MBA OPERATIONS MANAGEMENT DUE IN 4 HOURS

ffirs.indd iffirs.indd i 29/09/12 9:27 AM29/09/12 9:27 AM

Operations Management for MBAs

ffirs.indd iffirs.indd i 01/10/12 5:25 PM01/10/12 5:25 PM

ffirs.indd iiffirs.indd ii 01/10/12 5:25 PM01/10/12 5:25 PM

Operations Management for MBAs

Fifth Edition

Jack R. Meredith

Scott M. Shafer Wake Forest University

ffirs.indd iiiffirs.indd iii 01/10/12 5:25 PM01/10/12 5:25 PM

VICE PRESIDENT & EXECUTIVE PUBLISHER George Hoffman EXECUTIVE EDITOR Lisé Johnson PROJECT EDITOR Brian Baker ASSOCIATE DIRECTOR OF MARKETING Amy Scholz MARKETING MANAGER Kelly Simmons MARKETING ASSISTANT Marissa Carroll PRODUCT DESIGNER Allison Morris MEDIA SPECIALIST Ethan Bernard SENIOR CONTENT MANAGER Lucille Buonocore SENIOR PRODUCTION EDITOR Anna Melhorn PHOTO DEPARTMENT MANAGER Hillary Newman DESIGN DIRECTOR Harry Nolan COVER DESIGNER Wendy Lai PRODUCTION MANAGEMENT Ingrao Associates

This book was set in 10/12 ITC Garamond light by MPS Limited and printed and bound by RRD/Jefferson City. The cover was printed by RRD/Jefferson City. This book is printed on acid free paper.

Copyright © 2013, 2010, 2007, 2002, 1999 John Wiley & Sons, Inc. All rights reserved.

No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning or otherwise, except as permitted under Sections 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authoriza- tion through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc. 222 Rosewood Drive, Danvers, MA 01923, website www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030-5774, (201)748-6011, fax (201)748-6008, website http://www.wiley.com/go/permissions.

To order books or for customer service, please call 1-800-CALL WILEY (225-5945).

ISBN 13 9781118369975

Printed in the United States of America

10 9 8 7 6 5 4 3 2 1

ffirs.indd ivffirs.indd iv 01/10/12 5:25 PM01/10/12 5:25 PM

This book is dedicated to the Newest Generation:

Avery, Mitchell, and Ava

J. R. M.

Brianna, Sammy, and Kacy

S. M. S.

ffirs.indd vffirs.indd v 01/10/12 5:25 PM01/10/12 5:25 PM

ffirs.indd viffirs.indd vi 01/10/12 5:25 PM01/10/12 5:25 PM

Contents

Chapter 1: Operations Strategy and Global Competitiveness/1

Operations/6 Customer Value/15 Strategy and Competitiveness/26

Chapter 2: Process Planning and Design/47

Forms of Transformation Systems/51 Selection of a Transformation System/73

Chapter 3: Controlling Processes/91

Monitoring and Control/94 Process Monitoring/95 Process Control/102 Controlling Service Quality/110

Chapter 4: Process Improvement: Minimizing Variation Through Six Sigma/119

Approaches for Process Improvement/124 Business Process Design (Reengineering)/125 Six Sigma and the DMAIC Improvement Process/129 Example Six Sigma Project/132 The Define Phase/133 The Measure Phase/138 The Analyze Phase/149 The Improve Phase/155 The Control Phase/158 Six Sigma in Practice/158

vii

ftoc.indd viiftoc.indd vii 25/09/12 4:16 PM25/09/12 4:16 PM

viii C o n t e n t s

Chapter 5: Process Improvement: Reducing Waste Through Lean/167

History and Philosophy of Lean/171 Traditional Systems Compared with Lean/173 Specify Value/180 Identify the Value Stream/181 Make Value Flow/186 Pull Value Through the Value Stream/191 Pursue Perfection/194 Benefits of Lean/196 Lean Six Sigma/197

Chapter 6: Managing Process Improvement Projects/203

Defining a Project/206 Planning the Project/208 Scheduling the Project/218 Controlling the Project: Earned Value/233

Chapter 7: Supply Chain Management/241

Defining Supply Chain Management/246 Supply Chain Strategy/249 Supply Chain Design/253 Outsourcing and Global Sourcing/259 Inventory Management/265 Role of Information Technology/270 Successful Supply Chain Management/276 Chapter 7 Supplement A: The Beer Game/283 Chapter 7 Supplement B: The Economic Order Quantity Model/289

Chapter 8: Capacity, Scheduling, and Location Planning/297

Long-term Capacity Planning/301 Location Planning Strategies/307 Locating Pure Services/315 Effectively Utilizing Capacity Through Schedule Management/316

ftoc.indd viiiftoc.indd viii 25/09/12 4:16 PM25/09/12 4:16 PM

ixC o n t e n t s

Short-term Capacity Planning/328 Chapter 8 Supplement: Forecasting/353 Forecasting Purposes and Methods/354 Time Series Analysis/357 Causal Forecasting with Regression/367

Cases/379 BPO, Incorporated: Call Center Six Sigma Project/381 Peerless Laser Processors/394 United Lock: Door Hardware Division (A)/399 Heublein: Project Management and Control System/413 D. U. Singer Hospital Products Corp./424 Automotive Builders, Inc.: The Stanhope Project/427

Index/435

ftoc.indd ixftoc.indd ix 25/09/12 4:16 PM25/09/12 4:16 PM

ftoc.indd xftoc.indd x 25/09/12 4:16 PM25/09/12 4:16 PM

xi

Preface

The enthusiasm of the users of this MBA-oriented book has been greatly rewarding for us and we thank them for their comments, suggestions, criticism, and support. Although the book is not the massive seller that an undergraduate textbook can become, it is clear that there is, as we felt, a need for a solely MBA-level text. The book was originally written because of the express need we felt in our many MBA programs at Wake Forest University for an operations management textbook directed specifically to MBA students, and especially to those who had some real- world experience. We tried all of the current texts but found them either tomes that left no time for the cases and other materials we wanted to include, or shorter but simplistic quantitative books. Moreover, all the books were so expensive they did not allow us to order all the cases, readings, and other supplements and class activi- ties (such as the “Beer Game”; see Chapter 7 Supplement) that we wanted to include in our course.

What we were looking for was a short, inexpensive book that would cover just the introductory, basic, and primarily conceptual material. This would allow us, as the professors, to tailor the course through supplementary cases and other mate- rials for the unique class we would be teaching: executive, evening, full-time, short course, and so on. Although we wanted a brief, supplementary-type book so that we could add other material, we have colleagues who need a short book because they only have a half-semester module for the topic. Or they may have to include another course (e.g., management science or statistics) in the rest of the quarter or semester. In addition, we didn ’t need the depth of most texts that have two exten- sive chapters on supply chain management, two long chapters on scheduling, two chapters on quality, and so on; one chapter on each topic would be sufficient for our needs.

C H A N G E S I N T H I S F I F T H E D I T I O N The response to our changes in the previous edition was very positive, especially the addition of a second color, the new organization with fewer but more main- stream chapters, and the inclusion of multiple cases in the rear of the book. One of the major additions in this 5 th edition has been to elaborate the fl ow of the chapters in the book and how these topics contribute to the competitiveness of every organi- zation. Hence, each chapter starts with a diagram depicting where we are in the fl ow of the text rather than the details of the topics within each chapter, and is then fol- lowed by a discussion of how the topic is related to competitiveness. Another major change has been the addition of many new examples to open each chapter that illustrate how the topic is crucial to competitiveness.

Within the chapters we have added materials where the reviewers indicated more, or new, discussion was warranted. For example, in Chapter 1 we discuss the current trends in operations such as technology (e.g., RFID) and the green/sustainability

fpref.indd xifpref.indd xi 25/09/12 4:17 PM25/09/12 4:17 PM

xii P r e f a c e

movement. We also moved the defi nition of some generic, but crucial to operations, terminology into this fi rst chapter such as effi ciency and productivity . And in Chapter 2 we added the processes of assemble-to-order and engineer-to-order and added a discussion of servicescapes . In Chapter 6, we added a discussion of where to place a project in the organizational structure, and the major criteria for selecting a project manager. And fi nally, in Chapter 7 we updated material about supply chain management in a number of places. We also added some new mini-cases to the rear of three chapters that only had one mini-case previously.

In revising the book, we have kept the elements of our earlier philosophy. For example, we kept the other majors such as marketing and finance in mind—what did these students need to know about operations to help them in their careers? And we still minimize the heavier quantitative material, keeping only discussions and examples that illustrate a particular concept since fi nance and marketing majors would not be solving operations problems. Moreover, even operations managers probably wouldn ’t themselves be solving those problems; more likely, they would be assigned to an analyst. For those chapters in which exercises are included, they are intended only to help illustrate the concept we are trying to convey rather than make experts of the students.

We continued to add service examples throughout the text, since the great majority (over 80 percent these days!) of our students would be, or are already, employed in a service organization. And since these students will be working and competing in a highly global economy, we employ many international examples. We also kept the textual flow of material in the chapters away from the current under- graduate trend of fracturing the material flow with sidebars, examples, applications, solved problems, and so forth, in an attempt to keep the students ’ interest and atten- tion. Given the maturity of MBA students we instead worked these directly into the discussions to attain a smoother, clearer flow. As noted below, the Instructor ’s Manual includes suggestions for readings, cases, videos, and other course supplements that we have found to be particularly helpful for MBA classes since this book is intended to be only a small part of the MBA class.

S U P P L E M E N T S Our approach to supplementary MBA-level material here is to reference and annotate in the Instructor ’s Manual additional useful cases, books, video clips, and readings for each of the eight textbook chapters. The annotation is intended to help the instruc- tors select the most appropriate materials for their unique course. Although we have added some of our own and our colleagues ’ cases to the rear of this edition, we also rely on our favorite Harvard, Darden, Western Ontario, and European cases, plus Harvard Business Review readings to fully communicate the nature of the chapter topic we are covering. Although we didn ’t think that Test Bank Questions or PowerPoint slides would be used by most MBA instructors, these materials are avail- able from the publisher also. For that matter, the publisher can also custom bind selected content from this text, our larger undergraduate (or any other) Web text, along with cases and articles, should this approach be of interest to the professor. Please contact your local Wiley representative for more details.

fpref.indd xiifpref.indd xii 25/09/12 4:17 PM25/09/12 4:17 PM

xiiiP r e f a c e

Y O U R I N P U T S A P P R E C I A T E D We would once again like to encourage users of this book to send us their com- ments and suggestions. Tell us if there is something we missed that you would like to see in the next edition (or the Instructor ’s Manual or web site) or if there is per- haps material that is unneeded for this audience. Also, please tell us about any errors you uncover, or if there are other elements of the book you like or don ’t like. We hope to continue keeping this a living, dynamic project that evolves to meet the needs of the MBA audience, an audience whose needs are also evolving as our economy and society evolve and change.

We want to thank the many reviewers of this book and its previous editions: Alexander Ansari, Seattle University; Satya Chakravorty, Kennesaw State University; Okechi Geoffrey Egekwu; Michael H. Ensby, Clarkson University; James A. Fitzsimmons, University of Texas; Lawrence D. Fredendall, Clemson University; William C. Giauque, Brigham Young University; Mike Godfrey, University of Wisconsin-Oshkosh; Damodar Golhar, Western Michigan University; Suresh Kumar Goyal, Concordia University, Canada; Hector Guerrero, The College of William & Mary; Robert Handfield, North Carolina State University; Mark Gerard Haug, University of Kansas; Janelle Heineke, Boston University; David Hollingworth, Rensselaer Polytechnic Institute; James L. Hoyt, Troy State University; Kendra Ingram, Texas A&M University-Commerce; Mehdi Kaighobadi, Florida Atlantic University; Casey Kleindienst, California State University-Fullerton; Archie Lockamy III, Samford University; Manoj Malhotra, University of South Carolina; Gus Manoochehri, California State University-Fullerton; Robert F. Marsh, Sacred Heart; Ron McLachlin, University of Manitoba; Ivor P. Morgan, Babson College; Seungwook Park, California State University-Fullerton; Ranga V. Ramasesh, Texas Christian University; Jaime S. Ribera, IESE-Universidad de Navarra, Spain; Gary D. Scudder, Vanderbilt University; Sue Perrott Siferd, Arizona State University; Samia Siha, Kennesaw State Unversity; Donald E. Simmons, Ithaca College; William J. Tallon, Northern Illinois University; Asoo J. Vakharia, University of Florida; and Jerry C. Wei, University of Notre Dame.

For this edition we thank the following reviewers: Dennis Battistella, Florida Atlantic University; Linda Brennan, Mercer University; David Cadden, Quinnipiac University; Zhimin Huang, Hofstra University; Jonatan Jelen, NUY-Poly; Mehdi Kaighobadi, Florida Atlantic University; Rob Owen, Thunderbird School of Global Management; Forrest Thornton, River College; Richard Vail, Colorado Mesa University; Jack Zhang, Hofstra University.

Jack Meredith Scott Shafer

Schools of Business Schools of Business

Wake Forest University, P.O. Box 7959 Wake Forest University, P.O. Box 7959

Winston-Salem, NC 27109 Winston-Salem, NC 27109

[email protected] [email protected]

www.mba.wfu.edu/faculty/meredith www.mba.wfu.edu/faculty/shafer

336.758.4467 336.758.3687

fpref.indd xiiifpref.indd xiii 25/09/12 4:17 PM25/09/12 4:17 PM

fpref.indd xivfpref.indd xiv 25/09/12 4:17 PM25/09/12 4:17 PM

�

1

Operations Strategy and Global

Competitiveness

C H A P T E R 1

ROLE OF OPERATIONS IN ORGANIZATIONS’ COMPETITIVENESS

Operations Strategy

Process Design and Control

Process Improvement

Process Execution

Ch. 1: Operations Strategy and Global

Competitiveness

Ch. 2: Process Planning and

Design

Ch. 3: Controlling Processes

Ch. 7: Supply Chain Management

Ch. 8: Capacity, Scheduling, and

Location Planning

Ch. 4: Process Improvement:

Six Sigma

Ch. 5: Process Improvement: Lean

Ch. 6: Managing Process

Improvement Projects

c01.indd 1c01.indd 1 25/09/12 4:18 PM25/09/12 4:18 PM

2

The diagram above illustrates the crucial role that operations plays in the global competitiveness of all organizations. It achieves this through the exe- cution of an operations strategy (discussed here in Chapter 1) devoted to designing, improving, and then executing the production process by which the organization ’s services and products are cre- ated. We discuss the design and control of the proc- ess in terms of its planning and design (Chapter 2) and the design of its control (e.g., quality) proce- dures (Chapter 3). For a fi rm to stay competitive in the global marketplace, the process must be con- tinually improved by reducing the inherent varia- bility of its processes (Chapter 4) and eliminating any and all forms of waste (Chapter 5), which is typically achieved through improvement projects (Chapter 6). Finally, we must execute the process, primarily by managing the supply chain that pro- vides the inputs and outputs for the organization (Chapter 7) and the internal details of the process dealing mainly with capacity and scheduling.

Chapter 1, Operations Strategy and Global Competitiveness, describes the global competitive environment, what customers value (the benefi ts received at minimal cost), the evolution of strategy and supporting functional strategies, and, fi nally, the major strategic frameworks used in operations. Chapter 2, Process Planning and Design, describes the various ways of organizing the transformation process and each of their advantages and disad- vantages. The choice of an appropriate transfor- mation process is one of the major functions of operations. Chapter 3, Controlling Processes, describes the control element of the production

process, with special attention to quality control as our primary example. The implementation of the organization ’s production process involves the day-to-day running of the organization; carefully monitoring and controlling this production proc- ess to keep it functioning as intended is crucial.

Improving business processes is also critical to staying competitive in the global marketplace; Chapter 4, Process Improvement: Minimizing Variation Through Six Sigma, focuses on the use of the Six Sigma approach to constantly reduce the variations in the processes. Chapter 5, Process Improvement: Reducing Waste Through Lean, offers techniques to further improve business processes by eliminating all forms of waste, thereby saving cost, effort, and time. And to effectively conduct these process improvement projects, Chapter 6, Managing Process Improvement Projects, helps us to execute our improvement plans by using the proven proce- dures of project management. We illustrate the pro- cedures with a process improvement example, but project management can be applied to many other activities that organizations undertake, especially activities involving change.

Finally, we reach the application of all this plan- ning, where the processes are executed. Chapter 7, Supply Chain Management, covers a range of top- ics related to effective supply chain execution, such as management of inventories, enterprise requirements planning (ERP) systems, logistics, and purchasing. Chapter 8, Capacity, Scheduling, and Location Planning, covers the major internal details of the production process by which the organization grows and remains competitive.

2

C H A P T E R I N P E R S P E C T I V E

I N T R O D U C T I O N • No discussion of global competitiveness would be complete without the inclusion of

Apple Inc. ’s amazing comeback from its near-death experience over a decade ago. Under the futuristic vision of the late Steve Jobs, the fi rm has innovated in the

c01.indd 2c01.indd 2 25/09/12 4:18 PM25/09/12 4:18 PM

electronics market like no fi rm has ever done before, with high quality and reason- able pricing to bring magical capabilities to small gadgets and overwhelm its competitors.

Since 2007, Apple ’s share price has risen 500 percent, while the S&P 500 has lost 5 percent. For example, among the 100 largest Nasdaq-listed companies, known as the Nasdaq-100 index, Apple ’s 17 percent weighting is more than that of Google, Intel, and Amazon.com combined! And Apple ’s share price recently shot past $500, resulting in a market capitalization of almost half a trillion dollars and a lock on its position as the largest company in the United States.

This example of Apple ’s uniqueness shows how important operations capabilities in areas such as innovation, quality, customization, and cost can be to an organiza- tion ’s global competitiveness (Cheng and Intindola, 2012).

• As in sports, numerous intense rivalries exist in the world of business, such as the rivalries between Visa and MasterCard, Microsoft and Apple, Ford and General Motors, Energizer and Duracell, and Nike and Reebok. Certainly any list of top busi- ness rivalries would be incomplete without Coke and Pepsi. Interestingly, while these two fi rms compete in the same industry, one has had considerable success on the important dimension of share price performance while the other ’s performance has been rather dismal. More specifi cally, over the fi ve-year period ending in January 2012, Pepsi ’s stock price was down while Coke ’s increased by 51 percent. The result was that Coke ’s market capitalization increased $153.5 billion while Pepsi ’s remained relatively fl at at $101 billion. This difference in market capitalization is even more dramatic when one factors in the fact that Pepsi ’s sales are signifi cantly higher than Coke ’s—$57.8 billion versus $36.7 billion in 2010.

A question that naturally arises is: What accounts for these very different out- comes? One explanation offered by analysts and critics is that Pepsi simply took its eye off the ball. In particular, while Coke focused its attention on beverages, Pepsi has been distracted by attempting to develop nutritious snacks. One result is that Pepsi Cola went from being the number-two soda to the number-three soda behind Coke and Diet Coke. To address its weakened performance, Pepsi ’s board of directors initiated a strategic review of the company. A variety of opinions have been offered regarding what the outcome of Pepsi ’s strategic review will be, from reducing its pay- roll to free up additional resources for marketing its soft drink products to breaking up the company into a beverage company and a snack food company (Esterl 2012).

