. They are obvious sources of risk if this is the case.
An issue of considerable importance is that socially responsible organisations need to
avoid a position of competitive disadvantage by promoting the kind of explicit government
insurance approach just discussed, flushing out and removing from the market all those
competitors who are not prepared or not able to work on these terms, to establish ‘a level
playing field’ for ethical organisations and to exclude organisations society should not allow
into the game.
Acting on a clear understanding that ‘the bad drives out the good unless the good is
promoted and protected, and the bad is contested and constrained’, should be seen as
central in a market regulation context. If bad behaviour organisations are allowed to get
away with it, this will put those who do not copy their bad behaviour out of business. All
regulators need to ensure this does not happen. Regulators need to take a very aggressive
and proactive approach to contesting and constraining any behaviour which the regulators
have reasonable grounds to suspect is bad behaviour, including pre- emptive strikes to design
out this kind of behaviour. Regulators also need to take a well- founded and strong proactive
approach to the promotion and protection of good behaviour.
If regulators wait until bad behaviours dominate good behaviours, and guilt can be
proved, it may be far too late – the organisations which stick to high ethical standards may
be out of business, and if catastrophic or disaster incidents have not already happened, they
may now be inevitable.
In some cases, the best way forward may be a variant of Sophie’s ‘new deal’ approach to
the regulation of markets involving single players like NER, or many players like a mixture of
national and international energy companies seeking to develop a county’s energy sources.
The key may be organisations being required to develop enlightened plans which demon-
strate that they can be trusted whenever developments of very sensitive nature are involved.
One set of motives for using an explicit insurance premium approach to environmental
risk is making the premiums a function of government assessments of the risk involved, to
punish inappropriate safety and security planning, and to reward good safety and security
planning. Another is making sure that appropriate parties pay for the insurance of last resort
provided by governments. A third is making sure the market price of the energy or other
natural resources involved reflects the full cost of its production, net of any separate and
explicit subsidy or penalty adjustment. Governments often choose to subsidise some forms
of energy and apply penal taxes to others for sound reasons, to encourage renewable or low
carbon sources for example. But transparency about all aspects of these subsidies including
implicit insurer of last resort costs is arguably essential in an open economy.
If any country is interested in an EP approach of this kind to their own natural resource
areas, they may also need to protect their country from the implications of other countries
being abused by organisations no country should be prepared to tolerate. To protect their
own country, they may find it essential to spread the message that it is in everyone’s collec-
tive interest to spread this kind of practice internationally to collaboratively put the ‘bad’
international organisations out of business.
Nuclear power raises some special and important concerns in this area. In principle, it
should be feasible to further extend the kind of thinking just discussed in terms of railway
incidents, and then oil well blowouts, but the multidimensional framing implied by the
context may become exceedingly complex. The multiple- generation impacts, making future
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
576 Employing planning tools in practice
generations the losers in relation to present generation gains, and the huge uncertainty
about very complex sets of assumptions addressing human factor errors, computer system
failures and hostile cybersecurity attacks, low probability events like earthquakes with epi-
centres close to nuclear power stations, and so on may put well- founded trust beyond the
bounds of viable approaches directly linked to those of this book. Well- founded trust about
the use of potentially very dangerous technologies clearly needs a related form of thought.
It is crucial to remember that any issues which escape formal planning will require informal treatment – they will not go away. Very difficult issues associated with greater uncertainty and underlying complexity are involved, but not confronting them is not acceptable.
My personal concerns about nuclear power issues began with work in 1974 for Acres
related to Canadian and US nuclear power station seismic risk issues, when exploring the
framing assumptions associated with seismic events having an epicentre under or near a
nuclear power station. They grew during 1990s work for UK Nirex on nuclear waste stor-
age concerns, touched on in Chapter 7. The 1990s report on Ontario Hydro’s approach
to strategic planning underlying Chapter 8 and the drafting of Chapter 8 reinforced these
concerns. In the Canpower electricity utility context of Chapter 8, four points which need
thinking about in the light of this chapter’s approach are as follows.
One, both opportunities and risks associated with the nuclear generation of electricity are crucial. Leaving the long- term planning and futures planning which shapes a nation’s
approach to these concerns to market- driven private sector organisations offers no obvious
robust advantages and many obvious disadvantages. If nuclear power is a viable part of the
portfolio of sources of energy which a country needs, goals planning, futures planning, and
long- term planning ought to be a public sector decision- making domain as a default choice
unless there is a very convincing case otherwise.
Two, the Canpower framework developed in Chapter 8, extended to deal with Chap-
ter 9 concerns, seems the ideal default framework for all related planning, in my view. The
portfolio decision making required for long- term planning needs to address operational
constraints on both downside risk aspects and upside opportunity aspects. The risks have
to include nuclear contamination associated with both accidental and malicious attacks and
very broad political and military concerns, including risks associated with possible cyber-
warfare and more conventional warfare scenarios. Crucial intergenerational concerns need
direct and explicit attention. The opportunities are also important, and the balancing of dif-
ferent kinds of risks and opportunities affecting different generations is a public issue which
markets cannot possibly deal with.
Three, some aspects of operating nuclear power stations via private sector organisations
may offer cost– benefit advantages, but allowing incentives linked to profitability which is
not defined to reflect full costs including nuclear waste disposal and the risk of catastrophic
nuclear accidents does not look like a sensible proposition from an EP perspective.
Four, from an EP perspective what has happened to the UK energy supply industry since
privatisation is a national tragedy, with no viable case for arguing otherwise. UK press dis-
cussion broadly agrees with this view but for a range of very different reasons, driven by
political and economic theory assumptions which are highly variable. It is tempting to blame
the politicians and the party- political dogmas driving the politicians, but there is a reason-
able argument for suggesting that it is the political system which is broken and needs fixing.
Assuming that the system was ever up to the very long- term planning and associated decision
making required for this kind of issue, including unbiased full disclosure of the information
a well- informed public need, is clearly a contentious assumption. The operational features
of the democratic systems which we currently depend upon may need adapting. Fixing the
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Building well- founded trust example 577
very complex interconnected sets of problems which define current political messes ought
to be a key priority. We need to target all the best features of all alternative political perspec- tives, avoiding the worst. We need to exploit modern communication technology but avoid
simplistic assumptions about the wisdom of majority public opinions which may not be well
founded and may prove transitory. In contexts like the UK energy market, UK citizens seem
to be getting most of the worst features with very few of the best, although some organisa-
tions clearly prosper from this muddle, as should be expected. ‘It was ever thus’ some people
will argue, but change for the better is not impossible, and improvements are unlikely to be
found unless systematic searches are undertaken.
Some organisations may have comparable but different issues which do not involve cata-
strophic incidents in the same sense which are worth looking at in this chapter’s framework.
Some may be simpler than those confronted by NER. For example, cybersecurity for banks or other financial institutions might address a scenario range where S is a single metric
related to the severity of the problems which may arise, and TS is a transformation to yield
a $m equivalent, recognising the non- financial implications of what is involved. Lives lost
and serious physical injuries may not be involved, but customers may suffer in very impor-
tant non- financial ways which are very difficult to assess using a monetary metric. Explicitly
confronting some of the trade- offs using a variant of Sophie’s NER approach might prove
useful as part of addressing what are usually seen as reputational concerns. However, if the
non- financial implications are very diverse in nature, a single metric may not be effective,
and the complications may be comparable to those associated with oil well blowouts, per-
haps involving a need for even more separate metrics. The EP framework of this chapter
may prove a useful starting place, but important new concerns may need a lot of attention.
A striking feature of the Final Report on the Investigation of the Macondo Well Blowout by the Deepwater Horizon Study Group (CCRM, 2011), reviews of this report, and other
related literature, is that many of the same messages were clearly communicated earlier in
the context of a wide range of both similar and different but comparable contexts, and they
are still being communicated. For example, a 1990 report by Lord Cullen on the 1988 Piper
Alpha disaster in the North Sea which killed 167 workers was highly critical of manage-
ment’s training and safety culture – search the web using ‘Piper Alpha’ for access to copies of
this report and many other related and comparable reports. There are similar observations
in reports on Australia’s 2010 Montara platform blowout, the 2003 Columbia space shut- tle failure, the 1986 Challenger space shuttle explosion, and the 2011 Fukushima nuclear power station catastrophe. These kinds of incidents are not new. But collectively we keep
making the same kinds of mistakes, again and again. Any organisations directly involved
with this kind of issue needs to work out why and make a convincing case that they will not
contribute to maintaining this tradition of failure.
As a contemporary example of somewhat different issues, linked to the Ford Pinto dis-
cussion earlier in this chapter in terms of an automobile context, the development of the
Volkswagen diesel emissions controversy discussed in a Der Spiegel article involving collu- sion among VW, BMW and Daimler, which has been widely reported in the British press
(e.g. see Crossland, 2017), clearly demonstrates that too many people are still unaware of
basic aspects of the Ford Pinto controversy. In the German motor industry context, a mis-
judgement about the size of the tank needed for effective emission treatment seems to have
triggered wilful deceptions about emissions. This is clearly comparable to the misjudged fuel
tank design for the US Ford Pinto, as was the attempt to ignore the fatalities which would
result from software which cheated the emissions test procedures. The fatalities involved in
the emissions context are orders of magnitude larger than in the Ford Pinto case but not
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
578 Employing planning tools in practice
linked directly to the car purchaser, delayed, and not directly identifiable, raising significant
new kinds of complexity. There are also international regulation failures involved which are
very complex to address but arguably at the root of the problem because regulations which
encourage cheating by facilitating it are obviously incompetent. In addition, European gov-
ernments which encouraged the use of diesel failed to effectively address the trade- offs
between CO2 emissions, other noxious emissions, and fuel- efficiency concerns. The current
acceleration of plans to eliminate all automobile diesel use as part of a movement towards
electric and hybrid cars adds to the scope of what is becoming involved. Further still, at a
government level, all these issues are clearly interdependent with energy market concerns,
including electricity generation portfolio policy. As a consequence, it ought to be very clear
that no government can leave the markets to take care of all of this in the best interests of
their country’s citizens.
As an example of a very different kind, in 2017 the British press widely reported very
favourably on a Samaritan’s suicide prevention course taken by 15,000 Network Rail staff
and Transport Police, which was part of the reason for reducing the annual number of sui-
cides involving trains to 237, the lowest level since 2010 (e.g. see Purves, 2017). Particu-
larly interesting was the response of Network Rail staff who stated they had been sceptical
prior to the training but became strong advocates after the training because they had put
it into effect and were convinced that it had transformed their ability to prevent a needless
tragedy. As of November 2018, the suicide levels are still failing, railway passengers have
been involved as well as railway staff and Transport Police, and both the press and television
coverage is encouraging broader use of the basic ideas, beyond railways. This is a very good
example of the kind of imaginative bottom- up thinking which an EP approach ought to
encourage, within the coherent top- down framework discussed in this chapter. Early in my
1990s strategic review of Railtrack’s approach to safety, a feature which struck me as worth
special attention was the very large number of suicides involving trains every year. Suicides
were clearly very different from accidental fatalities in a number of respects, but apart from
the tragedy for the person killed and their family, they are extremely distressing for the train
driver involved, the emergency services who have to deal with the consequences may be
traumatised too, and there are serious knock- on delay impacts which may affect thousands
of railway passengers. The top- down focus of my review meant that the special nature of
this issue got no further attention, other than noting that a single T value independent of S
clearly had to deal with a very wide range of different kinds of circumstances, as well as the
numbers of fatalities involved. The idea of a Samaritan’s led course to address this specific
kind of incident would not have occurred to me given the focus of my concerns when ini-
tially drafting this chapter. But while this chapter arguably should have given greater empha-
sis to the power of a ‘big team’ approach which encourages people to contribute ideas like
this Samaritan’s training course, it is a very useful example to mention at this point, as a late
but hopefully useful attempt to adjust the balance of emphasis needed in practice. Within a
coherent top- down approach led by the board, all staff at all levels need to think creatively in bottom- up approach terms, with a focus on opportunities to make the system they are
collectively interested in safer and more secure, including directly involving customers and
the wider public. This is the case whatever the context – it is not just true of railway safety
contexts.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
10 Immediate and longer term ‘what needs to be done’ priorities
Starting with a comprehensive view of corporate goals
Moving on to a comprehensive view of futures planning
Developing and testing other planning capability areas
Developing specific planning capability aspects
Overall strategic level change management planning
11 Ongoing enhancement of strategic clarity and tactical clarity
Teamwork and collaboration as key capability- culture issues
Biased decision making as a ubiquitous concern
Contingency planning as a constant background concern
Corporate planning
Operations planning
Project planning
Process planning
Safety and security planning
Risk management
Lean and agile approaches
The HAT framework and Monte Carlo simulation approaches
Regulation and related issues
Politics and associated public pressure
Using ‘smiles’ as an indicator
Part 3
Further synthesising and reflecting
Chapters 10 and 11 contents indicating sections within chapters
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
All organisations seeking a significantly more enlightened approach to planning need to
plan the first few steps with a reasonable degree of strategic clarity about what the rest of
the journey might involve, including a broad understanding of both the opportunities and
the obstacles likely to be encountered along the way. It is important to begin by developing
change management plans which adopt an enlightened approach to clarifying ambiguity at
strategic and tactical levels. As part of this, everybody directly involved needs to become
comfortable with significant levels of the ‘right kind’ of ambiguity as quickly as possible,
with an enlightened intolerance of the ‘wrong kind’ of ambiguity, and a mindset which
embraces the need for significant ongoing change plus significant ‘learning by doing’ as
progress is made. Some crucial early changes may be quite radical and fundamental.
In Parts 1 and 2 ‘what needs to be done’ concerns were considered using a layered
approach to gradually building understanding which employed a sequence of topics and
areas of focus designed to explain what was involved as simply as possible. ‘What needs
to be done’ concerns are considered in this chapter in terms of an overview of key priori-
ties when an EP approach is initially introduced – precedence concerns ignored earlier are
addressed with an explicit focus on what needs to be done early on if there are no obvious
crises needing urgent attention. Longer term priorities are addressed when they are relevant
to immediate plans.
The next section addresses starting with a comprehensive view of corporate goals with
all inappropriate stealth assumptions exposed and revised to provide a dependable framing
assumption perspective. Following sections then build towards an overview of enlightened
plans for the management of change at a strategic level. Along the way inserted asides may
seem to depart from this book’s focus on planning for commercial organisations, to consider
some wider inferences about regulation and associated politics. However, there are impor-
tant implications from the perspective of most of the commercial organisations involved in
the tales of Part 2 and all comparable organisations you may have an interest in. There are
further important implications from the perspective of customers and employees of these
organisations plus other citizens who may be directly affected positively or adversely. In
terms of your overall understanding of an EP perspective, the interdependence of all these
perspectives is a fundamental aspect of what everyone needs to consider.
Starting with a comprehensive view of corporate goals
From the corporate planning perspective discussed in Chapter 8 and portrayed by Fig-
ure 8.1, clarity about all relevant corporate goals is the obvious place to start. All aspects of
all planning by an organisation involve a set of top- down flows of dependencies within the
goals– plans relationships structure as well as bottom- up interdependencies.
10 Immediate and longer term ‘what needs to be done’ priorities
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
582 Further synthesising and reflecting
A systematically developed comprehensive view of corporate goals from a top- down per-
spective ought to be an early high priority, because it is the logical place to start in terms of
key decision- making precedence relationships. The opening set of concerns is ‘Who owns
the organisation, and from the perspective of the organisation’s owners, what are the key
goals which clarify the vision or mission statement that the organisation should be focused
on achieving?’
The scope of this question means that those acting on behalf of the owners of the organi-
sation need to address complexities raised by all relevant interests of the organisation’s own-
ers, plus the interests of all other parties which need attention by the owners in an effective
manner. This includes accommodating the direct and indirect implications of other parties
impacting the pursuit of the owner’s goals in both positive and negative ways. Customers
and employees obviously need attention, but regulators and further interested parties like
bond funders may also prove crucial. Transforming potential adversaries into allies is part of
this, as explored in most of the Part 2 tales, Chapter 8, in particular.
To provide examples of what a comprehensive view of corporate goals might involve from
an EP perspective, consider the NER railway safety context of Chapter 9, then the other Part
2 example contexts, in a reverse ordering.
A comprehensive view of corporate goals for NER
A fair return on their financial investment in NER for shareholders was the obvious starting
point for an NER board level overall top- down view of corporate goals if the board adopted
the EP perspective developed in Parts 1 and 2. This observation was not made in Chapter 9,
because the focus was well- founded trust associated with safety, but the financial return goal
for shareholders had to be treated as inseparable for operational safety and security man-
agement purposes, fully integrated with minimising fatalities in a very general manner via
a ‘fatality efficiency’ concept. The need to perceive return for shareholders and safety and
security for those travelling on trains as inseparable was treated as a key framing assump-
tion, demanding a carefully structured approach to associated working assumptions, making
use of an S dependent TS concept. However, for the purposes of Chapter 10 and all overall
corporate planning purposes which go beyond safety and security, it is clearly important to
separate the shareholder return financial goal for NER, recognising all key interdependen- cies with all corporate goals, not just those associated with safety and security.
As an initial set of illustrative examples, preserving NER private sector status and avoid-
ing a hostile takeover might be seen as a component part of a decomposition structure for
return on shareholder investment goals, but it would be very important to view these kinds
of issues as separate interdependent goals with important additional nuances and interde-
pendencies. For comparable reasons, well- founded trust associated with good employee,
customer, and regulator relationships should be seen as directly related to both safety and
the return on shareholder investment goals, but for many practical purposes it would be
important to treat them as further separate interdependent goals, because of particular
nuances needing focused attention.
As a more specific example, a collapse in industrial relations leading to on- board railway
staff strikes might be linked to safety issues but strongly driven by other very different
issues, and responses designed to address all relevant issues might be very important. These
other issues might drive a collapse in passenger confidence in NER, also linked to service
failures like lack of punctuality, overcrowding, and excessive fare rises, which may all go
well beyond the strikes and the safety concerns, triggering hostile regulator reactions. Put
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Immediate and longer- term priorities 583
slightly differently, complex sets of closely interrelated goals may require treatment as if they
were separable in an iterative framework to ensure a comprehensive and balanced overall
treatment. That is what clarity efficient decomposition of goals and objectives is about when
very complex interdependencies are crucial – making effective use of separability to deal
with different relevant nuances of closely coupled issues in an effective and efficient manner.
This is an important underlying concern in the context of the tales about NER, Canpower,
WSL, Astro, and TLC, and it is a concern needing careful attention in all other commercial
organisations which may be of interest to you.
As an example, to put these issues in a realistic contemporary context, since privatisation
in the 1990s, the UK railway system as a whole has demonstrably failed to look after the best
interests of the railway travelling public. As of 2018, some of the companies operating trains
are experiencing serious levels of adverse public reactions to these kinds of interconnected
issues, with further interdependent issues involving Network Rail, the UK public sector
operator of the track and stations. The extent of associated concerns about the system as a
whole is reopening discussion of the 1990s privatisation of British Rail, the UK public sector
operator of the whole system until its breakup – initially into private sector components. As
of 2018, two- thirds of the UK railway system is back in public sector ownership, some of it
involving Network Rail’s UK public sector ownership of tracks and stations because of the
bankruptcy of the private sector Railtrack organisation it replaced, some of it involving non-
UK state owned train operators who are replacing UK private sector competitors in a very
complex franchise- based marketplace. One of the contributing drivers of current change
is that UK private sector train operators have a higher cost of capital than non- UK public
sector competitors, but some of the non- UK public sector operators are themselves about
to be privatised according to press reports. Another driver of current change is the role of
preferred risk– return trade- offs for train operators which are in part driven by ambiguous
regulation. According to the CEO of Network Rail (Carne, 2017), the recent switch from
off- balance sheet to on- balance sheet government funding for Network Rail will now lead
to third party funding of public sector assets and selling off public sector assets for reasons
not stated. This will obviously lead to further change with a direction of travel and implica-
tions which are debatable to say the least. The muddled intersection of party- politics and
market- driven commercial interests involved is a case study mess well beyond the scope of
this book, but it was worth one paragraph in this subsection, in part for reasons developed in
linked asides in the next two subsections, in addition to clarifying the nature of the complex-
ity which may be involved in formalising goals planning and then moving towards futures
planning.
In the absence of a safety crisis driven by two recent accidents, NER ought to begin the management of major changes in overall planning within their variant of an EP framework
like that outlined for Canpower in Chapter 8 by thinking about safety and security planning
goals at a strategic level within a comprehensive view of all the relevant interdependent corporate goals. NER might start with a mission statement like ‘The best railway travelling
experience in Europe, delivered to our passengers by railway staff who understand and care
about what good service means, including comfort, safety, punctuality and convenience at
a cost which is good value’. NER might then view ‘goals planning’ in terms of the devel-
opment of a suitable operational structure for formalising this mission statement in terms
of profits for shareholders linked to a service for travellers which is economical, punctual,
comfortable, and safe, weighted to reflect long- term railway passenger priorities as well as long- term regulatory and shareholder concerns, then move on towards a closely coupled
formal approach to ‘futures planning’, then move on to all other aspects of planning. But
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
584 Further synthesising and reflecting
in the Chapter 9 situation, given a crisis induced by a second accident was pressing, putting
many aspects of this starting point on hold for the moment, and picking them up as soon as
possible, was a reasonable approach. In the short-term, priorities have to adjust to immedi-
ate crisis events, although avoiding being deflected too much from longer term concerns is
very important.
Sophie’s recommended EP approach for NER was designed to address all these issues in
an integrated manner, but they were not discussed to avoid being distracted from the main
messages of Chapter 9. How this might be done could draw on both Canpower and WSL
goals planning ideas discussed earlier and in the next two subsections, without any interest
in ‘privatisation risk’ but a concern for ‘nationalisation risk’ comparable in some ways to that
faced by WSL, very different in others because of their national coverage as an integrated
monopoly like British Rail prior to privatisation.
A comprehensive view of corporate goals for Canpower
In Chapter 8, the need for a comprehensive treatment of corporate goals for Canpower
in the Figure 8.1 framework triggered the identification of a need to significantly shift the
focus for the way Canpower ought to serve its owners, who were owners/customers for
most purposes, as discussed towards the end of the chapter. Arguably, the need to address
the wishes of Canpower’s owners/customers in this way should have been obvious to the
board from the outset – without any need to be worried about privatisation risk, and with-
out any need to explore incompetent medium- term and long- term planning first. Indeed,
from the perspective of the citizens of the province owning Canpower, the provincial gov-
ernment officers representing their interests by determining who was on the Canpower
board ought to have insisted that those appointed to the board would look after the best
interests of Canpower’s owners in the way Chapter 8 suggested, explicitly holding the chair-
man of the board directly accountable for doing so effectively.
Furthermore, in the context of a clear strategy to move Canpower to a portfolio of
power sources which was predominantly nuclear in terms of new base- load generation,
the nuclear power safety and security issues addressed using the Chapter 9 NER frame-
work ought to have been a further explicit focal point in the Chapter 8 tale – for both the Canpower board and the provincial government. The complete absence of a board level advocate of the kind of EP approach to nuclear safety and security touched on in Chap-
ter 9 was arguably an important shortcoming in the composition of the Canpower board
as portrayed in Chapter 8. If Clive did not identify and respond to this problem fairly
quickly, Canpower’s failure to address it could have been viewed as an ongoing serious
failure on his part. Ultimately, this shortcoming was arguably attributable to the board
chairman and the government officers responsible for appointing him or her. But in prin-
ciple, and in practice, all the other board members were arguably also culpable for not
being fully aware of the issue or doing anything effective about it. To keep the Chapter 8
tale as simple as possible it was convenient to avoid in- depth treatment of the Chapter 9
concerns, to maintain focus on what was already a broad agenda. But in practice, overlook-
ing well- founded trust about nuclear safety and security when contemplating a portfolio
of base- load electricity generation sources dominated by nuclear power stations would not
involve enlightened planning on the part of any of those either on the board or responsible
for who was on the board.
As an elaboration of this point which also serves as a brief aside linked to the last and next
subsection’s asides, consider a broadening of the current discussion to address a consistent
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Immediate and longer- term priorities 585
treatment of nuclear power safety and security issues by all utilities in all countries in order
to serve the best interests of the all citizens of all the countries involved.
France is the only country I am aware of which has adopted an explicit national govern-
ment preference, over an extended period, for a portfolio of electric power sources involving
a largely nuclear base- load capability. This is in the context of a long history of signifi-
cant public sector involvement in electricity generation and distribution plus the design
and construction of nuclear power stations. Base- load generation is about 75% nuclear as
I understand it. In the past, the French approach has been viewed as a significant success
by many French people and others. However, a growing range of recent concerns suggest
opinions may be changing radically. In the US, the Three Mile Island incident early in the
development of nuclear power, and a sustained national preference for minimal public sector
involvement in all utilities, has led to a much more cautious approach to expanding nuclear
power generation of electricity. Arguably Ontario Hydro was looking to the French model
in the 1990s, with a very different policy after its plan for ten new nuclear power stations
was rejected. Prior to privatisation of the CEGB, the UK had a reasoned and balanced mid-
range preference exposed to public scrutiny. It is unclear what the UK policy has been since
privatisation, and its rationale is equally ambiguous.