• General Motors ’ market share had been in a long downward decline from about 45 percent in 1980 to about 20 percent in 2008 when the entire automotive industry got hit with a powerful one-two punch, throwing all the weakened American automo- bile producers into chaos. First, in early 2008, extreme gasoline prices killed the truck and SUV market, and then the sudden credit crisis and recession killed the rest of the automobile market. The high cost of debt, unionized labor, and unfunded lia- bilities (pensions and health care) forced GM and Chrysler to go begging to the gov- ernment for bailouts, with GM getting a $50 billion lifeline from U.S. taxpayers, for example. By late 2008, GM was burning through billions of dollars of cash every month. One industry analyst calculated that GM ’s obligations in March of 2009

I n t r o d u c t i o n 3

c01.indd 3c01.indd 3 25/09/12 4:18 PM25/09/12 4:18 PM

4 C h a p t e r 1 : O p e r a t i o n s S t r a t e g y a n d G l o b a l C o m p e t i t i v e n e s s

amounted to $62 billion, 35 times its market capitalization (Denning 2009, p. C10)! Finally, both GM and Chrysler had to fi le for a prepackaged structured bankruptcy. The bankruptcy helped GM to cut its labor costs, get rid of a lot of its debt, get rid of some of its pension and health care obligations, and cut the number of models it was offering to the public.

So how did the restructuring work out? In 2011, GM had the largest annual profi t, at $7.6 billion, in its 103-year history, up 62 percent from 2010. GM ’s revenues were up 13 percent on sales of 1.37 million cars (Chrysler ’s sales were up 26 percent), and GM had hired 100,000 workers in each of the previous fi ve months! GM ’s car sales are growing quickly in China as well as North America, and the company now has very little debt, over $38 billion in liquidity, and minimal taxes (as a part of their bankruptcy agreement). This represents a tremendous turnaround in the competi- tiveness of the U.S. automobile industry.

But GM ’s European business is in trouble, having lost $747 million in 2011 (but $2 billion in 2010). So they are currently working with European labor unions to cut capacity and costs and may possibly close factories in England and Germany after 2014 (Bennett 2012; Terlep 2012).

• In contrast to GM, Ford, and Chrysler, Toyota has been plagued with problems for the last couple of years. First, Toyota got hit with multiple quality complaints, forcing global recalls in 2010; then the March 11, 2011, Japanese earthquake and tsunami plus the fl ooding in Thailand severely crippled their capacity to produce, resulting in a 6 percent decrease in sales in 2011 to 7.9 million cars worldwide.

Another diffi cult problem for Toyota, as well as all Japanese exporters, has been the sharp increase in the strength of the yen due to the fi nancial crisis and a global rush into strong, reliable currencies such as the dollar, the yen, and the Swiss franc. Amazingly, the yen strengthened from 120 yen to the dollar prior to the crisis to less than 80 yen, a strengthening of over one-third. Imagine trying to sell cars for a third more than you ’ve been selling them for! Obviously, Toyota and other exporters are trying to hold down their price increases, but at tremendous cost to their profi ts (Bennett, 2012; Takabashi, 2011).

• Having rung up combined profi ts of $8 billion in 2004, manufacturers of fl at-panel TVs appeared to be especially optimistic about the profi t potential for the TV market in the years ahead. Indeed, a battle of epic proportions was brewing in the consumer electronics industry. On one side was a group of Asian manufacturers that spent $35 billion adding fl at-panel capacity in 2004 and 2005. The Asian players included a joint venture between LG Electronics and Royal Philips Electronics that invested $5.1 billion to build the world ’s largest liquid-crystal display factory, a $2 billion joint ven- ture between Sony and Samsung to produce LCDs, and Matsushita Electronics ’ new $1.3 billion plant for producing chips for thin TVs. On the other side, North America ’s Dell was attempting to leverage its streamlined supply chain and direct- sales model and thereby shift the basis of competition from features to price. For example, in the fall of 2004 Dell introduced a high-defi nition 42-inch plasma TV for under $3,000 while the similar offerings of its Asian competitors were still priced above $4,000. As a result, Dell was able to capture 10 percent of the market in a

c01.indd 4c01.indd 4 25/09/12 4:18 PM25/09/12 4:18 PM

5I n t r o d u c t i o n

span of only a couple of months. So would the Asian strategy based on product innovation and appealing designs win out over Dell ’s strategy that seeks to commod- itize the market and thereby shift the basis of competition to price? (Einhorn 2005).

Fast forward to 2007 and the fl at-panel sets have now overtaken the CRT sets, with LCD sets taking a commanding lead of 58 percent of the market by the fourth quarter of 2008, CRTs with a 34 percent share, and plasma with an 8 percent share. In addition, the market jumped from an $11 billion industry in 1998 to $102 billion in 2007—just ten years! And the winners are Samsung and Sony, each with about a 13 percent share of the market. Vizio is a close runner-up with 11 percent, and then Sharp with 8 percent. However, the future does not look quite as rosy as it did in 2004. Most of the consumers who wanted a fl at-panel set now have one, multiple low-end producers have entered the market and kept prices low, and the recession resulted in a 3 percent drop in sales as of January 2009 (http://news.cnet.com/8301- 10784_3-9891583-7.html).

These brief examples highlight the diversity and importance of operations while providing a glimpse of two themes that are central to operations: customer satisfac- tion and competitiveness . They also illustrate a more subtle point—that improve- ments made in operations can simultaneously increase customer satisfaction and lower costs. The Apple example demonstrates how a company obtained a substan- tial competitive advantage by improving their innovation capability, their production process, and their supply chain. The American automobile industry example shows how losing an operations focus can drive a fi rm into bankruptcy but how, through restructuring, the fi rm can regain its operational competitiveness. The Pepsi example illustrates a fundamental principle in strategy and competitiveness—namely, that organizations that focus on doing a few things well usually outperform organizations that lack this focus. The Toyota example further illustrates how losing your focus on your strengths, plus bad luck in terms of global fi nancial and natural crises, can dam- age your competitiveness. And the Apple and fl at-panel TV examples demonstrate how quickly technology can upend an industry and change the major players and their competitiveness.

Today, in our international marketplace, consumers purchase their products from the provider that offers them the most “value” for their money. To illustrate, you may be doing your course assignments on a Japanese notebook computer, driving a German automobile, or watching a sitcom on a TV made in Taiwan while cooking your food in a Korean microwave. However, most of your services—banking, insur- ance, personal care—are probably provided domestically, although some of these may also be owned by, or outsourced to, foreign corporations. There is a reason why most services are produced by domestic fi rms while products may be produced in part, or wholly, by foreign fi rms, and it concerns an area of business known as operations.

A great many societal changes that are occurring today intimately involve activi- ties associated with operations. For example, there is great pressure among compet- ing nations to increase their exports. And businesses are intent on building effi cient and effective supply chains, improving their processes through “Six Sigma” and suc- cessfully applying the precepts of “lean management” and other operations-based programs.

c01.indd 5c01.indd 5 25/09/12 4:18 PM25/09/12 4:18 PM

6 C h a p t e r 1 : O p e r a t i o n s S t r a t e g y a n d G l o b a l C o m p e t i t i v e n e s s

Another characteristic of our modern society is the explosion of new technology, an important aspect of operations. Technologies such as smart phones, e-mail, note- book computers, tablets, and the Web, to name a few, are profoundly affecting busi- ness and are fundamentally changing the nature of work. For example, many banks are shifting their focus from building new branch locations to using the Web as a way to establish and develop new customer relationships. Banks rely on technology to carry out more routine activities as well, such as transferring funds instantly across cities, states, and oceans. Our industries also rely increasingly on technology: robots carry and weld parts together, and workerless, dark “factories of the future” turn out a continuing stream of products. And soft operations technologies, such as “supply chain management” and “lean production” (Feld 2000; Womack and Jones 2003) have transformed world markets and the global economy.

This exciting, competitive world of operations is at the heart of every organization and, more than anything else, determines whether the organization survives in the international marketplace or disappears into bankruptcy or a takeover. It is this world that we will be covering in the following chapters.

O P E R A T I O N S Why do we argue that operations be considered the heart of every organization? Fundamentally, organizations exist to create value, and operations is the part of the organization that creates value for the customer. Michael Hammer (2004) maintains that operational innovation can provide organizations with long-term strategic advantages over their competitors. Regardless of whether the organization is for- profi t or not-for-profi t, primarily service or manufacturer, public or private, it exists to create value. Thus, even nonprofi t organizations like the Red Cross strive to create value for the recipients of their services in excess of their costs. Moreover, this has always been true, from the earliest days of bartering to modern-day corporations.

Consider McDonald ’s as an example. This fi rm uses a number of inputs, including ingredients, labor, equipment, and facilities; transforms them in a way that adds value to them (e.g., by frying); and obtains an output, such as a chicken sandwich, that can be sold at a profi t. This conversion process, termed a production system , is illustrated in Figure 1.1 . The elements of the fi gure represent what is known as a system 1 : a purpose- ful collection of people, objects, and procedures for operating within an environment .

Note the word purposeful ; systems are not merely arbitrary groupings but goal- directed or purposeful collections. Managing and running a production system effi - ciently and effectively is at the heart of the operations activities that will be discussed in this text. Since we will be using this term throughout the text, let us formally defi ne it. Operations is concerned with transforming inputs into useful outputs according to an agreed-upon strategy and thereby adding value to some entity; this constitutes the primary activity of virtually every organization.

Not only is operations central to organizations, it is also central to people ’s per- sonal and professional activities, regardless of their position. People, too, must oper- ate productively, adding value to inputs and producing quality outputs, whether

1Note the word system is being used here in a broad sense and should not be confused with more narrow usages such as information systems, planning and control systems, or performance evaluation systems.

c01.indd 6c01.indd 6 25/09/12 4:18 PM25/09/12 4:18 PM

7O p e r a t i o n s

those outputs are information, reports, services, products, or even personal accom- plishments. Thus, operations should be of major interest to every reader, not just professionally but also personally.

Systems Perspective As Figure 1.1 illustrates, a production system is defi ned in terms of the environment, a strategy, a set of inputs, the transformation process, the outputs, and some mecha- nism for controlling the overall system. The strategy includes such elements as what customers value, the vision and mission of the organization, an appropriate frame- work to execute this vision, and the core capabilities of the organization. We discuss the strategy in detail a bit later. The environment includes those things that are out- side the actual production system but that infl uence it in some way. Because of its infl uence, we need to consider the environment, even though it is beyond the con- trol of decision makers within the system.

For example, a large portion of the inputs to a production system are acquired from the environment. Also, government regulations related to pollution control and

ENVIRONMENT

TRANSFORMATION PROCESSES • Facilitating

goods • Services

• Alteration • Transportation • Storage • Inspection

INPUTS

CONTROL

• Capital • Materials • Equipment • Facilities • Suppliers • Labor • Knowledge • Time

• Measure • Compare • Plan improvements • Implement improvements

OUTPUT

ActionActionAction Data DataData

• Customers • Government

• Competitors • Technology

• Suppliers • Economy

STRATEGY

• Customer value • Vision/Mission • Strategic frameworks • Core capabilities

Figure 1.1 The production system.

c01.indd 7c01.indd 7 25/09/12 4:18 PM25/09/12 4:18 PM

8 C h a p t e r 1 : O p e r a t i o n s S t r a t e g y a n d G l o b a l C o m p e t i t i v e n e s s

workplace safety affect the transformation system. Think about how changes in cus- tomers ’ needs, a competitor ’s new product, or a new advance in technology can infl uence the level of satisfaction with a production system ’s current outputs. As these examples show, the environment exerts a great deal of infl uence on the production system.

Because the world around us is constantly changing, it is necessary to monitor the production system and take action when the system is not meeting its strategic goals. Of course, it may be that the current strategy is no longer appropriate, indicat- ing a need to revise the strategy. On the other hand, it may be found that the strat- egy is fi ne but that the inputs or transformation processes, or both, should be modifi ed in some way. In either case, it is important to continuously collect data from the environment, the transformation processes, and the outputs; compare that data to the strategic plan; and, if substantial deviations exist, design and implement improvements to the system, or perhaps the strategy, so that results agree with the strategic goals.

Thinking in terms of systems provides decision makers with numerous advan- tages. To begin with, the systems perspective focuses on how the individual compo- nents that make up a system interact. Thus, the systems perspective provides decision makers with a broad and complete picture of an entire situation. Furthermore, the systems perspective emphasizes the relationships between the various system com- ponents. Without considering these relationships, decision makers are prone to a problem called suboptimization . Suboptimization occurs when one part of the sys- tem is improved to the detriment of other parts of the system and, perhaps, the organization as a whole. For example, if a retailer decides to broaden its product line in an effort to increase sales, this could actually end up hurting the retailer as a whole if it does not have suffi cient shelf space or service personnel available to accommodate the broader product line. Thus, decisions need to be evaluated in terms of their effect on the entire system, not simply in terms of how they will affect one component of the system.

In the remainder of this section, we elaborate on inputs, the transformation proc- esses, and outputs. In later sections and chapters, we further discuss both strategy and elements of the control system in more detail.

Inputs The set of inputs used in a production system is more complex than might be sup- posed and typically involves many other areas such as marketing, fi nance, engineer- ing, and human resource management. Obvious inputs include facilities, labor, capital, equipment, raw materials, and supplies. Supplies are distinguished from raw materials by the fact that they are not usually a part of the fi nal output. Oil, paper clips, pens, tape, and other such items are commonly classifi ed as supplies because they only aid in producing the output.

Another very important but perhaps less obvious input is knowledge of how to transform the inputs into outputs. The employees of the organization hold this knowledge. Finally, having suffi cient time to accomplish the operations is always critical. Indeed, the operations function quite frequently fails in its task because it cannot complete the transformation activities within the required time limit.

c01.indd 8c01.indd 8 25/09/12 4:18 PM25/09/12 4:18 PM

9O p e r a t i o n s

Transformation Processes The transformation processes are the part of the system that adds value to the inputs. Value can be added to an entity in a number of ways. Four major ways are described here:

1. Alter: Something can be changed structurally. That would be a physical change, and this approach is basic to manufacturing industries, where goods are cut, stamped, formed, assembled, and so on. We then go out and buy the shirt, or computer, or whatever the good is. But it need not be a separate object or entity; for example, what is altered may be us . We might get our hair cut, or we might have our appendix removed.

Other, more subtle, alterations may also have value. Sensual alterations, such as heat when we are cold, or music, or beauty may be highly valued on certain occasions. Beyond this, even psychological alterations can have value, such as the feeling of worth from obtaining a college degree or the feeling of friendship from a long-distance phone call.

2. Transport: An entity, again including ourselves, may have more value if it is located somewhere other than where it currently is. We may appreciate hav- ing things brought to us, such as fl owers, or removed from us, such as garbage.

3. Store: The value of an entity may be enhanced for us if it is kept in a pro- tected environment for some period of time. Some examples are stock certifi - cates kept in a safe-deposit box, our pet boarded at a kennel while we go on vacation, or ourselves staying in a hotel.

4. Inspect: Last, an entity may be more valued because we better understand its properties. This may apply to something we own, plan to use, or are con- sidering purchasing, or, again, even to ourselves. Medical exams, elevator certifi cations, and jewelry appraisals fall into this category.

Thus, we see that value may be added to an entity in a number of different ways. The entity may be changed directly, in space, in time, or even just in our mind. Additionally, value may be added using a combination of these methods. To illustrate, an appliance store may create value by both storing merchandise and transporting (delivering) it. There are other, less frequent, ways of adding value as well, such as by “guaranteeing” something. These many varieties of transformations, and how they are managed, constitute some of the major issues to be discussed in this text.

Outputs Two types of outputs commonly result from a production process: services and products. Generally, products are physical goods, such as a personal computer, and services are abstract or nonphysical. More specifi cally, we can consider the char- acteristics in Table 1.1 to help us distinguish between the two.

However, this classifi cation may be more confusing than helpful. For example, consider a pizza delivery chain. Does this organization produce a product or

c01.indd 9c01.indd 9 25/09/12 4:18 PM25/09/12 4:18 PM

10 C h a p t e r 1 : O p e r a t i o n s S t r a t e g y a n d G l o b a l C o m p e t i t i v e n e s s

provide a service? If you answered “a service,” suppose that instead of delivering its pizzas to the actual consumer, it made the pizzas in a factory and sold them in the frozen-food section of grocery stores. Clearly the actual process of making pizzas for immediate consumption or to be frozen involves basically the same tasks, although one may be done on a larger scale and use more automated equipment. The point is, however, that both organizations produce a pizza, and defi ning one organization as a service and the other as a manufacturer seems to be a little arbitrary. In addition, both products and services can be produced as commodities or individually customized.

We avoid this ambiguity by adopting the point of view that any physical entity accompanying a transformation that adds value is a facilitating good (e.g., the pizza). In many cases, of course, there may be no facilitating good; we refer to these cases as pure services .

The advantage of this interpretation is that every transformation that adds value is simply a service, either with or without facilitating goods! If you buy a piece of lumber, you have not purchased a product. Rather, you have purchased a bundle of services, many of them embodied in a facilitating good: a tree-cutting service, a sawmill service, a transportation service, a storage service, and perhaps even an advertising service that told you where lumber was on sale. We refer to these services as a bundle of “benefi ts,” of which some are tangible (the sawed length of lumber, the type of tree) and others are intangible (courteous salesclerks, a convenient location, payment by charge card). Some services may, of course, even be negative, such as an audit of your tax return. In sum- mary, services are bundles of benefi ts, some of which may be tangible and others intangible, and they may be accompanied by a facilitating good or goods.

Firms often run into major diffi culties when they ignore this aspect of their opera- tions. They may think of, and even market themselves as, a “lumberyard” and not as providing a bundle of services. They may recognize that they have to include certain tangible services (such as cutting lumber to the length desired by the customer) but ignore the intangible services (charge sales, having a suffi cient number of clerks). Another reason for not making a distinction between manufacturing and services is that when a company thinks of itself as a manufacturer, it tends to focus on measures of internal performance such as effi ciency and utilization. But when companies con- sider themselves as providing services they tend to focus externally and ask questions such as “How can we serve our customers better?” This is not to imply that improving internal performance measures is not desirable. Rather, it suggests that improved cus- tomer service should be the primary impetus for all improvement efforts. It is generally

T A B L E 1 . 1 • Charac te r i s t i c s o f Produc t s and Ser v i ce s Products Services

Tangible Intangible

Minimal contact with customer Extensive contact with customer

Minimal participation by customer in the delivery

Extensive participation by customer in the delivery

Delayed consumption Immediate consumption

Equipment-intense production Labor-intense production

Quality easily measured Quality diffi cult to measure

c01.indd 10c01.indd 10 25/09/12 4:18 PM25/09/12 4:18 PM

11O p e r a t i o n s

not advisable to seek internal improvements if these improvements do not ultimately lead to corresponding improvements in customer service and customer satisfaction.

In this text we will adopt the point of view that all value-adding transformations (i.e., operations) are services, and there may or may not be a set of accompanying facilitating goods. Figure 1.2 illustrates how the tangible product (or facilitating good) portion and the intangible service portion for a variety of outputs contribute to the total value provided by each output. The outputs shown range from virtually pure services to what would be known as products. For example, the Plush restau- rant appears to be about 75 percent service and 25 percent product. Although we work with “products” as extensively as with services throughout the chapters in this book, bear in mind that in these cases we are working with only a portion of the total service, the facilitating good. In general, we will use the nonspecifi c term out- puts to mean either products or services.

100 50 0

% Services

Magazine purchase

Flour purchase

% Products

50 100

Plush restaurant

Theatrical performance

Travels

Auto repair

Hand-made suit

Video rental

Medical examination

Figure 1.2 The range from services to products.

c01.indd 11c01.indd 11 25/09/12 4:18 PM25/09/12 4:18 PM

12 C h a p t e r 1 : O p e r a t i o n s S t r a t e g y a n d G l o b a l C o m p e t i t i v e n e s s

One particular type of output that is substantially different from products and many other types of services is that of knowledge or information. These outputs often have the characteristic that the more they are used, the more valuable they become. For example, in a network, the more entities that belong to the network, the more useful it may be. If you are on Facebook® or use e-mail, the more other people that are also there, the more valuable it is to you. And when you share this output, you don ’t lose anything, you gain. Some other characteristics of information or knowledge that differ from normal goods and services are listed below.

• Giving or selling the information/knowledge to someone doesn ’t mean you can ’t give or sell it to someone else.

• The information/knowledge doesn ’t wear out. • The information/knowledge isn ’t subject to the law of diminishing returns. • The information/knowledge can be replicated at minimal cost and trouble. • The more the knowledge is used, the more valuable it becomes.

Control Suppose that in our production system we make a mistake. We must be able to observe this through, for example, accounting records (measurement data), compare it to standard to see how serious the error is, and then, if needed, plan and imple- ment (usually via a project) some improvements. If the changes are not signifi cantly affecting the outputs, then no control actions are needed. But if they are, manage- ment must intercede and apply corrective control to alter the inputs or the transfor- mation processes and, thereby, the outputs. The control activities illustrated in Figure 1.1 are used extensively in systems, including management systems, and will be encountered throughout this text.

Table 1.2 lists a few examples of some components of the production system for a variety of common organizations.

Operations Activities Operations include not only those activities associated specifi cally with the produc- tion system but also a variety of other activities. For example, purchasing or procure- ment activities are concerned with obtaining many of the inputs needed in the production system. Similarly, shipping and distribution are sometimes considered marketing activities and sometimes considered operations activities. Because of the important interdependencies of these activities, many organizations are attempting to manage these activities as one process commonly referred to as supply chain management .

As organizations begin to adopt new organizational structures based on business processes and abandon the traditional functional organization, it is becoming less important to classify activities as operations or nonoperations (e.g., sales, marketing, accounting). However, to understand the tasks more easily, we divide the fi eld of operations into a series of subject areas as shown in Table 1.3 . These areas are quite

c01.indd 12c01.indd 12 25/09/12 4:18 PM25/09/12 4:18 PM

13O p e r a t i o n s

interdependent, but to make their workings more understandable, we discuss them as though they were easily separable from each other. In some areas, a full-fl edged department may be responsible for the activities, such as quality control or schedul- ing, but in other areas the activities (such as facility location) may be infrequent and simply assigned to a particular group or project team. Moreover, some of the subar- eas such as supply chain management or maintenance are critically important because they are a part of a larger business process or because other areas depend on them. Finally, since we consider all operations to be services, these subject areas are equally applicable to organizations that have traditionally been classifi ed as man- ufacturers and services.

Trends in Operations As has been previously discussed in this chapter and will be further emphasized in the remaining chapters, an organization ’s operations play a critical role in its overall competitiveness and long-term success. Given the critical role played by operations, it is important to stay abreast of the signifi cant trends in the operations area as well as general trends that may impact the operations function.

T A B L E 1 . 2 • Example s o f Produc t ion Sy s t em Component s

Organization Strategy Inputs Transformation

Process Outputs Control Environment

Post offi ce Regular Consistent Dependable

Labor Equipment Trucks

Transportation Printing

Mail deliveries Stamps

Weather Mail volumes Sorting/loss errors

Transportation network Weather Civil service

Bank Secure Trustworthy Responsive Informative

Checks Deposits Vault ATMs

Safekeeping Investment Statement preparations

Interest Electronic transfer Loans Statements

Interest rates Wage rates Loan default rates

Federal Reserve Economy

Cinema Enjoyable Variety Timely

Films Food People Theater

Film projection Food preparation

Entertainment Snacks

Film popularity Disposable incomes

Economy Entertainment industry

Manufacturer Reliable Affordable Quality Variety

Materials Equipment Labor Technology

Cutting Forming Joining Mixing

Machines Chemicals Consumer goods Scrap

Material fl ows Production volumes

Economy Commodity prices Consumer market

School Knowledge Safe Trustworthy Friendly

Books Teachers Facility Students

Learning Counseling Motivating

Educated students Skills research

Demographics Grievances

State and county boards Tax system

c01.indd 13c01.indd 13 25/09/12 4:18 PM25/09/12 4:18 PM

14 C h a p t e r 1 : O p e r a t i o n s S t r a t e g y a n d G l o b a l C o m p e t i t i v e n e s s

As in other disciplines, technology is having a signifi cant impact on the practice of operations. For example, communication technologies such as the Internet are greatly facilitating the ability of organizations to share real-time information with their suppliers and customers. Having more timely information enhances the oppor- tunities for supply chain partners to coordinate and integrate their operations, which ultimately leads to a more effective and effi cient supply chain that benefi ts both the end customer and the trading partners in the supply chain.

One exciting technology that promises to greatly enhance the ability of organiza- tions to have real-time information on their inventory and other assets is radio

T A B L E 1 .3 • Ma jor Sub jec t A reas in Opera t ions Chapter Subject Area

1 Strategy : Determining the critical operations tasks to support the organiza- tion ’s overall mission.

1 Output planning : Selecting and designing the services and products the organization will offer to customers, patrons, or recipients.

1 Reliability and maintenance : Determining how the proper performance of both the output and the transformation process itself is to be maintained.

2 Transformation process design : Determining the physical transformation aspects of the production activities.

2 Facility layout : Devising an appropriate material fl ow and equipment layout within the facility to effi ciently and effectively accommodate the transforma- tion activities.

3 Quality control : Determining how quality standards are to be developed and maintained.

4,5 Process improvement : Applying process design techniques to improve the fl ow and effi ciency of production systems.

5 Lean management : Using techniques from the Toyota Production System and JIT to eliminate waste and non-value-added activities.

6 Project management : Learning how to plan and control project activities to meet specifi cations for performance, schedule, and cost.

7 Supply chain management : Organizing the activities from the customer ’s order through fi nal delivery for speed, effi ciency, and quality.

7 Inventory management : Deciding what amounts of raw materials, work-in- process, and fi nished goods to hold.

7 Enterprise and material requirements planning : Using information manage- ment systems to coordinate enterprise-wide activities, especially for ordering or producing materials to meet a master delivery schedule.

8 Capacity planning : Determining when to have facilities, equipment, and labor available and in what amounts.

8 Facility location : Deciding where to locate production, storage, and other major facilities.

8 Schedule planning : Anticipating the yearly needs for labor, materials, and facilities by month, week, or day within the year.

c01.indd 14c01.indd 14 25/09/12 4:18 PM25/09/12 4:18 PM

15C u s t o m e r V a l u e

frequency identifi cation, or RFID; RFID tags are attached to individual inventory items, and these tags transmit identifi cation and location information. For example, by attaching an RFID tag to a part, its progress through the production process can be monitored and, when fi nished, its location in the warehouse tracked.

RFID tags are classifi ed as passive or active. Passive RFID tags contain no power source and therefore rely on the power source of an RFID reader to transmit their information. Active RFID tags contain a power source such as a battery and use this power source to periodically transmit a signal that provides identifi cation informa- tion. Perhaps the greatest challenge to greater adoption of RFID tags is the cost of the tags themselves. As with other technologies, the cost of RFID has decreased dra- matically and is expected to continue on this trajectory. In 2011 the cost of passive RFID tags ranged from $0.05 to $5.00, depending on the volume of tags purchased and the environmental factors they were designed to withstand. The cost of active RFID tags typically ranges between $50 and $100. Thus, at present, the costs of active RFID tags are justifi ed only for tracking expensive assets such as a rail car or delivery truck.

Beyond technology, another important trend impacting the practice of operations management is increasing levels of concern for the environment. Addressing envi- ronmental concerns impacts virtually all aspects of operations management from the design of the organization ’s output to the sourcing of parts, the distribution of the product, and even the disposal or recycling of the product or its components once it reaches the end of its useful life. Green sourcing, for example, seeks to iden- tify suppliers in such a way that the organization ’s carbon footprint and overall impact on the environment is minimized.

Reducing the waste associated with products is another top priority of organiza- tions that seek to minimize the negative impact they have on the environment. In this case, organizations can deploy a strategy often referred to as the three R ’s: reduce, reuse, and recycle. As its name suggests, the reduce strategy seeks to decrease the amount of waste associated with a product. One way to accomplish this is to minimize the amount of product packaging used. In services, switching to electronic copies of documents helps reduce waste, such as when a bank switches to electronic statements. Reuse is a second strategy for minimizing waste. The idea underlying reuse is to identify alternative uses for an item after its initial use. For example, there are kits available for converting old computer monitors into fi sh aquariums. Finally, recycling involves using the materials from old prod- ucts to create new products. For example, many greeting cards are made from recycled paper.

C U S T O M E R V A L U E Costs

In the Introduction to this chapter, we mentioned that customers support the pro- vider of goods and services who offers them the most “value.” In this section, we elaborate on this concept. The equation for value is conceptually clear:

Value 5 perceived benefi ts/costs

c01.indd 15c01.indd 15 25/09/12 4:18 PM25/09/12 4:18 PM

16 C h a p t e r 1 : O p e r a t i o n s S t r a t e g y a n d G l o b a l C o m p e t i t i v e n e s s

The perceived benefi ts can take a wide variety of forms, but the costs are usually more straightforward:

• The upfront monetary investment • Other monetary life-cycle costs of using the service or product, such as

maintenance

• The hassles involved in obtaining the product or service, such as travel required, obtaining fi nancing, the friendliness of service, and so on

The cost to the customer is, of course, the price paid, but this is usually highly cor- related with the cost of producing the service or product, which is itself largely based on the “effi ciency” of the production process. Effi ciency is always measured as output/ input; for example, a standard automobile engine that uses gasoline is usually about 15 to 20 percent effi cient (that is, the energy put into the engine in terms of gasoline vs. the energy put out in terms of automobile motion). However, electric and jet engines are more effi cient, and rocket engines can reach almost 70 percent effi ciency.

The primary method of attaining effi ciency in production is through high produc- tivity , which is normally defi ned as output per worker-hour. This defi nition of pro- ductivity is actually what is known as a partial factor measure of productivity, in the sense that it considers only worker-hours as the productive factor. Although in the past, labor often constituted as much as 50 percent of the cost of a product—or even more for a service—it is now frequently as little as 5 percent, so labor produc- tivity is no longer a good measure of effi ciency. Clearly, labor productivity could easily be increased by substituting machinery for labor, but that doesn ’t mean that this is a wise, or even cost-saving, decision. A multifactor productivity measure uses more than a single factor, such as both labor and capital. Obviously, the different factors must be measured in the same units, such as dollars. An even broader gauge of productivity, called total factor productivity, is measured by including all the factors of production—labor, capital, materials, and energy—in the denominator. This measure is to be preferred in making any comparisons of productivity for effi - ciency or cost purposes.

Last, we also frequently hear of “effectiveness,” which is a measure of the achieve- ment of goals; where effi ciency is sometimes considered to be “doing the thing right,” effectiveness is instead considered to be “doing the right thing” or being focused on the proper task or goal.

Benefi ts In contrast to the role of costs in the customer ’s value equation, the benefi ts can be multiple. We will consider fi ve of these in detail: innovativeness, functionality, qual- ity, customization, and responsiveness.

Innovativeness Many people (called “early adopters” in marketing) will buy products and services simply because they are so innovative, or major improvements over what has been available formerly. It is the fi eld of research and development (known as R&D) that is primarily responsible for developing innovative new product and service ideas.

c01.indd 16c01.indd 16 25/09/12 4:18 PM25/09/12 4:18 PM

17C u s t o m e r V a l u e

R&D activities focus on creating and developing (but not producing) the organiza- tion ’s outputs. On occasion, R&D also creates new production methods by which outputs, either new or old, may be produced.

Research itself is typically divided into two types: pure and applied. Pure research is simply working with basic technology to develop new knowledge. Applied research is attempting to develop new knowledge along particular lines. For exam- ple, pure research might focus on developing a material that conducts electricity with zero resistance, whereas applied research could focus on further developing this material to be used in products for customers. Development is the attempt to utilize the fi ndings of research and expand the possible applications, often consist- ing of modifi cations or extensions to existing outputs to meet customers ’ interests. Figure 1.3 illustrates the range of applicability of development as the output becomes more clearly defi ned. In the early years of a new output, development is oriented toward removing “bugs,” increasing performance, improving quality, and so on. In the middle years, options and variants of the output are developed. In the later years, development is oriented toward extensions of the output that will prolong its life.

Unfortunately, the returns from R&D are frequently meager, whereas the costs are great. Figure 1.4 illustrates the mortality curve (fallout rate) associated with the con- current design, evaluation, and selection for a hypothetical group of 50 potential chemical products, assuming that the 50 candidate products are available for consid- eration in year 3. (The fi rst three years, on the average, are required for the necessary research preceding each candidate product.) Initial evaluation and screening reduce the 50 to about 22, and economic analysis further reduces the number to about 9. Development reduces this number even more, to about 5, and design and testing reduce it to perhaps 3. By the time construction (for production), market development, and a year ’s commercialization are completed, there is only one successful product left. (Sometimes there are none!) One study found that, beyond this, only 64 percent of the new products brought to market were successful, or about two out of three.

Time

Discovery

E ff

o rt

Development

Maturity

Variants

Saturations

Extensions

Decline

Death

Options

Idea incubation

Idea refinement

Idea examination

and evaluation

Improving performance

Output selection

Full marketing

Acceptance testing,

modification

Pure Research Applied

Growth

Figure 1.3 The development effort.

c01.indd 17c01.indd 17 25/09/12 4:18 PM25/09/12 4:18 PM

18 C h a p t e r 1 : O p e r a t i o n s S t r a t e g y a n d G l o b a l C o m p e t i t i v e n e s s

Two alternatives to research frequently used by organizations are imitation of a proven new idea (i.e., employing a second-to-market strategy) or outright pur- chase of someone else ’s invention. The outright purchase strategy is becoming extremely popular in those industries where bringing a new product to market can cost huge sums, such as pharmaceuticals and high technology. It is also employed in those industries where technology advances so rapidly that there isn ’t enough time to employ a second-to-market strategy. Although imitation does not put the organization fi rst in the market with the new product or service, it does provide an opportunity to study any possible defects in the original product or service and rapidly develop a better design, frequently at a better price. The second approach—purchasing an invention or the inventing company itself—eliminates the risks inherent in research, but it still requires the company to develop and market the product or service before knowing whether it will be successful. Either route spares the organization the risk and tremendous cost of conducting the actual research leading up to a new invention or improvement.

In addition to product research (as it is generally known), there is also process research , which involves the generation of new knowledge concerning how to pro- duce outputs. Currently, the production of many familiar products out of plastic (toys, pipe, furniture, etc.) is an outstanding example of successful process research. Motorola, to take another example, extensively uses project teams that conduct process development at the same time as product development.

Functionality Many people confuse functionality with quality (discussed next). But functionality involves the activities the product or service is intended to perform, thereby

21 3

10

20

30

40

4 5 Year

N u

m b

er o

f ca

n d

id at

es

6 7 8 9

Research

Customer Value

Screening

Construction

Economic analysis

Product development

Process development

Design and testing

Market development

Commercialization

Figure 1.4 Mortality curve of chemical product ideas from research to commercialization. Source : Adapted from This Is Dupont 30 , Wilmington, DE, by permission of DuPont de Nemours and Co.

c01.indd 18c01.indd 18 25/09/12 4:18 PM25/09/12 4:18 PM

19C u s t o m e r V a l u e

providing the benefi ts to the customer. A contemporary example is the ubiquitous cell phone. These days it is probably rare to fi nd a cell phone that is only a phone; many phones include a camera and a way to send its picture to another person, or provide access to the Internet, as well as a myriad of other functions.

However, many products, especially electronics, but also some services, may be advertised to provide purchasers with a new, unique function and they may do so, but it may not work well or for long . The former involves performance and the latter has to do with reliability. Clearly, these are different attributes of the output, and one can be well addressed while others disappoint. Our discussion of quality, next, elab- orates a bit more on the distinction between these attributes.

Quality Quality is a relative term, meaning different things to different people at different times. Moreover, quality is not an absolute but, rather, is based on customers ’ per- ceptions. Customers ’ impressions can be infl uenced by a number of factors, includ- ing brand loyalty and an organization ’s reputation.

Quality Dimensions

Richard J. Schonberger has compiled a list of multiple quality dimensions that cus- tomers often associate with products and services:

1. Conformance to specifi cations. Conformance to specifi cations is the extent to which the actual product matches the design specifi cations, such as a pizza delivery shop that consistently meets its advertised delivery time of 30 minutes.

2. Performance. Customers frequently equate the quality of products and services with their performance. (Note, however, that this dimension may in some cases actually refer to functionality.) Examples of performance include how quickly a sports car accelerates or the battery life of a cell phone.

3. Features. Features are the options that a product or service offers, such as side impact airbags or leather seats in automobiles. (Again, however, this dimension may also be confused with functionality.)

4. Quick response. Quick response is associated with the amount of time required to react to customers ’ demands. However, we consider this to be a separate benefi t, discussed further below.

5. Reliability. Reliability is the probability that a product or service will per- form as intended on any given trial or for some period of time, such as the probability that a car will start on any given morning.

6. Durability. Durability refers to how tough a product is, such as a notebook computer that still functions after being dropped or a knife that can cut through steel and not need sharpening.

7. Serviceability. Serviceability refers to the ease with which maintenance or a repair can be performed.

c01.indd 19c01.indd 19 25/09/12 4:18 PM25/09/12 4:18 PM

20 C h a p t e r 1 : O p e r a t i o n s S t r a t e g y a n d G l o b a l C o m p e t i t i v e n e s s

8. Aesthetics. Aesthetics are factors that appeal to human senses, such as the taste of a steak or the sound of a sports car ’s engine.

9. Humanity. Humanity has to do with how the customer is treated, such as a private university that maintains small classes so students are not treated like numbers by its professors.