In the context of the possible use of nuclear power, for all the countries and all the asso-
ciated electricity generation and distribution utilities plus their customers, whatever the
party- political preference issues, developing a clear understanding of what really matters
seems long overdue.
The international interactions are numerous. As one example, the Chernobyl incident was
catastrophic locally and a disaster in the broader geographic region, but there was also a wide
range of international repercussions, as diverse and distant as making Welsh lamb unsafe to
eat for a while. As another interrelated set of examples, current French nuclear power station
component safety problems are now linked to a large number of US nuclear power stations
built some time ago, as well as French, UK, and other European power stations currently
operating, under construction, or proposed. The Hinkley Point power station in the UK,
which started construction in 2017 with an estimated cost of £19.6 billion, has planned
French technology with significant Chinese involvement, potentially implying that if the
French technology fails, Chinese technology might be a fall- back option. My earlier under-
standing was that the present thrust of Chinese nuclear power technology development was
smaller modular nuclear power units than the current norm, the focus of effort of Siemens
in Germany when it asked me to provide advice on project risk management concerns in the
1990s, an interesting option. But my current understanding is a Chinese ambition to part
ownership of existing UK power stations of all kinds has finally raised the obvious concerns
about significant involvement in UK electricity supply of other countries with potential
geopolitical security implications, which arguably needed systematic and effective attention
a long time ago. Germany’s decision to shut down all nuclear power following the Fuku-
shima disaster in 2011 and switch to burning coal, which is available from domestic mines
but particularly polluting, was seen by many at the time as an obvious overreaction without
thinking through the implications properly. However, there is a growing perception (e.g.
see Pagnamenta, 2017), that current cost overruns and other problems could effectively
shut down the international nuclear industry, apart from clean- up contracts. Pagnamenta
cites the US example of a half- built VC Summer plant at Jenkinsville, SC, which already had
$4.7 billion spent on it, cancelled because revised estimates involved a completion date now
slipped five years and cost estimates requiring a further $7 billion, more than double the
original $5.1 billion budget. He links this to Hinkley Point in the UK, French, and Finnish
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
586 Further synthesising and reflecting
projects overrunning and uses the article heading Fukushima disaster is still radiating fallout nuclear industry wishes to avoid, opening the article by discussing the current $188 billion estimate for clean- up costs which could be interpreted as a very optimistic plausible mini-
mum – no one knows what the final bill will be.
A positive case for nuclear power may still be tenable. However, the issues which really
matter are extremely complex. They go well beyond the scope of this book’s focus, but they
are far too important to carry on failing to address them in a comprehensive manner which
every citizen with a stake in the implications can understand at an appropriate level.
A comprehensive view of corporate goals for WSL
In the WSL tale of Chapter 7, focused on project planning, Paul did not want to be diverted
from his primary concerns by involvement in the exploration of public versus private sector
issues which his EP approach triggered. In Chapter 7 exploring the contentious nature of
public sector ownership versus private sector ownership and related issues for WSL would
have been a serious distraction in terms of the central messages and focus of the tale of that
chapter. However, these issues were far too important to just ignore, and this subsection
provides some comments directly linked to the tale of Chapter 7 which build on both the
main messages and the asides in the last two subsections which should be useful for most
target readers in terms of consolidating an EP perspective on public and private ownership
issues.
In addition to illustrating a broadly defined EP approach to goals planning for any organi-
sation using WSL as a specific example, this subsection clarifies some of the issues which
ought to drive private sector versus public sector choices for any organisation. It uses what
might be seen as further asides focused on private versus public sector concerns which have
been integrated with the primary messages about the need for a comprehensive view of all corporate goals of all kinds in all organisations that has no blind spots in terms of components or interdependencies.
To put the WSL discussion of this subsection in context, in Chapter 8 it was argued that
privatisation of the nationalised UK electricity industry was a serious mistake, and provincial
government ownership of a comparable Canadian electricity utility should have been proac-
tively defended as a central part of the Canpower corporate plan because public ownership
was economically and socially desirable for the citizens of Canpower’s province. However,
in Chapter 9 a case was made for preserving the private sector status of a national railway
system operated as a coherent fully integrated system, because the operational decision-
making problems could be resolved effectively and the capability- culture liabilities could
be transformed into appropriate assets without altering the current private sector nature of
NER. The tale’s working assumption was public sector ownership of NER was not worth
even considering if the ‘national treasure’ nature of the organisation could be preserved.
The positions taken in Chapters 8 and 9 may have seemed inconsistent before reading
both chapters. The tales should make it reasonably clear why this is not the case, but there
is a need for a set of further key clarifications considered now.
In the Canpower context, the starting position for goals planning was ownership of Can-
power by the citizens of the province, but in the NER context the starting position for goals
planning was ownership by NER shareholders. This difference in perspective matters, with
political implications which are complex but inescapable, and it involves crucial working
assumptions within a framing assumption that allows for any form of ownership so long
as its implications are clarified. One key set of issues is driven by citizens having a vote to
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Immediate and longer- term priorities 587
determine the nature of the government setting the regulators’ agendas, the need for the
government and the regulators to show enlightenment, and the need for both private and
public sector organisations to heed the implications.
To build on this perspective, and explore how to exploit it by a private sector organisation,
consider in outline aspects of a ‘desirable future’ which might have been developed by WSL.
The basic purpose of this exploration is further illustrating the nature of a comprehensive
view of goals planning which is a suitable basis for moving into formal futures planning, but
there are wider implications.
Following his discussions with Paul, Richard, and Curt as outlined in Chapter 7, Frank
might have begun a dialogue with Richard, Curt, Michael, and Charles with a view to help-
ing WSL to secure lower cost bond funding in addition to a broader WSL agenda. This
might have led to further discussions triggering three closely coupled propositions.
The first proposition might have been a Local Assets Company (LAC), set up by WSL as
a separate company. The purpose of this LAC might be to manage bonds used to fund water
treatment and pipeline distribution networks, sewage collection and treatment systems, and
all other bond funded WSL assets. These bond funded assets might include reservoirs used
for recreational purposes by local residents, plus recreational areas and parks made available
to local residents which are an integral part of the property used to accommodate water stor-
age, treatment, or extraction works or sewage works. This LAC concept might be viewed as
a ‘WSL localisation’ concept to link WSL to the local public who were also their customers.
The LAC board might include local people with a reputation for good local citizenship. It
might provide bond rate long- term loan funding secured by LAC ownership of the assets if
WSL becomes bankrupt, directly comparable to a mortgage for these long- life ‘local assets’
which local investors might see as safe and secure locally beneficial investments that they
would like to have a financial stake in for sound financial reasons as well as possible ‘feel
good’ reasons. Genuine altruism might or might not be involved, and fostering local owner-
ship of local assets in this form need not be concerned about enlightened self- interest serv-
ing the same goals as genuine altruism. The corporate design, including board membership
structure, might be very carefully planned at the design stage and subsequently controlled
by WSL, but this corporate design, and the associated broader approach taken by WSL to
the LAC concept as a whole, might aim to make local people feel that they had a measure of
ownership and control over a set of local assets which was on balance ‘as good as or better
than more conventional public ownership’. It would certainly not be ‘public ownership’ in
the ‘nationalisation’ sense, but from a local population perspective it could provide better
focus on local concerns.
One key objective might have been to secure funding for WSL long- term assets at the
lowest feasible cost, not only Frank’s central concern but also a local population concern,
because it would help to keep water and sewage charges low. Frank might have been ada-
mant that bond funding from outside the region should be part of the LAC proposition if
the market made that significantly cheaper and no serious adverse effect for local people was
involved, not only consistent with the central focus of his role as Finance Director as well as
his general governance role, but also a concern for local people in terms of their interest in
low cost water and sewage.
A second key objective might have been protection against nationalisation and privatisa-
tion threats plus exploitive non- UK ownership of local assets by providing the population
served by WSL with a better alternative, not only Richard’s central concern but also a local
population concern. Richard might have been adamant that they should not just impress the
regulators by their LAC proposition. WSL should aim to make their regulators firm allies
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
588 Further synthesising and reflecting
in terms of a constructive long- term relationship, without losing sight of the constructive
tensions to be managed, consistent with the central focus of his role as Regulation Director
as well as his general governance role.
A third key objective might have been enhancing customer relationships by giving the
local population a new form of direct input to the management of assets like reservoirs and
parks used for public recreational purposes, Curt’s central concern, and a local population
concern. Curt might have been adamant that a ‘big team’ approach to WSL customers via
the LAC proposition had to be based on a genuine concern for collaborative relationships
clearly demonstrated by a willingness to spend money and effort making sure the LAC
worked effectively from the perspective of the best interests of all local people, consistent
with the central focus of his WSL board role as Customer Relations Director as well as his
general governance role.
Michael and Charles might have both been adamant that the overall best interests of WSL
and all its shareholder owners should be effectively served, consistent with their central and
general governance WSL board roles.
However, meeting reasonable demands from all five directors and aligning all five sets
of concerns, plus those raised earlier by Paul and Ollie, ought to have been feasible. They
ought to have been able to reach agreement on a collective ‘revealed preference’ which
balanced their concerns in a way the WSL board as a whole could support. Reducing the
risk of nationalisation in the usual sense, or separate privatisation of some of the assets they
owned (like reservoirs), or used (like rivers), ought to have been a welcome and very impor-
tant spinoff, but these spin- offs could have been seen as simply the results of enlightened
collaborative top- down planning at the board level by a set of directors with interrelated
concerns and a willingness to work towards corporate revealed preferences which balanced
all relevant concerns. Some negotiated version of this first proposition ought to have been
a win– win package.
The second proposition might have been a National Asset Company (NAC), a further
separate company set up and led by WSL. This NAC concept might have been a ‘WSL
nationalisation’ concept, designed to link WSL to the public in the rest of the country to
the extent that doing so might prove mutually useful. Its central purpose might have been
developing a water distribution system and wholesale water supply market between different
adjacent regions served by different water companies, starting with WSL’s immediate neigh-
bours but perhaps growing to complete national coverage. The corporate design including
board membership structure might have been carefully planned and controlled by WSL, but
mutual benefits for different regions might have been an explicit goal so that directors of
collaborating water supply utilities in other regions might have been asked to serve on the
board. Some regions are relatively rich in low cost water sources but have a limited customer
base, while other regions are relatively poor in terms of low cost water supply with excessive
demand. This implies that a degree of collaboration with other water companies ought to
be ‘socially optimal’ from a UK national perspective, and ‘commercially optimal’ from the
perspective of the private sector water suppliers. The goal for WSL might have been WSL
advantage in cost or revenue or water- quality terms, to improve all aspects of WSL perfor-
mance, plus further protection against nationalisation in the usual sense. To achieve these
goals the benefits of collaboration would have to be shared with water suppliers in other
regions, potentially benefiting the populations of these other regions as well, perhaps the
whole country benefiting on a regional basis.
The third proposition might have been a quiet and unobtrusive lobby initiative that
was aimed at making Ofwat and the other water regulators indirect SWL allies in terms of
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Immediate and longer- term priorities 589
persuading them to impose regulatory rules which would stop all of WSL’s water and sew-
age industry competitors using practices which were profitable for those competitors but
unhelpful for their customers or other citizens of the UK. This might be based upon WSL
adopting a quiet but effectively managed strategy characterised by:
The good ought to be able to drive out the bad if the good is managed effectively and
creatively. WSL needed to ensure that they were good enough to drive out or takeover
all competitors who might negatively affect successful service provision for all WSL
customers, ensuring that this objective was closely coupled to a good return for WSL
shareholders. Being ‘good enough’ to achieve these goals must include a constructive relationship with their regulators, who must be motivated to provide an appropriate
level playing field.
As part of this strategy, WSL might have indirectly tried to persuade the regulators to
use the WSL business model as a regulatory aspirational target for the water and sewage
industry, with a view to WSL making sure that WSL had ‘a level playing field’ that it could
perform on better than their competitors. That is, WSL might have taken a strong and overt
partnership approach to its own customers and to adjacent water companies, making their
water and sewage services cost as little as possible as part of an approach based on short- term
profit rates which were modest but fair for exemplary service, with a view to appropriate
long- term profits driven by outstanding service provision provided in a cost- effective man-
ner. They might have made this business model so attractive the regulators would look bad
or even negligent if they did not actively encourage its use by other water and sewage com-
panies, actively contest and constrain WSL competitors which tried to exploit their custom-
ers. This would protect WSL’s own market position and long- term profitability, and it might
allow expansion of WSL territory where appropriate in the longer term. It would not be
sensible for WSL to advertise or promote this strategy to other water and sewage companies
or make it obvious to anyone else that it was adopting this approach. But it might be very
important to get regulator support, with a carefully prepared case and demonstrable benefits
exposed to its regulators in a carefully planned manner.
For example, WSL might begin by explaining to Ofwat why its LAC approach was good
for WSL customers in terms of lower costs but also make sure the regulator clearly under-
stood that the LAC approach generated UK taxes on income paid to UK LAC bond inves-
tors, while some water and sewage companies not using this model were getting lower
capital costs by just avoiding UK tax. WSL might then argue ‘surely the regulator could see
the political advantages of stopping a practice that would let “the bad drive out the good”
in a classic market failure sense before the press got hold of the idea and before some of the good companies were driven out by the bad, never mind waiting until all the good compa-
nies had been eliminated’. WSL might carefully avoid any sense of direct threats but make
sure the regulators understood that if they did not act on these ideas WSL would have to
ensure that the government and the public understood the implications in terms of the lack
of competence of the regulators as WSL saw it. This might be underscored by very publicly
seeking tax exemption for LAC investors involving or comparable to ISA (zero income tax)
status for UK taxpayers on the grounds that both the regulator and the government would
otherwise be condoning a very unbalanced and unfair tax treatment for UK- based LAC
investors relative to foreign investors.
This particular version of a desirable future for WSL might not attract favourable atten-
tion from post- 2020 UK water and sewage companies, UK governments, or UK regulators,
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
590 Further synthesising and reflecting
and its development in this book is not aimed at the water and sewage industry in particular.
However, it provides an example which illustrates how a private sector organisation with a
local natural monopoly might, with appropriate regulator encouragement and a broad view of goals planning, look after its own commercial interests in an enlightened manner. The
intention is illustrating an overall approach capable of delivering outcomes which approxi-
mate to a ‘social optimum’ comparable to or better than a more conventional public sector
approach or an inappropriately regulated private sector approach. The details are purely
illustrative.
An EP framing assumption is that a ‘social optimum’ may exist in a clear and well- defined
form in the judgement of each member of the relevant set of citizens, but even if this is the
case, it may be different for each citizen, so any operational overall ‘social optimum’ concept
involves appropriate working assumptions about aggregated revealed preferences, like an
overall consensus given appropriate information and consultation processes. Ambiguity in
some peoples’ minds just adds to the ambiguity associated with revealed preferences for any
aggregate of different peoples’ views.
Assuming a ‘social optimum’ exists in this WSL context means assuming that ‘the revealed
preferences of the local population and other local users of WSL services would favour this
choice, relative to all alternative feasible options, provided they were informed about the
issues and then consulted in an effective manner, and provided there were no compelling
reasons not to serve their wishes in favour of other commercial or political interests’.
There is a good case for arguing that this kind of desirable future for WSL would cer-
tainly have been achievable for all UK water and sewage utilities pre- privatisation of the
UK water and sewage industry, if the UK government had considered the best interests of
consumers as the government’s primary objective in an enlightened manner when structur-
ing the outcomes and mandating the regulators. This would have required treating the
interests of commercial lobbies and political dogma as secondary concerns, to be managed
and constrained, perhaps a non- feasible expectation for current governments. However,
future governments might consider this kind of approach if a large enough group of people
understood the issues and took effective political action.
To finish the WSL discussion on a contemporary note, an October 2017 article with the
headline ‘Water Firms Risking their own Demise’ (Lee, 2017b) by the Times Industrial Editor reporting on ‘a withering attack on the sector’ by the industry’s regulator ends with
‘Asked afterwards whether she had made the comments because she was leaving the sector
for a job with BT, Ms Ross replied “I said it because it needed saying and the sector needs
to hear it. The sector is treating (the threat of nationalisation) as if it was an exogenous risk,
as if there is nothing they can do about it” ’.
Both private and public sector organisations can be ‘national treasures’, ‘national trag-
edies’, or anything in between. Taking the NER, WSL, and Canpower approaches together
arguably illustrates how both private and public sectors could deliver a balanced approach to a ‘socially optimal’ choice which is very different from the common practice private versus
public sector choices portrayed as the only choices by those with narrow political beliefs
which push for either a ‘free market with minimal regulation’ option or a ‘full lock- stock
and barrel traditional nationalisation’ option. From an EP perspective, one key working
assumption is that relevant regulators, fully backed by their controlling governments as
necessary, will act in a competent manner in the best interests of the customers for the
services and other citizens involved. This will be the case for private sector industries and
public sector industries in the traditional sense, including NER. It will also be the case for
novel private and public sector organisations, to accommodate organisations like the LAC
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Immediate and longer- term priorities 591
and NAC which WSL might set, and the form Canpower might transform itself into plus
its associated novel partnership ventures. A public or private sector ownership starting posi-
tion matters. The competence of those responsible for goals and futures planning within the
organisations directly involved is also crucial. But the citizens who are customers of all local,
provincial, national, and international organisations should press their governments for a
social optimum defined by their interests in a suitable revealed preference form. Unyield-
ing resistance from those determined to limit discussion to traditional extremes might be
interpreted as evidence of stealth assumptions driven by goals not shared by the majority. It
is the majority whose interests ought to prevail in an open democracy, given visible working
and framing assumptions most people would support.
This book is not about definitive answers to very complex social optimisation issues with
highly emotive political, ethical, and economic dimensions. It is about the conceptual and
operational tools that might be used by private and public sector commercial organisa-
tions to seek better decision making within their planning frameworks. But there are wider
implications which could be pursued elsewhere, and encouraging those with an interest in
doing so might prove useful for all of us. For example, a novel framework developed in this
spirit might provide a more desirable future for the UK rail system, addressing it from the
perspective of rail travelling customers plus other UK citizens. This approach might use a
nationalised organisation structure which was a variant of Canpower plus British Rail as the
model or a privatised organisation which was a variant of NER plus WSL or a very different-
looking hybrid based on features of an EP approach plus other relevant ideas. Whatever
the approach adopted, it would have to be tailored to the context, including recent history.
When considering these broader implications, a saying attributed to Winston Churchill is
Some people regard enterprise as a predatory tiger to be shot. Others look on it as a cow they can milk. Not enough people see it as a healthy horse, pulling a sturdy wagon.
Elaborating Churchill’s notions in terms of public enterprise as well as private enterprise in
a post- 2020 context, the key issues are ‘Whose wagons are we talking about?’ and ‘What
is the best way to look after the horses pulling all the “good” wagons, while appropriately
ensuring that the horses pulling all the “bad” wagons put their efforts to better purpose?’
This seems to be a key opportunity and an important challenge worth taking very seriously.
Doing so successfully would obviously not be a simple matter. It would need to embrace
complex long- term cultural issues and shifts in people’s value systems, as well as enlightened
approaches to short- term ‘carrots and sticks’ which go well beyond crude financial incen-
tives or prison sentences for serious misdemeanours. Failing to try to do so, or failing to
even understand why trying to do so might be a good idea, can be seen as fundamental
social risks which involve lost opportunities that really matter. Lost opportunities present
particular difficulties because their perception is never easy, with different people inevitably
seeing them very differently. But a focus on avoiding lost opportunities provides a usefully
positive perspective.
A comprehensive view of corporate goals for Astro
In Chapter 6 Martha was operating at the level of a marketing manager for a regional office
of Astro UK, part of the US- based Astro Inc, an international organisation comparable to
IBM. Martha was sensitive to the national and international competition issues as well as
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
592 Further synthesising and reflecting
their local implications, but it was convenient for the purposes of the tale to limit the focus
to the regional situation. Encouraging other regional offices of Astro to move in the same
direction that Martha was advocating with coordinated variants of comparable approaches
is one obvious potential component of the overall corporate goal sets needed by Astro Inc
and Astro UK corporate goals planning. However, viewed from the top down in overall
corporate goals planning terms, Astro Inc and Astro UK face tasks which have to visualise
Martha’s concerns as relatively low- level issues. The high level issues at international and
national levels are very different from those just discussed for WSL, Canpower, and NER,
but the basic concerns are comparable and some of the EP conceptual and operational tools
are clearly relevant. It would not be helpful to speculate here on how an organisation like
Astro Inc might set about structuring its goals planning, as a basis for futures planning in the
Chapter 8 sense. However, successful reinvention of itself by IBM on more than one occa-
sion since the 1950s is a good illustration of what can be done in practice. The complexities
IBM has faced were obviously very different from those associated with NER, Canpower,
and WSL, but the degree of difficulty was arguably comparable or greater, the required
creativity was arguably significantly greater, and the scope for using all the EP toolset along
with many other toolsets should be reasonably clear.
A comprehensive view of corporate goals for TLC
In Chapter 5 the need for a comprehensive view of corporate goals was triggered by Nicola’s
low level tactical analysis of an optimal order quantity for lawnmower engine speed controls.
Nicola’s efforts initiated a sequence of propositions with some features common to those
just discussed for WSL. Ajit’s contribution provided a fairly comprehensive starting point for
TLC goals planning, and the propositions generated by the others provided a fairly compre-
hensive starting point for related futures planning. However, even a very small family- owned
organisation like TLC really ought to begin with the kind of nuanced initial goals planning
associated with Ajit at the tale’s conclusion and have in mind the generation and shaping of a
comprehensive portfolio of possible propositions prior to moving fully into futures planning
along some of the lines explored for Canpower in Chapter 8.
A comprehensive view of corporate goals for all organisations of interest to you
The defining characteristics of most organisations of interest to you are unlikely to directly
match any of those explored in Part 2, and specific aspects of each context will significantly
drive the most effective and efficient approach for any given specific organisation. General
advice at a detailed level for all contexts is not a viable proposition, and even overview advice
is a very difficult proposition, because a ‘one size suits all’ approach is seriously inappropri-
ate. But a top- down UP perspective making use of some variant of a Figure 8.1 perspective
and many of the interdependent EP conceptual and operational tools does seem universally
useful. Specific process variants of the soft OR variety designed to explore and help struc-
ture the ‘mess’ of interconnected ‘problems’ associated with strategic issues may be helpful,
viewed as particular tools within a UP approach. Testing the robustness of early portrayals
as rigorously as possible is clearly essential, as a core part of the process adopted. The way
the Chapter 8 tale identified Canpower’s need to reinvent its approach to avoid privatisation
but missed the full implications of the Chapter 9 nuclear power issues as explored in this sec-
tion may be a helpful reminder of the need to keep pushing the scope of what is addressed,
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Immediate and longer- term priorities 593
avoiding crucial blind spots. The way this section addressed the privatisation and nationalisa-
tion concerns for WSL may also serve as an illustration of one way to push the scope of what
is addressed, bearing in mind that organisations which are more closely related to Astro or
TLC will have very different issues needing this kind of creativity in a form relevant to their
specific contexts.
Moving on to a comprehensive view of futures planning
Effectively linking goals to plans in operational terms needs an operational form for all five
of the categories of planning within the four planning horizons discussed in Chapter 8 and
portrayed by Figure 8.1, adapted to the context of the organisation involved. This opera-
tional framework will have to include a gateway structure for approving corporate changes
to link bottom- up and top- down proposition generation and overall management of change
decision making. This, in turn, needs intersecting operations planning and project planning
as outlined in Chapters 5, 6, 7, and 9. Testing the robustness of what has been achieved
during goals planning, and then moving on to a comprehensive view of futures planning,
with all other aspects of planning to follow, is the obvious next step.
The boundary between futures planning and long- term, medium- term, and short- term
planning was defined in Chapter 8 in terms of currently feasible and non- feasible options.
There is a particularly important need to avoid approaches which fail to distinguish futures
planning from the long- term planning of currently feasible options. Considerable clarity
enhancement is provided by using explicit contingency plans to embrace opportunities that
might arise if futures planning provides new opportunities by offering new feasible options
in the long- term planning framework, with follow- on medium- term and short- term con-
tingency plans. Futures planning focused on designing desirable futures to aim for, given
a comprehensive view of corporate goals, then planning to make feasible any key potential
opportunities that were both desirable and open to cost- effective research and develop-
ment or other transition mechanisms, needs explicit separate consideration. The realisation
of a transformation to feasibility can feed into long- term planning using prior contingency
plans when relevant. Successful realisation of a feasibility transformation might also have an
impact on medium- term or short- term planning, and awareness of this possibility could be
important follow- on concerns.
Arguably a crucial and central aspect of early ‘futures planning’ for all organisations is
ensuring that all the key corporate capability- culture assets needed to translate corporate
goals into futures plans are in place, and any key liabilities are dealt with. The questions
raised are not really separable from those associated with the last section, and the interde-
pendencies are complex. For example, in Chapter 8, Carl’s recommendations to replace
Larry and his corporate planning director role with several board level posts and a new asso- ciated management structure was a crucial first step towards a radically different approach to
corporate planning. It was directly linked to a broadened view of corporate goals planning
and a broadened view of futures planning, with toolset, skill set, and mindset implications
for all the key players involved. As another example, in the much simpler TLC context of
Chapter 5, Ajit indicated the need for new people with appropriate skill sets in addition
to new skill sets for existing staff. But even in this relatively simple context there were
complexities of importance. As an illustration, picking up on the inherent interdependen-
cies between marketing and product innovation, TLC needed to move away from a highly
production- led perspective towards a more market- led perspective which embraced this
interdependence.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
594 Further synthesising and reflecting
Because the nature of the frameworks needed for extensive formal futures planning is
highly dependent on the context, a detailed exploration of futures planning in Chapter 8
would not have been very instructive for other contexts, in addition to being a distraction
in terms of the main messages of the Canpower tale. In this section it was assumed that the
most appropriate approach to get you started in the contexts of interest to you was a very
limited list of observations, as follows.
First, in the rapidly changing environment that most organisations now have to live with,
effective futures planning needs to be seen as central to corporate survival as well as corpo-
rate prosperity.
Second, it is crucial not to muddle futures planning with long- term planning or follow- on
medium- term and short- term planning. Exploring what is not currently feasible but might
be made feasible makes futures planning very different from follow- on long- term, medium-
term, and short- term planning.
Third, the span of issues futures planning could usefully address might be very wide, with
considerable depth needed in many areas. Technology based and driven organisations like
Astro need to confront an enormous span and depth of complexity. Manufacturing based
and driven organisations like TLC might anticipate relatively simple issues, even if the scale
of their operation is quite large.
Fourth, there is a sizable relevant literature, but it needs synthesis in the specific contexts
of interest to you with considerable care taken to clarify the framing assumptions and work-
ing assumptions made by the authors of the approaches being discussed.
Developing and testing other planning capability areas
A number of other planning areas need clear separation from goals and futures planning in
conceptual and operational terms, plus careful early attention in terms of requisite capabili-
ties. Going beyond emphasising a few key issues in this section would not be very produc-
tive, but you may find the following points worth reflecting on if they have not received
your attention already.
Long- term, medium- term and short- term corporate planning capability
The ability to translate the results of goals and futures planning into long- term, medium-
term and short- term plans in terms of top- down and integrative corporate planning capabil-
ity is clearly an early essential. In particular, having people with the requisite toolsets, skill
sets, and mindsets to begin developing and testing appropriate long- term and medium- term
planning within the corporate planning function is an early priority, perhaps just behind the
need for people to revise goals and futures planning approaches, but probably best thought
about at the same time. Short- term planning is usually less of a concern.
Developing and testing project planning capability
As soon as any organisation starts to address the management of change, project planning
in the broad Chapter 7 frameworks becomes directly relevant. The related capability- culture
issues are core concerns. Project planning capability is obviously essential in terms of detailed
change management plans as and when tactical concerns start to emerge in all different spe-
cific areas of an organisation. But it is particularly relevant at an overall strategic level from
the outset. All change management is usefully perceived as a breed of project management
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Immediate and longer- term priorities 595
dealing with very high levels of ambiguity as well as very high levels of other aspects of
uncertainty, which implies some very big potential opportunities and risks. Ensuring those
involved have the requisite toolsets, skill sets and mindsets is obviously a crucial issue.
This point was made in an initial low- key way by Ajit in Chapter 5, and it was developed
significantly by Paul in Chapter 7. It suggests a very early focus on hiring someone with
Paul’s capabilities if an organisation does not already have a seasoned Paul equivalent or
finding a suitable ‘prototype Paul’ and ensuring well- informed access to appropriate training
along with appropriate mentoring.
Developing and testing operations planning capability
Even if no immediate changes in operations management areas is the initial working
assumption, ongoing operations need planning in the sense of a generalisation of the areas
addressed in Chapters 5, 6, and 9, and integrative operations planning concerns as addressed
in Chapters 7 and 8 are also important. Again, the availability and quality of directly related
toolsets, skill sets, and mindsets should be the central initial focus. For example, are the
inherent interdependencies between demand side marketing, product design, production
technology, and supply chain concerns being addressed effectively, including all relevant
safety and environmental concerns, and the way operations management concerns in par-
ticular areas feed into corporate planning and management of change project planning?
Developing specific planning capability aspects
A number of specific planning capability aspects may need early focused effort. As for the
last section, going beyond emphasising a few key issues in this section would not be very
productive, but observations in the following subsections should be worth you thinking
about in your own context as a starting point.
Development of key corporate tools and initial use in a coordinated manner
The priority ordering of corporate, operations, and project planning concerns will vary with
the context, and any one of these three areas might be judged the most important in terms
of the need for immediate change. However, all these areas of planning need integration,
and a coherent approach to all new frameworks, processes, and other tools need early atten- tion in terms of overall coordination to achieve compatibility and coherence.
Project planning as discussed in Chapter 7 uses four frameworks, the four Fs for project
planning:
1 a lifecycle framework,
2 a seven Ws framework,
3 a goals– plans relationships framework, and
4 an SP for projects framework.
It is useful to see these four frameworks plus the embedded models as the project planning
versions of a four Fs ‘corporate tools’ concept which has related variants in operations and
corporate planning contexts.
For example, the project lifecycle stage structure equivalent for corporate planning is the
goals, futures, long- term, medium- term and short- term planning structure. In an operations
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
596 Further synthesising and reflecting
planning framework like that of Chapter 6, Martha’s three bidding stages are the equivalent
of project lifecycle stages until the bid is won or lost, but other equivalents may be involved
in different operations management contexts.
As another example, the way the seven Ws framework was adapted to mesh with the
lifecycle framework in Chapter 7 needs an equivalent treatment in the Chapter 8 corporate
strategy formation structure of Figure 8.1 for goals, futures, long- term, medium- term and
short- term planning. There will not only be important commonalities but also important
differences with a nature dependent upon the context details. In the operations planning
framework of Chapter 6, the seven Ws needs adapting to Martha’s three bidding stages in a
reasonably obvious way, but other adaptations will be required in different contexts, and the
differences will be highly context dependent.
As a further example, the way the SP for projects concept was adapted to mesh with the
goals– plans relationships via the seven Ws frameworks plus the lifecycle framework in Chap-
ter 7 also needs equivalent adaptations in a corporate planning context, and the notion of a
common overall framework for project planning and operations planning explored briefly in
Chapter 7 needs further generalisation.
Chapter 8 suggested a ‘Processes Director’ for Canpower, who might be responsible for
these kinds of issues at the board level. However, a generalised version might be someone
with appropriate skills and authority having ‘corporate EP tools’ responsibility, covering the
coordinated development and initial use of all the new conceptual and operational tools
required by the organisation, with a job title suitable for the specific organisation and its
history of related positions. An internal consulting group with EP implementation planning
and supporting roles might be one useful way to see this role.
Some organisations may need help from external consultants, as well as appointing new
staff to new roles and helping existing staff to understand the direction of change plus
acquiring new toolsets, skill sets and mindsets. It is important to avoid letting action on this
front get delayed because new territory is involved and the uncertainty associated with dif-
ficult choices is very high. But it is also important to avoid expensive false starts which take
the organisation in the wrong direction. Furthermore, there is a need for ongoing vigilant
attention to a constructive tension between taking externally provided advice and develop-
ing internal capability.
As with all early management of change, effectiveness rather than efficiency needs to be
the early emphasis. Initially doing the right things is more important than doing them right.
Getting approximately correct answers to the right questions in an appropriate timeframe is
another way of looking at the same concern.
Addressing trade- offs between precedence relationships and pressing priorities
The ordering of concerns addressed in terms of the trade- offs between priorities associ-
ated with major or intermediate crises (like a second serious railway accident), and impor-
tant precedence relationships (like being clear about all key objectives before investing heavily in dependent plans), needs coordination by someone, or a small team, with signif-
icant insight and authority from the outset. This is arguably as crucial at the outset as any
other urgent concerns. This aspect of managing change was implicit in Part 2 chapters.
The implications of key choices may be very uncertain for complex reasons. Predomi-
nantly informal planning by people with the requisite insight and sound instincts may be
what is needed in many contexts. Formal planning approaches may not be appropriate,
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Immediate and longer- term priorities 597
and directly related experience may not be relevant even if it is available, depending on
the circumstances.
Developing and implementing change incrementally
In the Astro tale of Chapter 6, Martha’s approach to involving Trevor has a number of fea-
tures worth testing for relevance in your organisation’s context. For example, using one par-
ticular development case example, with a carefully selected player like Trevor contributing to
development, testing, demonstration, and roll- out into other areas, illustrated the effective
capture of opportunities and knowledge in a number of areas which Trevor was particularly
well placed to contribute to. His role in the incremental roll- out planned by Martha might
be particularly relevant in your organisation.
In the 1990s, I developed a consultancy relationship with NatWest Bank which lasted
several years, helping two of their staff develop what they referred to as a ‘benefits risk man-
agement’ approach for a very large information system development programme involving
the computerisation of all NatWest branch banking in the UK. The benefits management
framework NatWest used had been developed with consultants based at Cranfield University
advocating a ‘balanced scorecard’ view of business cases for change, driven by a portfolio
of benefits. The Cranfield consultants, in turn, had drawn on the Harvard- based authors
Kaplan and Norton (1992, 1993a and b, 1996). The two NatWest staff I worked with used
the balanced scorecard starting point to overlay a prototype of the Chapter 7 approach then
available to develop their prototype specific process for managing uncertainty. One of the
key contributing factors to the significant overall success of their efforts was rolling out the
initial branch implementation with a very small number of carefully selected branches, then
using the branch staff involved to refine the approach before rolling out a revised approach
to a slightly bigger set of branches, only moving on to a general roll- out when all the early
concerns had been resolved effectively.
This illustrates a practical example of a way to focus on being effective first, then moving
on to efficiency concerns as experience was gained. It also illustrates one way to move from
a focus on a limited number of metrics like cost or profit, and associated uncertainty, to a
broader set of objectives portrayed as ‘benefits’ which may not have simple metrics.
A specific interest group of the APM (Association for Project Management) and many
others have developed these ideas since. From an EP perspective, it is clear that trade- offs
between benefits have implicit if not explicit shadow price interpretations linked to the TS
concept explored in Chapter 9, possibly an area worth further exploration.
This was long before NatWest became part of the Royal Bank of Scotland (RBS), part of
a series of developments which ended in UK government rescue to prevent UK financial
market meltdown during the 2007– 2008 crisis. As part of the run- up to these problems,
both of the NatWest people I worked with moved to other employers. RBS might have been
well advised to have retained and made effective use of them, employing a generalised vari-
ant of their own 1990s NatWest benefit risk management approach at the RBS board level.
Testing actual outcomes and learning from experience, recognising limitations
Testing actual outcomes against expectations and learning how to do better next time was
clearly part of the strategy adopted by Astro’s Martha as discussed in Chapter 6 and Nat-
West’s benefit risk management developers as discussed in the last subsection, but there
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
598 Further synthesising and reflecting
are wider implications which also need addressing. For example, the way BP used formal
assessment of how cost and duration outcomes compared with plans in the early 1980s was
a robustness test and a learning process at a corporate level, and it was both useful and effec-
tive in terms of what it addressed. But what it did not address, which with hindsight might
have been extremely useful, is the extent to which using simpler approaches once lessons
have been learned may lead to overconfidence, and the extent to which ongoing robustness
testing ought to be wide enough and deep enough to test for potential issues which are ‘off
the radar’ because they have not yet been encountered.
The Chapter 9 discussion made a point of avoiding bias and blind spots associated with
too much focus on what happened in the past and not enough attention to very different
things that might happen in the future. The very long return periods of catastrophic inci-
dents make it particularly obvious that the relevant future may not resemble the past. But in
a very different context, the early parts of Chapter 3 made similar points about the need to
treat the past as a very limited lens for thinking about the future. All intermediate chapters
involved related thinking because this issue is a very general concern. The management
of change needs leaders who are very clear about the importance of learning from experi-
ence and the pressures imposed by understandable concerns about evidence- based decision
making but who are also very comfortable about creative planning for uncertain issues of
importance which have not been encountered in the past. This kind of competence is not as
common as it needs to be.
Key opportunities and risks associating with outsourcing
In Chapter 5, TLC’s in- house manufacturing operation used lawnmower engines and other
components which were outsourced, but the possible option of outsourcing the manufac-
ture of complete lawnmowers or other garden machinery was new territory. In Chapter 6,
outsourcing some of the current roles of Sian’s Systems group, including moving existing
Astro employees into new companies and current strategic partners, was also new territory.
In Chapter 7 Paul suggested WSL keep strategic design in- house but continue a policy of
outsourcing detailed design to their contractors, with no radical changes. But a much more
flexible approach to construction outsourcing was proposed. In Chapter 8, outsourcing
a range of current in- house activities was a central aspect of the broadened approach to
goals and futures planning developed by Carl. In Chapter 9, continuing to rely on external
maintenance contractors and moving to a larger number of possibly smaller organisations
to reduce costs for NER was recommended by Sophie, comparable in part to some aspects
of the Chapter 7 approach. In all these cases important opportunities were being identified,
and exploiting these opportunities was recommended, despite a lack of earlier experience of
this kind in some particularly important cases.
It is obvious that outsourcing can involve strategic levels of risk as well as tactical risk
concerns. The sources of these strategic level risks need early identification and treatment
in the goals and futures planning stages, as well as ongoing treatment at long , medium and
short- term levels.
Outsourcing first raised concerns for me in the context of BP outsourcing significant
aspects of design associated with offshore North Sea projects in the 1980s. Earlier outsourc-
ing of some aspects of cost uncertainty and associated risk for activities like pipe- laying by seeking fixed price contracts, which outsourced detailed plan management including cost–
duration trade- offs, did not worry me at the time, but perhaps it should have. A linked issue
in the 1970s and 1980s was separate risk management in a safety sense by BP, which was
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Immediate and longer- term priorities 599
simply well beyond my remit and off my radar. Arguably, the 2010 Deepwater Horizon inci-
dent was a direct result of a mix of many things, including inappropriate outsourcing of cost
risk, and inappropriate management of other sources of risk in separate silos. Outsourcing
risk and uncertainty via contractual arrangements is a much broader interpretation of out-
sourcing than may be common, but it is a useful perspective, for reasons clearly illustrated by
many misguided organisations with a cultural bias towards outsourcing risks to contractors
which the contractors would take to get the job but prove unwilling or unable to meet if it
was realised.
All enlightened planning needs an effective uncertainty management process which deals
with outsourcing opportunities and risks in a very broad sense, fully integrated with all
related contract planning. Decisions to outsource or keep key functions or issues or sources
of risk and opportunity in- house are central to the public versus private sector status of
organisations and a wide range of hybrid approaches. But they are also central to organisa-
tions like TLC and Astro, plus all other private sector organisations, when possible public
sector status is nowhere near the agenda.
Investing in future learning
Early investment in corporate learning, via development and demonstration projects for exam-
ple, can be exceedingly valuable central aspects of moving an organisation towards fully imple-
mented EP based approaches to planning. Examples associated with BP, National Power and
IBM UK were mentioned earlier. Each illustrate some of the tactical options available, but a
broad strategy also needs consideration as part of change- management planning.
Future learning strategy has to address a very wide range of issues in a systematic and
coherent manner. For example, the different learning needs of different groups of people
within an organisation need addressing, as well as common needs.
Approaches to learning which address this, like the Biggs and Tang (2011) ‘constructive
alignment’ approach, contain ideas which might be adopted by organisations other than
universities and issues as diverse as apprenticeships for early development of technical skills
and mentoring of new board members, or those being prepared for new senior management
roles at other organisational levels, are all part of the big picture needing attention.
Much of the associated planning has to be informal, but enlightened formal planning
based on a rich understanding of a wide range of alternative approaches can shape what hap-
pens informally in crucial ways.
Overall strategic level change management planning
To introduce any significant changes in any organisation, there is an obvious need to evolve
an overall strategic level change management plan. If the significant changes include a shift
to a variant of an EP perspective which is substantially different from current approaches,
the plans will have to cope with an emerging sense of direction and priorities, with flexible
and robust provisional plans, anticipating a wide range of significant surprises, some entirely
unpredictable. The plans will have to ensure that the organisation responds effectively
to all surprises. This is a very demanding kind of project planning, which needs effective
integration with corporate and operations planning needs, and addressing it is particularly
challenging.
The frameworks provided by Chapter 7 are a useful basis, and this chapter has emphasised
some key issues, but there is a vast literature on the management of change which also needs
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
600 Further synthesising and reflecting
to be drawn on. What needs to be done has to include making effective use of a compatible
and coherent version of what an organisation currently knows, plus what it can usefully learn
from all other sources, testing all aspects for robustness.
Someone within the organisation needs to take or be given responsibility for coordinating
the acquisition and the ongoing coherent application of corporate management of change
capability. It may be very important for this person to be at the board level. What other
roles they might play and what they are called will clearly have to depend on the context.
Widespread capability- culture asset and liability changes may be integral aspects of what is
involved, touching all parts of the organisation. Making these kinds of significant changes to
an organisation in a successful manner obviously requires competent management of change
leadership with appropriate support from everyone else. Associated formal planning can and
should be approached as a supporting framework, a vehicle for testing and communicating
ideas, not just a prescription for action.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
The next three sections of this chapter consider broad aspects of EP which need ongoing
attention by all participants in an EP approach. These sections address important aspects of
strategic and tactical clarity which are recurrent themes in all earlier chapters: teamwork and
wider collaboration, biased decision making, and contingency planning aspects of EP which
may underlie some aspects of biased decision making but have much wider implications.
Enhancement of strategic clarity and tactical clarity at personal and corporate levels depends
on ongoing attention to these three broad concerns.
Further sections then consider broad planning areas, perspectives, frameworks or
approaches which most readers ought to have some interest in which may be a particular
focus of interest for subsets of readers: overall corporate planning, operations planning,
project planning, process planning, safety and security planning, risk management, lean and
agile approaches, the histogram and tree (HAT) framework and Monte Carlo simulation
approaches. The primary purpose of these sections is exploring tactical clarity issues which
underpin ongoing strategic clarity enhancement for all readers. But an important secondary
purpose is helping readers with interests in specific areas to take their personal tactical clarity
enhancement further.
A section addresses regulation which many organisations may have to deal with and the
relevance of EP concepts to regulation. A following section very briefly addresses politics
and associated public pressure as part of the environment which all organisations have to
consider when planning and all of us as citizens have ultimate responsibility for.
The final section provides an overview summary which uses ‘smiles’ as an indicator of EP
effectiveness in terms of corporate progress with capability- culture concerns. It employs an
influence diagram structure evolved over many years which has been updated in this book to
incorporate ‘frowns’, a ‘red face’ special case of ‘frowns’, and a ‘halo’ special case of ‘smiles’.
Teamwork and collaboration as key capability- culture issues
Teamwork plus wider collaboration is a central feature of all five Part 2 tales. You may have
always regarded teamwork and wider collaboration as crucial, but formally recognising their
value as key capability- culture assets, with direct implications for both formal and informal
approaches to planning, seemed a very useful feature to embody in an EP approach via the
UP concept. Some approaches to planning do not exploit the importance of teamwork
and wider collaboration enough, and some seem to make inappropriate stealth assumptions
about teamwork and collaboration which need systematic identification and testing. A lack
of effective teamwork and insufficient attention to broader collaboration are capability-
culture liabilities which often require effective resolution and at the very least need effective
11 Ongoing enhancement of strategic clarity and tactical clarity
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
602 Further synthesising and reflecting
accommodation. At worst, a lack of effective teamwork and collaboration may prove a fatal
flaw for an organisation, and realisation of associated risk can be catastrophic.
The need to make effective use of an organisation’s capability- culture assets in terms of
big and little team working in an explicit manner, as well as either eliminating or accom-
modating all relevant associated liabilities, is always crucial in the context of deciding
how an organisation should approach formal and interrelated informal planning. Put
the other way around, assumptions about what teamwork and collaboration are needed
in a planning context should depend on the nature of the plans in terms of their infor-
mal planning implications as well as their formal content, and the competence of every-
one involved is always crucial. These aspects of EP are operationally inseparable, and
they need nuanced and well- grounded operational attention, recognising all relevant
interdependencies.
One illustrative key implication is the need for careful and enlightened consideration
of an organisation’s approach to formal contractual relationships and the closely coupled
central role of the operational interpretation of formal contractual relationships. Work
with Stephen Ward in this area, stimulated by a Science and Engineering Research Coun-
cil (SERC) grant which funded two years of research undertaken by Bernard Curtis, led
to a report published as a book in 1991 which Stephen and I have built on. Curtis, Ward,
and Chapman (1991); Ward, Chapman and Curtis (1991); Ward and Chapman (1994,
1995, 2008); and Chapman and Ward (1994, 2008) might usefully supplement other
relevant literature to help you develop your views of the formal and associated informal
concerns which need to be addressed. A core theme of this cited work is that contracts
are not just legal constructs designed to protect particular players. They are frameworks
for a collaborative approach to sharing risk and reward which require interpretation by
people who have a nuanced understanding of ‘reciprocation’ concepts, a practical aspect
of successful collaboration. All the parties and people involved need to be prepared to
‘give’ as well as ‘take’ in relationships founded on reciprocation of generosity and well-
founded trust. As one of our particularly enlightened experienced practitioner SERC
grant contributors put it, ‘as soon as “the contracts are out of the drawers”, everybody
is in trouble’. This is a comment Martha attributed to her uncle Paul in Chapter 7 and
used to clarify her approach to clients on associated issues. Contracts which are confron-
tational may be common, and they may suit the style of many organisations, but avoiding
confrontation when feasible is usually a more enlightened approach for most parties.
Simply refusing to work with parties who are confrontational whenever possible is usually
a sound strategy for a complex set of reasons, including attracting and retaining staff with
the ‘right stuff ’. Most really good staff with the right stuff will not put up with routinely
working with people who cannot be trusted to be collaborative if they have alternative
job opportunities.
Other key implications include the need to very carefully assess the extent to which formal relationship planning beyond contractual planning is currently needed and effective in an
organisation. This assessment should consider the nature of all associated informal planning, the way related positive and negative feedback loops work, and the way these issues interact
with corporate policies designed to address attracting new staff plus the ongoing retain-
ing, training, and motivating of all staff. Well- motivated people who can work together as
effective members of a team with all the other requisite capabilities may be drawn into an
organisation and retained if that organisation fosters those characteristics by light- touch for-
mal planning systems and associated governance with strong and consistent informal plan-
ning approaches both encouraged and facilitated. Inappropriate formal planning approaches
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Ongoing strategic clarity and tactical clarity 603
which inhibit skilled and creative informal planning may attract people with seriously inap-
propriate attitudes, and drive out the people who really matter. The interdependencies
which need thoughtful attention are very complex.
Both the opportunities and the challenges involved are significant, and a focus on these
issues using the capability- culture asset and liability concepts relationship to both the UP
and all dependent SPs may be useful.
Being open to approaches to collaborate for mutual benefit and proactive in seeking out
new possibilities for cooperation has been a core feature of the Part 2 tales. Generating and
sustaining a successful collaboration track record within and beyond a generalised form
of ‘big team’ approach is a capability- asset worth investing in. It may not happen without
explicit facilitation.
I found The Road to Co- operation: Escaping the Bottom Line (Pearson, 2012) useful back- ground reading when writing this book, and you might find it useful as part of your further
reading in this area. Although Pearson’s approach and mine are very different, because the
context and concerns are different, you are likely to find a wide range of perspectives and
approaches relevant, and this is just one illustrative example. The Gift (Hyde, 2006) also addresses ‘reciprocation’ from a perspective you may find useful, and there is a developing
literature addressing related ‘reciprocal altruism’ concepts as well as genuine altruism plus
other related and relevant approaches – see Altruism: The Science and Psychology of Kindness (Ricard, 2018) for a very interesting synthesis. What is particularly relevant about Matthieu
Ricard’s book is its well- argued modern science– based critique of all approaches to philoso-
phy and psychology which are grounded on egocentric selfish behaviour, including that of
Ayn Rand, Sigmund Freud and Machiavelli, coupled to well- argued endorsement of writers
like Charles Darwin, with direct implications for economic and environmental concerns as
well as greater life satisfaction for those who embrace Ricard’s ideas.
There is a very rich literature on teamwork, collaboration, and associated culture con-
cerns. Most of it is more specifically focused than the very broad references provided when
introducing the capability- culture concept in Chapter 2, and quite different from the very
specific references to my work on contracting with Curtis and Ward and the other refer-
ences just discussed. For example, in a website paper Matthew Leitch looks at the nature of
‘risk culture’ and how to change it (Leitch, 2013), building on ideas in his book Intelligent Internal Control and Risk Management (Leitch, 2008). You may find it useful to explore a quite wide range of how to do it literature relevant to your specific concerns to deepen your
strategic clarity and build tactical clarity.
As with many other aspects of the proposed EP approach, the teamwork and collabora-
tion opportunities and challenges relevant to the organisations of interest to you may be
somewhat different from those explored in this book. However, making these issues an
explicit part of the basic frameworks used for thinking about planning via the capability-
culture concept seems a useful step forward. Responding effectively to all the complex ques-
tions raised by the capability- culture concept is a significant how to do it challenge, and the
role of teamwork and wider collaboration for you and your organisation may be a good
place to start to get to grips with how to do it concerns. If nothing else, it is a reminder to
ask some useful questions which need thinking about in most organisations.