It is worth noting that not all the dimensions of quality are relevant to all products and services. Thus, organizations need to identify the dimensions of quality that are relevant to the products and services they offer. Market research about customers ’ needs is the primary input for determining which dimensions are important. Of course, measuring the quality of a service can often be more diffi cult than meas- uring the quality of a product or facilitating good. However, the dimensions of qual- ity described above apply to both.

Quality ’s Benefi ts and Costs

Many benefi ts are associated with providing products and services that have high quality. Obviously, customers are more pleased with a high-quality product or service. They are more apt to encourage their friends to patronize the fi rm, as well as giving the fi rm their own repeat business. Top quality also establishes a reputation for the fi rm that is very diffi cult to obtain in any other manner, and it allows the fi rm to charge a premium price. And, typically, high-quality products and services are not only the most profi table but also garner the largest market shares. High quality also tends to protect the fi rm from competitors, who may have to offer competing outputs at an especially low price (and low margins). It also enhances the attractiveness of follow-up products or services so that their chances of success are much improved. And, of course, high quality minimizes risks to safety and health and reduces liability for the fi rm.

Traditionally, it was thought that making products and services of excellent qual- ity would translate into higher costs. Of course, this view neglects the negative con- sequences of gaining a reputation for producing shoddy outputs. Also, the Japanese have demonstrated that it is often possible to improve quality and lower costs at the same time. One explanation for this phenomenon is that it is simply cheaper to do a job right the fi rst time than to try to fi x it or rework it later. And if quality is built into the production system, it improves workers ’ morale, reduces scrap and waste, smoothes work fl ows, improves control, and reduces a wide variety of other costs. As a result, Philip Crosby, a well-known quality consultant, maintains that “quality is free,” as in the title of his book, Quality Is Free (1979), which sold approximately 1 million copies. Crosby estimates that fi rms can lose up to 25 percent of the amount of their sales because of poor quality.

Two primary sets of costs are involved in quality: control costs and failure costs. The aggregate of these costs runs between 15 and 35 percent of sales for many U.S. fi rms. Traditionally, these costs are broken down into four categories: prevention costs (including planning, training, product design, maintenance); appraisal costs (measuring, testing, test equipment, inspectors, reports); internal costs of defects (extra labor and materials to repair, scrap, rework interruptions, expediting); and external costs of defects (ill-will, complaints, quick response to correct, warranties, insurance, recalls, lawsuits). The fi rst two costs are incurred in attempting to control quality, and the last two are the costs of failing to control quality. Costs of defects (or nonconformance) can run from 50 to 90 percent of the total cost of quality.

c01.indd 20c01.indd 20 25/09/12 4:18 PM25/09/12 4:18 PM

21C u s t o m e r V a l u e

Deming promised the Japanese that if they followed his advice, they would be able to compete with the West within just a few years. They did! Now the most pres- tigious industrial quality award given in Japan each year is named the Deming Prize. But the Japanese did not stop there. They tied the concept of quality control directly into their production system—and now they have even tied it into their entire economy through inspections to guarantee the quality of exports. The natural inclinations of Japanese culture and traditions were exploited in this quality crusade. After nearly two decades of a national emphasis on quality, Japan ’s reputation for producing shoddy goods was totally reversed. And when high quality is combined with competitive pricing—another strength of the Japanese system—the result is extremely strong competition for existing producers.

Evolution of Quality: Japan vs. America

Although you might think that “made in Japan” signifi es a product of superior qual- ity, it may surprise you to learn that many of the techniques and philosophies Japanese companies employ today were actually developed in the United States, many around the end of World War II. Unfortunately, the sentiment among U.S. manufacturers at the end of World War II was that they already produced the highest-quality products in the world at the lowest cost. Thus, they were not particu- larly interested in or concerned with improving quality.

Japan was an entirely different story. Its products had a reputation for poor quality, and after it lost the war, its economy was a shambles. As a result, Japanese manufac- turers were eager for help related to quality improvement. In 1950 the Japanese gov- ernment invited W. Edwards Deming (then a professor at New York University) to give a series of lectures on quality control to help Japanese engineers reindustrialize the country. But Deming insisted that the heads of the companies attend the talks, too. As a result, the top Japanese managers were also invited, and they all showed up.

According to Deming (1986), the major cause of poor quality is variation . Thus, a key tenet of Deming ’s approach is to reduce variability in the process. (This topic is discussed further in Chapter 4.) Deming stressed that improving quality was the responsibility of top management. However, he also believed that all employees should be trained in the use of problem-solving tools, especially statistical techniques. Deming believed that improvements in quality create a chain reaction in which improved quality leads to lower costs, which then translate into higher productivity. In contrast to Deming, Crosby focused more on management, organizational proc- esses, and changing corporate culture than on the use of statistical techniques.

DILBERT: ©Scott Adams/Dist. by United Feature Syndicate, Inc.

c01.indd 21c01.indd 21 25/09/12 4:18 PM25/09/12 4:18 PM

22 C h a p t e r 1 : O p e r a t i o n s S t r a t e g y a n d G l o b a l C o m p e t i t i v e n e s s

Flexibility

However, to offer customization demands fl exibility on the part of the fi rm. Professor David Upton (1994), formerly of the Harvard Business School, defi nes fl exibility as “the ability to change or react with little penalty in time, effort, cost, or performance” (p. 73). There are more than a dozen different types of fl exibility that we will not pursue here—design, volume, routing through the production system, product mix, and many others. But having the right types of fl exibility can offer the following major competitive advantages:

• Faster matches to customers ’ needs because changeover time from one prod- uct or service to another is quicker

• Closer matches to customers ’ needs

A more recent concept (similar to zero defects) that the Japanese and some American fi rms have embraced is called total quality management (TQM) or total quality control (TQC). The basic idea of TQM is that it is extremely expensive to “inspect” quality into a company ’s outputs and much more effi cient and effective to produce them right in the fi rst place. As a result, responsibility for quality has been taken away from the quality control department and placed where it belongs—with the workers who produce the parts or provide the service in the fi rst place. This is called quality at the source . It is the heart of statistical quality control (SQC), some- times called statistical process control (SPC), which we discuss further in Chapter 3.

Customization Customization refers to offering a product or service exactly suited to a customer ’s desires or needs. However, there is a range of accommodation to the customer ’s needs, as illustrated in Figure 1.5 . At the left, there is the completely standard, world-class (excellence suitable for all markets) product or service. Moving to the right is the standard with options, continuing on to variants and alternative models, and ending at the right with made-to-order customization. In general, the more customization the better—if it can be provided quickly, with acceptable quality and cost.

Standard world-class

Increasing customization

Increasing standardization

Standard with options

Variants Alternate models

Customization

Figure 1.5 Continuum of customization.

c01.indd 22c01.indd 22 25/09/12 4:18 PM25/09/12 4:18 PM

23C u s t o m e r V a l u e

• Ability to supply the needed items in the volumes required for the markets as they develop

• Faster design-to-market time to meet new customer needs • Lower cost of changing production to meet needs • Ability to offer a full line of products or services without the attendant cost of

stocking large inventories

• Ability to meet market demands even if delays develop in the production or distribution process

Mass Customization

Until recently, it was widely believed that producing low-cost standard products (at the far left in Figure 1.5 ) required one type of transformation process and producing higher-cost customized products (far right) required another type of process. However, in addition to vast improvements in operating effi ciency, an unexpected by-product of the continuous improvement programs of the 1980s was substantial improvement in fl exibility. Indeed, prior to this, effi ciency and fl exibility were thought to be tradeoffs. Increasing effi ciency meant that fl exibility had to be sacri- fi ced, and vice versa.

Thus, with the emphasis on continuous improvement came the realization that increasing operating effi ciency could also enhance fl exibility. For example, many man- ufacturers initiated efforts to reduce the amount of time required to set up (or change over) equipment when switching from the production of one product to another. Obviously, all time spent setting up equipment is wasteful, since the equipment is not being used during this time to produce outputs that ultimately create revenues for the organization. Consequently, improving the amount of time a resource is used produc- tively directly translates into improved effi ciency. Interestingly, these same reductions in equipment setup times also resulted in improved fl exibility. Specifi cally, with shorter equipment setup times, manufacturers could produce economically in smaller-size batches, making it easier to switch from the production of one product to another.

In response to the discovery that effi ciency and fl exibility can be improved simul- taneously and may not have to be traded off, the strategy of mass customization emerged (see Pine 1993; Gilmore and Pine 1997). Organizations pursuing mass cus- tomization seek to produce low-cost, high-quality outputs in great variety. Of course, not all products and services lend themselves to being customized. This is particu- larly true of commodities, such as sugar, gas, electricity, and fl our. On the other hand, mass customization is often quite applicable to products characterized by short life cycles, rapidly advancing technology, or changing customer requirements. However, recent research suggests that successfully employing mass customization requires an organization to fi rst develop a transformation process that can consist- ently deliver high-quality outputs at a low cost. With this foundation in place, the organization can then seek ways to increase the variety of its offerings while at the same time ensuring that quality and cost are not compromised.

In an article published in the Harvard Business Review , Gilmore and Pine (1997) identifi ed four mass customization strategies:

1. Collaborative customizers. These organizations establish a dialogue to help customers articulate their needs and then develop customized outputs

c01.indd 23c01.indd 23 25/09/12 4:18 PM25/09/12 4:18 PM

24 C h a p t e r 1 : O p e r a t i o n s S t r a t e g y a n d G l o b a l C o m p e t i t i v e n e s s

to meet these needs. For example, one Japanese eyewear retailer developed a computerized system to help customers select eyewear. The system com- bines a digital image of the customer ’s face, and then various styles of eye- ware are displayed on the digital image. Once the customer is satisfi ed, the customized glasses are produced at the retail store within an hour.

2. Adaptive customizers. These organizations offer a standard product that customers can modify themselves, such as fast-food hamburgers (ketchup, etc.) and closet organizers. Each closet-organizer package is the same but includes instructions and tools to cut the shelving and clothes rods so that the unit can fi t a wide variety of closet sizes.

3. Cosmetic customizers. These organizations produce a standard product but present it differently to different customers. For example, Planters packages its peanuts and mixed nuts in a variety of containers on the basis of specifi c needs of its retailing customers, such as Wal-Mart, 7-Eleven, and Safeway.

4. Transparent customizers. These organizations provide custom products without the customers knowing that a product has been customized for them. For example, Amazon.com provides book recommendations based on information about past purchases.

Example: Hewlett-Packard Faced with increasing pressure from its customers for quicker order fulfi llment and for more highly customized products, Hewlett-Packard (HP) wondered whether it was really possible to deliver mass-customized products rapidly, while at the same time con- tinuing to reduce costs (Feitzinger and Lee 1997). HP ’s approach to mass customization can be summarized as effectively delaying tasks that customize a product as long as possible in the product supply process. It is based on the following three principles:

• Products should be designed around a number of independent modules that can be easily combined in a variety of ways.

• Manufacturing tasks should also be designed and performed as independent modules that can be relocated or rearranged to support new production requirements.

• The product supply process must perform two functions. First, it must cost- effectively supply the basic product to the locations that complete the cus- tomization activities. Second, it must have the requisite fl exibility to process individual customers ’ orders.

HP has discovered that modular design provides three primary benefi ts. First, com- ponents that differentiate the product can be added during the later stages of produc- tion. This method of mass customization, generally called postponement , is one form of the assemble-to-order production process, discussed in more detail in Chapter 3. For example, the company designed its printers so that country-specifi c power sup- plies are combined with the printers at local distribution centers and actually plugged in by the customer when the printer is set up. Second, production time can be signifi - cantly reduced by simultaneously producing the required modules. Third, producing in modules facilitates the identifi cation of production and quality problems.

c01.indd 24c01.indd 24 25/09/12 4:18 PM25/09/12 4:18 PM

25C u s t o m e r V a l u e

Responsiveness The competitive advantages of faster, dependable response to new markets or to the individual customer ’s needs have occasionally been noted in the business media (Eisenhardt and Brown 1998; Stalk 1988; Vessey 1991). For example, in a study of the U.S. and Japanese robotics industry, the National Science Foundation found that the Japanese tend to be about 25 percent faster than Americans, and to spend 10 percent less, in developing and marketing new robots. The major difference is that the Americans spend more time and money on marketing, whereas the Japanese spend fi ve times more than the Americans on developing more effi cient production methods.

Table 1.4 identifi es a number of prerequisites for and advantages of fast, depend- able response. These include higher quality, faster revenue generation, and lower costs through elimination of overhead, reduction of inventories, greater effi ciency, and fewer errors and scrap. One of the most important but least recognized advan- tages for managers is that by responding faster, they can allow a customer to delay an order until the exact need is known. Thus, the customer does not have to change the order—a perennial headache for most operations managers.

Faster response to a customer also can, up to a point, reduce the unit costs of the product or service, sometimes signifi cantly. On the basis of empirical studies reported by Meredith et al. (1994) and illustrated in Figure 1.6 , it seems that there is about a 2:1 (i.e., 0.50) relationship between response time and unit cost. That is, starting

T A B L E 1 .4 • P re requ i s i t e s fo r and Advantages o f Rap id Response 1 Sharper focus on the customer . Faster response for both standard and custom-

designed items places the customer at the center of attention.

2 Better management . Attention shifts to management ’s real job, improving the fi rm ’s infrastructure and systems.

3 Effi cient processing . Effi cient processing reduces inventories, eliminates non-value- added processing steps, smoothes fl ows, and eliminates bottlenecks.

4 Higher quality . Since there is no time for rework, the production system must be suf- fi ciently improved to make parts accurately, reliably, consistently, and correctly.

5 Elimination of overhead . More effi cient, faster fl ows through fewer steps eliminate the overhead needed to support the remaining steps, processes, and systems.

6 Improved focus . A customer-based focus is provided for strategy, investment, and gen- eral attention (instead of an internal focus on surrogate measures such as utilization).

7 Reduced changes . With less time to delivery, there is less time for changes in product mix, engineering changes, and especially changes to the order by the customer who just wanted to get in the queue in the fi rst place.

8 Faster revenue generation . With faster deliveries, orders can be billed faster, thereby improving cash fl ows and reducing the need for working capital.

9 Better communication . More direct communication lines result in fewer mistakes, oversights, and lost orders.

10 Improved morale . The reduced processing steps and overhead allow workers to see the results of their efforts, giving a feeling of working for a smaller fi rm, with its greater visibility and responsibility.

c01.indd 25c01.indd 25 25/09/12 4:18 PM25/09/12 4:18 PM

26 C h a p t e r 1 : O p e r a t i o n s S t r a t e g y a n d G l o b a l C o m p e t i t i v e n e s s

from typical values, an 80 percent reduction in response time results in a corre- sponding 40 percent reduction in unit cost. The actual empirical data indicated a range between about 0.60 and 0.20, so for an 80 percent reduction in response time there could be a cost reduction from a high of 0.60 3 80 percent 5 48 percent to a low of 16 percent.

This is an overwhelming benefi t because if corresponding price reductions are made, it improves the value delivered to the customer through both higher respon- siveness and lower price. The result for the producer is a much higher market share.

If the producer chooses not to reduce the price, then the result is both higher margins and higher sales, for signifi cantly increased profi tability.

S T R A T E G Y A N D C O M P E T I T I V E N E S S Competitiveness can be defi ned in a number of ways. We may think of it as the long- term viability of a fi rm or organization, or we may defi ne it in a short-term context such as the current success of a fi rm in the marketplace as measured by its market share or its profi tability. We can also talk about the competitiveness of a nation, in the sense of its aggregate competitive success in all markets. The U.S. President ’s Council on Industrial Competitiveness gave this defi nition in 1985:

Competitiveness for a nation is the degree to which it can, under free and fair market conditions, produce goods and services that meet the test of international markets while simultaneously maintaining and expanding the real incomes of its citizens.

Global Trends The United States provides a graphic example of global trade trends. The trend in merchandise trade for the United States is startling. Although some might think that

20 40 60

Percentage change in response time

P er

ce n

ta ge

c h

an ge

i n

c o

st

80 100

20

40

60

80

100

Approximation Lower range Upper range

Figure 1.6 Cost reductions with decreases in response time.

c01.indd 26c01.indd 26 25/09/12 4:18 PM25/09/12 4:18 PM

27S t r a t e g y a n d C o m p e t i t i v e n e s s

foreign competition has been taking markets away from U.S. producers only in the past decade, U.S. merchandise imports have grown considerably for over 30 years. Although exports have increased over this period as well, they have not increased as fast as imports; the result is an exploding trade defi cit with foreign countries. Partly as a result of this defi cit, the United States is now the biggest debtor nation in the world, with a cumulative defi cit of about $5 trillion, nearly half of the U.S. annual gross domestic product (GDP), and an annual defi cit running about 6 percent of GDP. However, these values hold only for the period up to mid-2008, when the glo- bal fi nancial/credit/recession crisis started. It now appears that all these fi gures will become much worse—not for just the United States, but globally.

Another important issue relating to the fi nancial crisis involves the exchange rate between currencies. Let ’s consider in more detail what it means when a country ’s currency declines in value relative to foreign currencies. A weaker currency means that citizens in that country will have to pay more for products imported from for- eign countries. Meanwhile, the prices for products produced in that country and exported to foreign countries will decline, making them more desirable. Thus, a decline in the value of a country ’s currency is a double-edged sword. Such a decline makes imported goods more expensive for citizens to purchase but at the same time makes exports less expensive for foreign consumers, increasing the demand for domestic products.

As an example, let ’s consider the American dollar. In the fi nancial crisis of 2008, the dollar grew stronger as Americans sold foreign assets and foreigners rushed to hold assets in the dollar, the world ’s strongest currency, as well as a “reserve” (com- modities are priced in dollars) currency. However, given the massive amount of dol- lars the U.S. government borrowed and created to overcome the fi nancial crisis, there is widespread concern that the dollar may weaken or even collapse in the future.

According to economic theory, a stronger dollar should make American products less desirable (or competitive) in foreign markets and imports more desirable in American markets. However, some market actions that governments and businesses often take to keep from losing customers can alter this perfect economic relation- ship. For instance, in the 1990s, when the price of Japanese products in the United States started increasing in terms of dollars, Japanese fi rms initiated huge cost- cutting drives to reduce the cost (and thereby the dollar price) of their products, to keep from losing American customers, which was largely successful. Similarly, China controls the exchange rate of its currency, the renminbi, to stay at about 7 to the dol- lar (though they have been letting it strengthen recently), so it always sells its goods at a competitive price.

In the last decade, particularly with the economic rise of China and India, global markets, manufacturers, and service producers have evolved in a dramatic fashion. With the changes occurring in the World Trade Organization (WTO), international competition has grown very complex in the last two decades. Previously, fi rms were domestic, exporters, or international. A domestic fi rm produced and sold in the same country. An exporter sold goods, often someone else ’s, abroad. An international fi rm sold domestically produced as well as foreign-produced goods both domestically and in foreign countries. However, domestic sales were usually produced domesti- cally, and foreign sales were made either in the home country or in a plant in the foreign country, typically altered to suit national regulations, needs, and tastes.