Biased decision making as a ubiquitous concern
Unbiased estimation of key parameters for decision- making purposes is a ubiquitous con-
cern in Parts 1 and 2 – because biased parameter estimation is an obvious basic reason for
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
604 Further synthesising and reflecting
biased decision making, and biased decision making in a general sense has a very broad
range of forms and causes which all need attention.
For reasons outlined in Chapter 3, and illustrated by later chapters, parameter estima-
tion approaches based on a point estimate framing assumption ought to be scrapped in all contexts. Point estimates are simplistic, in the sense that they are usually too simple in the
‘wrong way’, and keeping it simple in the ‘right way’ is important. A ‘minimum clarity’
approach can be a useful place to start, and there is a rich range of ways available to add more
clarity in an efficient manner whenever more clarity is worthwhile.
There is a rich literature on controlling unconscious bias in parameter estimates which
some readers may need detailed knowledge of. For example, a classic reference which I first
used with university students in the 1980s is Judgement Under Uncertainty: Heuristics and Biases (Kahneman, Slovic, and Tversky, 1982, especially chapters 21 [Alpert and Raiffa] and 23 [Fischoff]). Recently reading Thinking, Fast and Slow (Kahneman, 2012), and then The Undoing Project (Lewis, 2017) about the Kahneman and Tversky partnership, greatly enhanced my understanding of these issues, and working backwards with these three refer-
ences might play a useful part in your exploration of the literature in this area.
There is a growing literature on conscious (deliberate) bias which all readers should at
least be aware of, and some may need to engage with. Megaprojects and Risk: An Anatomy of Ambition (Flyvbjerg, Bruzelius, and Rothengatter, 2003) is a good starting point, but be aware that the way much of this literature is often used is highly controversial.
Even those who do not need to understand the conscious or unconscious bias literature
in detail do need a clear understanding of how to deal with the broader issues underlying
‘a conspiracy of optimism’ and ‘strategic misrepresentation’. They also need to understand
that biased decision making viewed in general terms involves issues which go well beyond
conventional interpretations of biased estimation of component parameters, for a range of
reasons touched on in Part 1 and addressed throughout Part 2.
As an opening example, Nicola’s Chapter 5 pass three treatment of the cost of capital asso-
ciated with a credit period when an order quantity arrives but has not yet been paid for can
be seen as an approach to eliminating bias in the economic order quantity model overlooked
for many decades, with significant implications. Nicola’s pass two plans, based on an over-
estimated cost of capital, reduced an initial order quantity of 2,000 to 1,500. But her pass 3
plans, which recognised the value of using other people’s money to avoid her pass 2 biased
estimate of the cost of capital, increased the order quantity to 7,000. This triggered Bob’s
very different pass 4 approach, which then triggered a whole string of bottom- up strategic-
change proposals. You may prefer to see this as a biased approach to decision making which
goes beyond biased parameter estimates used in models, addressing biases in the models
themselves, and the processes used to construct or select and adapt the generic models.
Martha’s Chapter 6 approach to avoiding biased estimation of costs and probabilities via
explicit qualitative analysis using a traffic light approach is also an illustration of a biased
decision- making issue with implications of importance which go well beyond the usual
unbiased parameter estimates concerns. This kind of concern is not usually thought of as
a biased parameter estimation concern, or addressed effectively in any other way. You may
prefer to see it as an illustration of biased decision making driven by the need for qualita-
tive analysis of important issues which quantitative analysis cannot deal with, a very basic
decision- making process concern.
Biased discount rates for NPV calculations, as addressed by Paul in Chapter 7, provides
another example of important bias inherent in the framing assumptions of the decision-
making approach itself, not just a matter of biased estimation of parameters. Bias associated
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Ongoing strategic clarity and tactical clarity 605
with approaches to NPV calculations is somewhat buried in Chapter 7. However, the dis-
posal of nuclear waste example, discussed in more detail by Chapman, Ward and Klein
(2006), suggests a very serious bias in NPV calculations, of the order of billions of pounds
in a context when a hundred million pounds was the determining feature of a crucial choice.
Ignoring objectives which arguably mattered more than those addressed probably further
increased the implications of the biased decision making involved, and may have led to a
national strategic mistake with 1990s ‘present value’ consequences involving many billions
of pounds and ongoing consequences which will matter even more this century and the
next. There are directly equivalent concerns for private sector organisations, demonstrated
as clearly as possible by the Chapter 7 discussion, in terms of shaping the nature of projects
as well as influencing which projects get selected. And this is just part of the biased decision
making which can be driven by an inappropriate approach to project planning and interde-
pendent operations planning, the primary focus of Chapter 7.
Biased estimates of nuclear power expected costs and ranges, as explored by Carl in the
context of corporate planning for Canpower in Chapter 8, was a relatively minor part of the
biased corporate strategic planning approach adopted by Canpower. Biased management
decision making in relation to goals planning, futures planning, long- term planning, and
medium- term planning was the real issue in Chapter 8.
Bias associated with estimating ‘value of an avoided fatality’ concepts as the number of
fatalities in a single incident varies as addressed by Sophie is the focus of Chapter 9, and bias
in this sense also matters greatly, especially if we broaden the set of relevant objectives to
embrace ongoing health and environmental concerns in a wide range of relevant contexts.
These concerns might include very low probability major international events, like nuclear
catastrophes, but they may also include issues we have to face for certain, like the implica-
tions of toxic hydrocarbon combustion emissions and the full and very complex implications
of the current international movement towards electric motive power.
These Part 2 biased decision- making issues involving framing assumption approaches as
well as parameter values are all ongoing concerns, some with complex political and ethical
dimensions requiring pragmatic resolution strategies which are inherently very difficult to
address. What is needed is testable working assumptions that facilitate properly grounded
management decision making at all levels, avoiding the wrong kind of simplicity, and facili-
tating well- informed debate whenever this is appropriate. Early and ongoing how to do it
attention by your organisation which addresses all sources of biased decision making is very important, and you may be able to make key contributions. Plans built upon seriously biased
parameters are inherently unsound. But if the basic frameworks which underlie peoples’
understanding of the way decision making should be approached are biased, the implica-
tions can be orders of magnitude more serious. There are huge opportunities to make much better management decisions associated with overcoming these bias issues.
A particularly difficult challenge is finding effective ways to confront wilful bias, whatever
its form, whatever its motivation, whatever its justification, and whatever the power and
influence of those perpetrating it.
Contingency planning as a constant background concern
In Chapters 3, 4, and 7 contingency planning ideas based on the SCERT (Synergistic Con-
tingency Planning Evaluation and Review Technique) approach were discussed directly.
A generalisation of this approach underlies EP as a whole, and you may find it helpful to see
the EP approach to contingency planning as a synthesis of seven central ideas.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
606 Further synthesising and reflecting
First, it can be important to address sources of uncertainty in terms of one or more of five
different partially overlapping portrayals of uncertainty: event uncertainty, variability uncer-
tainty, ambiguity uncertainty, capability- culture uncertainty, and systemic uncertainty. The
way we frame our contingency planning is affected by the uncertainty portrayals we employ,
so we need to choose appropriate portrayals carefully with this in mind.
Second, it is often important to consider proactive responses which involve changing the current base plan A1 to a better base plan A2 or A3 plus developing pre- planned reactive responses in the form of a plan B, C, or D and so on to be used if the base plan does not
work out or new opportunities are identified. These contingency plans may need their own higher order contingency plans with obvious practical limits on this kind of complexity.
Third, it is generally important to consider the use of responses which are specific to indi- vidual sources of uncertainty plus further broadly defined general responses which deal with combinations of sources net of the residual implications of a specific response, including a
widely defined view of the ‘unknown unknowns’.
Fourth, it is crucial to understand that general responses are essential to deal with the
unknown unknowns which are ‘unknowable’, but general responses are also very useful for
dealing with any unknowns which are simply not worth the effort or cost involved in iden-
tification and follow- on planning.
Fifth, as soon as we start planning with options in mind, risk efficiency should be a central
concern, and contingency planning is one route into addressing planning options.
Sixth, it is always important to see opportunity efficiency as the motivation for all aspects
of planning in a very general sense, including identifying and capturing the important
opportunities associated with simplifications of the right kind and avoiding serious risks
associated with simplifications of the wrong kind. Put a little differently, there are uncertain-
ties associated with the way we choose to plan which require their own particular kind of
contingency planning.
Seventh, managing the extent to which we have yet to unravel ambiguity uncertainty
associated with contingency planning because we have not yet addressed it, although we
might now address some aspects of it if doing so looked worthwhile, is always essential.
Eighth, understanding that ‘strategic clarity’ involves a good grasp of how all these ideas
can be made operational in any given context without losing any key opportunities or realis-
ing any key risks is the essence of an EP approach to contingency planning.
A key working assumption when we use these central ideas explicitly in a high clarity
approach, or implicitly in a low clarity approach, is that the assumed general responses are
effective and efficient enough to cope with all relevant sources of uncertainty net of the effect
of the specific responses. Anything general responses cannot cope with appropriately needs
effective and efficient specific responses, and vice versa. Both specific and general responses
can incorporate higher order responses if primary responses fail, but this kind of higher order
contingency planning has practical limitations and costs which must be addressed effectively
and efficiently. It is always important to have sufficient understanding of what is involved in order to make a judgement about the validity of this key working assumption.
This generalised high and low clarity view of contingency planning is central to EP as a
whole in all contexts. In the early stages of decision- making processes only a modest degree
of contingency planning may be desirable to reduce potential bias associated with expected
outcome estimates and associated variability estimates, so those involved know what they
are doing in strategic clarity terms before making commitments. Sometimes organisations should wait until strategy has some stability and commitments become necessary before
investing in detailed contingency planning. This involves a ‘clarified ambiguity’ approach to
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Ongoing strategic clarity and tactical clarity 607
complexity which project management has to address via carefully planned and controlled
lifecycle stage structures, and all planning needs to draw on this aspect of project planning within comparable structures appropriate to the context. In the very early stages, sometimes
a very simple minimum clarity approach may be suitable, completely ignoring any explicit
formal contingency planning. But when this approach is adopted, someone who under-
stands the implied assumptions should ensure that the simplicity currently assumed remains
robust overall, and effective and efficient in terms of each component of the planning.
Depending on the specific concerns that you and your organisation have to address,
exploring suitable corporate perspectives on contingency planning, plus associated robust-
ness and resilience concerns, may pay significant dividends. I cannot offer any useful general
advice on where best to start, but each of the very different tales of Part 2 may offer some
clues relevant to the contexts of interest to you and your organisation.
Corporate planning
Several interrelated key messages for readers interested in top- down, bottom- up, and inte-
grative corporate planning as part of a more general corporate management perspective
were central to Chapter 8. However, some important messages were introduced late in
Chapter 8 without a lot of emphasis or were left until Chapters 9 and 10.
In Chapter 8, Canpower’s need for formal corporate planning involved a very much big-
ger set of concerns than those embodied in the role Larry had been given as the ‘Corporate
Planning Director’. An implication was central aspects of the formal corporate planning
concerns which Canpower had to address probably needed at least two separate directors,
with a carefully defined form of separability in their relationships so that the board was
presented with a balanced perspective when making decisions about the portfolio of change
projects defining the organisation’s future evolution. A ‘balanced perspective’ in this context
implies an unbiased portrayal. A degree of separation between plan creation and enhance-
ment initially and the follow- on integration of plans plus governance are central bias control
concerns at an overall corporate decision- making process level.
Assuming these roles are executed along the lines suggested in Chapter 8 by capable
directors, formulating corporate strategy also involves a capable board led by a capable
managing director and chairman in terms of managing creative bottom- up option genera-
tion as well as some direct top- down thinking and overall governance. It may be useful to
explicitly associate the label ‘corporate management’ with the board as a whole, supported
by the crucial overall leadership roles of the CEO and board chair but also jointly led by
a number of directors with explicit formal corporate management roles, fully integrating
relevant emergent strategy roles for all the ‘operations’ directors plus interrelated change-
management roles for one or more ‘projects’ directors.
More generally, from an EP perspective corporate planning involves central roles for the
director or directors responsible for ‘corporate planning’ in a traditional direct sense, but
other members of the board also need to take a very close interest, perhaps with clearly
defined and formally identified corporate management roles. This flows from seeing cor-
porate planning as part of a holistic corporate, operations and project planning trio which
embraces all planning within an organisation. Chapter 8 is a fairly rich example, but the
treatment is necessarily very context dependent, so other contexts may require significant
differences in approach.
The tale of Chapter 9 suggested the possible need for a board level ‘Director of Safety
and Security’ to carry forward Sophie’s ideas, with sufficient power to play a full role in
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
608 Further synthesising and reflecting
top- down corporate strategy formulation as well as bottom- up planning in this area, with
related implications for Chapter 8. For organisations facing any significant safety and secu-
rity issues, this is an area of complexity which must be addressed effectively. If the potential
catastrophic nuclear incident issues associated with the electricity utility context of Chap-
ter 8 are replaced by a possible set of ‘nuclear catastrophe equivalents’ for other organisa-
tions when relevant, there may be important analogous implications. Nuclear catastrophe
equivalents might include cybersecurity attacks with safety, reputational and financial impli-
cations or disruptive new technologies or disruptive political changes driven by social and
economic changes, for example.
The way all these directors and associated senior managers need to work together as a
team clearly needs understanding and acting upon by everyone involved. Opportunity effi-
ciency for the organisation as a whole depends on how well this is done. The tales of Part 2
all contribute some ideas relevant to the contexts of each chapter, but putting all these ideas
together with adaptation to other contexts is clearly a significant challenge.
Key aspects of any organisation’s capability- culture asset and liability set will be driven by
these considerations and will interact with the planning processes needed. To take a very
simple example, even after Nicola’s arrival, TLC as discussed in Chapter 5 had very lim-
ited formal planning process capability, but under George’s leadership with Ajit’s guidance
TLC functioned very effectively. This effectiveness included responding to the opportunities
Dave triggered recognition of, using strategies created, enhanced, shaped, and tested by
Nicola initially, and then by all the other relevant players. However, as Ajit advised, if they
were going to effectively grasp the opportunities associated with all the identified proposi-
tions, they needed to build on their current capability- culture assets and add more requisite
skills, especially in a corporate planning context from a marketing perspective.
There is an extensive literature on corporate and associated strategic planning which any-
one who is an effective reflective manager in this area clearly needs to draw on. Some of it
has MS/OR aspects. For example, the UK Strategic Planning Society was originally started
as a study group of the OR Society, soft OR authors have been active in the strategic plan-
ning area, and forecasting as discussed in Chapter 8 is MS/OR territory as well as that of
economics and econometrics. However, mainstream corporate and associated strategic plan-
ning is very much wider, and a coherent, broad and internally consistent view of the whole
field is well beyond the scope of my grasp of this area. I have acquired some familiarity over
the years with a few of the well- known mainstream standard texts, like Implanting Strategic Management (Ansoff, 1984), The Rise and Fall of Strategic Planning (Mintzberg, 1994) and Strategic Safari (Mintzberg, Ahlstrand and Lampel, 1998), but I do not have an up- to- date knowledge of strategic planning or broader corporate management and associated
governance in sufficient depth to offer useful advice about how to approach it here. I do
know enough about it to suggest that a very serious challenge is integrating all the relevant aspects of it into an internally consistent set of approaches which can be used in a coherent
manner with all the enlightened planning ideas explored in this book.
Within every organisation this is a challenge which needs addressing in an effective man-
ner. If it is not directly relevant to your interests, some of the key ‘big picture’ concerns
addressed in Chapter 8 are still worth thinking about in the context of your organisation
because of the interdependencies with your concerns. If it is relevant to your interests,
Chapter 8 should provide some useful frameworks and ideas for integrating with other
frameworks and ideas, but they will need adapting to the contexts of interest to you along
with careful linking to consistent frameworks and ideas from the significant relevant litera-
ture. If you are a director with expertise in this area, you will already have a much richer
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Ongoing strategic clarity and tactical clarity 609
background in this area than I have, and testing your framing assumptions with a view to
developing a more general set of framing assumptions embracing EP concepts will be a chal-
lenge, with difficulties dependent on the scope of the desirable changes. If you are a student
at an early stage in your career, in a sense your challenge is easier, because you can use the
EP frameworks as a starting point, and test both your reading and your experience against
that framework as your views mature.
Wherever you are in terms of the director, expert and student categories of target read-
ers it would not be helpful for me to try to provide more detailed guidance focused on
corporate planning here. However, it may be useful to end this section by encouraging you
to see an interest in corporate planning as closely connected with both operations planning
and project planning – for some purposes they need to be treated as separable in working
assumption terms, but the interdependencies are crucial because they are not really separa-
ble, and your framing assumptions need to explicitly acknowledge the importance of this
lack of separability.
Operations planning
In Chapter 5, the operations planner who Nicola saw herself supporting initially was Pete,
TLC’s production manager. However, as the tale evolved, the buyer Bob became involved
and supply chain management in general became an issue, then TLC’s sales manager Surinder
and demand chain management in even broader terms became relevant. George and Pete
then had to confront complex interdependencies between operations concerns involving
product design driven by marketing as well as supply chain and production concerns, with
broad overall corporate strategy implications. Finally, corporate strategy requiring top- down
corporate planning of operations management issues starting from basic goals for TLC’s
owners became an obvious concern which Ajit picked up when addressing finance. The
opportunities for all those involved in operations planning were wide in range and rich in
possibilities.
In Chapter 6, the operations planner involved directly was Martha, the Astro Marketing
Manager for Wales. However, as the tale evolved Martha was clearly going to be a key player
in reshaping the approach Astro Wales took to strategic partners and related internal market
structure concerns as part of a much bigger national and international redefinition of the
Astro business model. There was a lot of scope for more effective and efficient operations
planning, including a much greater depth of understanding about quite basic issues like
‘adding value’ to requirements specified by customers to increase profitability with possible
international portability of key insights and processes.
In Chapter 7, Ollie was an operations planner with a more traditional Operations Manager
role and an Operations Director title. He was only involved in the tale indirectly because the
focus of the tale was Paul’s approach to project management. However, Ollie’s supporting
role was crucial because he was a key player in terms of initiating projects to invest in some
of the new facilities WSL corporate planning had to consider and Paul had to deliver as Pro-
jects Director, and even more important was their joint role at the front- end of all project
lifecycles. As Operations Director, Ollie needed a much closer relationship with Paul than
was the case with Paul’s predecessor to provide more effective input to corporate planning,
an opportunity of strategic importance to WSL as well as to Ollie and Paul.
In Chapter 8, Dick was in a traditional Operations Director role, but Larry seemed to
pre- empt his bottom- up strategy formation role. Carl’s approach to corporate planning
meant that operations planning of all kinds should be seen as a source of emergent strategy
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
610 Further synthesising and reflecting
formulation, a bottom- up approach to corporate planning needing effective integration
with top- down corporate planning. This implies a collective opportunity for all bottom-
up corporate planners addressed individually in terms of specific operations management
examples earlier. It also suggests other operations directors need integrated treatment, like
‘finance directors’ and ‘human resources directors’.
In Chapter 9, Oscar was concerned with all operations planning as part of his Operations
Director responsibilities for NER. Oscar needed support from Sophie, a new Head of Safety
whose responsibilities were extended to security. As an operations planner Sophie might
need direct board level representation to influence corporate strategy appropriately.
The opportunities for operations planning and associated operations management are
clearly immense in terms of the range of possibilities and the richness and depth of what
might be done. Seeing operations management in these very broad terms should not be an
issue – the labels ‘operations planning’ and ‘operations management’ can mean different
things in different contexts, as long as everyone is clear about how key terms are used.
The key challenges are everyone in the organisations involved having a common under-
standing of what can be done, what should be done, and how to do it effectively and effi-
ciently, working together as both small teams and big teams, with broader collaboration
when appropriate. These are non- trivial challenges, but overcoming them is worth a lot of
thought and effort, perhaps starting with the easy bits unless a difficult bit is ‘mission criti-
cal’ for your organisation.
The literature on the range of areas operations planners and operations managers have
to address is several orders of magnitude bigger than the corporate planning literature as
I understand both. As in the context of the last section on corporate planning, I do not
have an up- to- date knowledge of enough of it to provide useful detailed guidance on how
to approach it, and you will have to be selective according to the contexts of interest and
where you are in the director, expert, and student space. But all relevant aspects of interest
to you and your organisation need internally consistent integration with the enlightened
planning ideas if joint use is going to work, part of the overall challenge to be faced which
is well worth rising to. If you are not directly interested in rising to it, you still may need to
take advice from those who are doing so because the impacts have very wide implications.
Project planning
The tale of Chapter 7 suggested that Paul’s predecessor was lacking competence in some
areas which Paul provided. Paul’s approach outlines the portfolio of key conceptual and
operational tools all project planning teams need, using his WSL projects in the water and
sewage industry as an example context. The broad ‘management of change’ interpretation
of project management which this book adopts made Paul’s approach directly relevant to
the tales of Chapters 8 and 9. It was also relevant to Chapters 5 and 6. This broad interpre-
tation also makes it a basis for your management of change planning if you and your col-
leagues want to introduce an enlightened planning approach into your organisation.
There is much literature to draw on in the project planning and associated project man-
agement area. I do know this literature better than those considered in the last two sections.
The perspective outlined in How to Manage Project Opportunity and Risk (Chapman and Ward, 2011), updated and broadened by this book’s Chapter 7 approach as developed by
Paul, is a good starting point. Work outwards in a way reflecting your interests and priori-
ties, reviewing your earlier understanding of issues as well as addressing unfamiliar territory
within the framework this provides. This approach should avoid a good deal of potential
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Ongoing strategic clarity and tactical clarity 611
confusion when trying to develop internal consistency across the whole spectrum of issues
that project planning in a very broad management- of- change sense has to address.
The scope for potential confusion is high because different people use very different
framing assumptions without identifying them – never mind clearly testing the robustness
of crucial assumptions – like what they mean by ‘opportunity’, ‘risk’, and ‘uncertainty’. But
if you can learn how to deal with the sources of this kind of confusion in a project planning
context, there will be immediate and wider pay- offs.
As you work outwards, make sure you take an early look at influential project manage-
ment thinkers who from a practical perspective argue complimentary positions to those
adopted by this book, about the need to embed project management from the concept stage
for example, as in Morris (2009, 2013) and Williams, Samset, and Suannevag (2009). It is
particularly important to resist arguments directed at keeping the focus of ‘project manage-
ment’ narrow because a broad management of change perspective is crucially important.
Even in a traditional area like that confronting Paul in Chapter 7, a broad set of conceptual
and operational tools is essential. It is also essential to maintain a very broad view of ‘oppor-
tunity management’ grounded on ‘uncertainty management’.
You might then move on to more general project management literature like Muller and
Turner (2010), Turner (1992, 2014) and Pinto (2016).
You might both extend and test your understanding by reading books with many par-
ticularly useful ideas worth reading about in a primary source but with some underlying
premises which differ from mine in ways which matter to various degrees, like Lichten-
berg (2000); Hillson (2004, 2009); Cooper, Grey, Geoffrey, and Walker (2005, 2014); and
Hopkinson (2011, 2016). These examples involve some authors who I view as good friends
as well as colleagues with original ideas and approaches that should be cherished, but key
framing or working assumptions include terminology choices where we sometimes take a
different path.
You may also find it very useful to take a critical look at the wide range of key guides and
standards – for example APM (1997, 2004), ICE and IFoA (1998, 2005), PMI (2008,
2009, 2013, 2017a) and ISO 31000 (International Standard, 2009). Even if like me you do
not agree with important aspects of what these guides say, you are likely to find it useful and
perhaps very important to understand what they are saying and the perspective portrayed.
Process planning
Chapter 8 suggested a ‘Process Planning Director’ was needed, to give board level lead-
ership to the corporate use of UP variants and their application to the development and
testing all SPs that were needed, including a set of specific processes for projects like those
developed by Paul in Chapter 7. Chapter 10 raised this issue again, considering generalisa-
tion to cover all relevant toolsets and skill sets. The conceptual basis of the Process Planning
Director’s group, if a very general approach is taken, could be an internal consultancy ser-
vice which includes all kinds of planning process facilitation plus problem solving and mess
resolution services, drawing on external consultants as an ‘enlightened customer’ whenever
this is worthwhile, perhaps also providing external consulting to other organisations for a
range of reasons.
My inaugural address as President of the Operational Research Society (ORS), published
as ‘My Two Cents Worth on How OR Should Develop’ (Chapman, 1992b), urged the MS/
OR profession to see themselves as ‘process planners’ who could play key roles in this kind
of group. They are not a new idea, and there have been some very successful examples in
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
612 Further synthesising and reflecting
many contexts and countries. In the 1990s there was no discernible rush of ORS members
to respond to my inaugural address, and from a current perspective this might have been
anticipated, because ORS members have a wide range of different agendas and many of them
do not align with this process planning role in any obvious ways. However, I know some are
very interested, and if even a few members of the ‘broad church’ of MS/OR in all countries,
plus everyone else interested in process planning from a perspective comparable to EP, took
an interest and began a collaborative conversation about what they were doing, I believe this
could become a useful area to develop in a more coherent manner. As long as a viable propor-
tion of those involved share a moderately aligned set of goals and some common language,
there is no need for concern if different people have different perspectives and priorities.
We don’t even need common labels, like ‘Planning Process Director’, except for particu-
lar applications and discussions. But somebody has to be responsible for getting board level agreement about a coherent and effective approach to all corporate planning and com-
ponent decision- making processes and then delivering what is needed in an opportunity
efficient manner, and an EP approach can clarify aspects of their overall aspirational goals.
Safety and security planning
Safety planning has a long history and a rich literature which I first engaged with in a serious
manner looking at US nuclear power stations (NUREG, 1975) with a view to the need to
draft very specific aspects of design standards related to seismic risk associated with Canadian
nuclear power stations. It seemed to me then that key players had a deep understanding
of the importance of assumptions like excluding common-mode failures via inappropriate
independence assumptions. However, there were important areas where new generalisations
would provide new insights. One example was building effective contingency planning capa-
bility using decision trees into otherwise well- developed fault and event tree concepts. My
thinking has been influenced by following some of the many useful contributions by authors
like Perrow (1984, 1994), Reason (1990, 1997), Wynne (1992) and Woods et al (2010),
but I am not up to date on this literature.
One big opportunity in this area from an EP perspective flows from initiating analysis with
a top- down corporate strategy concern for all very high impact incidents, starting with a
maximum plausible impact scenario perspective to avoid a complex set of bias and blind spot
concerns. Another is seeing a key aspect of risk as departures from expectations with all the
conceptual and operational tools which flow from the clarity, risk, and overall opportunity-
efficiency aspects of an EP approach, including enlightened prudence concerns. The rel-
evant incidents may be accidental or malicious. Using this perspective we can decompose
and structure all associated safety and security concerns, working from the most important
towards the least important. All relevant stealth assumptions and other unknown unknowns need to be accommodated in the analysis at all relevant levels. Particularly crucial is ensur-
ing that the unknown unknowns are addressed in terms of robust and resilient plans which
include appropriate provisions and prudence considerations. This is the approach Chapter 9
outlines, using a railway safety and security context with a top- down focus on catastrophic
scenarios to avoid bias and blind spots as well as ensuring full integration at a strategic level.
But it is relevant in many other contexts, and in all these contexts contributing bottom- up
analysis has to play a very important role. For example, the application of contingency plan-
ning in the SCERT driven approach to estimation discussed in Chapter 3, and its elabora-
tion in Chapman and Ward (2011), can deal with unknown unknowns of various kinds at
all levels.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Ongoing strategic clarity and tactical clarity 613
This perspective and the implied scope of the relevant underlying literature separates
it from a mainstream ERM (Enterprise Risk Management) approach like that suggested
by Duckert (2010) or AIRMIC/ALARM/IRM (2010), and much other mainstream ‘risk
management’ as discussed next. However, there is no reason all the compatible ideas from all sources cannot be synthesised. Adopting the Chapter 9 perspective opens a rich set of
opportunities for further development, but it can also embrace all the useful features of cur-
rent approaches.
Risk management
From an EP perspective, anyone using the term ‘risk management’ in their current job or
departmental label or professional affiliations needs to address several key issues, and every-
one else needs an overview understanding of these issues.
One is the toolset and associated skill set plus mindset implications driven by the defini-
tions adopted for ‘risk’, ‘opportunity’, ‘uncertainty’ and ‘complexity’. The broadly defined
‘risk management community’ uses the term ‘risk’ to mean radically different things to
different people in different contexts, with nuances associated with these interpretations
changing over time, and complex associated implications for what is meant by ‘opportunity’,
‘uncertainty’ and ‘complexity’. This is a source of serious confusion with crucially important
and very complex implications in many different areas and contexts.
Another is the mindset implications of using the terms ‘risk management’ and ‘risk man- ager’ when ‘risk management planning’ or ‘risk management coordination’ or ‘risk manage- ment enabling’ might be a better term because clarifying the need for others to actually manage the decision making directly involved would more accurately reflect the collab-
orative teamwork required. Although ‘risk managers’ in some contexts may have crucial
duty of care and legal compliance concerns to manage, the most effective ways to achieve
organisation- wide compliance with a best practice approach may be best seen as ‘enabling’
in a ‘team coach’ sense, not ‘blocking’ in a ‘compliance enforcement’ sense, with requisite
compliance concerns addressed in a deliberately unobtrusive manner. The spirit of ‘pro-
moting and protecting the good as well as contesting and constraining the bad’ is a useful
general perspective.
These key issues raise interdependent problematic concerns which are overdue for effec-
tive consideration and resolution by all the key players in ‘the risk management community’
broadly defined.
If we start with the meaning of ‘risk’, it may be worth emphasising that this is by no means
a new problem, nor am I the only one concerned. Arguments about appropriate boundaries
for ‘risk’ concepts have been going on for at least a century. Some attack very specific issues,
like Cox (2008). Some people are so fed up with the whole business they have suggested,
not entirely in jest, that the basic word risk is itself not really necessary (Dowie, 1999). The EP position is that ‘risk’ is a crucially important word, and a plain English basic interpreta-
tion linked to compatible interpretations of ‘opportunity’, ‘uncertainty’, ‘complexity’ and
other associated terms plus clear working assumptions within very broad framing assump-
tions is the only plausible way forward, as discussed in detail in Chapter 1. This view is
unlikely to attract unanimous support. You will have to make your own judgement and act
accordingly based on your own circumstances.
Parts 1 and 2 explored the implications of the overall EP toolset in some detail at a stra-
tegic clarity level. You will need more tactical clarity to fully understand how to do it in
specific areas where that is your direct responsibility, but a suitable set of conceptual and
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
614 Further synthesising and reflecting
operational tools to address ‘what needs to be done’ is a key aspect of what this book as a
whole is about. Skill set and mindset issues build on the toolset issues in an interdependent
manner. They are very complex, but important, and the focus of capability- culture aspects
of the EP perspective.
Chapter 7 was explicitly critical of a current mainstream project risk management per-
spective which sees project risk management as an ‘add- on’ to the rest of project planning
and project management more generally. The tale of Chapter 7 built directly on an ‘add- in’
perspective developed in more detail by Chapman and Ward (2011) to provide a ‘designed-
in’ approach. What the additional development of Chapter 7 involved in terms of important
details had not been fully developed and thought through until writing Chapter 7 of this
book, although the strategy adopted was explicitly recommended in the 2011 book plus
its earlier 2003 and 1997 editions. Indeed, it was actually implicit in my 1970s and 1980s
work with BP and other clients. In more implicit terms this critique applies to operations
and corporate planning and management contexts, especially the latter if an event- based
approach is involved. Furthermore, safety and security as addressed in Chapter 9 involve
some related themes.
The key opportunity for those directly interested in risk management as a discipline
or as a professional label is embracing the full implications of responding to a general
need for a much broader scope for their agenda to appropriate aspects of opportunity as
well as risk, plus underlying uncertainty and complexity, working with others more col-
laboratively to address this much broader scope of work. In many contexts this includes
abandoning the idea that risk managers manage the risk so others in the organisation can leave risk to the risk managers and recognising that everybody has to take direct appropri- ate responsibility for the full interplay of opportunity, risk, uncertainty and complexity
associated with their role.
Risk Management Organisation and Context (Ward, 2005) is an Institute of Risk Man- agement (IRM) textbook which, unlike most professional society guides and standards,
discusses the controversial nature of definitions of risk and uncertainty with clarity in a com-
prehensive manner. But all professional organisations with risk management concerns need
clear statements about their particular approaches in relation to the wide range of alternative
views taken by all relevant players.
The key challenge for all readers and everyone affected is dealing with the implications if risk managers and their supporting professional societies do not see these toolset, skill set,
and mindset changes as an opportunity and resist them as a challenge. If the risk manage-
ment community does not rise to the challenge of dealing with these issues, the risk man-
agers in some organisation may themselves become identifiable risks which you and your
colleagues might have to deal with effectively.
Lean and agile approaches
‘Agile’ approaches using operational tools derived from ‘lean’ enterprise approaches were not
discussed earlier, but if we adopt plain English interpretations of ‘agile’ and ‘lean’, arguably
being agile and employing a lean perspective are central characteristics of much of EP. A sec-
tion addressing lean and agile approaches in this final chapter is relevant, in part, because
there is a current very widespread resurgence in interest in these approaches; in part, because
they overlap with a wide range of issues which an EP perspective has to address; and, in part
because they provide one useful example of a wide range of alternative perspectives which
need embracing by any approach to planning which is general by design in the EP sense.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Ongoing strategic clarity and tactical clarity 615
‘Lean enterprises’ as explored in books like When Lean Enterprises Collide: Competing Through Confrontation (Cooper, 1995) have their origins in fundamental changes in the way organisations operate, initially developed by Toyota and a number of other leading
Japanese corporations that transformed Japanese industry in the 1950s and 1960s. Lean
enterprise conceptual and operational tools have since become embraced worldwide. For
example, most managers will have heard of, and perhaps used, aspects of ‘Total Quality
Management’.
Some of these lean tools are specific techniques or models, like ‘Just- In- Time’ approaches
to supply chain inventories, mentioned in Chapter 5, and treated as a special case of the
‘Your- Place- My- Time’ variants discussed for TLC.
Other lean tools are very general conceptual models, like ‘the survival triplet’, which
involves cost price plus functionality plus quality, assuming carefully defined distinctions
between cost and price and a decomposition of functionality, as discussed in the book by
Cooper (1995) cited earlier. This approach is consistent with the Chapter 6 approach to
bidding when Astro addresses a search for value added.
Futures planning, as portrayed by Figure 8.1 and discussed in Chapters 8 and 10, is
one area where being agile is particularly important. A central EP concern is being ‘agile’
when considering the way futures planning needs to be driven by goals planning, and the
way futures planning needs to then interface with short- term planning, as well as medium-
and long- term planning, with direct implications for overall corporate planning, operations
planning, and project planning. In Chapter 8 a key working assumption was long- term,
medium- term, and short- term base plans should focus on what was assumed to be feasible, but associated contingency plans should address possible new opportunities that futures plan- ning plus projects driven by making currently non- feasible options feasible might deliver.
This is about highly proactive and ambitious ‘agile planning’ in a plain English sense, main-
taining a robust feasible solution fall- back position while systematically trying to do better.
It is an approach which should generalise, bearing in mind different contexts may involve
much shorter time horizons than those relevant to an electricity utility. But it is not an
explicit focus for agile planning or related lean approaches in the technical senses used by
the many authors addressing these topics in the literature.
Early in 2017 I became interested in better understanding the relationship between agile
planning tools directly linked to lean organisations and EP tools. Several developments then
fostered this interest, and a few related comments here may relevant to your concerns.
The initial trigger was my younger son Andy asking me to read Lean Enterprise: How High Performance Organisations Innovate at Scale (Humble, Molesky, and O’Reilly, 2015), and let him know what I thought about the efficacy of agile approaches to project plan-
ning. He was encouraging several key players managing a software programme to adopt a
more formal approach to planning, and they were resisting, one of them citing this book
as the rationale for what could be interpreted as an absence of the useful discipline which
the programme’s owner wanted Andy to instil. The agile approach in question seemed to
promote less focus on clearly defined project deliverables and associated cost and duration
outcomes. Comparing an EP approach and an agile perspective in the lean approach sense
looks like an interesting area for further work from both perspectives, and it is near the top
of my ‘to do’ list.
Associated agile approaches to project planning have a wide range of both promoters
and detractors. Their current growth has significant and widespread implications. As one
immediately useful illustrative example of what is involved, from an agile perspective all
project management methodologies are often put on a spectrum from ‘plan- driven’ to
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
616 Further synthesising and reflecting
‘Scrum’ interpretations of ‘agile’, the latter using a series of ‘sprints’ in place of more con-
ventional detailed tactical plans which elaborate strategic plans on a rolling basis. A simple
characterisation of a plan- driven approach to project planning is a traditional ‘waterfall’
approach. PMI (2017b) explores linking this end of the plan- driven to Scrum spectrum to
agile approaches. The most recent edition of the PMBOK project management guide (PMI,
2017a) also addresses these issues, as does a 2017 APM publication. Scrum approaches
(Rubin, 2013) can vary significantly, but a simple example is a software development project
which explicitly avoids predefined criteria for outcome goals, assuming that incremental
evolution in a very flexible planning framework is more effective than attempting to elimi-
nate what some writers on agile approaches call ‘end uncertainty’. Some advocates of agile
approaches claim that risk and opportunity, as well as agility and discipline, can be balanced
to make agile approaches highly resilient, issues explored by Boehm and Turner (2003,
2004) for example. However, a broad study of agile project risk management literature by
Gold and Vassell (2015) suggests that there are serious ongoing problems which can only be
overcome by risk management approaches which are more compatible philosophically and
practically than current common project risk management practice. It seems reasonably clear
from an EP perspective that SP for projects adaptations could address these agile perspective
concerns, but it is not immediately obvious to me what the best way forward might be.
One possible approach would involve exploiting the advantages of the ‘uncertainty manage-
ment’ aspects of an EP approach which Black (2017) attributes to my uncertainty management
work with Stephen Ward over several decades. In Black’s words, ‘What makes Chapman &
Ward “Risk Originals” is they have developed a risk control approach that focuses on maturing
controls in order to reduce the uncertainty that creates risk, rather than attempting to actually
predict the risks’. That is, deal with the ‘end uncertainty’ by replacing a plan- driven concern
for end- point clarity with a more flexible ‘maturing controls’ approach.
However, the opportunity efficiency driven framing assumptions underlying an EP
approach suggests acknowledging an enlightened view about the ambiguous nature of end
of project goals at a detailed tactical planning level within a sprint framework, but sustaining
a focus on ‘clarified ambiguity’ during the strategic planning stages. From an EP perspec-
tive, a sustained search for clarity about what will be delivered and associated cost and dura-
tion outcomes ought to remain a concern in the sense that strategic goals need articulation
and systematic ongoing attention to maintain a clear sense of direction, even if the details
of what the goals are remain unclear, and how they might be achieved also remains unclear.
Just how this might be done, whether or not an intermediate D&A strategy progress stage
tailored to agile planning would be needed, and whether or not some of the portfolio analy-
sis ideas discussed in Chapter 8 might be relevant to hedge bets about what the achievement
of end goals might involve, are not clear to me. What is clear is the interface between futures
planning and other planning in the Chapter 8 sense may be central, context specific issues
would inevitably be particularly crucial, and thinking through how best to adjust the SP for
projects as discussed for WSL in Chapter 7 would require considerable care. Indeed, syn-
thesising ‘agile’, ‘Scrum’, and related non- conventional methodological perspectives with
a sound overall EP approach driven by strategic clarity is a good example of the challenges
facing any attempt to develop a fully comprehensive ‘universal SP for projects’.
Linking agile approaches to project planning to the SP for projects of Chapter 7 and the
broader issue of linking EP as a whole to lean approaches to organisations remains a work in
progress for me. If this area is of special interest to you, it may be useful to begin by seeing
different people’s definitions of ‘lean’ and ‘agile’ and related terms as raising problems which
are directly comparable to different people using very different technical definitions for ‘risk’,
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Ongoing strategic clarity and tactical clarity 617
‘opportunity’, ‘uncertaint’ and ‘complexity’. The exponential acceleration of the rate of
change over the past few decades is a key reason for considering the use of an EP frame-
work for carefully integrating our collective wisdom about agility in a lean sense plus agility
defined in more general terms from all relevant perspectives.
The HAT framework and Monte Carlo simulation approaches
The basic question addressed in this section is:
If an analysis involves multiple sources of uncertainty which require probabilistic modelling, what is the best way to estimate probability distributions for each source and then compute and display the combinations required, given an extensive range of possible assumptions about both probability distribution specifications for each source and interdependency pat- terns between sources?
This section briefly consolidates what most readers need to understand for strategic clarity,
and it also provides some initial information for those readers with further tactical clarity
concerns. It links the HAT framework used throughout this book to the Monte Carlo simu-
lation approaches many readers will be familiar with and addresses computer software choice
concerns of importance in all practice contexts.
Prior to about 1960, many practitioners and academics assumed methods of moments
was the best general answer to the question this section addresses, usually using a mean–
variance approach. For example, Markowitz (1959) used a mean– variance model for finan-
cial investment portfolio planning purposes, and quadratic programming was developed to
exploit this model. Initial 1950s’ approaches to PERT used a different very simple mean–
variance approach, assuming independence between activities on the critical path, with no
subcritical paths becoming critical. A mean– variance approach can be clarity efficient if the
key working assumptions required hold, but they rarely do.
A somewhat separate school of thought, which was well developed prior to 1960, used
discrete probability tree approaches, sometimes with specific forms like fault trees and event
trees for safety analysis purposes, sometimes with a focus on decision trees for ‘decision
analysis’ or ‘decision theory’ purposes. This school of thought continues to command a lot
of interest, unlike approaches based on moments.
After about 1980, most practitioners assumed stochastic models of any size required
computer software based on Monte Carlo simulation or some other variant of sampling to
assess the combined effect of several sources of uncertainty. For example, by this time most
people using PERT models employed PERT software which used Monte Carlo simulation
to address ‘criticality’ in stochastic model terms, and many decision analysis texts recom-
mended Monte Carlo simulation for complex probability tree structures, referring to this
option as ‘risk analysis’. Some people had used these Monte Carlo simulation methods
manually before computers, there are some ‘non- reducible’ combinations which cannot be
solved by sequential discrete probability distribution methods, and some people interested
in simulation modelling simply did not worry about reducibility issues. However, it was
the power and speed of modern computers plus the availability of reasonably inexpensive
off-the-shelf simulation packages which changed common practice and made approaches to
stochastic (probabilistic) modelling using simulation the norm.
In the 1970s, when the SCERT model and its prototypes were being developed, my
modelling preference was a prototype HAT framework because interdependences between
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
618 Further synthesising and reflecting
sources had to be modelled effectively and efficiently, and causal dependence structures
driven by decision tree structures were central to my concerns and those of my clients. I still
believe that a HAT approach conceptual basis is by far the best way to understand these
issues. That is the rationale for the HAT based approach taken in Parts 1 and 2 of this book.
In the early 1980s, I still believed that the kind of HAT software developed by BP, and
used by Acres for several Canadian and US studies, was the best way forward in operational
terms, because apart from the conceptual power of the HAT framework it was about 1,000
times faster than a sampling equivalent for a comparable level of computational precision/
error. That is why the BP/Acres collaboration which facilitated Acre’s use of BP software
was very important to me. Sampling approaches involve sampling error, normally reduced
by larger samples, although sampling stratification approaches are also possible. Discrete
probability arithmetic HAT approaches involve a computational error associated with using
a discrete representation of variables which are actually continuous, discussed in Chapter 3
in terms of just adding two sources of uncertainty defined by uniform distributions. When
rectangular histograms involving n histogram classes are involved and n = 10 or more this
computational error is reduced dramatically, but it can be further reduced by a correction
based on one- eighth the difference between adjacent class probabilities being moved from
the higher probability class to the adjacent lower probability class, as discussed in Cooper
and Chapman (1987) and other earlier publications. That is, functional integration tech-
niques can be used to size and correct the computational error. Furthermore, this software
approach is not limited to addition operations. It can deal with multiplication, division and
‘greatest’ operations, the latter necessary for merge events in a network model.
By the late 1980s I had given up the idea of discrete probability arithmetic HAT based
computer software becoming readily available on a commercial basis, and advised all users of
HAT based approaches to employ commercially available Monte Carlo simulation software
appropriate to their application areas, with several key provisos.
First, they needed to understand and model all relevant interdependencies, avoiding a
common tendency to assume independence, usefully stated as a 0% dependence assumption.
Second, usually perfect positive correlation was a better default assumption than inde-
pendence if a low clarity analysis was appropriate, usefully stated as a 100% positive depend-
ence assumption.
Third, sometimes a 70% or 80% positive dependence assumption was a reasonable and
very useful improvement on 0% or 100% assumptions, defined by interpolation, an approach
tested empirically by studies for BP concerned with the cost of refineries.
Fourth, it was crucial to have the capability to use much more complex dependence
structures whenever this was clarity efficient, drawing on a range of approaches illustrated
in the Part 2 tales.
Fifth, it was crucial to provide a full set of relevant sensitivity diagrams, like Figure 4.1
or simpler variants, using whatever nested structure best suited the working assumptions
about interdependences. Pairwise separability was a particularly useful basis for thinking
about specifying estimates of uncertainty incorporating interdependence, computing results
incorporating interdependence, and then portraying of the way uncertainty accumulates net
of all responses of both specific and general forms.
Put slightly differently, it doesn’t really matter what computational basis drives the soft-
ware employed, provided the capabilities discussed in Parts 1 and 2 of this book are feasible. Simulation capabilities when non- reducible computational sequences need confronting may
prove an invaluable further capability in some particularly complex contexts. An understand-
ing of clarity efficient software choices if and when a lot of modelling and computation is
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Ongoing strategic clarity and tactical clarity 619
involved may also prove very useful. However, if your organisation does a limited amount
of this kind of modelling, then clarity efficiency will dictate the initial use of off the shelf
software. You can later consider alternatives more fully when you are clear what you need.
Just be very careful that the initial software choice does not constrain your organisation’s understanding of what really matters. You may have to work at ensuring that sensitivity diagrams and decision diagrams are produced as required, and dependence is modelled
as required. Bear in mind the simulation structure has to reflect any pairwise separability
assumed for sensitivity diagram purposes, more general separability assumptions may be
important in some contexts, and some simulation packages may make separate packages
necessary for a flexible approach to sensitivity diagrams and decision diagrams.
If you are particularly interested in software, it might be useful to start with the most flex-
ible Monte Carlo simulation software package available. As a first step, build input modules
which facilitate direct and easy specification of sensitivity diagram structures within a pair-
wise separable structure when non- reducible structures are not involved. Then add a hybrid
approach to computation which exploits reducible combinations using computation error
corrected probability arithmetic when this is feasible, if this capability looks cost- effective.
The most important issue is the flexibility to facilitate employing the most effective forms
of modelling in the contexts of interest. Almost as important is making it easy to obtain the
form of output needed so that everyone involved can understand what the analysis is saying,
as it evolves during the iterative analysis process, and later when it is used by senior decision
makers. Modern computers make computational efficiency issues a relatively unimportant
tertiary concern in most contexts.
Regulation and related issues
Most of the private and public sector commercial organisations which are the focus of this
book need to consider taking a proactive approach to the regulation of their sector, for
a variety of reasons. Some reasons are obvious, some perhaps less obvious. One reason
worth emphasising here is the need for ‘good’ organisations to collaborate to ensure that
regulators do their best to maintain a level playing field for good organisations, proactively
promoting and protecting all good organisations, contesting and constraining, if not pre-
emptively driving out, all ‘bad’ organisations.
Some regulators might make use of some of the ideas explored in this book and work in
a collaborative manner with the organisations they regulate where and when this is in the
best interests of the citizens they serve. They might also work with other regulators to make
‘enlightened regulation’ using ‘systematic simplicity’ an area of serious development, feeding
into and leading linked political developments – not just waiting for politicians to direct them.
Crucially, politicians and the layers of government between them and regulators need to
actively promote and protect ‘good’ regulators who act in this way, contest and constrain
‘bad’ regulators who manage ineffectively or collude with those they are supposed to regu-
late, showing enlightened governance of the regulation system.
In his first public address as a recently appointed chairman of the UK Health and Safety
Executive (HSE), Martin Temple made several points of relevance to this book’s perspective
on all aspects of regulation, some very directly and highly relevant to the EP perspective
developed in this book (Temple, 2016).
First, people working at all levels in a regulatory organisation can and should start with a
passion for looking after the best interests of those they serve – employees in UK organisa-
tions and the general public in the HSE case.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
620 Further synthesising and reflecting
Second, looking after the best interests of those they serve successfully is most effectively
achieved via an ‘enabling’ approach rather than a ‘blocking’ approach to those they need
to regulate – primarily industry in the HSE case. Enlightened industry leaders will see the
value of aligned objectives. However, if any industry leaders do not demonstrate requisite
enlightenment, firm enforcement is also required.
Third, embedding core values at a board level which are communicated to and embraced
by all an organisation’s employees and contractors is crucial.
Fourth, some of the challenges take society into very new territory which needs great
care and insight – like the impact of growing cybersecurity concerns in relation to chemical
industry facilities and other potentially dangerous contexts in the HSE case.
Fifth, very difficult issues to deal with, like long- term health effects, are not being addressed as effectively as more obvious issues like fatalities, but even cursory use of what data is avail-
able suggests they are several orders of magnitude more important than the current focus of
attention of the HSE, industry, and society more generally.
My invitation to attend Martin Temple’s address arose because I had been an invited
speaker (Chapman, 2002) for an earlier session in the same series of annual lectures, and
I went because the summary of Temple’s paper provided with the invitation suggested it
might be relevant to this book, which it certainly was. There is a very good case for arguing
that variants of all the ideas he outlined as summarised above need addressing by all regula-
tors, with strong support from all relevant layers of the governments responsible for their
governance. There is also a strong case for linking this perspective to the growing use across
the world of very big fines by a wide variety of regulators for a wide variety of behaviours
deemed unacceptable. And there is a good case for arguing that regulators need to be agile
and enlightened users of sound goals planning, futures planning, and long- term planning
which anticipates and avoids regulatory problems before they become serious crises, with the
powers, attitude, courage, and resilience needed to take on big international corporations
which abuse their power. The whole business of regulation can and should embrace massive
change to confront a wide range of challenges, with important international implications.