Now, however, there are global fi rms, joint ventures, partial ownerships, foreign subsidiaries, and other types of international producers. For example, Canon is a

c01.indd 27c01.indd 27 25/09/12 4:18 PM25/09/12 4:18 PM

28 C h a p t e r 1 : O p e r a t i o n s S t r a t e g y a n d G l o b a l C o m p e t i t i v e n e s s

global producer that sells a standard “world-class” camera with options and add-ons available through local dealers. And automobile producers frequently own stock in foreign automobile companies. Mazak, a fast-growing machine tool company, is the U.S. subsidiary of Yamazaki Machinery Company of Japan. Part of the reason for cross-ownerships and cross-endeavors is the spiraling cost of bringing out new products. New drugs and memory chips run in the hundreds of millions to billions of dollars to bring to market. By using joint ventures and other such approaches to share costs (and thereby lower risks), fi rms can remain competitive.

Whether to build offshore, assemble offshore, use foreign parts, employ a joint venture, and so on is a complex decision for any fi rm and depends on a multitude of factors. For example, the Japanese have many of their automobile manufacturing plants in foreign countries. The reasons are many and include: to circumvent foreign governmental regulation of importers, to avoid the high yen cost of Japanese- produced products, to avoid import fees and quotas, and to placate foreign consum- ers. Of course, other considerations are involved in producing in foreign countries: culture (e.g., whether women are part of the labor force), political stability, laws, taxes, regulations, and image.

Other complex arrangements of suppliers can result in hidden international com- petition. For example, many products that bear an American nameplate have been totally produced and assembled in a foreign country and are simply imported under a U.S. manufacturer ’s or retailer ’s nameplate, such as Nike shoes. Even more confus- ing, many products contain a signifi cant proportion of foreign parts or may be com- posed entirely of foreign parts and only assembled in the United States (e.g., toasters, mixers, hand tools). This recent strategic approach of fi nding the best mix of pro- ducers and assemblers to deliver a product or service to a customer has come to be known as “supply chain management,” a topic we discuss in detail in Chapter 7.

Strategy The organization ’s business strategy is a set of objectives, plans, and policies for the organization to compete successfully in its markets. In effect, the business strategy specifi es what an organization ’s competitive advantage will be and how this advan- tage will be achieved and sustained. As we will see, a key aspect of the business strategy is defi ning the organization ’s core competencies and focus. The actual stra- tegic plan that details the business strategy is typically formulated at the executive committee level (CEO, president, vice presidents). It is usually long range, in the neighborhood of three to fi ve years.

In fact, however, the decisions that are made over time are the long-range strategy. In too many fi rms, these decisions show no pattern at all, refl ecting the truth that they have no active business strategy, even if they have gone through a process of strategic planning. In other cases these decisions bear little or no relationship to the organization ’s stated or offi cial business strategy. The point is that an organization ’s actions often tell more about its true business strategy than its public statements.

The general process of formulating a business strategy is illustrated in Figure 1.7 . Relevant inputs to the strategic planning process include the organization ’s vision/ mission statement, a variety of factors external to the organization, and a range of factors internal to the organization. One school of thought—the resource-based view —considers the set of resources (an internal factor in Figure 1.7 ) available to the

c01.indd 28c01.indd 28 25/09/12 4:18 PM25/09/12 4:18 PM

29

organization as the primary driver of the business strategy. For further discussion of this topic and its impact on the development of corporate strategy, consult Barney and Clark (2007) or Collis and Montgomery (2005).

After these inputs are collectively considered, strategic planning is often initiated by developing a vision statement, a mission statement, or both. Vision statements are used to express the organization ’s values and aspirations. Mission statements express the organization ’s purpose or reason for existence. In some cases, organizations may choose to combine the vision and mission statements into a single statement. Regardless of whether separate statements or combined statements are developed, the intent is to communicate the organization ’s values, aspirations, and purpose so that employees can make decisions that are consistent with and support these objectives.

Effective vision and mission statements tend to be written using language that inspires employees to high levels of performance. Further, to foster employees ’ com- mitment, it is advisable to include a wide variety of employees in the development of the vision or mission statement, rather than enforcing top management ’s view by edict. Once the vision and mission statements are developed for the organization as a whole, divisions, departments, process teams, project teams, work groups, and so on can develop individual vision/mission statements that support the organization ’s overall statement. For example, after a university develops its overall vision/mission statement, each college could develop its own unique statements specifying the role that it will play in supporting the overall mission of the university. Likewise, once each school develops its own vision/mission statement, the departments within the school can develop unique statements. Having each organizational unit develop its own unique statements promotes wider participation in the process, helps employees think in terms of how their work supports the overall mission, and results in statements that are more meaningful to a select group of employees. An example of an actual vision and mission statement is given in Figure 1.8 .

Business Strategy

External Forces

Environmental Competitors Technology Customers

Internal Forces

Resources Core Competencies/Capabilities

Culture Weaknesses

Vision/Mission Statement

Business Unit Strategies

Business Model

Figure 1.7 Strategy formulation.

S t r a t e g y a n d C o m p e t i t i v e n e s s

c01.indd 29c01.indd 29 25/09/12 4:18 PM25/09/12 4:18 PM

30 C h a p t e r 1 : O p e r a t i o n s S t r a t e g y a n d G l o b a l C o m p e t i t i v e n e s s

THE CENTRAL INTELLIGENCE AGENCY Vision

One Agency. One Community. An Agency unmatched in its core capabilities, functioning as one team, fully integrated into the Intelligence Community.

Mission

We are the nation’s first line of defense. We accomplish what others cannot accomplish and go where others cannot go. We carry out our mission by:

• Collecting information that reveals the plans, intentions, and capabilities of our adversaries and provides the basis for decision and action.

• Producing timely analysis that provides insight, warning, and opportunity to the President and decision makers charged with protecting and advancing America’s interests.

• Conducting covert action at the direction of the President to preempt threats or achieve U.S. policy objectives.

Figure 1.8 An example of a vision and mission statement.

In addition to the vision/mission statement, other important inputs in the formula- tion of the business strategy are categorized as forces external and forces internal to the organization in Figure 1.7 . Although both sets of forces are considered to some extent in formulating the vision/mission statement (as shown by the dotted lines in the fi gure), they are considered at a more detailed level and more directly when developing the business strategy. Important external forces include the environment (e.g., the economy, government regulations, climate), competitors (e.g., new prod- uct introductions, industry consolidation, new entrants from outside the industry), the technology available, and customer requirements. Relevant internal forces include organizational resources, the organization ’s core competencies/capabilities, its culture, and its weaknesses. As shown, there is a bidirectional relationship between the organization ’s business strategy and both the internal and external forces. For example, an action by a key competitor may impact the organization ’s strategy just as its business strategy may force a reaction by a key competitor.

Overall, as seen in Figure 1.7 , strategy is primarily concerned with making sets of choices that result in a “business model” that provides the tools to help further develop and communicate the strategy. In particular, the business model represents the organ- ization ’s underlying core logic and strategic choices for creating and capturing value within a network. As such, the business model helps an organization verify that the elements of the strategy are consistent with one another, that they are logical, and that they are mutually reinforcing. Business models typically include expanded verbal discussions of key elements of the strategy as well as quantitative projections for important operational, marketing, and fi nancial aspects of the business.

To help further understand the distinction between strategy and a business model, consider the construction of a custom home. Initially, the architect consults with the future homeowners to understand how they envision the home and their life within it. The architect then creates a design to fulfi ll this vision. This corresponds to strategy.

c01.indd 30c01.indd 30 25/09/12 4:18 PM25/09/12 4:18 PM

31

Next, the architect prepares a detailed fl oor plan and elevation based on the choices made during the design process. These correspond to the business model. Just as a business model can be used to help analyze and communicate strategic choices, the fl oor plan can be used to help understand, analyze, and communicate the design choices that were made.

Once the business strategy has been developed and the resulting business model analyzed, the fi nal step in strategy formulation is the development of business unit strategies. At this stage, each business unit develops its own strategy to guide its activities so that they are consistent and support the organization ’s overall business strategy. Although formulating the business strategy is displayed as rather straightfor- ward in Figure 1.7 , in reality it is very iterative.

Strategic Frameworks We now move to a discussion of the business unit strategies box in Figure 1.7 . Clearly, there will be a marketing strategy, a fi nancial strategy, an R&D strategy, and so on. Here, of course, we are interested in the operations strategy. As it happens, there are a number of fairly well defi ned such strategies. One that is common to many of the functional areas is related to the life cycle of the organization ’s products or services.

The Life Cycle

A number of functional strategies are tied to the stages in the standard life cycle of products and services, shown in Figure 1.9 . Studies of the introduction of new prod- ucts indicate that the life cycle (or stretched-S growth curve , as it is also known) provides a good pattern for the growth of demand for a new output. The curve can be divided into three major segments: (1) introduction and early adoption, (2) acceptance and growth of the market, and (3) maturity with market saturation. After market saturation, demand may remain high or decline, or the output may be improved and possibly start on a new growth curve.

Introduction

D em

an d

Growth Maturity

Time

Figure 1.9 The life-cycle curve.

S t r a t e g y a n d C o m p e t i t i v e n e s s

c01.indd 31c01.indd 31 25/09/12 4:18 PM25/09/12 4:18 PM

32 C h a p t e r 1 : O p e r a t i o n s S t r a t e g y a n d G l o b a l C o m p e t i t i v e n e s s

The length of product and service life cycles has been shrinking signifi cantly in the last decade or so. In the past, a life cycle might have been fi ve years, but it is now six months. This places a tremendous burden on the fi rm to constantly monitor its strategy and quickly change a strategy that becomes inappropriate to the market.

The life cycle begins with an innovation —a new output or process for the mar- ket, as discussed earlier. The innovation may be a patented product or process, a new combination of existing elements that has created a unique product or proc- ess, or some service that was previously unavailable. Initial versions of the product or service may change relatively frequently; production volumes are small, since the output has not caught on yet; and margins are high. As volume increases, the design of the output stabilizes and more competitors enter the market, frequently with more capital-intensive equipment. In the mature phase, the now high-volume output is a virtual commodity, and the fi rm that can produce an acceptable version at the lowest cost usually controls the market.

Clearly, a fi rm ’s business strategy should match the life-cycle stages of its products and services. If a fi rm such as Hewlett-Packard is good at innovation, it may choose to focus only on the introduction and acceptance phases of the product ’s life cycle and then sell or license production to others as the product moves beyond the intro- duction stage. If its strength is in high-volume, low-cost production, the company should stick with proven products that are in the maturity stage. Most common, per- haps, are fi rms that attempt to stick with products throughout their life cycle, chang- ing their strategy with each stage.

One approach to categorizing an organization ’s business strategy is based on its timing of introductions of new outputs. Two researchers, Maidique and Patch (1979), suggest the following four product development strategies:

1. First-to-market. Organizations that use this strategy attempt to have their products available before the competition. To achieve this, strong applied research is needed. If a company is fi rst to market, it has to decide if it wants to price its products high and thus skim the market to achieve large short- term profi ts or set a lower initial price to obtain a higher market share and perhaps larger long-term profi ts.

2. Second-to-market. Organizations that use this strategy try to quickly imitate successful outputs offered by fi rst-to-market organizations. This strategy requires less emphasis on applied research and more emphasis on fast devel- opment. Often, fi rms that use the second-to-market strategy attempt to learn from the mistakes of the fi rst-to-market fi rm and offer improved or enhanced versions of the original products.

3. Cost minimization or late-to-market. Organizations that use this strategy wait until a product becomes fairly standardized and is demanded in large volumes. They then attempt to compete on the basis of costs as opposed to features of the product. These organizations focus most of their research and development on improving the production process, as opposed to focusing on product development.

4. Market segmentation. This strategy focuses on serving niche markets with specifi c needs. Applied engineering skills and fl exible manufacturing systems are often needed for the market-segmentation strategy.

c01.indd 32c01.indd 32 25/09/12 4:18 PM25/09/12 4:18 PM

33

As illustrated by the points A, B, and C, improvement on one dimension can usu- ally only be attained by sacrifi cing performance on another dimension. For example, as shown in Figure 1.10 , increasing output variety may result in higher unit costs. In effect, this curve represents the level of performance that organizations in an indus- try can achieve across two dimensions given the technology available at a given

Be aware that a number of implicit tradeoffs are involved in developing a strategy. Let us use the fi rst-to-market strategy to demonstrate. A fi rst-to-market strategy requires large investments in product development in an effort to stay ahead of the competition. Typically, organizations that pursue this strategy expect to achieve rela- tively higher profi t margins, larger market shares, or both as a result of initially having the market to themselves. The strategy is somewhat risky because a competitor may end up beating them to the market. Also, even if a company succeeds in getting to the market fi rst, it may end up simply creating an opportunity for the competition to learn from its mistakes and overtake it in the market. To illustrate, although Sony introduced its Betamax format for VCRs in 1975, JVC ’s VHS format—introduced the following year—is the standard that ultimately gained widespread market acceptance.

Such tradeoffs are basic to the concept of selecting a business strategy. Although specifi c tasks must be done well to execute the selected strategy, not everything needs to be particularly outstanding—only a few things. And, of course, strategies based on anything else—acquisitions, mergers, tax loss carry-forwards, even streams of high-technology products—will not be successful if the customer is ignored in the process.

Performance Frontiers

As we know from the earlier Customer Value section, there are a wide range of ben- efi ts and costs that organizations can compete on and various groups of customers value. If, say, n of these factors are important for an organization to consider, we might then conceive of a graph or space with n dimensions on it showing the organ- ization ’s measures on each of the n factors as well as their competitors ’ measures. The curve connecting all these measures would then be called the organization ’s performance frontier (Clark 1996). For simplicity, let us use just two factors, say cost and variety, as shown in Figure 1.10 , with the performance frontier curve labeled 1.

Output variety

U n

it c

o st

C B

A

1

Figure 1.10 Example performance frontier.

S t r a t e g y a n d C o m p e t i t i v e n e s s

c01.indd 33c01.indd 33 25/09/12 4:18 PM25/09/12 4:18 PM

34 C h a p t e r 1 : O p e r a t i o n s S t r a t e g y a n d G l o b a l C o m p e t i t i v e n e s s

Suppose you were employed at company A and company B chose to operate at point B

1 . In effect, company B can now offer a wider variety of outputs and at lower

unit costs. What are your options? As it turns out, there are two generic options or improvement trajectories company A could try to follow. One improvement trajec- tory would be for company A to streamline its operations and make cost–variety tradeoffs, moving down curve 1 toward company C. Upon streamlining its operati- ons, company A could then attempt to adopt the new technology and choose a posi- tion on the new frontier. A second improvement trajectory would be for company A to attempt to directly adopt the new technology and move to the new frontier with- out streamlining its current operations.

There are advantages and disadvantages associated with both trajectories. An advantage of streamlining operations fi rst is that this might provide a better under- standing of current processes. In turn, this better understanding might increase com- pany A ’s options in choosing a location on the new frontier and might even better position it to adopt the new technology. One drawback of streamlining its current operations fi rst is that the knowledge gained might be irrelevant when the new tech- nology is eventually adopted, and delaying the adoption of the new technology might mean reduced market share and profi ts. Another important factor is the amount of time required to execute the improvement trajectory and get to the new

point in time. According to the fi gure, company A is apparently pursuing more of a customization strategy than the two other competitors shown, offering a wider vari- ety of outputs but incurring greater cost. We might think of J.C. Penney as perhaps fi tting point A. Company C, perhaps Costco, seems to be pursuing a standardization strategy, offering a smaller range of outputs but incurring lower unit costs.

An interesting use of this framework is to investigate and evaluate the impact of a change in technology or operational innovation (Hammer 2004). For example, in Figure 1.11 , assume a new innovation such as “cross-docking” has been developed by company B, perhaps represented by Wal-Mart, shifting its performance frontier to curve 2. In this case, company B could hold its unit price constant and offer higher output variety than company A and at lower unit cost (position B

1 ). Alternatively,

company B could maintain its current level of output variety and lower its unit cost to levels below company C ’s (position B

2 ) or perhaps choose a position somewhere

between points B 1 and B

2 .

Output variety

U n

it c

o st

C B

B2

B1

A 2

1

Figure 1.11 Development of new technology results in shift in the performance frontier.

c01.indd 34c01.indd 34 25/09/12 4:18 PM25/09/12 4:18 PM

35

position on the new performance frontier. However, although it might appear that streamlining the current operation fi rst before adopting the new technology should take more time than immediately adopting the technology, when ease of implemen- tation is considered, the former approach might in fact be more expedient.

On a more practical note, Kmart some years ago tried to challenge Wal-Mart on low prices but was unsuccessful. Then Sears and Kmart merged instead, but that didn ’t seem to work well either; now both seem to be in trouble.

One fi nal point. In Figure 1.11 it was assumed that the result of the new technology/ innovation was simply a shift in the performance frontier. It is also important to be aware of the possibility that a new technology can change the shape as well as the location of the performance frontier. Such a change in shape can have important implications regarding choosing a location on the new frontier as well as the nature of the tradeoff facing the industry. In either case, the way to beat your competition is through developing or using new technology to move to a new frontier.

Focus

In the past, fi rms primarily competed on one factor, such as low cost or innovation, because that was what they were good at. Obviously, they could not ignore the other factors of competition, which they had to do acceptably on, but their heavy attention to their one strength was based on a strategic framework called focus (Skinner 1974).

McKinsey & Company, a top management consulting fi rm, studied 27 outstanding fi rms to fi nd their common attributes. Two of the major attributes reported in Business Week are directly related to focus:

1. Stressing one key business value. At Apple, the key value is developing innovative new products that are easy to use; at Dana Corporation, it is improving productivity.

2. Sticking to what they know best. All the outstanding fi rms defi ne their core capabilities (or strengths) and then build on them. They resist the temp- tation to move into new areas or diversify.

When an organization chooses to stress one or two key areas of strength, it is referred to as a focused organization . For example, IBM is known for its customer service, General Electric for its technology, and Procter & Gamble for its consumer marketing. In general, most but not all areas of focus relate to operations. Some fi rms, such as those in the insurance industry, focus on fi nancial strength and others focus on marketing strengths. For example, Harley-Davidson considers its strength to be in building relationships with its dealers and motorcycle owners. And many health care organizations are achieving signifi cant operational effi ciencies by focusing on a nar- row range of ailments. For example, by treating only long-term acute cases, Intensiva HealthCare has been able to reduce its costs to 50 percent of those of a traditional intensive-care ward. Clearly, adopting a focus strategy means knowing not only what customers to concentrate on but also knowing what customers you do not want.

Table 1.5 identifi es several areas of focus that organizations commonly choose when forming their competitive strategy; all are various forms of differentiation. Recent competitive behavior among fi rms seems to be dividing most of the factors in

S t r a t e g y a n d C o m p e t i t i v e n e s s

c01.indd 35c01.indd 35 25/09/12 4:18 PM25/09/12 4:18 PM

36 C h a p t e r 1 : O p e r a t i o n s S t r a t e g y a n d G l o b a l C o m p e t i t i v e n e s s

Table 1.5 into two sets that Terry Hill (2000), an operations strategist and researcher in England, calls order qualifi ers and order winners . An order qualifi er is a charac- teristic of the product or service that is required if the product is even to be consid- ered or in the running. In other words, it is a prerequisite for entering the market. An order winner is a characteristic that will win the bid or the purchase. These qualifi ers and winners vary with the market, of course, but some general commonalties exist across markets. For example, response time, performance, customization, innova- tion, and price seem to be frequent order winners, and the other factors (e.g., qual- ity, reliability, and fl exibility) tend to be order qualifi ers. Working with marketing and sales to properly identify which factors are which is clearly of major strategic importance.