‘Regulation’ can be defined and addressed in very broad terms, going well beyond tradi-
tional government directed regulation. It can include forms of self- regulation which those
who favour minimal government interference will clearly want to encourage. However, the
limitations as well as the strengths of all approaches to regulation clearly need address-
ing, scrutiny by a proactive investigative press and other investigative media is crucial, and
policing the overall effectiveness and efficiency of all regulation is ultimately a government
responsibility.
A particular current concern which illustrates one aspect of the complexities involved is
the need for a generalised view of ‘level playing fields’ which involves national government
responses to the taxation of multinational companies which ensure that no honest and com-
petent businesses operating in that nation are disadvantaged by any businesses indulging in
aggressive tax avoidance, regardless of the physical or nominal domicile locations of these
organisations. Related legal system responses may also be essential. National governments
have to look after the interests of all that nation’s businesses as well as all that nation’s con- sumers in a balanced manner, as part of looking after all that nation’s citizens. Due regard to international collaboration at a range of levels, some not necessarily involving governments
directly, has to be a part of this approach.
A famous quote from the Melian dialogue by Thucydides (a Greek historian who lived
about 460– 400 BC) is ‘the strong do what they can and the weak suffer what they must’.
Regulators and their sponsoring governments might do well to also consider a quote from
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Ongoing strategic clarity and tactical clarity 621
Thucydides’s History of the Peloponnesian War – ‘when one is deprived of one’s liberty, one is right blaming not so much the man who put the shackles on as the one who had the power
to prevent him, but did not use it’. In the present context ‘liberty’ needs interpretation to
cover the interests which regulators might address of all the citizens they serve, as both
‘buyers’ of services or products in a private consumer or corporate context, and as ‘sellers’
of labour, skills, capital, services, or products in a private citizen or corporate context. Being
at liberty to unfairly exploit markets will always be a clear goal for some people and some
organisations, and some will try to protect this liberty by promoting a minimal government
involvement in regulation or public sector provision of services. Good regulators, with the
support of good governments, will successfully ‘promote and protect “good” citizens and
organisations and contest and constrain “bad” citizens and organisations’.
Regulation which embraces in an effective and efficient manner what is fair and reason-
able, as well as what is economically efficient, is essential to make a capitalist system desirable
for the majority as well as acceptable for everyone else. What is ‘fair’ and ‘reasonable’ is
debatable, and managing associated ambiguity is an inherent part of the planning problems
which have to be confronted. But there is growing evidence that a very complex set of driv-
ers of change in public expectations about regulation and associated governance by their
controlling governments is swinging public opinion and political agendas radically in direc-
tions which are quite new and fundamentally different in the UK, in Europe generally, in the
US and Canada, and in many other parts of the world.
For example, in September 2017 a Populus poll of UK citizens showed massive support
for the nationalisation of various industries, from 83% for water through 50% for banks to
23% for travel agents. As Hugo Rifkind commented in an article citing these results with the
title ‘For Generation Rent, Owning Things Is History’ (Rifkind, 2017), ‘what on earth is
going on?’ Rifkind, like me, was particularly surprised by 23% wanting travel agents nation-
alised. He offered some very interesting suggestions, but several other interesting articles
in the same edition addressed directly related concerns, and the wide range of views being
expressed in the UK media as a whole suggests we clearly need a much more coherent
understanding of the questions which need to be addressed before attempting coherent
answers for any given country of interest.
One thought which may interest some regulators and other readers interested in regula-
tion is that emergent government policy on regulation led by bold and innovative regula-
tors seems to be gaining ground, and provided it is led by good regulators who are both
competent and motivated to serve ‘the right social interests’, this looks like a very promising
development. In particular, it suggests a nuanced separation of the leadership and govern-
ance of regulation from some aspects of direct political pressures which may be very helpful.
However, regulators are watchdogs that need to be empowered by governments, and they
need effective teeth as well as fearsome growls and tails that wag which are not feasible with-
out coherent long- term government support.
It would be naïve to think that appropriate government support might be achieved with-
out having to overcome exceedingly difficult obstacles in some areas, and even if govern-
ments are determined to resolve the issues, reaching a consensus on key concerns is going to
prove very difficult in many areas of regulation that really matter. As an example, in his 2016
book The End of Alchemy: Money, Banking and the Future of the Global Economy, Mervyn King, governor of the Bank of England from 2003 until 2013, a distinguished economist
with international credentials, makes a number of important and practical suggestions about
the need for financial regulation which less concerned about the best interests of those who
are part of the finance industry they are supposed to regulate and more concerned about
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
622 Further synthesising and reflecting
non- financial industry organisations and most of their country’s citizens. At the heart of
King’s recommendations is public sector ownership of the money supply, private sector
ownership of all financial industry organisations, including banks and organisations operat-
ing like banks. Crucially, the government and its central bank would refuse to continue to
play the role of ‘lender of last resort’ (or ‘insurer of last resort’) with respect to any financial
industry organisations. There would be much simpler regulation of the financial industry,
which addresses all the ‘right questions’, with a view to what is in the best interests of the
county as a whole. The government’s central bank would provide a ‘pawnbroker for all sea-
sons’ role. Related regulation would require finance industry organisations to pre- position
collateral which would become government central bank property if liquidity loans pro-
vided by the central bank were not repaid. The pre- positioning agreements would involve
‘haircuts’ to ensure that the realisable value of the assets used as collateral exceeded the
loans, as required by pawnbrokers (or prudent mortgage providers including banks). This
looks eminently sensible and practical in an enlightened planning sense as far as I can judge,
although I do not profess expertise in this area. This kind of thinking is certainly worth very
careful consideration in the context of a clear need for a new approach to financial regulation
which is effective from the point of view of the population as a whole. But King is not opti-
mistic about the rate at which progress might be made because it would be firmly resisted
by a powerful set of people with influential allies. In effect, there is a widely understood
argument that the financial sector is currently being subsidised by everyone else, a less well-
understood argument that regulation is overcomplicated in the ‘wrong way’ because it is
addressing the ‘wrong questions’. King has useful ideas about avoiding the current subsidy
situation which would be less expensive for the banks than the more extreme capitalisation
requirements which have earlier been assumed to be the only solution available, but King’s
suggestions, as well as the more expensive alternatives, will be resisted by a financial sector
which wants to preserve its current advantageous position.
The underlying economics driving bank runs that led to the 2008 financial crisis are
even more complicated and difficult to deal with. The nature of some of this complexity is
also explored by King, from a perspective which has an EP flavour in some very interesting
respects. However, convincing evidence of effective solutions likely to be forthcoming from
current macroeconomic thinking was not provided from my perspective. The macroeco-
nomic problems involved are not going to go away, and resolving them matters greatly to
everyone, but the first priority is a consensus on regulation which is effective even if there is
no consensus on a coherent understanding of the underlying economic issues.
A review by Gerard Baker (Baker, 2018) of Capitalism in America: A History (Greenspan and Wooldridge, 2018) begins by noting that an August 2018 Fox News survey reported
36% of US citizens said it would be a good thing to “move away from capitalism to more
socialism”. Alan Greenspan has played a central banking role in the US which has perhaps
been even more prominent than Mervyn King’s comparable role in the UK, and the review
approaches its conclusion noting that Greenspan and Wooldridge argue that a robustly free-
market approach with minimal government involvement is the American tradition, with a
long history of success despite some bumps along the way, and betting against this argument
might be unwise.
I suspect Greenspan and most minimal government advocates living in the US or any
other country would not be keen on strong and effective regulation of the financial sector
or any other sector in what an EP perspective might judge a ‘socially optimal’ manner. How-
ever, significant direct involvement of governments in simple but effective regulation, and encouraging some use of novel public sector organisations comparable to the recommended
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Ongoing strategic clarity and tactical clarity 623
Canpower approach of Chapter 8 to provide government leadership of issues like energy
market management, addressing the right questions, has to be part of any enlightened gov- ernment approach in my view. Encouraging novel private sector organisations like the WSL
local and national asset companies discussed in Chapter 10, building on the Chapter 7 dis-
cussion of a UK private sector water and sewage utility, may also be useful. A completely
unconstrained free- market approach with no coherent, centralised government decision-
making input into issues like financial markets and energy markets does not seem viable for
a host of reasonably obvious reasons, and however limited the regulation level you believe
to be appropriate, an EP perspective should be useful.
Chapter 10 addresses effective ways a private sector organisation like a water and sew-
age utility in the UK might set up novel private sector organisations to help serve local and
national citizens in ways which are beneficial for the local and national private and corpo-
rate customers they serve. The issues involved are certainly not simple, but an enlightened
approach based on a consensus view of what really matters seems viable. The Canpower
discussion of Chapter 8 explores how an organisation in the electricity utility industry in
Canada which was owned by the citizens it served but was prepared to transform itself from
a traditional public sector organisation into something very different might operate. Again,
the issues involved are certainly not simple, but an enlightened approach based on a consen-
sus view of what really matters seems viable.
The final section in Chapter 9 builds on a railway safety and security context discussion to
explore a number of important private and public sector contexts where a consensus view
on the best way forward seems much more difficult. All these contexts require well- founded
trust supported by high clarity analysis of exceedingly complex issues. They all also involve
pressures from interested parties with conflicting narratives tailored to the messages they
want to sell which will tend to ensure ongoing controversy and confusion.
All these contexts require enlightened regulators with strong and effective government
backing, and relatively simple regulatory rules addressing the right questions, instead of
relatively complex rules addressing the wrong questions. In all these contexts we need effec-
tive regulatory and other direct government action which is not dependent on achieving
a consensus on a full detailed understanding of underlying factors. Key issues include the
effective regulation of organisations which might be directly responsible for large numbers
of fatalities or ongoing health difficulties because of pollution of many kinds, including
nuclear contamination. Example concerns include the very complex issues associated with
the current movement towards electric and hydrogen motive power and linked reduced use
of fossil fuels for other purposes, for a range of health and environmental reasons, in con-
junction with the need to address coherent and coordinated plans for supplying the result-
ing increased demand for electricity. A common central argument is current approaches are
clearly unenlightened when they involve leaving key aspects of decision making to private
sector organisations in contexts which very clearly need a holistic national perspective and
an integrated overall approach to achieve outcomes in the best interests of the majority of
the population. The basic problems arising from a lack of requisite government controlled
direct management are often aggravated by a lack of effectively coordinated regulation, and
in some cases by an absence of requisite international cooperation. This seems to me an
inescapable fact of life, whatever the nature of government you prefer. Even in these very
difficult cases, effective coordinated decision making and regulation is clearly essential, and
some variants of the novel public and private sector approaches just mentioned in conjunc-
tion with other EP concepts might provide a contribution to an enlightened way forward.
Assuming that doing nothing new is appropriate is certainly not a viable option.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
624 Further synthesising and reflecting
Politics and associated public pressure
Politics, and the public pressure that drives political decision making, is an inescapable aspect
of reality which all commercial organisations have to deal with. This includes, but is not
limited to, the way governments approach regulation.
All organisations adopting any plausible variant of an EP approach have to accommodate
the implications of their political environment, and associated public pressure, as effectively
and efficiently as possible. Several of the Part 2 tales illustrate what might be involved. The
key is making working assumptions which are realistic, systematically testing the robustness
of all associated assumptions, looking for opportunities as well as threats, and adopting a
proactive ‘designing desirable futures’ approach whenever this may be useful. A traditional
‘predict and prepare’ mindset is not opportunity efficient. Proactively seeking opportunities
as well as looking for and avoiding threats is crucial, and collaboration essential.
Individual citizens addressing broad social concerns, not organisations or pressure groups,
arguably ought to determine government policy. However, collaboration via single issue or
multiple issue pressure groups is undoubtedly more effective than exercising lone voices,
and wider collaboration via political parties which successfully align a coherent set of objec-
tives is undoubtedly the most sensible approach for most purposes.
Some hypotheses for putting together and managing coherent political groups might
be inferred from this book’s EP approach. However, the context is very different, with
important implications. For example, if we begin with the working assumption ‘the major-
ity’ ought to define the ‘central ground’ in a way which may incorporate most political
parties but not their extremes, one obvious complexity is the complications associated with
adversarial approaches by key political parties or factions within them having a dysfunctional
effect on majority best interests, including the viability of this working assumption. An even
more fundamental example complication is political views are founded on framing assump-
tions we might debate in stealth assumption terms, but they are also based on personal con-
text and cultural values which are not usually debatable in any direct or objective manner.
These issues are well beyond my areas of experience, and it would not be useful for me to
speculate on them in this book, but in the context of all broad social concerns as well as all
commercial organisations of interest to you they should not just be put in a ‘too difficult’
box, because they matter greatly.
Using ‘smiles’ as an indicator of EP effectiveness
When I was giving public seminars and short courses on project risk management processes
in the 1980s, building on work with BP and linked work with other clients, a challenge
which had to be addressed was people who were acclimatised to very informal and unstruc-
tured approaches to planning feeling that the structured and formalised processes I was
advocating, like SCERT, would restrict their ability to be creative, reducing job satisfaction.
Part of the response was developing and explaining a ‘smiles test’, which has gradually
evolved into Figure 11.1. You might like to study the form and content of Figure 11.1 very
briefly before its background and interpretation are explored.
When I went into an organisation where those involved in planning didn’t smile very
much, one possible working assumption, soon viewed as a hypothesis worth testing,
was that they were unhappy with their work or working environment for reasons which
might involve important capability- culture liabilities needing exploration. A linked work-
ing assumption which seemed to hold true was that if they became familiar with formal
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Ongoing strategic clarity and tactical clarity 625
using an
enlightened
planning
approach
more ‘pleasure’ +
more trust more openness more ‘halos’ better communication more congruent objectives more lateral thinking more creativity better ‘bonding’ better teamwork and partnering more ‘empowerment’ more training, less directing better ‘management of expectations’ better use of diverse experiences good management recognised
less ‘pain’
less wasted effort less frustration fewer ‘red faces’ fewer ‘frowns’ fewer crises less confrontation fewer witch-hunts bad luck recognised
more ‘smiles’ =
more reward higher-quality staff
fewer ‘disasters’ more opportunities
more ‘smiles’
less wasted effort uncertainty as opportunities
enlightened gambles
enlightened planning horizons, estimates
and controls
enlightened caution and
enlightened prudence
opportunity efficiency
fewer ‘red faces’ and ‘frowns’ plus
more ‘halos’
Figure 11.1 Corporate benefits of using an enlightened planning approach with a focus on smiles and associated capability- culture concerns, including the way these concerns drive ‘reward’ (profit and other positive objectives).
processes of the prototype EP kind being advocated and used them effectively, they would
not be constrained and bored by the kind of discipline involved – quite the opposite. They
would be set free, enabled, and encouraged to ask the right questions and then answer
the right questions in the most effective manner feasible within the time available. They
could do so by being innovative when generating the most appropriate responses, greatly
increasing their creativity and job satisfaction. This would tend to mean good people were
attracted to and stayed with the planning groups in these organisations, and because their
roles became more demanding as the scope of the jobs increased, people who were not so
good at it would tend to leave and be replaced by people who were better at it. A ‘virtuous
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
626 Further synthesising and reflecting
circle’ would be established – and a routine smiles test would confirm whether this virtuous
circle effect was occurring.
This perspective, based on direct observation including feedback from client employees,
has proved robust in practice ever since, provided the senior management involved understand what is required, promoting and protecting the good and contesting and constraining the bad as part of endorsing a broad agenda for the development prototype variants of an EP approach.
This perspective led Stephen Ward to develop a prototype of Figure 11.1 used in Chap-
man and Ward (1997, 2002, 2011) employing the influence diagram and cognitive map
ideas of authors like Colin Eden (1988).
Figure 11.1 addresses the corporate benefits of an enlightened planning approach with
a focus on capability- culture interdependences and feedback relationships. The basic direc-
tion of flow is from the bottom upwards, towards ‘more reward’ from ‘using an enlightened
planning approach’. It uses an ‘influence diagram’ format in the top portion of the diagram.
‘More smiles’ in the top portion of the diagram are elaborated in the bottom portion as
‘more pleasure’ and ‘less pain’, with both ‘pleasure’ and ‘pain’ covering a range of relevant
cultural issues not portrayed in the influence diagram but discussed in earlier chapters.
The top portion of Figure 11.1 can be read starting at the bottom node labelled ‘using
an enlightened planning approach’. The arrow on the left- hand outside suggests ‘more
reward’, the top node on the left side, will be a direct result of ‘opportunity efficiency’
for a number of reasons not worth individual identification – a residual of everything not
addressed to the right. ‘Reward’ can be associated with the positive objectives discussed in
relation to Figure 3.8 plus later extensions and clarifications in Parts 1, 2 and 3 in a general
unified sense.
The arrow from ‘using an enlightened planning approach’ to ‘fewer red faces and frowns
plus more halos’ feeds directly into ‘more reward’, but crucially it also feeds into ‘fewer dis-
asters’ which feeds into ‘more opportunities’, ‘more smiles’ and ‘higher- quality staff ’, with
implicit virtuous feedback cycle implications.
The arrow from ‘using an enlightened planning approach’ directly to ‘fewer disasters’
plus the arrow from ‘fewer disasters’ to ‘more reward’ suggests ‘fewer disasters’ will directly
increase profit and other positive aspects of ‘reward’ for a range of reasons not worth indi-
vidual identification to the right, because disaster events are very costly in many different
respects. The arrows between these three nodes which also include ‘enlightened caution
and enlightened prudence’ suggest this four- node route is also worth separate considera-
tion because it is particularly important to recognise the role of enlightened caution and
prudence even if the concern for avoided disasters is not directly relevant. Both routes send
arrows to ‘more smiles’, which sends an arrow to ‘higher- quality staff ’.
‘Higher- quality staff ’ is at the same top level as ‘more reward’ because an important
top- level two way relationship is involved, again with implicit virtuous feedback cycle
implications.
All the other arrows tell the same sort of story. ‘Fewer disasters’, ‘more opportunities’
and ‘more smiles are all at roughly the same level, just below the top level, but moving to
the right involves even more dependencies, including some important two way dependen-
cies. ‘Less wasted effort’ and ‘uncertainty as opportunities’ are at the next level down, both
feeding upwards. ‘Enlightened gambles’ are at the same level as enlightened caution and
enlightened prudence’, with slight differences in implications. ‘Enlightened planning hori-
zons, estimates, and controls’ have different implications again. The two way arrows involve
feedback as well as feedforward, the basis of explicit virtuous circles in this case.
The summary boxes associated with ‘more smiles = more pleasure + less pain’ draws
together associated messages scattered through this book.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Ongoing strategic clarity and tactical clarity 627
You might like to regard the ‘soft’ or systematic approach to qualitative concerns por-
trayed by Figure 11.1 as an illustration of the need to go well beyond quantitative analysis,
understanding all the key qualitative analysis features in a framework which encourages and
facilitates thinking about non- separable qualitative issues, before you even consider starting
a quantitative analysis.
You might wish to revise the top portion of this diagram, and the linked lists below, and
see Figure 11.1 as just a starting point for exploring your own interpretation of what using
an enlightened planning approach can achieve if the capability- culture concerns are com-
prehensively addressed along with all other aspects of an EP approach. More generally, you
need to see my variant of EP as a starting point for your own framework for considering
capability- culture interdependencies in a ‘systematic simplicity’ framework relevant to you
and your organisation, taking from this book whatever is most relevant to your organisation
and integrating everything else of relevance in a coherent framework.
There are many ways you might use Figure 11.1. But the key intention is providing you
with one possible basis for thinking about an overall conceptual framework for applying all
the ideas in this book to help you and those you work with to smile in the sense that Fig-
ure 11.1 portrays and keep on smiling.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Ackoff, R.L. (1974) Redesigning the Future. John Wiley and Sons, Chichester. Ackoff, R.L. and Sasieni, M.W. (1968) Fundamentals of Operations Research. John Wiley and
Sons, New York. AIRMIC/ALARM/IRM (2010) A Structured Approach to Enterprise Risk Management (ERM)
and the Requirements of ISO 31000. Association of Insurance and Risk Managers (AIRMIC), Association of Local Authority Risk Managers (ALARM) and Institute of Risk Managers (IRM), London.
Ambrose, S.E. (1994) D- Day, June 6, 1944: The Climatic Battle of World War II. Touchstone, New York.
Ansoff, H.I. (1984) Implanting Strategic Management. Prentice- Hall International, Englewood Cliffs, NJ.
APM (1997) PRAM Project Risk Analysis and Management Guide. Association for Project Man- agement (APM), Norwich.
APM (2004) PRAM Project Risk Analysis and Management Guide, second edition. Association for Project Management (APM), Norwich.
Baker, G. (2018) US capitalism: not dead yet. The Times, 27 October, Saturday review, 14. Becker, H.S. (1983) Scenarios: a tool of growing importance to policy analysts in government
and industry. Technological Forecasting and Social Change, 23, 95– 120. Biggs, J. and Tang, C. (2011) Teaching for Quality Learning at University: What the Student
Does. Society for Research into Higher Education & Open University Press, McGraw Hill, Maidenhead.
Black, W. (2017) Originals: How Non- Conformists Will Ultimately Disrupt the World of Risk Management. Downloaded from www.linkedin.com/pulse/orginals- how- non- conformists- ultimately- disrupt- world- warren- black, 12 May 2017.
Boehm, B. and Turner, R. (2003) Using risk to balance agile and plan- driven methods. Com- puter, 36(6), 57– 66.
Boehm, B. and Turner, R. (2004) Balancing agility and discipline: evaluating and integrating agile and plan- driven methods. Proceedings of the 26th International Conference on Software Engineering.
Byrne, L. (2017) Tech Kings must learn from our great entrepreneurs. The Times, 10 October, 26.
Carne, M. (2017) The state of rail. Letters to the Editor, The Times, 14 January, 26. Carroll, L. (1865) Alice’s Adventures in Wonderland, commonly contracted to Alice in Wonder-
land, Lewis Carroll being the pseudonym of Charles Lutwidge Dodgson. Use the web for a range of current publishers.
CCRM (2011) Final Report on the Investigation of the Macondo Well Blowout. Deepwater Hori- zon Study Group (DHSG) of the Centre for Catastrophic Risk Management (CCRM), The University of California at Berkeley, available online as a pdf.
References
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
References 629
Chapman, C.B. (1974) Modular portfolio selection: an introduction, in Dickinson, J.P. (editor), Portfolio Analysis: Book of Readings. Saxon House/Lexington Books, Farnborough.
Chapman, C.B. (1975) Modular Demand Analysis: An Introduction in the Context of a Theoreti- cal Basis for Consumer Demand Analysis. Saxon House/Lexington Books, Farnborough.
Chapman, C.B. (1979) Large engineering project risk analysis. IEEE Transactions on Engineer- ing Management, EM- 26, 78– 86.
Chapman, C.B. (1988) Science, engineering and economics: OR at the interface. Journal of the Operational Research Society, 39(1), 1– 6.
Chapman, C.B. (1992a) Risk Management: Predicting and Dealing with an Uncertain Future. Exhibit #748, Province of Ontario Environmental Assessment Board Hearings on Ontario Hydro’s Demand/Supply Plan, submitted by the Independent Power Producers Society of Ontario, 30 September.
Chapman, C.B. (1992b) My two cents worth on how OR should develop. Journal of the Opera- tional Research Society, 43(7), 647– 664.
Chapman, C.B. (2002) Constructively Simple Probabilities: Taking the Guesswork out of Risk Man- agement. Eleventh Annual Symonds Lecture in Association with the Health and Safety Board of the Institution of Civil Engineers (ICE), London.
Chapman, C.B. and Cooper, D.F. (1983) Parametric discounting. Omega – International Jour- nal of Management Science, 11(3), 303– 310.
Chapman, C.B. and Cooper, D.F. (1985) A programmed equity redemption approach to the finance of public projects. Managerial and Decision Economics, 6(2), 112– 118.
Chapman, C.B., Cooper, D.F. and Cammaert, A.B. (1984) Model and situation specific OR methods: risk engineering reliability analysis of an L.N.G. facility. Journal of the Operational Research Society, 35, 27– 35.
Chapman, C.B., Cooper, D.F., Debelius, C.A. and Pecora, A.G. (1985) Problem solving meth- odology design on the run. Journal of the Operational Research Society, 36(9), 769– 778.
Chapman, C.B., Cooper, D.F. and Page, M.J. (1987) Management for Engineers. John Wiley and Sons, Chichester.
Chapman, C.B. and Harwood, I. (2011) Optimal risk taking and risk mitigation, in Cochran, J.J. (editor in chief), Wiley Encyclopaedia of Operations Research and Management Science. John Wiley and Sons, New York.
Chapman, C.B. and Howden, M. (1997) Two phase parametric and probabilistic NPV calcula- tions, with possible deferral of disposal of UK Nuclear Waste as an example. Omega, Interna- tional Journal of Management Science, 25(6), 707– 714.
Chapman, C.B. and Ward, S.C. (1992) Financial control of portfolio management decisions, in Ezzamel, M. and Heathfield, D. (editors), Perspectives on Financial Control: Essays in Memory of Kenneth Hilton. Chapman and Hall, London.
Chapman, C.B. and Ward, S.C. (1994) The efficient allocation of risk in contracts. Omega – The International Journal of Management Science, 22(6), 537– 552.
Chapman, C.B. and Ward, S.C. (1996) Valuing the flexibility of alternative sources of power generation. Energy Policy, 24(2), 129– 136.
Chapman, C.B. and Ward, S.C. (1997) Project Risk Management: Processes, Techniques and Insights. John Wiley and Sons, Chichester.
Chapman, C.B. and Ward, S.C. (2002) Managing Project Risk and Uncertainty: A Constructively Simple Approach to Decision Making. John Wiley and Sons, Chichester.
Chapman, C.B. and Ward, S.C. (2003) Project Risk Management: Processes, Techniques and Insights, second edition. John Wiley and Sons, Chichester.
Chapman, C.B. and Ward, S.C. (2008) Developing and implementing a balanced incentive and risk sharing contracts. Construction Management and Economics, 26(6), 659– 669.
Chapman, C.B., Ward, S.C. and Bennell, J.A. (2000) Incorporating uncertainty in competitive bidding. International Journal of Project Management, 18(5), 337– 347.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
630 References
Chapman, C.B., Ward, S.C., Cooper, D.F. and Page, M.J. (1984) Credit policy and inventory control. Journal of the Operational Research Society, 8(12), 1055– 1065.
Chapman, C.B., Ward, S.C. and Klein, J.H. (2006) An optimized multiple test framework for project selection in the public sector, with a nuclear waste disposal case- based example. Inter- national Journal of Project Management, 24, 373– 384.
Chapman, C.B. and Ward, S.C. (2011) How to Manage Project Opportunity and Risk: Why Uncertainty Management Can Be a Much Better Approach than Risk Management. John Wiley and Sons, Chichester.
Checkland, P.B. (1981) Systems Thinking, Systems Practice. John Wiley and Sons, Chichester. Checkland, P.B. and Scholes, J. (1990) Soft Systems Methodology in Action. John Wiley and Sons,
Chichester. Churchman, C.W., Ackoff, R.L. and Arnoff, E.L. (1957) Introduction to Operations Research.
John Wiley and Sons, New York. Coleman, A. (2017) When doing good is great for business: socially responsible franchising is a
growing trend which can benefit business and the community. Raconteur Insert in The Times, 5 September.
Cooper, R. (1995) When Lean Enterprises Collide: Competing Through Confrontation. Harvard Business School Press, Boston.
Cooper, D.F., Bosnich, P.M., Grey, S.J., Purdy, G., Raymond, G.A., Walker, P. and Wood, M.J. (2014) Project Risk Guidelines: Managing Risk with ISO 31000 and IEC 62198. John Wiley and Sons, Chichester.
Cooper, D.F. and Chapman, C.B. (1987) Risk Analysis for Large Projects: Models, Methods and Cases. John Wiley and Sons, Chichester.
Cooper, D.F., Grey, S., Geoffrey, R. and Walker, P. (2005) Project Risk Guidelines: Managing Risk in Large Projects and Complex Procurements. John Wiley and Sons, Chichester.
Cox, L.A. (2008) What’s wrong with risk matrices? Risk Analysis, 28(2), 497– 512. Crossland, D. (2017) Carmakers ‘colluded on technology for years’. The Times, 22 July, 53. Curtis, B., Ward, S.C. and Chapman, C.B. (1991) Roles, Responsibilities and Risks in Manage-
ment Contracting (Special Publication 81). Construction Industry Research and Information Association (CIRIA), London.
Dasgupta, A. and Dasgupta, S. (1966) Crop- Planning in a Risky Environment. Read to the Euro- pean Meeting of the Econometric Society, Warsaw.
DoE (1994) Review of Radioactive Waste Management Policy Preliminary Conclusions: A Con- sultative Document, Radioactive Substances Division. Department of the Environment (DoE), London.
Dowie, J. (1999) Against risk. Risk Decision and Policy, 4(1), 57– 73. Dublin, C.I. and Lotka, A.J. (1930) The Money Value of a Man. Ronald Press, New York. Duckert, G.H. (2010) Practical Enterprise Risk Management: A Business Process Approach. John
Wiley and Sons, Chichester. Eden, C. (1988) Cognative mapping, a review. European Journal of Operational Research, 4,
1– 13. Fischoff, B. (1982) For those condemned to study the past: heuristics and biases in hindsight,
Chapter 23, in Kahneman, D., Slovic, P. and Tversky, A. (editors), Judgment Under Uncer- tainty: Heuristics and Biases. Cambridge University Press, New York.
Flood, R.L. (1999) Rethinking the Fifth Discipline: Learning Within the Unknowable. Routledge, London and New York.
Flyvbjerg, B., Bruzelius, N. and Rothengatter, W. (2003) Megaprojects and Risk: An Anatomy of Ambition. Cambridge University Press, Cambridge.
Friedman, L. (1956) A competitive bidding strategy. Operations Research, 4, 104– 112. Gawande, A. (2011) The Checklist Manifesto: How to Get Things Right. Profile Books Ltd, London. Gold, B. and Vassell, C. (2015) Using risk management to balance agile methods: a study of
the Scrum process. 2015 2nd International Conference on Knowledge- Based Engineering and Innovation (KBEI).
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
References 631
Goodwin, P. and Wright, G. (2014) Decision Analysis for Management Judgement, fifth edition. John Wiley and Sons, Chichester.
Greenspan, A. and Wooldridge, A. (2018) Capitalism in America: A History. Allen Lane, London. Grey, S. (1995) Practical Risk Assessment for Project Management. John Wiley and Sons,
Chichester. Harari, Y.N. (2014) Sapiens: A Brief History of Humankind. Harvill Secker, London. Hardy, C. and Macquire, S. (2016) Organising risk: discourse, power and “riskification”. Acad-
emy of Management Review, 41(1), 80– 108. Hawkins, C.J. and Pearce, D.W. (1971) Capital Investment Appraisal. Macmillan Studies in
Economics, Macmillan, London. Hertz, D.B. (1964) Risk analysis in capital investment. Harvard Business Review, 42(1), 95– 106. Hillson, D. (2004) Effective Opportunity Management for Projects: Exploiting Positive Risk. Mar-
cel Dekker, Inc, New York. Hillson, D. (2009) Managing Risk in Projects. Gower, Abingdon, Oxon. Hillson, D. (2012) History of the Special Interest Group on Project Risk Management of the Associa-
tion for Project Management. APM Press Limited, High Wycombe. HM Treasury (2003a) The Green Book: Appraisal and Evaluation in Central Government. HM
Treasury, 1 Horse Guards Road, London SW1A 2HQ. HM Treasury (2003b) The Green Book Supplementary Guidance – Optimism Bias. Downloaded
from www.hm- treasury.gov.uk, November 2010. HM Treasury (2014) Improving Infrastructure Delivery: Project Initiation Route Map Hand-
book. Downloaded from www.gov.uk/government/organisations/infrastucture- uk, February 2015.
Hopkinson, M. (2011) The Project Risk Maturity Model: Measuring and Improving Risk Manage- ment Capability. Gower, Farnham, Surrey.
Hopkinson, M. (2016) Net Present Value and Risk Modelling for Projects. Gower, Farnham, Surrey.
Hopkinson, M., Close, P., Hillson, D. and Ward, S. (2008) Prioritising Project Risks: A Short Guide to Useful Techniques. Association for Project Management (APM), Princes Risborough, Bucks.
Humble, J., Molesky, J. and O’Reilly, B. (2015) Lean Enterprise: How High Performance Organi- sations Innovate at Scale. O’Reilly Media, Sebastopol, CA.
Hyde, L. (2006) The Gift: How the Creative Spirit Transforms the World. First published in the US and Canada in 1983 by Random House, Cannongate, Edinburgh.
ICE and IFoA (1998) RAMP Risk Analysis and Management for Projects. Institution of Civil Engineers (ICE) and the Institute and Faculty of Actuaries (IFoA). Thomas Telford, London. ICE and IFoA (2005) RAMP Risk Analysis and Management for Projects – A Strategic Frame- work for Managing Project Risk and its Financial Implications, second edition. Institution of Civil Engineers (ICE) and the Institute and Faculty of Actuaries (IFoA). Thomas Telford, London.
International Standard (2009) ISO 31000 Risk Management – Principles and Guidelines. ISO, Switzerland.
Johnson, L. (2017) Every board needs a member of the awkward squad. The Sunday Times, 7 May, 12.
Johnson, L. (2018) Bosses discover it pays dividends to put workers first. The Sunday Times, 24 June, 9.
Kahneman, D. (2012) Thinking, Fast and Slow. Penguin Books, London. Kahneman, D., Slovic, P. and Tversky, A. (1982) Judgment Under Uncertainty: Heuristics and
Biases. Cambridge University Press, New York. Kaplan, R.S. and Norton, D.P. (1992) The balanced scorecard – Measures that drive perfor-
mance. Harvard Business Review, 70(1), 71– 79. Kaplan, R.S. and Norton, D.P. (1993a) Putting the balanced scorecard to work. Harvard Busi-
ness Review, 71(5), 134– 147.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
632 References
Kaplan, R.S. and Norton, D.P. (1993b) Using the balanced scorecard as a strategic management system. Harvard Business Review, 71(1), 75– 85.
Kaplan, R.S. and Norton, D.P. (1996) The Balanced Scorecard: Translating Strategy into Action. Harvard Business School Press, Boston.
Kennedy, P. (2014) Engineers of Victory: The Problem Solvers Who Turned the Tide in the Second World War. Penguin Books, London.
Keeney, R.L. and Raiffa, H. (1976) Decisions with Multiple Objectives. John Wiley and Sons, New York.
King, M. (2016) The End of Alchemy: Money, Banking and the Future of the Global Economy. Lit- tle, Brown Book Group, London.
King, M. and Mercer, A. (1985) Problems in determining bidding strategies. Journal of the Operational Research Society, 36, 915– 923.
King, M. and Mercer, A. (1991) Distributions in competitive bidding. Journal of the Operational Research Society, 42(2), 151.
Knight, F. (1921) Risk, Uncertainty and Profit. Houghton Mifflin, Boston. Lea, R. (2017a) Chinese in the driving seat with deal for Volvo Lorries. The Times, 28 December,
43. Lea, R. (2017b) Water firms risking their own demise. The Times, 18 October, 37. Leitch, M. (2008) Intelligent Internal Control and Risk Management. Gower, Aldershot,
Hants. Leitch, M. (2013) Working in uncertainty: What is ‘risk culture’ and how can ‘risk culture’
be Changed. Downloaded from www.workinginuncertainty.co.uk/risk_culture.shtml, 14 November 2013.
Lewis, M. (2017) The Undoing Project: A Friendship that Changed the World. Penguin Random House, London.
Lichtenberg, S. (2000) Proactive Management of Uncertainty using the Successive Principle. Pol- yteknisk Press, Copenhagen.
MacNulty, A.A.R. (1977) Scenario development for corporate planning. Futures, 9(2), 128– 138. Markowitz, H. (1959) Portfolio Selection: Efficient Diversification of Investments. John Wiley and
Sons, New York. Marsh, B. (1991) Coping with uncertainty: rational investment. Paper presented at ‘Chaos Fore-
casting and Risk Assessment’, Join Strategic and Operational Research Society Conference. London, February.
Marshall, T. (2015) Prisoners of Geography: Ten Maps That Tell You Everything You Need to Know About Global Politics. Elliot and Thompson Limited, London.
Mintzberg, H. (1994) The Rise and Fall of Strategic Planning. Prentice Hall, New York. Mintzberg, H., Ahlstrand, B. and Lampel, J. (1998) Strategy Safari. Prentice Hall Europe,
Hemel Hempstead. Moder, J.J. and Philips, C.R. (1970) Project Management with CPM and PERT. Van Nostrand,
New York. Morris, P.W.G. (2009) Implementing strategy through project management: the importance
of managing the project front end. Chapter 2, in Williams, T.M., Samset, K. and Suannevag, K.J. (editors), Making Essential Choices with Scant Information: Front End Decision Making in Major Projects. Palgrave Macmillan, Basingstoke.
Morris, P.W.G. (2013) Reconstructing Project Management. John Wiley and Sons, Chichester. Muller, R. and Turner, J.R. (2010) Project- Oriented Leadership. Gower, Abingdon, Oxon. Nichols, M. (2007) Review of Highways Agency’s Major Roads Programmer: Report to the Secre-
tary of State for Transport. Nichols Group London. Available on the Department for Trans- port (UK) website at www.dft.gov.uk/pgr/roads/nicholsreport/NUREG (1975) US Nuclear Regulatory Commission Reactor Safety Study – An Assessment of Accident Risk in US Commer- cial Power Plants. WASH- 1400, NUREG- 75/014.
Overman, S. (2014) The Conscience Economy: How a Mass Movement for Good is Great for Busi- ness. Bibliomotion, Brookline, MA 02445.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
References 633
Pagnamenta, R. (2017) Fukushima disaster is still radiating fallout nuclear industry wishes to avoid. The Times, 3 August, 41.
Pearce, I.F. (1964) A Contribution to Demand Analysis. Oxford University Press, Oxford. Pearson, G. (2012) The Road to Co- operation: Escaping the Bottom Line. Gower, Farnham Surrey
and Ashgate Publishing Company, Burlington Vermont USA. Perrow, C. (1984) Normal Accidents: Living with High Risk Systems. Basic Books, New York. Perrow, C. (1994) Accidents in high risk systems. Technology Studies, 1(1). Pidd, M. (1996) Tools for Thinking: Modelling in Management Science. John Wiley and Sons,
Chichester. Pinto, J.K. (2016) Project Management: Achieving Competitive Advantage, fourth edition. Pear-
son, USA. PMI (2008) Project risk management, Chapter 11, in A Guide to the Project Management Body
of Knowledge (PMBOK® Guide), fourth edition. Project Management Institute (PMI) Inc, Newtown Square, Pennsylvania.
PMI (2009) Practice Standard for Project Risk Management. Project Management Institute (PMI) Inc, Newtown Square, Pennsylvania.
PMI (2013) Project risk management, Chapter 11, in A guide to the Project Management Body of Knowledge (PMBOK® Guide), fifth edition. Project Management Institute (PMI) Inc, New- town Square, Pennsylvania.
PMI (2017a) Project risk management, Chapter 11, in A guide to the Project Management Body of Knowledge (PMBOK® Guide), sixth edition. Project Management Institute (PMI) Inc, New- town Square, Pennsylvania.
PMI (2017b) Waterfall Methodology Agile Approach. Downloaded from www.pmi.org/learning/ library/waterfall- methodology- agile- approach- 5821, 17 April 2017.
Purves, L. (2017) Small acts of kindness that can save a life. The Times, 3 July, 27. Rail Safety and Standards Board (2006) RSSB T430 – Definition of VPF & the Impact of Societal
Concerns – Final 30/1/2006. Oxford Risk Research and Analysis, Clarendon Enterprise Cen- tre, Oxford.
Raiffa, H. (1968) Decision Analysis: Introductory Lectures on Choices Under Uncertainty. Addison Wesley, Reading, MA.
Reason, J. (1990) Human Error. Cambridge University Press, Cambridge. Reason, J. (1997) Managing the Risks of Organizational Accidents. Ashgate, Farnham Surrey. Ricard, M. (2018) Altruism: The Science and Psychology of Kindness. Atlantic Books, London.
Earlier hardback edition 2015, French edition 2013. Rifkind, H. (2017) For generation rent, owning things is history. The Times, 3 October, 31. Rivett, P. (1994) The Craft of Decision Modelling. John Wiley and Sons, Chichester. Rivett, B.H.P. and Ackoff, R.L. (1967) A Manager’s Guide to Operational Research. John Wiley
and Sons, Chichester. Rosenhead, J. (1989) Rational Analysis for a Problematic World: Problem Structuring Methods for
Complexity, Uncertainty and Conflict. John Wiley and Sons, Chichester. Rosenhead, J. and Mingers, J. (2001) Rational Analysis for a Problematic World Revisited: Prob-
lem Structuring Methods for Complexity, Uncertainty and Conflict, second edition. John Wiley and Sons, Chichester.
Rubin, K. (2013) Essential Scrum. Addison- Wesley, Upper Saddle River, NJ. Ruthkopf, M.H. (1983) Auctions, Bidding and Contracting: Uses and Theory. New York Univer-
sity Press, New York. Samuelson, W. (1986) Bidding for contracts. Management Science, 32(12), 1533– 1550. Sasieni, M.W., Yaspan, A. and Friedman, L. (1959) Operations Research: Methods and Problems.
John Wiley and Sons, New York. Schoemaker, P.J.H. (1992) How to link strategic vision to core capabilities. Sloan Management
Review, 34(1), 67– 72. Schoemaker, P.J.H. (1995) Scenario planning: a tool for strategic thinking. Sloan Management
Review, 36(2), 25– 40.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
634 References
Senge, P.M. (1990) The Fifth Discipline: The Art and Practice of the Learning Organization. Doubleday, New York.
Taleb, N.N. (2007) The Black Swan: The Impact of the Highly Improbable. Allen Lane and Pen- guin Books Ltd, London.
Tamiz, M., Jones, D. and Romero, C. (1998) Goal programming for decision making: an over- view of the current state- of-the- art. European Journal of Operational Research, 111, 569– 581.
Taylor, F.W. (1911) The Principles of Scientific Management. Harper and Brothers, New York and London.
Temple, M. (2016) First Public Address as Chair of the Health and Safety Executive: Twenty Fifth Capita Health and Safety Lecture. The Royal College of Surgeons, London.
Toffler, A. (1970) Future Shock. Random House, London. Turner, J.R. (1992) The Handbook of Project Based Management: Improving Processes for Achiev-
ing Your Strategic Objectives. McGraw- Hill, New York. Turner, J.R. (2014) Gower Handbook of Project Management, fifth edition. Gower Press Aldershot. Tweedley, N. (1995) Winning the Bid: A Manager’s Guide to Competitive Bidding. Financial
Times Pitman, London. van der Heijden, K. (1996) Scenarios: The Art of Strategic Conversations. John Wiley and Sons,
Chichester. Ward, S.C. (1989) Arguments for constructively simple models. Journal of the Operational
Research Society, 40(2), 141– 153. Ward, S.C. (2005) Risk Management Organisation and Management. Institute of Risk Manage-
ment (IRM) Series. Witherby & Co. Ltd, London. Ward, S.C. and Chapman, C.B. (1994) Choosing contractor payment terms. International Jour-
nal of Project Management, 12(4), 216– 221. Ward, S.C. and Chapman, C.B. (1995) Evaluating fixed price incentive contracts. Omega – The
International Journal of Management Science, 23(1), 49– 62. Ward, S.C. and Chapman, C.B. (2008) Stakeholders and uncertainty management in projects.
Construction Management and Economics, 26(6), 563– 578. Ward, S.C., Chapman, C.B. and Curtis, B. (1991) On the allocation of risk in construction pro-
jects. International Journal of Project Management, 9(3), 140– 147. Williams, T.M. (2002) Modelling Complex Projects. John Wiley and Sons, Chichester. Williams, T.M., Samset, K. and Suannevag, K.J. (2009) Making Essential Choices with Scant
Information: Front End Decision Making in Major Projects. Palgrave Macmillan, Basingstoke. Williams, T.M. and Samset, K.J. (2012) Project Governance: Getting Investments Right. Palgrave
Macmillan, Basingstoke. Woods, D.D., Dekker, S., Cook, R., Johannesen, L. and Starter, N. (2010) Beyond Human
Error, second edition. Ashgate Publishing Ltd, Farnham Surrey. Woolsey, R.E.D. (2003) Real World Operations Research: The Woolsey Papers. Abe Books, USA. Wynne, B. (1992) Uncertainty and environmental learning – Reconceiving science and policy in
the preventative paradigm. Global Environmental Change, June, 111– 127. Yamane, T. (1973) Statistics: An Introductory Analysis, third edition. Harper & Row, New York. Zenger, T. (2016) Beyond Competitive Advantage. Harvard Business Review Press, Cambridge,
MA.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Website information
The website associated with this book may be accessed using www.enlightenedplanning.uk
website address.
The short- term plan is to launch this website shortly after the book’s publication, using it
to support those who want to use aspects of this book when leading professional courses and
workshops for practitioners and their managers, including board-level managers, or teach-
ing university courses. Initial support focuses on perspectives and areas of application where
I have relevant experience to share in terms of slides, case studies, and other material which
might be useful. I plan to provide papers on a few particular aspects of enlightened planning
and more general systematic simplicity concerns.
I also plan to indicate websites with relevant papers and other material created by others
that I am aware of with a view to contributing to a wider discussion of systematic simplicity
approaches. As one example, my colleague Matthew Leitch, who provided a contribution to
the ‘Comments by other colleagues’ section in the front of this book, has included his web-
site address, and one paper available on it has been cited and included in this book’s refer-
ences, illustrating one website address with papers directly relevant to enlightened planning
which I plan to list on the book’s website. As another example, Stephen Cresswell, who also
provided a contribution to the ‘Comments by other colleagues’ section and included his
website address, has papers plus other material you may find helpful available on his website,
illustrating further kinds of material which others might make available in a similar manner
via their website addresses on this book’s website.
Suggestions for further material directly relevant to enlightened planning or systematic
simplicity and an address to access it will be sought via this book’s website. The long- term
plan is this book’s website evolving to provide access to material provided by myself and
others which I am aware of and can help to make others aware of in a manner which is as
balanced and non- judgemental as possible on a simple website.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Ackoff, Russell 59 – 60, 67, 400, 415 ambiguity 20 – 24, 76 – 77, 108 – 109, 202 – 203,
225 – 226, 276 – 277; clarified 311 – 313, 329, 376, 378, 387, 606 – 607, 616; and a goals-plan relationships framework 342 – 344, 351 – 352; and key corporate clarity concerns 387 – 388; and a project lifecycle framework 319 – 320, 324 – 325; and regulation risk 364 – 365; resolvable 560 – 561; and the seven Ws framework 328 – 329; and the strategy gateway stages 354 – 355; and technical workshops 369 – 370; see also ambiguity uncertainty
ambiguity uncertainty 19 – 20, 22 – 23, 108 – 109, 342 – 344, 369 – 370; as key aspect of estimates 371 – 378, 374, 377
analysis: formal 296, 561 – 566; probability of winning analysis 230, 232 – 234, 232, 238, 261 – 272, 264, 264 – 266, 267, 269; safety and security analysis 483, 505, 566 – 568; scope of 5, 25; see also under uncertainty
Association for Project Management (APM) 309 – 310, 366, 597, 611, 616
assumptions: framing assumptions 10 – 18, 31 – 33, 206 – 208, 394 – 396, 547 – 549, 611 – 613; front-end 196 – 197; stealth assumptions 15 – 17, 177 – 180, 372, 396 – 398, 499, 515 – 516
Baker, Gerard 622 bidding 33 – 34, 61 – 62, 212 – 216, 216,
296 – 306; bidding strategy and tactics development 237 – 240, 238; bid evaluation and development 274 – 275; and change management 287 – 292; and credibility 292 – 296; and enhanced cost uncertainty analysis 240 – 261, 241, 243, 246, 249, 252, 257; and enhanced probability of winning analysis 261 – 272, 264, 264 – 266, 267, 269; and enhanced uncertainty syntheses
272 – 274, 273; and identifying no-hopers 229 – 237, 230, 232, 235; and identifying non-starters 224 – 229; and a preliminary proposal document 216 – 221, 218 – 219; and the rationale for separate stages 237; and workshop presentation of the SP 221 – 224, 275 – 287, 279, 282, 284 – 285
Black, Warren xix, 616 black swans 20, 179 boundary concepts: clarity efficient 105 – 107,
109 – 110 BP International 20, 27, 31; and building
trust 562, 572 – 573; and clarity efficient boundary concepts 105 – 107, 109 – 110; and confronting complexities 132 – 133, 136 – 138, 157, 159; and corporate strategy formulation 425, 433, 435, 437; and enhancement of clarity 614, 618, 624; and enlightened prudence 548, 551; and generalised foundations for safety and security 482; and generic processes 307, 310, 347 – 349, 382 – 383, 386 – 388, 392 – 393, 397; and a HAT approach 84, 87, 94 – 95; and low to high clarity approaches 74, 104 – 105; and multiple sources uncertainty 97 – 104; and opportunity efficiency 129; and priorities 598 – 599; and probability distribution 91; and revealed preferences 555; and risk efficiency 110 – 119; and sensitivity diagrams 552 – 553; and specific and general responses 120 – 121, 124, 126 – 127; and specific process 252, 268, 270 – 271, 292, 297; and using risk efficiency 535, 544
capability-culture uncertainty 19 – 20, 22 – 23, 108 – 109, 369 – 370
Central Limit Theorem 94 – 95 Centre for Operational Research,
Management Science and Information Systems (CORMSIS) xli
Index
Note: page numbers in italic indicate a figure and page numbers in bold indicate a table on the corresponding page.