In addition to the advantages of being focused, there are also some dangers. A narrowly focused fi rm can easily become uncompetitive in the market if the cus- tomers ’ requirements change. In addition to being focused, a fi rm must also be fl ex- ible enough to alter its focus when the need changes and to spot the change in time. Frequently, a focus in one area can be used to an advantage in another way if there is enough time to adapt—for example, to move into a new product line or alter the application of the focus. Moreover, as products go through their life cycle, the task of operations often changes, as shown in Figure 1.12 , from being fl exible enough to accept changes in design, to meeting the growing demand in the marketplace, to cutting costs. Throughout this life cycle, the focus of the organization has to change if it stays with the same output. Many fi rms, however, choose to compete at only one stage of the life cycle and abandon other stages so that they can keep the strength of their original focus.

An organization can also easily lose its focus. For example, in the traditional func- tional organization, purchasing may buy the cheapest materials it can. This requires buying large quantities with advance notice. Scheduling, however, is trying to reduce inventories, so it orders materials on short notice and in small quantities. Quality

T A B L E 1 .5 • Common Areas o f Organ iza t iona l Focus Innovation . Bringing a range of new products and services to market quickly

Customization . Being able to quickly redesign and produce a product or service to meet customers ’ unique needs

Flexibility of products and services . Switching between different models or variants quickly to satisfy a customer or market

Flexibility of volume . Changing quickly and economically from low-volume production to high volumes and vice versa

Performance . Offering products and services with unique, valuable features

Quality . Having better craftsmanship or consistency

Reliability of the product or service . Always working acceptably, enabling customers to count on the performance

Reliability of delivery . Always fulfi lling promises with a product or service that is never late

Response . Offering very short lead times to obtain products and services

After-sale service . Making available extensive, continuing help

Price . Having the lowest price

c01.indd 36c01.indd 36 25/09/12 4:18 PM25/09/12 4:18 PM

37

Introduction

Sa le

s

Growth

Time

Maturity

Design changes Quality Performance

Volume Capacity

Emphasis required

Price

Figure 1.12 Product life cycle: stages and emphasis.

control is trying to improve the output, so it carefully inspects every item, creating delays and extensive rework. In this example, each functional department is pursu- ing its own objectives but is not focusing on how it can support the organization ’s overall business strategy.

However, the most common reason a fi rm loses its focus is simply that the focus was never clearly identifi ed in the fi rst place. Never having been well defi ned, it could not be communicated to the employees, could therefore not gain their support, and thus was lost. Sometimes a focus is identifi ed but not communicated throughout the organization because management thinks that lower-level employees don ’t need to know the strategic focus of the fi rm in order to do their jobs.

The Sand Cone

For many organizations that relied on the focus framework of strategy, the traditional view was that competing on one competitive dimension required trading off per- formance on one or more other dimensions (e.g., higher quality results in higher costs). However, research suggests that, at least in some cases, building strengths along alternative competitive dimensions may in fact be cumulative and that build- ing a strength on one dimension may facilitate building strengths on other dimen- sions (Ferdows and De Meyer 1990).

Furthermore, according to this research, there is a preferred order in developing strengths on various competitive dimensions. According to the sand cone model (as it is called), shown in Figure 1.13 , organizations should fi rst develop the capability to produce quality outputs. Once an organization has developed this profi ciency, it is next appropriate to address the issue of delivery dependability. Next, according to the model, the competitive dimensions of speed and cost should be addressed, respectively.

In addition to providing guidance to organizations regarding the order in which to focus their attention and initiatives, the model has intuitive appeal. For example, it makes little sense to focus on improving delivery dependability before an organization

S t r a t e g y a n d C o m p e t i t i v e n e s s

c01.indd 37c01.indd 37 25/09/12 4:18 PM25/09/12 4:18 PM

38 C h a p t e r 1 : O p e r a t i o n s S t r a t e g y a n d G l o b a l C o m p e t i t i v e n e s s

can provide a consistent level of quality. In today ’s competitive marketplace, provid- ing defective outputs in a timely fashion is not a recipe for long-term success.

Likewise, organizations should achieve consistent quality levels and delivery dependability before attempting to reduce lead times. Of course, the model is not set in stone (remember it is called the sand cone) and organizations facing different circumstances may choose to address the competitive dimensions in a different order. We will return to these critical competitive factors in the fi nal section of this chapter.

Core Capabilities One important result of developing a business strategy is identifying the organiza- tion ’s core competencies and capabilities that provide those product/service dimensions important to customers and hence are the source of customer value. Core competencies (Prahalad and Hamel 1990) are the collective knowledge and skills an organization has that distinguish it from the competition. In effect, these core competencies become the building blocks for organizational practices and business processes, referred to as core capabilities (Stalk et al. 1992). (Hereafter we will refer to both of these simply as “core capabilities.”) The importance of these core capabilities derives from their strong relationship to an organization ’s ability to integrate a variety of technologies and skills in the development of new prod- ucts and services. Clearly, then, one of top management ’s most important activities is the identifi cation and development of the core capabilities the organization will need to successfully execute the business strategy.

In effect, core capabilities provide the basis for developing new products and services and are a primary factor in determining an organization ’s long-term com- petitiveness. Hammer (2004) points out the importance of “operational innovation” in the organization as one basis for sustained competitive advantage, the clear result of a core capability. Therefore, two important parts of strategic planning are identifying and predicting the core capabilities that will be critical to sustaining and enhancing the organization ’s competitive position. On this basis, an organization can also assess its suppliers ’ and competitors ’ capabilities. If the organization fi nds that it is not the leader, it must determine the cost and risks of catching up with the best versus the cost and risks of losing that core capability.

Cost

Speed

Dependability

Quality

Figure 1.13 The sand cone model. (Adapted from Ferdows and De Meyer 1990, p. 175.)

c01.indd 38c01.indd 38 25/09/12 4:18 PM25/09/12 4:18 PM

39

Hayes and Pisano (1994) stress the importance of a fi rm not looking for “the” solu- tion to a current competitive problem but rather the “paths” to building one or two core capabilities to provide the source of customer value for the indefi nite future. Moreover, the fi rm should not think in terms of “tradeoffs” between core capabilities (e.g., moving from fl exibility as a strength to low cost), but rather of “building” one capability on top of others and determining which set will provide the most customer value.

Often, it is more useful to think of an organization in terms of its portfolio of core capabilities, rather than its portfolio of businesses or products. For instance, Sony is known for its expertise in miniaturization; 3M for its knowledge of substrates, coat- ings, and adhesives; Black and Decker for small electrical motors and industrial design; Boeing for its ability to integrate large-scale complex systems; and Honda for engines and power trains. Had Sony initially viewed itself as primarily a manufac- turer of Walkmans, rather than as a company with expertise in miniaturization, it might have overlooked several profi table opportunities, such as entering the cam- corder business. As another example, Boeing has successfully leveraged its core capability related to integrating large-scale systems in its production of commercial jetliners, space stations, fi ghter-bombers, and missiles.

As these examples illustrate, core capabilities are often used to gain access to a wide variety of markets. Cannon used its core capabilities in optics, imaging, and electronic controls to enter the markets for copiers, laser printers, cameras, and image scanners. In a similar fashion, Honda ’s core capabilities in engines and power trains comprise the basis for its entry into other businesses: automobiles, motorcycles, lawn mowers, and generators.

In addition to providing access to a variety of markets, a core capability should be strongly related to the benefi ts provided by the product or service that customers value. In Sony ’s case, its expertise in miniaturization translates directly into impor- tant product features such as portability and aesthetic designs. Alternatively, suppose Sony developed a core competence in writing understandable user manuals. Since people who purchase an HD TV or a camcorder rarely base their purchase decision on the quality of the user manual (when was the last time you read a user manual?), this core capability would provide little if any competitive advantage.

Another characteristic of core capabilities is that they should be diffi cult to imi- tate. Clearly, no sustainable competitive advantage is provided by a core capability that is easily imitated. For example, Sony ’s expertise in miniaturization would mean little if other electronics manufacturers could match it simply by purchasing and tak- ing apart Sony ’s products (this is called reverse engineering ). Bartmess and Cerny (1996) identify three elements of a core capability that hinder imitation:

• It is complex and requires organizational learning over a long period of time. • It is based on multiple functional areas, both internal and external to the

organization.

• It is a result of how the functions interact rather than the skills/knowledge within the functions themselves.

The topic of core capabilities is also strongly related to the recent surge in out- sourcing and offshoring. Outsourcing —an approach increasingly common—involves subcontracting out certain activities or services. For example, a manufacturer might outsource the production of certain components, the management and maintenance of its computer resources, employee recruitment, or the processing of its payroll.

S t r a t e g y a n d C o m p e t i t i v e n e s s

c01.indd 39c01.indd 39 25/09/12 4:18 PM25/09/12 4:18 PM

40 C h a p t e r 1 : O p e r a t i o n s S t r a t e g y a n d G l o b a l C o m p e t i t i v e n e s s

When we consider the concept of core capability, it is important to recognize that not all parts, services, and activities are equal. Rather, these activities and parts can be thought of as falling on a continuum ranging from strategically important to unimportant. Parts and activities are considered strategically important when:

• They are strongly related to what customers perceive to be the key character- istics of the product or service.

• They require highly specialized knowledge and skill, a core capability. • They require highly specialized physical assets, and few other suppliers pos-

sess these assets.

• The organization has a technological lead or is likely to obtain one.

Activities that are not strategic or core are candidates for outsourcing. These parts or activities are not strongly linked to key product characteristics, do not require highly specialized knowledge, do not need special physical assets, and the organiza- tion does not have the technological lead in this area. Thus, if it is benefi cial to out- source these parts or activities—perhaps because of lower cost or higher quality—no loss in competitiveness should result. On the other hand, when a fi rm ’s strategic parts and activities have been outsourced, particularly to a foreign supplier, called offshoring , the fi rm has become hollow ( Jonas 1986). As we have stated, the wise fi rm will outsource only nonstrategic, simple, relatively standard parts and processes such as screws or types of processes that are not worth the time for the fi rm to pro- duce itself; the complex, proprietary parts and processes that give their products an edge in the marketplace are produced internally. If the fi rm outsources these parts and processes as well, it soon fi nds that the engineering design talent follows the production of the part outside the fi rm, too, and its core capabilities have been lost. Then the fi rm has been hollowed out , becoming merely a distributor of its supplier ’s products.

Given the huge potential effects of outsourcing, both positive and negative, a fi rm should consider such a move very carefully. Management needs to think about both the long-term and short-term effects. They also need to consider the impact of this decision on their core capabilities and everything else they do within the company. Such a major decision as outsourcing will affect other decisions as well, such as sourcing materials, hiring/releasing labor and management, marketing, fi nance, and a wide range of other areas.

So what is the problem? If a supplier can deliver the parts at lower cost and better quality when they are needed, why not use the supplier? The problem is that the supplier gains the expertise (and core capabilities) to produce the critical parts you need, and as Hayes and Pisano (1994), among others, note, organizations quickly forget how they produced those critical parts. After a while, when the supplier has improved on the process and you have forgotten how to make the parts, it is likely to start competing with you, producing the products you have been selling and drop- ping you as a customer. This is even more dangerous if, as already noted, the product and transformation system has also been hollowed out, following the production activities to the supplier. This happened extensively in the television industry, where the Japanese learned fi rst how to produce and then how to engineer black-and-white and, later, color television sets. They then started tentatively introducing their own brands, to see if U.S. customers would buy them. Their products were inexpensive,

c01.indd 40c01.indd 40 25/09/12 4:18 PM25/09/12 4:18 PM

41

of high quality, and caught on quickly in the free-enterprise American markets. The Japanese, and now Koreans, virtually control this industry.

Outsourcing is a growing trend among U.S. manufacturers. The Big Three U.S. automakers are well-known examples of manufacturers that extensively outsourced for years. As other examples, Deere & Co. puts its name on midrange utility tractors produced by a Japanese company, and Agco Corp. outsources the production of almost all of the transmissions and engines used in its farm equipment. Of course, not all manufacturers are jumping on the outsourcing bandwagon. New Balance Athletic Shoes, for example, invested $25 million in its manufacturing facilities as part of an overall strategy to do more assembly in-house.

Kodak ’s Business Imaging Systems Division (BIS) is one fi rm that looked into off- shoring and decided against it (Bartmess and Cerny 1996). Initially they took the “traditional” perspective and discovered that overseas wage rates were 75 percent cheaper than domestic rates. (Recent research reveals that wages in China are now commonly 96 percent cheaper than domestic U.S. wages!) The traditional perspec- tive is one dominated by considering only current conditions with no thought about the future, a short-term response to a competitive threat, a heavy emphasis on cost (primarily labor), and a singular focus on one function, typically operations, to the exclusion of other functions such as engineering, marketing, and design.

But then BIS considered a capabilities perspective and discovered the following:

• Offshore productivity was also low, negating the benefi t of low wage rates. • Large overhead costs were primarily fi xed and would not shrink with overseas

labor.

• Engineering would also have to accompany manufacturing overseas, but off- shore engineering wages were almost equal to domestic wages. Moreover, BIS did not want to lose their domestic engineering competence.

• Over time, foreign wages would be increasing: “… once trained and experi- enced, labor does not stay cheap very long.”

• Cost advantages almost equal to the benefi t of low offshore wages were avail- able through product redesigns.

BIS therefore decided not to move their operation overseas, though they did decide to start an offshore plant for a low-end product for the foreign market, mainly to educate themselves in the advantages and disadvantages of offshore production as well as to learn ways to improve their internal low-cost manufacturing capabili- ties. (For another view on offshoring, see Markides and Berg 1988.)

Regarding the purpose for outsourcing, it is also important to be aware of another danger of outsourcing activities or parts primarily on the basis of “cost.” To illustrate, assume a manufacturer produces four product lines each with an annual volume of 100,000 units. Further assume that the company ’s overhead is £1.2 million (British pounds). Allocating this overhead evenly across the four product lines would result in each unit being allocated £3 in overhead charges. Now suppose that in the interest of lowering its cost and increasing its competitiveness, the manufacturer investigates out- sourcing products that can be produced at lower unit costs by external suppliers. In fact, suppose that a supplier is found for one of its product lines. What is the impact of out- sourcing this product line? Clearly, the company obtains the product at a lower unit cost. But what is the impact on the remaining product lines and the organization ’s overhead?

S t r a t e g y a n d C o m p e t i t i v e n e s s

c01.indd 41c01.indd 41 25/09/12 4:18 PM25/09/12 4:18 PM

42 C h a p t e r 1 : O p e r a t i o n s S t r a t e g y a n d G l o b a l C o m p e t i t i v e n e s s

1. Why is it so hard to increase productivity in the service sector?

2. Identify other major differences between services and products in addition to those listed in Table 1.1 .

3. Many foreign fi rms have been successful in the fol- lowing areas: steel, autos, cameras, and televisions. Are services more protected from foreign competi- tion? How?

4. It is commonly said that Japanese fi rms employ 10 times as many engineers per operations worker as U.S. fi rms and 10 times fewer accountants. What effect would you expect this to have on their com- petitiveness? Why?

5. How might the concept of a “facilitating good” alter the way we perceive a product? A service?

6. Is it wise for a fi rm to stick to what it knows best, or should it expand its market by moving into adjoin- ing products or services? How can it avoid losing its focus?

7. What do you think the result will be of the continu- ing escalation of the U.S. trade defi cit? Will a grad- ual devaluation of the dollar solve the “problem”? If it does, what do you think will be the resulting effect on the United States?

8. Can you think of any other areas of possible focus for a fi rm besides those identifi ed in Table 1.5 ?

9. What core capabilities do you think China pos- sesses? India? Japan? The United States?

10. According to K. Blanchard and N. V. Peale ( The Power of Ethical Management , New York: Morrow, 1988), the following three ethical tests may be useful: (1) Is it legal or within company policy? (2) Is it balanced and fair in the short and long term? (3) Would you be

proud if the public or your family knew about it? Apply these tests to the following situations:

a. A foreign fi rm subsidizes its sales in another country.

b. A foreign fi rm dumps its products (sells them for less than cost) in another country.

c. A country imports products that, had they been made domestically, would have violated domes- tic laws (e.g., laws against pollution).

11. In responding faster to customers ’ needs, where might the cost savings come from? What benefi ts would result?

12. Can you think of companies that have moved the performance frontier of their industries?

13. Why do Americans invest more in marketing new products while the Japanese invest more in engineer- ing? What advantages accrue to each investment?

14. Using new technologies, it is not uncommon for fi rms to cut their response times by a factor of 10. What effect would you expect this to have on their unit costs?

15. With the increasing trend of offshoring in the United States, although companies may get richer, what will happen to their workers? What will the future hold?

16. What are the order winners and order qualifi ers for Wal-Mart? Toyota? BMW? Sony?

17. Given the recent trends in products and services, does the focus strategy or sand cone strategy seem most applicable these days?

18. Why don ’t we see more mass customization in products and services?

More than likely, outsourcing one product line will not have a dramatic impact on total overhead; however, the amount of overhead each unit must now absorb increases from £3/unit to £4/unit. In effect, each unit now appears to be more expensive to produce internally. Thus, outsourcing other product lines now appears to be warranted and likely will be investigated. As you can see, this logic results in a vicious cycle commonly referred to as the creeping breakeven phenomenon . As out- puts are outsourced, the remaining outputs appear to be more expensive to produce in-house. This creates an incentive to outsource even more outputs. The logical con- clusion of this process is that the organization ends up producing no outputs and going bankrupt.

E X P A N D Y O U R U N D E R S T A N D I N G

c01.indd 42c01.indd 42 25/09/12 4:18 PM25/09/12 4:18 PM

43A p p l y Y o u r U n d e r s t a n d i n g

A P P L Y Y O U R U N D E R S T A N D I N G Taracare , Inc .

Taracare, Inc. operates a single factory in Ensenada, Mexico, where it fabricates and assembles a wide range of outdoor furniture for the U.S. market, including chairs, tables, and matching accessories. Taracare ’s primary production activities include extruding the aluminum furniture parts, bending and shaping the extruded parts, fi nishing and painting the parts, and then assembling the parts into completed furniture. Upholstery, glass tabletops, and all hardware are purchased from outside suppliers.

Jorge Gonzalez purchased Taracare in 2001. Before that, Jorge had distinguished himself as a top sales rep of outdoor furniture for the western region of one of the leading national manufacturers. However, after spending 10 years on the road, he wanted to settle down and spend more time with his family back in Mexico. After searching for a couple of months, he came across what he believed to be an ideal opportunity. Not only was it in an industry that he had a great deal of knowledge about, but he would be his own boss. Unfortunately, the asking price was well beyond Jorge ’s means. However, after a month of negotiation, Jorge convinced Jesus Garza, Taracare ’s founder, to maintain a 25 percent stake in the business. Although Jesus had originally intended to sell out completely, he was impressed with Jorge ’s knowledge of the business, his extensive contacts, and his enthusiasm. He therefore agreed to sell Jorge 75 percent of Taracare and retain 25 percent as an investment.

Jorge ’s ambition for Taracare was to expand it from a small regional manufacturer to one that sold to major national retailers. To accomplish this objective, Jorge ’s fi rst initiative was to triple Taracare ’s sales force in 2002. As sales began to increase, Jorge increased the sup- port staff by hiring an accountant, a comptroller, two new designers, and a purchasing agent.

By mid-2005, Taracare ’s line was carried by several national retailers on a trial basis. How- ever, Taracare was having diffi culty both in meeting the deliveries its sales reps were promis- ing and in satisfying the national retailers ’ standards for quality. To respond to this problem, Jorge hired Alfredo Diaz as the new manufacturing manager. Before accepting Jorge ’s offer, Alfredo was the plant manager of a factory that manufactured replacement windows sold by large regional and national retailers.

After several months on the job—and after making little progress toward improving on-time delivery and quality—Alfredo scheduled a meeting with Jorge to discuss his major concerns. Alfredo began:

I requested this meeting with you, Jorge, because I am not satisfi ed with the progress we are making toward improving our delivery performance and quality. The bottom line is that I feel I ’m getting very little cooperation from the other department heads. For example, last month purchasing switched to a new supplier for paint; and although it is true that the new paint costs less per gallon, we have to apply a thicker coat to give the furniture the same protection. I haven ’t actually run the numbers, but I know it is actually costing us more, in both materials and labor.

Another problem is that we typically run a special promotion to coincide with launching new product lines. I understand that the sales guys want to get the product into the stores as quickly as possible, but they are making promises about delivery that we can ’t meet. It takes time to work out the bugs and get things running smoothly. Then there is the problem with the designers. They are constantly adding features to the product that make it almost impossible for us to produce. At the very least, they make it much more expensive for us to produce. For example, on the new “Destiny” line, they designed table legs that required a new die at a cost of 250,000 pesos. Why couldn ’t they have left the legs alone so that we could have used one of our existing dies? On top of this, we have the accounting department telling us that our equipment utilization is too low. Then, when we increase our equipment utilization and make more products, the fi nance guys tell us we have too much capital tied up in inventory. To be honest, I really don ’t feel that I ’m getting very much support.

c01.indd 43c01.indd 43 25/09/12 4:18 PM25/09/12 4:18 PM

44 C h a p t e r 1 : O p e r a t i o n s S t r a t e g y a n d G l o b a l C o m p e t i t i v e n e s s

Rising from his chair, Jorge commented:

You have raised some important issues, Alfredo. Unfortunately, I have to run to another meeting right now. Why don ’t you send me a memo outlining these issues and your recommendations? Then perhaps I will call a meeting and we can discuss these issues with the other department heads. At least our production problems are really no worse than those of our competitors, and we don ’t expect you to solve all of our problems overnight. Keep up the good work and send me that memo at your earliest convenience.

Questions

1. Does Alfredo ’s previous experience running a plant that made replacement windows qualify him to run a plant that makes outdoor furniture?

2. What recommendations would you make if you were Alfredo? 3. Given Jorge ’s background and apparent priorities, how is he likely to respond to Alfredo ’s

recommendations? On the basis of this likely response, is it possible to rephrase Alfredo ’s recommendations so they are more appealing to Jorge?

I zmir Na t iona l Un ive r s i t y

Izmir National University (INU) was chartered in 2000 to facilitate Turkey ’s expected eventual entry into the economy of Europe, via the EU. To foster growth and development in the European economy, engineering, science, and business were deemed to be the institution ’s primary areas of intellectual endeavor. The university grew rapidly during its fi rst three years. By 2005, the enrollment had reached just over 9300 students. However, with this rapid growth came a number of problems. For example, because the faculty had to be hired so quickly, there was little real organization, and curriculum seemed to be decided on the basis of which adviser a student happened to consult. The administrative offi ces were often reshuffl ed, with vague responsibilities and short tenures.

The faculty of the new Business School was typical of the confusion that gripped the entire university. The 26 faculty members were mostly recent graduates of doctoral programs at major European and Turkish universities. There were 21 Assistant Docents and Lecturers, 3 Docents, and 2 full Professors, spread fairly evenly over the four Departments, each over- seen by a Kürsü professor (department head). In addition, funds were available to hire 3 ad- ditional faculty members, either assistant or regular Docents. The background of the newly recruited Dekan (administrative head, dean) of the Business School included fi ve years of teaching at a primarily Muslim university in Turkey and two years of departmental administra- tion at a large southern European university.

Upon arriving at the Business School, the Dekan asked the faculty to e-mail their concerns to her so that she could begin to get a handle on the major issues confronting the school. Her offi ce assistant selected the following comments as representative of the sentiments expressed:

• “Our student–teacher ratio is much higher than what it was at my former university. We need to fi ll those open slots as quickly as possible and ask the university to fund at least two more faculty positions.”

• “If we don ’t get the quality of enrollments up in the MBA program, the graduate school will never approve our application for a doctoral program. We need the doctoral program to attract the best faculty, and we need the doctoral students to help cover our courses.”

• “Given that research is our primary mission, we need to fund more graduate research assistants.”

• “The travel budget isn ’t suffi cient to allow me to attend the meetings I ’m interested in. How can we improve and maintain our visibility if we get funding for only one meeting per year?”

c01.indd 44c01.indd 44 25/09/12 4:18 PM25/09/12 4:18 PM

45B i b l i o g r a p h y

Barney , J. B. , and D. N. Clark . Resource-Based Theory: Creating and Sustaining Competitive Advantage . Oxford, UK : Oxford University Press , 2007 .

Bartlett , C. A. , and S. Ghoshal . “ Going Global: Lessons from Late Movers .” Harvard Business Review (March– April 2000 ): 132 – 142 .

Bartmess , A. , and K. Cerny . “ Building Competitive Advan- tage Through a Global Network of Capabilities .” Chapter 7 in R. A. Hayes, G. P. Pisano, and D. M. Upton: Strategic Operations: Competing Through Capabilities . New York : Free Press , 1996 .

Bennett , J. “U.S. Auto Sales Finish Year Strong.” Wall Street Journal ( January 5, 2012 ).

Bernstein , A. “ Backlash: Behind the Anxiety of Globali- zation .” Business Week (April 24, 2000 ): 38 – 44 .

Bowen , D. E. , R. E. Chase , and T. G. Cummings & Asso- ciates . Service Management Effectiveness . San Francisco : Jossey-Bass , 1990 .

Bower , J. L. , and C. M. Christensen . “ Disruptive Tech- nologies: Catching the Wave .” Harvard Business Review ( January–February 1995 ): 43 – 53 .

Cheng , J. , and B. Intindola . “Apple ’s Size Clouds Market.” Wall Street Journal (February 15, 2012 ).

Clark , K. “ Competing Through Manufacturing and the New Manufacturing Paradigm: Is Manufacturing Strategy Passé? ” Production and Operations Management , 5 ( 1996 ): 42 – 58 .

Collis , D. J. , and C. A. Montgomery . Corporate Strategy: A Resource Based Approach , 2nd ed. New York : McGraw-Hill/Irwin , 2005 .

Crosby , P. B. Quality Is Free: The Art of Making Quality Certain . New York : McGraw-Hill , 1979 .

Deming , W. E. Out of Crisis . Cambridge, MA : MIT Press , 1986 .

Denning , L. “The End of the Road is Nigh.” Wall Street Journal (March 31, 2009 ).

Einhorn , B. “ Your Next TV .” Business Week (April 4, 2005 ): 33 – 36 .

Eisenhardt , K. , and S. Brown . “ Time-Pacing: Competing in Markets that Won ’t Stand Still .” Harvard Business Review (March–April 1998 ): 59 – 69 .

Esterl , M. “ PepsiCo Board Stands by Nooyi .” Wall Street Journal ( January 13, 2012 ): B1 – B2 .

Evans , P. , and T. S. Wurster . Blown to Bits . Boston : HBS Press , 2000 .

Feitzinger , E. , and H. L. Lee . “ Mass Customization at Hewlett-Packard: The Power of Postponement .” Harvard Business Review ( January–February 1997 ): 116 – 121 .

Feld , W. M. Lean Manufacturing: Tools, Techniques, and How to Use Them . Boca Raton, FL : CRC Press , 2000 .

Ferdows , K. , ed. Managing International Manufactur- ing . New York : North-Holland , 1989 .

• “We need better staff support. Faculty members are required to submit their exams for copying fi ve days before they are needed. However, doing this makes it diffi cult to test the students on the material covered in class right before the exam, since it ’s diffi cult to know ahead of time exactly how much material we will cover.”

• “I think far too much emphasis is placed on research. We are here to teach.” • “Being limited in our consulting is far too restrictive. In Europe we were allowed one day

a week. How are we supposed to stay current without consulting?” • “We need a voice mail system. I never get my important messages.”

Questions

1. What do the comments by the faculty tell you about INU ’s strategy? 2. What would you recommend the Dekan do regarding the Business School ’s strategic

planning process? What role would you recommend the Dekan play in this process? 3. Productivity is defi ned as the ratio of output (including both goods and services) to the

input used to produce it. How could the productivity of the Business School be measured? What would the effect be on productivity if the faculty all received a 10 percent raise but continued to teach the same number of classes and students?

B I B L I O G R A P H Y

c01.indd 45c01.indd 45 25/09/12 4:18 PM25/09/12 4:18 PM

46 C h a p t e r 1 : O p e r a t i o n s S t r a t e g y a n d G l o b a l C o m p e t i t i v e n e s s

Ferdows , K. , and A. DeMeyer . “ Lasting Improvements in Manufacturing Performance: In Search of a New Theory .” Journal of Operations Management , 9 ( 1990 ): 168 – 184 .

Fitzsimmons , J. A. , and M. J. Fitzsimmons . Service Man- agement for Competitive Advantage , 5th ed. New York : McGraw-Hill , 2006 .

Gilmore , J. H. , and B. J. Pine II . “ The Four Faces of Mass Customization .” Harvard Business Review ( January– February 1997 ): 91 – 101 .

Goetsch , D. L. , and S. B. Davis . Quality Management: Introduction to Total Quality Management for Produc- tion, Processing, and Services , 3rd ed. Upper Saddle River, NJ : Prentice Hall , 2000 .

Goldstein , S. M. , R. Johnson , J. Duffy , and J. Rao . “ The Service Concept: The Missing Link in Service Design Re- search? ” Journal of Operations Management , 20 ( 2002 ): 121 – 134 .

Hammer , M. “ Deep Change: How Operational Innova- tion Can Transform Your Company .” Harvard Business Review (April 2004 ): 85 – 93 .

Hammer , M. , and S. Stanton . “ How Process Enterprises Really Work .” Harvard Business Review (November– December 1999 ): 108 – 120 .

Handfi eld , R. B. , and E. L. Nichols Jr. Introduction to Supply Chain Management . Upper Saddle River, NJ : Prentice-Hall , 1999 .

Hayes , R. H. , and G. P. Pisano . “ Beyond World-Class: The New Manufacturing Strategy .” Harvard Business Review ( January–February 1994 ): 77 – 86 .

Hayes , R. H. , G. P. Pisano , D. M. Upton , and S. C. Wheel- wright . Operations, Strategy, and Technology: Pursuing the Competitive Edge . New York : John Wiley , 2004 .

Hill , T. Manufacturing Strategy: Text and Cases , 3rd ed. Homewood, IL : Irwin , 2000 .

Jonas , N. “ The Hollow Corporation .” Business Week (March 3, 1986 ): 57 – 85 .

Kaplan , R. S. , and D. P. Norton .“ Having Trouble with Your Strategy? Then Map It .” Harvard Business Review (September–October 2000 ): 167 – 176 .

Kaplan , R. S. , and D. P. Norton . The Strategy Focused Organization . Boston : Harvard Business School Press , 2001 .

Maidique , M. A. , and P. Patch . Corporate Strategy and Technological Policy . Harvard Business School Case 9-679-033 , Boston , 1979 .

Markides , C. C. , and N. Berg . “ Manufacturing Offshore Is Bad Business .” Harvard Business Review (September– October 1988 ): 113 – 120 .

Meredith , J. R. , D. M. McCutcheon , and J. Hartley . “ Enhancing Competitiveness Through the New Market Value Equation .” International Journal of Operations and Production Management , 14 (November 11, 1994 ): 7 – 21 .

Pande , P. S. , R. P. Neuman , and R. R. Cavanagh . The Six Sigma Way . New York : McGraw-Hill , 2000 .

Pine , B. J. , II . Mass Customization: The New Frontier in Business Competition . Boston : Harvard Business School Press , 1993 .

Porter , M. “ What Is Strategy? ” Harvard Business Review (November–December 1996 ): 61 – 78 .

Porter , M. E. Competitive Advantage . New York : Free Press , 1985 .

Prahalad , C. K. , and G. Hamel . “ The Core Competence of the Corporation .” Harvard Business Review (May– June 1990 ): 79 – 91 .

Shafer , S. M. , H. J. Smith , and J. C. Linder . “ The Power of Business Models .” Business Horizons , 48 (May–June 2005 ): 199 – 207 .

Skinner , W. “ The Focused Factory .” Harvard Business Review (May–June 1974 ): 113 – 122 .

Slack , N. , and M. Lewis . Operations Strategy . New York: Financial Times Prentice-Hall, 2001 .

Stalk , G. “ Time—The Next Source of Competitive Advant- age .” Harvard Business Review (July–August 1988 ): 41 – 51 .

Stalk , G. , P. Evans , and L. E. Shulman . “ Competing on Ca- pabilities: The New Rules of Corporate Strategy .” Harvard Business Review (March–April 1992 ): 57 – 69 .

Takabashi , Y. “Toyota Forecasts Sharp Rebound After Tough 2011.” Wall Street Journal (December 23, 2011 .)

Terlep , S. “GM Squeezes After Bailout.” Wall Street Jour- nal , (February 17, 2012) .

Upton , D. M. “ The Management of Manufacturing Flexibil- ity .” California Management Review (Winter 1994 ): 72 – 89 .

Vessey , J. T. “ The New Competitors: They Think in Terms of ‘Speed-to-Market.’ ” Academy of Management Review , 5 (April 1991 ): 23 – 33 .

Womack , J. P. , and D. T. Jones . Lean Thinking: Banish Waste and Create Wealth in Your Corporation , rev. ed. New York : Free Press , 2003 .

c01.indd 46c01.indd 46 25/09/12 4:18 PM25/09/12 4:18 PM

47

� Process Planning and Design

C H A P T E R 2

ROLE OF OPERATIONS IN ORGANIZATIONS’ COMPETITIVENESS

Operations Strategy

Process Design and Control

Process Improvement

Process Execution

Ch. 1: Operations Strategy and Global

Competitiveness

Ch. 2: Process Planning and

Design

Ch. 3: Controlling Processes

Ch. 7: Supply Chain Management

Ch. 8: Capacity, Scheduling, and

Location Planning

Ch. 4: Process Improvement:

Six Sigma

Ch. 5: Process Improvement: Lean

Ch. 6: Managing Process

Improvement Projects

c02.indd 47c02.indd 47 25/09/12 4:20 PM25/09/12 4:20 PM

48

IN T R O D U C T I O N • Louis Vuitton, one of the luxury goods brands of giant LVMH Moët Hennessy Louis

Vuitton, has the unusual problem of making its products in volume for the world ’s rich, and aspiring rich, while maintaining an image that its products are limited- edition and exclusive, which many of its products are. Vuitton has classically used a job shop to make their famous runway pieces, bags, travel cases, wallets, and other items. In the old days, some workers would cut fabric, others would sew together leather panels, some would glue in linings, still others would attach handles, and so on.

These days technology has frequently replaced manual labor for higher produc- tivity and greater consistency, such as the robots that fetch shoe molds and the com- puter software that helps program how to cut leather to avoid fl aws in the skins. Another approach—cellular manufacturing—is also being used for some items. In these cells, small teams of six or more workers, each performing a few different tasks around a U-shaped cluster of workstations, pass their work around the cluster to produce a fi nished product in one-eighth the time the job shop required. In addition, fewer workers are needed and defects are caught sooner—returns of faulty handbags and wallets dropped by two-thirds after Vuitton implemented the cellular production process (Passariello 2006, 2011).

• The assembly line at IBM ’s plant in Charlotte, North Carolina, was unlike any other in the world. What made it unique was that it was designed to produce 27 signifi cantly different products. Indeed, the variety of products produced by the team of 40 workers who operated this line is astounding; these products included hand-held bar-code scanners, portable medical computers, fi ber-optic connectors, and satellite communica- tions devices. The assembly line operated by delivering to each worker a “kit” of parts based on the production schedule. Since each product required different assembly

Chapter 1 described some operations strategies to enhance an organization ’s competitiveness and identifi ed critical factors in providing value to the customer. As shown in the accompanying dia- gram, the next task is the selection and design of the transformation process that can deliver those factors—low cost, high quality, innovative outputs, customization, fast reponse, and so on—in an effi - cient and effective manner. If an organization is using the wrong transformation process, fre- quently because the organization has changed or

the market has changed over time, it will not be competitive on some of these critical value factors. The chapter begins with an overview of the fi ve major types of transformation processes and their respective advantages and disadvantages. Next, issues related to the selection of a competi- tive transformation process, such as considerations of volume, variety, and product life cycles, are discussed. Last, explicit attention is given to some of the unique aspects of designing service operations.

48

C H A P T E R I N P E R S P E C T I V E

c02.indd 48c02.indd 48 25/09/12 4:20 PM25/09/12 4:20 PM

procedures, each worker had a computer screen at his or her station that displayed updated assembly instructions for the current product (Bylinsky 1994).

• Rickard Associates, an editorial production company that produces magazines and marketing materials, was a pioneer in the mid-1990s of a new type of organizational structure, the “virtual” organization. Only two of its employees actually worked at its headquarters in New Jersey: the art director worked in Arizona; the editors were located in Florida, Georgia, Michigan, and the District of Columbia; and the freelancers were even more scattered. To coordinate work, the Internet and America Online were used. For example, art directors were able to submit electronic fi les of fi nished pages to headquarters in a matter of minutes using these computer networks (Verity 1994).

Today, many organizations use this method, or enhancements of it, for getting work done, either for their entire organization or, more frequently, for project teams that may be scattered around the globe. For instance, Intel, Inc. and other high-tech fi rms commonly pass work on a project from one continent to another—as one team goes home to bed and another wakes up and starts work—to keep progress moving on quick-response, high-importance projects (Meredith and Mantel, 2012).

• Martin Marietta ’s aerospace electronics manufacturing facility in Denver, Colorado, was initially set up as a job shop with numerous functional departments. As is typi- cal of most job shops, the Marietta plant had high levels of work-in-process and long lead times, and parts had to travel long distances throughout the plant to have their processing completed. Also, as is typical of functional organizations, departmental divisions created barriers to communication and often resulted in confl icting goals. To address these problems, Martin Marietta organized its plant into three “focused factories.” Each focused factory was completely responsible and accountable for building electronic assemblies for a particular application (e.g., fl ight, space, or ground use). The intent was to make each focused factory a separate business enterprise.

A factory manager was assigned to each focused factory. The factory managers then engaged in a sort of “NFL draft” to select employees for their teams. Workers not drafted had to fi nd other positions either inside or outside the company. Within the focused factories, product families were identifi ed; these were based on the tech- nology and processing requirements of the products. Next, standardized routings and sequences were identifi ed for each product family. The plant realized a number of improvements as a result of these and other changes, including seven consecutive months of production with no scrap, a 50 percent reduction in work-in-process inventory, a 21 percent average reduction in lead times, and a 90 percent reduction in overtime (Ferras 1994).

These examples illustrate several transformation systems. The Louis Vuitton fac- tory was originally a pure job shop that had specialized departments for cutting, stitching, glueing, and so on. Although some of its products are still produced in this manner, others (typically those with higher volumes) are produced in a cellular

I n t r o d u c t i o n 49

c02.indd 49c02.indd 49 25/09/12 4:20 PM25/09/12 4:20 PM

50 C h a p t e r 2 : P r o c e s s P l a n n i n g a n d D e s i g n

production process. Likewise, because work is organized by the task performed, Rickard Associates is also a job shop—even though the work is not performed in one location. Actually, and as mentioned in the example, companies like Rickard that rely on information technology to bring separated workers together are referred to as virtual organizations . Martin Marietta converted into focused factories . And assembly lines like the one IBM uses are referred to as fl ow shops.

As we noted in Chapter 1, the sand cone model of additive and complementary competitive strengths emphasizes operations that can deliver quality, delivery dependability, speed, and low cost. The most important ingredient in achieving these strengths is selecting the most appropriate transformation process design and layout for the organization ’s operations. There are various basic forms of transforma- tion process designs, each with its own layout, as well as myriad combinations and hybrids of them. This chapter describes these transformation systems, how the oper- ations are laid out for each of them, and how to select the most appropriate one for maximum competitiveness.

The fi ve basic forms of transformation systems are (1) continuous process, (2) fl ow shop, (3) job shop, (4) cellular, and (5) project. The continuous process industries are in many ways the most advanced, moving fl uid material continuously through vats and pipes until a fi nal product is obtained. Flow shops produce dis- crete, usually standardized, outputs on a continuous basis by means of assembly lines or mass production, often using automated equipment. Cellular shops produce “families” of outputs within a variety of fl ow cells, but numerous cells within the plant can offer a range of families of outputs. Job shops offer a wide range of pos- sible outputs, usually in batches, via individualized processing into and out of a number of functionally specialized departments. These departments typically consist of a set of largely identical equipment, as well as highly skilled workers. (Potentially, job shops could also produce unique—that is, one-of-a-kind—customized outputs, but job shops that do this are commonly called model shops or, in Europe, jobbers .) Finally, projects are temporary endeavors set up to achieve a unique outcome. The most commonly known projects are those performed on a massive scale when the labor and equipment are brought to each site rather than to a fi xed production facil- ity, such as dams, buildings, roadways, space launches, and so on.

The general procedure for selecting a transformation system is to consider all alternative forms and combinations to devise the best strategy for obtaining the desired outputs. The major considerations in designing the transformation system— effi ciency, effectiveness, volume, capacity, lead time, fl exibility , and so on—are so interdependent that changing the system to alter one will change the others as well. And the layout of the operations is another aspect that must be considered in the selection of the transformation system. The main purpose of layout analysis is to maximize the effi ciency (cost-orientation) or effectiveness (e.g., quality, lead time, fl exibility) of operations. Other purposes also exist, such as reducing safety or health hazards, minimizing interference or noise between different operational areas (e.g., separating painting from sanding), facilitating crucial staff interactions, or maximiz- ing customers ’ exposure to products or services.

In laying out service operations, the emphasis may instead be on accommodating the customer rather than on operations per se. Moreover, capacity and layout analy- ses are frequently conducted simultaneously by analyzing service operations and the wait that the customer must endure. Thus, waiting line (or queuing ) theory , a topic discussed in Chapter 8, is often used in the design of a service delivery system.

c02.indd 50c02.indd 50 25/09/12 4:20 PM25/09/12 4:20 PM

51F o r m s o f T r a n s f o r m a t i o n S y s t e m s

The layouts of parking lots, entry zones, reception rooms, waiting areas, service facilities, and other points of customer contact are of top priority in service-oriented organizations such as clinics, stores, nightclubs, restaurants, and banks.

In a frequently changing environment, the transformation system and its layout will have to be constantly monitored and occasionally redesigned to cope with new demands, new products and services, new government regulations, and new tech- nology. Technology, increasing global competition, sustainability, the green move- ment, and shortages of materials and energy are only a few examples of changes in the recent past that have forced organizations to recognize the necessity of adapting their operations.

F O R M S O F T R A N S F O R M A T I O N S Y S T E M S Continuous Process

The continuous transformation process is commonly used to produce highly stand- ardized outputs, usually fl uidic products, in extremely large volumes. In some cases these outputs have become so standardized that there are virtually no real differ- ences between the outputs of different fi rms. Examples of such commodities include water, gases, chemicals, electricity, ores, rubber, fl our, spirits, cements, petroleum, and milk. The term continuous process refl ects the typical practice of running these operations 24 hours a day, seven days a week. One reason for running these systems continuously is to spread their enormous fi xed cost over as large a volume as pos- sible, thereby reducing unit costs. This is particularly important in commodity mar- kets, where price can be the single most important factor in competing successfully. Another reason for operating these processes continuously is that stopping and start- ing them can be prohibitively expensive.

Continuous process industries constitute about half of the manufacturing industry in the United States. Although not all of this industry produces commodities, those are what is typically envisioned. The operations in these commodity industries are highly automated, with very specialized equipment and controls, often electronic and computerized. Such automation and the expense it entails are necessary because of strict processing requirements. Because of the highly specialized and automated nature of the equipment, changing the rate of output can be quite diffi cult. The facil- ity is typically a maze of pipes, conveyors, tanks, valves, vats, and bins. The layout follows the processing stages of the product, and the output rate is controlled through equipment capacity and fl ow and mixture rates. Labor requirements are low and are devoted primarily to monitoring and maintaining the equipment.

Research (Dennis and Meredith 2000), however, has shown that there is a much wider range of continuous process industries than just commodity manufacturers. In fact, these industries range all the way from intermittent forms akin to job shops to rigidly continuous fl ow shops (both described next). In fact, there appear to be at least seven clearly differentiable forms of continuous processes. Some run for a short time making one product and then switch over to make another product, largely on demand and to the specifi cation of individual customers, which is almost the opposite of commodity production. In addition to these two extremes, there are also blending types of continuous processes as well as unusual hybrids of both job and fl ow shops.

c02.indd 51c02.indd 51 25/09/12 4:20 PM25/09/12 4:20 PM

52 C h a p t e r 2 : P r o c e s s P l a n n i n g a n d D e s i g n

The major characteristic of processing industries, especially commodities, is that there is often one primary, “fl uid”-type input material (gas, wood, wheat, milk, etc.). This input is then often converted to multiple outputs, although sometimes there may be only one output (e.g., clean, chlorinated water). In contrast, in discrete production many types of materials are made or purchased and combined to form the output.

Although human variation in continuous processing fi rms does not usually create the problems it creates in discrete manufacturing, the demands of processing are usu- ally more critical. For example, chemical reactions must be accurately timed. The result is that the initial setup of equipment and procedures is even more complex and critical than it is for fl ow shops. Fixed costs are extremely high; the major variable cost is materials. Variable labor (excluding distribution) is usually insignifi cant.

Flow Shop The fl ow shop is a transformation system similar to the continuous process, the major difference being that in the fl ow shop there is a discrete product or service, whereas in continuous processes the end product is not naturally divisible. Thus, in continu- ous processes an additional step, such as bottling or canning, might be needed to get the product into discrete units. Like the continuous process, the fl ow shop treats all the outputs as basically the same, and the fl ow of work is thus relatively continu- ous. Organizations that use this form are heavily automated, with large, special- purpose equipment. The characteristics of the fl ow shop are a fi xed set of inputs, constant throughput times, and a fi xed set of outputs. Examples of the fl ow form for discrete products are pencil manufacturing, steelmaking, and automobile assembly, whereas for services, some examples include the car wash, the processing of insur- ance claims, and the ubiquitous fast-food restaurant.

An organization that produces, or plans to produce, a high volume of a small variety of outputs will thus probably organize its operations as a fl ow shop. In doing so, the organization will take advantage of the simplicity and the savings in variable costs that such an approach offers. Because outputs and operations are standard- ized, specialized equipment can be used to perform the necessary operations at low per-unit costs, and the relatively large fi xed costs of the equipment are distributed over a large volume of outputs.

Continuous types of materials-handling equipment, such as conveyors—again operating at low per-unit costs—can be used because the operations are standard- ized and, typically, all outputs follow the same path from one operation to the next. This standardization of treatment provides for a known, fi xed throughput time, giv- ing managers easier control of the system and more reliable delivery dates. The fl ow shop is easier to manage for other reasons as well: routing, scheduling, and control are all facilitated because each output does not have to be individually monitored and controlled. Standardization of operations means that fewer skilled workers can be used and each manager ’s span of control can increase.

The general form of the fl ow shop is illustrated in Figure 2.1 , which shows a pro- duction line . (If only assembly operations were being performed, as in many auto- motive plants, the line would be called an assembly line .) This production line could represent new military inductees taking their physical exams, small appliances being assembled, or double-decker hamburgers being prepared.

c02.indd 52c02.indd 52 25/09/12 4:20 PM25/09/12 4:20 PM

53F o r m s o f T r a n s f o r m a t i o n S y s t e m s

Note that both services and products can be organized as fl ow shops and can capitalize on the many advantages of this form of processing.

Advantages of the Flow Shop

The primary advantage of a fl ow shop is the low per-unit cost that is attainable owing to specialized high-volume equipment, bulk purchasing, lower labor rates, effi cient utilization of the facility, low in-process inventories, and simplifi ed mana- gerial control. In addition, with everyone working on all the required tasks simul- taneously, referred to as overlapping , product or service outputs are produced very quickly.

Because of the high rate of output, materials can often be bought in large quanti- ties at a signifi cant savings. Also, because operations are standardized, processing times tend to remain constant so that large in-process inventories are not required to queue up for processing. This minimizes investment in in-process inventory and queue ( buffer ) space. Furthermore, because a standardized product is produced, inventory control and purchasing decisions are routine.

Because the machines are specialized, less skilled operators are needed, and therefore lower wages can be paid. In addition, fewer supervisors are needed, fur- ther reducing costs. Since the fl ow shop is generally continuous, with materials handling often built into the system itself, the operations can be designed to per- form compactly and effi ciently with narrow aisles, thereby making maximum use of space.

The simplifi cation in managerial control of a well-designed fl ow shop should not be overlooked. Constant operations problems requiring unending managerial atten- tion penalize the organization by distracting managers from their normal duties of planning and decision making.

Shipping

Storage

Out

In

Figure 2.1 A generalized fl ow shop operation.

c02.indd 53c02.indd 53 25/09/12 4:20 PM25/09/12 4:20 PM

54 C h a p t e r 2 : P r o c e s s P l a n n i n g a n d D e s i g n

Disadvantages of the Flow Shop

Despite the important cost advantage of the fl ow shop, it can have some serious draw- backs. Not only is variety of output diffi cult to obtain, even changes in the rate of output are hard to make. Changing the rate of output may require using overtime, lay- ing off workers, adding additional shifts, or temporarily closing the plant. Also, because the equipment is so specialized, minor changes in the design of the product often require substantial changes in the equipment. Thus, important changes in product design are infrequent, and this could weaken the organization ’s competitive position.

A well-known problem in fl ow shops is boredom and absenteeism among the labor force. Since the equipment performs the skilled tasks, there is no challenge for the workers. And, of course, the constant, unending, repetitive nature of the manu- facturing line can dehumanize workers. Because the rate of work fl ow is generally set ( paced ) by the line speed, incentive pay and other output-based incentives are not possible.

The fl ow production line form has another important drawback. If the line should stop for any reason—a breakdown of a machine or conveyor, a shortage of supplies, and so forth—production may come to an immediate halt unless work-in-process (WIP) is stored at key points in the line. Such occurrences are prohibitively expensive.

Other requirements of the fl ow shop also add to its cost and its problems. For example, parts must be standardized so that they will fi t together easily and quickly on the assembly line. And, since all machines and labor must work at the same repetitive pace in order to coordinate operations, the workloads along the entire line are generally balanced to the pace of the slowest element. To keep the line running smoothly, a large support staff is required, as well as large stocks of raw materials, all of which also add to the expense.

Last, in the fl ow shop, simplicity in ongoing operation is achieved at the cost of complexity in the initial setup . The planning, design, and installation of the typically complicated, special-purpose, high-volume equipment are mammoth tasks. The equipment is costly not only to set up originally but also to maintain and service. Furthermore, such special-purpose equipment is very susceptible to obsolescence and is diffi cult to dispose of or to modify for other purposes.

Layout of the Flow Shop

The crux of the problem of realizing the advantages of a fl ow shop is whether the work fl ow can be subdivided suffi ciently so that labor and equipment are utilized smoothly throughout the processing operations. If, for example, one operation takes longer than all the others, this single operation (perhaps a machine) will become a bottleneck, delaying all the operations following it and restricting the output rate of the entire process.

Obtaining smooth utilization of workers and equipment across all operations involves assigning to groups tasks that take about the same amount of time to com- plete. This balancing applies to production lines, where parts or outputs are pro- duced, as well as to assembly lines, where parts are assembled into fi nal products.

Final assembly operations usually have more labor input and fewer fi xed- equipment cycles and can therefore be subdivided more easily for smooth fl ow. Either of two types of lines can then be used. A paced line uses some sort of con- veyor and moves the output along at a continuous rate, and operators do their work

c02.indd 54c02.indd 54 25/09/12 4:20 PM25/09/12 4:20 PM

55F o r m s o f T r a n s f o r m a t i o n S y s t e m s

as the output passes by them. For longer operations, the worker may walk or ride along- side the conveyor and then have to walk back to the starting workstation. The many disadvantages of this arrangement, such as boredom and monotony, are, of course, well known. An automobile assembly line is a common example of a paced line. Workers install doors, engines, hoods, and the like as the conveyor moves past them.

In unpaced lines, the workers build up queues between workstations and can then vary their pace to meet the needs of the job or their personal desires; however, average daily output must remain the same. The advantage of an unpaced line is that a worker can spend longer on the more diffi cult outputs and balance this with less time spent on the easier outputs. Similarly, workers can vary their pace to add vari- ety to a boring task. For example, a worker may work fast to get ahead of the pace and then pause for a few seconds before returning to the task.

There are some disadvantages to unpaced lines, however. For one thing, they cannot be used with large, bulky products because too much in-process storage space is required. More important, minimum output rates are diffi cult to maintain because short durations in one operation usually do not dovetail with long durations in the next operation. When long durations coincide, operators downstream from these operations may run out of in-process inventory to work on and may thus be forced to sit idle.

For operations that can be smoothed to obtain the benefi ts of a production line, there are two main elements in designing the most effi cient line. The fi rst is formu- lating the situation by determining the necessary output rate, the available work time per day, the times for operational tasks, and the order of precedence of the opera- tions. The second element is actually to solve the balancing problem by subdividing and grouping the operations into balanced jobs. To more clearly communicate the concept of a balanced production line, we will give an example that addresses both of these main elements. In reality, of course, one of a variety of computer packages would be employed.

Balancing the Production Line

We illustrate the formulation of the line balancing situation with an example. Longform Credit receives 1200 credit applications a day, on average. Longform com- petes on the basis of its ability to process applications within hours. Daily applica- tion processing tasks, average times, and required preceding tasks (tasks that must be completed before the next task) are listed in Table 2.1 .

The precedence graph for these tasks is shown in Figure 2.2 ; it is constructed directly from Table 2.1 . This graph is simply a picture of the operations (boxed) with arrows indicating which tasks must precede others. The number or letter of the operation is shown above the box, with its time inside.

In balancing a line, the intent is to fi nd a cycle time in which each workstation can complete its tasks. A workstation is usually a single person, but it may include any number of people responsible for completing all the tasks associated with the job for that station. Conceptually, at the end of this time every workstation passes its part on to the next station, and, of course, one item comes off the end of the line fully com- plete. (The lean term takt time is now commonly used in practice instead of cycle time . This switch in terminology has been made in part to help clear up the confu- sion created when the term cycle time was erroneously used to refer to the through- put time, or the time it takes to complete all the work to produce a fi nished item.)

c02.indd 55c02.indd 55 25/09/12 4:20 PM25/09/12 4:20 PM

56 C h a p t e r 2 : P r o c e s s P l a n n i n g a n d D e s i g n

Task elements are thus grouped for each workstation so as to utilize as much of this cycle time as possible but not to exceed it. Each workstation will have a slightly dif- ferent idle time within the cycle time.

Cycle time   5   available work time / demand

  5 8 hr   3   60 min/hr ________________

____

1200 applications  5   ( 0.4 min / application )  

The cycle time is determined from the required output rate. In this case, the aver- age daily output rate must equal the average daily input rate, 1200. If it is less than this amount, a backlog of applications will accumulate. If it is more than this, unnecessary

T A B L E 2 .1 • Tasks in Cred i t App l i ca t ion Proce s s ing

Task Average Time

(minutes) Immediately

Preceding Tasks

a Open and stack applications 0.20 none

b Process enclosed letter; make note of and handle any special requirements

0.37 a

c Check off form 1 for page 1 of application 0.21 a

d Check off form 2 for page 2 of application; fi le original copy of application

0.18 a

e Calculate credit limit from standardized tables according to forms 1 and 2

0.19 c, d

f Supervisor checks quotation in light of special processing of letter, notes type of form letter, address, and credit limit to return to applicant

0.39 b, e

g Administrative assistant types in details on form letter and mails

0.36 f

Total 1.90

Figure 2.2 Precedence graph for credit applications.

0.37

0.21 0.19

0.18

0.20 0.39 0.36

b

a c e f g

d

c02.indd 56c02.indd 56 25/09/12 4:20 PM25/09/12 4:20 PM

57F o r m s o f T r a n s f o r m a t i o n S y s t e m s

idle time will result. Assuming an eight-hour day, 1200 applications per eight hours means completing 150 every hour or one every 0.4 minute—this, then, is the cycle time.

Adding up the task times in Table 2.1 , we can see that the total is 1.9 minutes. Since every workstation will do no more than 0.4 minute ’s work during each cycle, it is clear that a minimum of 1.9/0.4 5 4.75 workstations are needed—or, always rounding up , fi ve workstations.

Number of theoretical workstations ,   N T   5   ∑  task times / cycle time

  5   1.9 ___ 0.4

  5   4.75 