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Index 637
change management 195, 210, 287 – 288, 294; and building trust 579; and corporate strategy formulation 465, 469, 474; and enhancement of clarity 607; and generic processes 315 – 319, 356 – 357, 386, 389 – 393; and priorities 581, 594 – 595, 599 – 600
change management planning 357, 599 – 600 Chapman, Chris 146, 501, 616; and Dale
Cooper 93, 104, 174, 392; and M. Howden 392; and S.C. Ward 22, 26, 31 – 32, 65 – 66, 139, 295, 310 – 311, 314, 327, 333, 347, 353, 359 – 360, 362, 365 – 366, 369 – 370, 372, 375, 378, 382, 385 – 386, 391 – 392, 394, 480, 552, 602, 605, 612, 614, 616
clarity: corporate 141, 204, 386 – 388; minimum clarity approach 71 – 76, 79 – 80, 89 – 90, 105 – 110, 373 – 375; see also clarity efficiency; strategic clarity; tactical clarity
clarity efficiency 27 – 32, 34 – 36; and building trust 478 – 479; clarity efficient boundary concepts 105 – 110, 106, 129; and corporate strategy formulation 464 – 465, 431 – 435; and generic processes 312 – 313; and low to high clarity approaches 71 – 75, 82 – 83, 95 – 96, 128 – 130; and specific process 222 – 223; and universal process 164 – 165, 201 – 205
closure with completeness 53 – 54, 66, 148, 263, 268; and building trust 510, 512, 520
contingency planning 3 – 4, 26, 38, 77 – 79, 82 – 84, 88; and building trust 557, 579; and clarity efficient boundary concepts 107 – 110; and confronting complexities 138, 148; and corporate strategy formulation 411 – 412, 417, 435, 445; and enhancement of clarity 601, 605 – 607, 612; and generic processes 343, 349, 352; and multiple sources uncertainty 96 – 98; and opportunity efficiency 129 – 131; and risk efficiency 110 – 112, 114; and specific and general responses 126; and specific process 277
Cooper, Dale 59, 93, 104 corporate clarity see under clarity corporate management 5 – 6, 9, 43, 96,
122, 125; and building trust 480; and confronting complexities 156, 160; and corporate strategy formulation 401; and enhancement of clarity 607 – 608; and generic processes 311, 314 – 315, 319 – 320, 366; and priorities 600; and specific process 212; and universal process 208
corporate planning 6, 9, 13, 35 – 37; and building trust 477, 571 – 572, 579; and confronting complexities 162 – 163;
corporate planning horizons 399, 407 – 419, 408; and corporate strategy formulation 399 – 401, 405, 441, 454, 465 – 466, 469 – 470, 472 – 474; and enhancement of clarity 601, 605, 607 – 610, 612, 614 – 615; and generic processes 315, 396 – 397; and low to high clarity approaches 72, 122, 125; and priorities 581 – 582, 593 – 596; and universal process 166, 183, 194 – 196, 203
correlation 93 – 94, 97 – 98, 428 – 430, 433 – 434, 508 – 513, 519 – 520; see also systematic uncertainty
Critical Path Method (CPM) 66, 72, 76, 97 – 98, 110; and generic processes 308 – 310, 333, 340, 342
decision making 3 – 4, 9 – 10, 38 – 39, 120 – 121, 124 – 125, 142 – 146; biased 139, 142, 208, 601, 603 – 605; and building trust 477 – 479, 487 – 490, 551 – 552; decision-making process goals 155 – 160; and enhancement of clarity 603 – 605, 623 – 624; and generic processes 312 – 313, 357 – 358
efficiency see clarity efficiency; opportunity efficiency; risk efficiency
enterprise risk management (ERM) 33, 366, 395, 478, 613
estimation-efficiency spectrum 36, 73, 120, 130 – 131
ethics 34, 51, 56, 131; and building trust 479, 495, 510, 514 – 515; corporate 299, 478; personal 569 – 570
event uncertainty 19, 22 – 24, 108 – 109, 367 – 371
Ford Pinto 492 – 496, 508, 519, 546, 577 formal planning 3 – 5, 10, 14, 37, 58; and
building trust 476 – 478, 535, 545 – 547, 572, 576; and confronting complexities 132, 153 – 154; and corporate strategy formulation 410, 412, 414, 418; and enhancement of clarity 602, 608; and generic processes 342, 391; and low to high clarity approaches 89, 96, 115 – 116; and priorities 596, 599 – 600; and specific process 212, 222
frowns test 227, 255 futures planning 203, 419 – 420, 435,
455 – 457; and building trust 571, 576, 579; and corporate planning horizons 408, 410; and corporate strategy formulation 400, 464, 472 – 473, 475; and enhancement of clarity 605, 615 – 616, 620; and organisational separability 466 – 469; and priorities 583, 587, 591 – 594, 598
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
638 Index
Gawande, Atul 19 – 20, 62 – 64 generic process 307 – 308; and basic problems
317 – 319; and change management strategy 389 – 391; and clarity inefficient toolsets 394 – 398; and credibility 391 – 394; and the discount rate 357 – 363; and enlightened approaches to governance, teams, and contracts 388 – 389; and a goals-plans relationships framework 332; and key corporate clarity concerns 386 – 388; and the need to clarify ambiguity 311 – 313; and a project lifecycle framework 319 – 325, 319, 321; and project management 314 – 315; and the projects group 315 – 317; and regulation risk 363 – 365; and the seven Ws framework 325 – 332, 326, 327; and the strategy gateway stages 353 – 357, 354; and the strategy progress stages 332 – 353, 334, 335, 337, 341, 348; and technical workshops 365 – 386, 367, 374, 377; used for project planning 308 – 311
goals planning 203; and building trust 571, 576; and corporate strategy formulation 400, 408 – 410, 419, 455, 457 – 458, 461 – 462, 464 – 473; and enhancement of clarity 605, 615, 620; and priorities 583, 586 – 587, 590, 592 – 593
Graphical Evaluation and Review Technique (GERT) 98 – 99, 101, 110, 125 – 126
Harari, Yurval 69 histogram and tree (HAT) 19, 32; and
building trust 486, 502, 538 – 539; and confronting complexities 134, 136, 138 – 140; and corporate strategy formulation 475; and enhancement of clarity 601; and long-term planning 420 – 424, 427 – 428, 431 – 435, 443, 448 – 449; and low to high clarity approaches 74, 88 – 92, 98 – 100, 104, 107 – 108, 128 – 129; and Monte Carlo simulation approaches 617 – 619; and short- term planning 451; and specific process 235, 263, 265 – 268; a synthesis underlying a HAT approach 92 – 96; a two-scenario HAT example 80 – 87, 81, 85, 85; and universal process 174, 199, 204
holding costs 169, 175, 176 – 177, 182 – 183 Hopkinson, Martin 392, 611
IBM 117 – 121, 124 – 125, 129, 146, 149, 156; and building trust 544, 548; and corporate strategy formulation 437, 471; and generic processes 308; and priorities 591 – 592, 599; and specific process 212, 252, 292 – 294, 296, 301; and universal process 200, 202
informal planning 3 – 5, 14, 37, 47, 69, 131; and confronting complexities 151 – 152, 155, 157; and corporate strategy formulation 401, 418, 475; and enhancement of clarity 602 – 603; and generic processes 342 – 343, 391; and priorities 596
initial goals planning 409 – 410, 592 invitation to tender (ITT) 215, 217 – 218,
225, 231 – 233, 253, 257 – 259, 300
John Lewis Partnership 193, 211 Johnson, Johnnie xli Johnson, Luke 211, 406
Kennedy, Paul 62 – 64 King, Mervyn 621 – 622 KISS (Keep It Simple Systematically) xx, xliii,
3 – 4, 11, 34 – 37, 145; and building trust 568; and corporate strategy formulation 404; and universal process 207
lean and agile approaches 601, 614 – 617, 620 Leitch, Matthew 22, 603 long-term planning 419 – 420, 428, 434,
437, 439 – 441, 447 – 449; and building trust 571, 576; and corporate planning horizons 408 – 419; and corporate strategy formulation 400, 402, 405, 463 – 465, 472; and enhancement of clarity 605, 615, 620; integrated short/medium/long-term planning 454 – 455; and organisational separability 465 – 468; and priorities 584, 593 – 594
management see change management; corporate management; operations management; project management; risk management
medium-term planning 400, 408, 411, 416 – 420, 439 – 441, 444, 446 – 449; and building trust 571; and corporate strategy formulation 472; and enhancement of clarity 605; and organisational separability 465, 468; and priorities 594
Monte Carlo simulation 32, 74, 82, 93 – 95, 104, 136; and building trust 539; and corporate strategy formulation 435, 448 – 449; and enhancement of clarity 601, 617 – 619; and generic processes 347, 350, 371; and specific process 252, 257, 265, 271
multiple objectives 132, 152 – 155; general conceptual framework for 143 – 146
My-Place-Your-Timing approach 184 – 185, 187, 201, 204, 209, 304, 343
nationalisation 363 – 365, 584 – 593 no-hopers 216, 221 – 222, 229 – 237 non-starters 216, 221 – 222, 224 – 230, 237 Net Present Values (NPV) calculations 142,
145, 311, 327, 357 – 363, 395, 604 – 605
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Index 639
Operational Research (OR) 45, 47, 54 – 55, 58, 59 – 60, 64, 66 – 70; in the 1930s and 1940s 55 – 57; in the 1950s and 1960s 57 – 59; and qualitative factors 60 – 61; and Russell Ackoff 59 – 60; and specific process 294 – 295; and universal process 165, 167 – 168, 180, 185, 196, 204
Operational Research Society (ORS) xliii, 611 – 612
operations management 5 – 6, 71 – 72, 106, 120; and building trust 572; and confronting complexities 137, 146, 160; and corporate strategy formulation 405, 475; and enhancement of clarity 609 – 610; and generic processes 313 – 315, 319 – 320, 379; and priorities 595 – 596; and specific process 212, 214; and universal process 199, 209
operations planning 6, 8, 36 – 37, 122, 164, 167; and building trust 562, 572; and enhancement of clarity 601, 605, 609 – 610, 615; and priorities 593, 595 – 596, 599
opportunity efficiency 27 – 36, 110 – 112, 128 – 131, 148 – 150, 154 – 156, 205 – 206; and building trust 491 – 492; and enhancement of clarity 625, 626; and enlightened prudence 120 – 122; and generic processes 340 – 344
Overman, Steven 210 – 211, 299
Pidd, Michael 204 planning: interpreting the plans 153,
172 – 173; planning application 5 – 9, 331; planning capability 38, 98, 190, 594 – 599; planning tools 44 – 45, 615; process planning 601, 611 – 612; testing the plans 153, 172 – 173; see also change management planning; contingency planning; corporate planning; formal planning; futures planning; goals planning; informal planning; long-term planning; medium- term planning; operations planning; planning context; planning process; project planning; safety and security planning; short-term planning
planning context 20, 24, 65 – 66; and building trust 528; and confronting complexities 154; contextual variables ix; and corporate strategy formulation 404, 413, 439, 441, 472; and enhancement of clarity 602, 608, 611; and generic processes 316, 325, 352, 396 – 397; and low to high clarity approaches 72, 110 – 111, 125, 127 – 128; and priorities 596; and universal process 164
planning process 5, 49 – 50, 58, 152, 182 – 183, 336; and building trust 572; and corporate strategy formulation 441,
446 – 449, 453, 455, 472; and enhancement of clarity 608, 611 – 612
planning tools see generic process; specific process; universal process
plans/models distinction 181 – 186 Preliminary Proposal Document (PPD)
217 – 223, 218, 231 – 232, 239 – 240 Probabilistic Project Planning (PPP)
101 – 103, 460 probability-impact graphs (PIGs) 299,
365 – 371, 367, 394 – 395, 397 probability of winning 229 – 239, 261 – 262,
266 – 267, 273 – 274, 282 – 284; see also under analysis
process models 57, 212 – 213, 305, 352 Program Evaluation and Review Technique
(PERT) 94, 97 – 99, 101 – 102, 107, 109 – 110, 125; and building trust 563; and enhancement of clarity 617; and generic processes 308 – 310, 325, 333, 340; and specific process 268
project lifecycle framework 307, 318 – 325, 319, 321, 327, 332
project management 5 – 6, 64; and building trust 480; and a change management strategy 392; and clarity inefficient toolsets 397; and confronting complexities 137, 153, 160, 162; and corporate strategy formulation 475; and enhancement of clarity 607, 609 – 611, 614 – 616; and generic processes 308 – 311, 319; and a goals-plan relationships framework 342, 353; and low to high clarity approaches 72, 105, 108, 112, 117, 120, 129; and priorities 594, 597; the role of 314 – 315; and the seven Ws framework 329; and technical workshops 366, 372, 385; and universal process 199
Project Management Institute (PMI) 310, 366, 611, 616
project planning 6 – 8, 11, 20, 37, 45, 65 – 67; and building trust 477, 491, 552, 562, 572, 579; and a change management strategy 356, 391; and clarity inefficient toolsets 396 – 397; and confronting complexities 152, 162; and corporate strategy formulation 399, 405, 441, 465; and enhancement of clarity 601, 605, 607, 609 – 611, 614 – 616; and generic processes 307 – 308, 310 – 311, 313, 318; and a goals-plan relationships framework 333, 336, 339 – 340, 342, 346, 352; and low to high clarity approaches 72 – 74, 97, 101 – 103, 110 – 111, 122, 125, 128 – 129; and priorities 586, 593 – 596, 599; and the seven Ws framework 325, 328; and technical workshops 371, 379, 386; and universal process 195 – 196
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
640 Index
Project Risk Analysis and Management Guide (PRAM) 309 – 311, 366
project risk management (PRM) 103 – 104, 141, 155, 307, 309 – 311, 317; and a change management strategy 392; and clarity inefficient toolsets 394 – 395, 397; and enhancement of clarity 614, 616, 624; and a goals-plan relationships framework 342 – 344, 346 – 347, 351 – 352; and priorities 585; and technical workshops 366, 370 – 372
red face test 226 – 227, 256, 296, 299, 302 – 303, 496
Risk Analysis and Management of Projects (RAMP) 309, 311
risk efficiency 27 – 34, 110 – 112, 114 – 125, 128 – 130, 163 – 165, 244 – 247; and building trust 534 – 535, 542 – 544; and corporate strategy formulation 431 – 432, 435 – 438
risk management 21 – 24, 33, 35, 38; and building trust 482, 491, 541, 561, 579; and a change management strategy 392 – 394; and clarity inefficient toolsets 394 – 397; and confronting complexities 141, 155; and corporate strategy formulation 439; and enhancement of clarity 601, 603, 613 – 614, 616, 624; and generic processes 307, 309 – 311; and a goals-plan relationships framework 342 – 344, 346 – 347, 351 – 352; and low to high clarity approaches 103 – 104, 110 – 111, 120, 125; and priorities 585, 597 – 598; and a project lifecycle framework 324 – 325; and the role of project management 316 – 317; and technical workshops 366 – 367, 369 – 372, 375, 378, 385; and universal process 207
safety and security planning 66; and building trust 483, 528 – 529, 532 – 533, 556, 562 – 564, 571 – 572, 575; and enhancement of clarity 601, 612 – 613; and priorities 583
sensitivity diagrams 139, 267 – 269, 267, 270; and building trust 496, 534, 550 – 555, 552, 562; and corporate strategy formulation 439; and enhancement of clarity 618 – 619; and generic processes 347 – 350, 348, 356, 368 – 369, 374, 375; as portals to further opportunities 132 – 137, 133; portrayal of option costs 552 – 555, 552
seven Ws framework 7 – 8, 47, 60, 65, 154, 325 – 332; and building trust 572; and a change management strategy 389, 393; and clarity inefficient toolsets 397; and corporate strategy formulation 470; and generic processes 307, 318; and a goals-plan
relationships framework 332 – 334, 339 – 340, 344, 351 – 353; and priorities 595 – 596; and technical workshops 377 – 379, 383
short-term planning 6, 203; and building trust 571; and corporate planning horizons 407 – 410, 412 – 414, 419; and corporate strategy formulation 400, 405, 449, 455, 457, 472; and enhancement of clarity 615; and organisational separability 468; and priorities 593 – 596
smiles test 227 – 228, 299, 303, 496, 579, 601, 624 – 626
source-response diagrams 137 – 139, 137, 159, 369
Southampton Business School (SBS) xli Southampton University Management School
(SUMS) xl Specific, Measurable, Achievable, Relevant,
and Time-framed (SMART) 151 specific process (SP) 12 – 13, 60, 101, 118;
and bidding process 212 – 216, 216, 275 – 286, 279, 282, 284 – 285, 296 – 306; bidding strategy and tactics development 237 – 240, 238; bid evaluation and development 274 – 275; and building trust 480; and change management 287 – 292; and confronting complexities 133, 153, 155; and corporate strategy formulation 403, 470, 474; and credibility 292 – 296; enhanced cost uncertainty 240 – 261, 241, 243, 246, 249, 252, 257; enhanced probability of winning analysis 261 – 272, 264, 264 – 266, 267, 269; enhanced uncertainty syntheses 272 – 274, 273; and enhancement of clarity 616; and generic processes 307 – 313, 319, 325, 333, 338, 352, 390, 393; identifying no-hopers 229 – 237, 230, 232, 235; identifying non-starters 224 – 229; and a preliminary proposal document 216 – 221, 218 – 219; and priorities 595 – 596; and the SPP concept 397 – 398; and universal process 197, 209; and UP phases 64 – 66; workshop presentation of 221 – 224; see also specific process, design of
specific process, design of 212 – 213, 223 – 224, 275 – 287, 279, 282, 284 – 285; bidding strategy and tactics development for 237 – 240, 238; bid evaluation and development for 274 – 275; change management for 287 – 292; context information for 213 – 216, 216; and corporate goals 591 – 592; and credibility 293 – 296; enhanced cost uncertainty for 240 – 261, 241, 243, 246, 249, 252, 257; enhanced probability of winning for 261 – 272, 264, 265 – 266, 267, 269;
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
Index 641
enhanced uncertainty for 272 – 274, 273; and formal analysis of uncertainty 296; identifying no-hopers for 229 – 237, 230, 232, 235; identifying non-starters for 224 – 229; Preliminary Proposal Document for 216 – 221, 218 – 219; and process iterations 300; and strong assumptions 297 – 298; and time spent 301; and top- down use of the UP concept 302 – 303, 305; and traffic light issues 299; workshop presentation of SP 221 – 223
specific process for projects (SPP) 325, 333 – 340, 343 – 347, 350 – 356; and a change management strategy 289 – 393; and clarity inefficient toolsets 396 – 398; and corporate strategy formulation 470; and technical workshops 365, 373, 376, 379 – 384
strategic clarity 35 – 37, 39 – 40, 142 – 143, 268 – 269, 276 – 277, 312; and biased decision making 603 – 605; and building trust 476 – 478; and contingency planning 605 – 607; and corporate planning 607 – 609; and lean and agile approaches 614 – 617; and multiple sources of uncertainty 617 – 619; and operations planning 609 – 610; and process planning 611 – 612; and project planning 610 – 611; and public pressure 624; and regulation 619 – 623; and risk management 613 – 614; and safety and security planning 612 – 613; and the smiles test 624 – 627, 625; and teamwork and collaboration 601 – 603
Synergistic Contingency Evaluation and Response Technique (SCERT) 99, 101 – 102, 104 – 106, 109 – 110; and building trust 551, 562; and confronting complexities 132 – 133, 136, 139, 155, 159; and enhancement of clarity 605, 612, 617, 624; and generic processes 382, 386, 393, 397; and opportunity efficiency 129 – 130; and risk efficiency 112, 114, 116, 119; and specific and general responses 126 – 127; and specific process 268
systematic simplicity 37, 619, 627 systemic uncertainty 19 – 24, 94, 108 – 109,
169, 204; and building trust 483; and corporate strategy formulation 418, 427; and enhancement of clarity 606; and generic processes 342, 369 – 370, 373, 394; and specific process 278, 286
tactical clarity 35 – 37; and biased decision making 603 – 605; and contingency planning 605 – 607; and corporate planning 607 – 609; and lean and agile approaches 614 – 617; and multiple sources
of uncertainty 617 – 619; and operations planning 609 – 610; and process planning 611 – 612; and project planning 610 – 611; and public pressure 624; and regulation 619 – 623; and risk management 613 – 614; and safety and security planning 612 – 613; and the smiles test 624 – 627, 625; and teamwork and collaboration 601 – 603
Thornton, Paul 299 trade-offs 29 – 32, 121 – 125, 128 – 129,
144 – 146, 206 – 207, 341 – 342; appreciation of 205; and building trust 477 – 478, 495 – 496, 529 – 530, 544 – 545, 559 – 561, 564 – 565, 577 – 578; and corporate strategy formulation 467 – 468; effective low clarity views of 146 – 149, 147; multiple objective trade-off approaches 31 – 32, 145; and priorities 596 – 598
UK Ministry of Defence (MoD) xxix – xxx, xxxiv – xxxv, 74, 88, 107, 114; and building trust 477, 567 – 568; and confronting complexities 141 – 142; and generic processes 308; and specific process 242
uncertainty: cost uncertainty analysis 231 – 232, 230, 238, 240 – 261, 241, 243, 246, 249, 252, 257; cost uncertainty- time profile graph 376 – 377, 377; uncertainty syntheses 230, 234 – 235, 235, 238, 272 – 274, 273; see also ambiguity uncertainty; capability-culture uncertainty; event uncertainty; systemic uncertainty; variability uncertainty
universal process (UP) 12 – 13, 16, 79 – 95, 132 – 133, 151 – 156; bottom-up 303 – 304; and building trust 480, 569 – 570; capability-culture concepts 49 – 53, 52; and capture the context assumptions 186 – 196; closure with completeness concept 53 – 54; and corporate strategy formulation 399, 403, 470, 472, 474; and creativity 180 – 181; and enhancement of clarity 601, 603, 611; EP clarification of 68 – 70; and front-end working assumptions 196 – 197; and generic processes 308 – 310; and a goals-plan relationships framework 333 – 334, 334, 336 – 339, 343 – 347, 350, 352 – 353; and initial model creation 168 – 172, 169, 171; initiating the process 166 – 168; key concepts 198 – 210; and looking wider or deeper 197 – 198; overview of 41 – 43, 42; and the plans/ models distinction 181 – 186; and priorities 592; process initiation and termination 43; process phase structure 43 – 49; provenance of 54 – 67; refining the analysis 173 – 176, 175; and relevant influences 176 – 180, 178,
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
642 Index
178; and specific process 213 – 215, 223, 294; SPP concept and 397 – 398; and the strategy gateway stages 355; testing and interpreting the plans 172 – 173; top-down use of 302 – 306; UP generality in principle and in practice 54; using a UP 164 – 166, 210 – 211
unknown unknowns 20 – 21, 102 – 105, 386 – 388
variability uncertainty 19, 22 – 24, 108 – 109, 169, 369 – 370, 543
Ward, S.C. 22, 26, 31 – 32, 65 – 66, 139, 295, 310 – 311, 314, 327, 333, 347, 353, 359 – 360, 362, 365 – 366, 369 – 370, 372, 375, 378, 382, 385 – 386, 391 – 392, 394, 480, 552, 602, 605, 612, 614, 616
workshops: for bidding overview 221 – 223, 275 – 277, 298, 302; bidding strategy and tactics development in 237 – 240, 238; bid evaluation and development in 274 – 275; and building trust 481; change management in 287 – 292; enhanced cost uncertainty analysis in 240 – 261, 241, 243, 246, 249, 252, 257; enhanced probability of winning analysis in 261 – 272, 264, 264 – 266, 267, 269; enhanced uncertainty syntheses in 272 – 274, 273; and generic processes 314, 317, 320, 344 – 349, 352, 356; identifying no-hopers in 229 – 237, 230, 232, 235; identifying non-starters in 224 – 229; IPMA Copenhagen advanced training workshop 480; leadership in 277 – 287; separate stages in 237; technical 365 – 386
Downloaded by Vignesh Sivadass ([email protected])
lOMoARcPSD|13426921
- Cover
- Half Title
- Title
- Copyright
- Dedication
- Contents
- Foreword by Stephen Ward
- Comments by other colleagues
- Preface
- Overview
- About the author
- Acknowledgements
- Part 1 Foundations
- 1 Why planning is usually vital but often difficult and frequently inept
- 2 A ‘universal planning uncertainty and complexity management process’ (UP)
- 3 Low to high clarity approaches and the ‘estimationefficiency spectrum’
- 4 Confronting challenging complexities usually needing more clarity
- Part 2 Employing planning tools in practice – five illustrative tales
- 5 Using a UP – an initially simple supply chain management example
- 6 Building ‘specific processes’ – a bidding process example
- 7 Adapting ‘generic processes’ – a project planning example
- 8 Corporate strategy formulation – an electricity utility example
- 9 Building wellfounded trust about complex concerns – a railway safety example
- Part 3 Further synthesising and reflecting
- 10 Immediate and longer term ‘what needs to be done’ priorities
- 11 Ongoing enhancement of strategic clarity and tactical clarity
- References
- Website information
- Index
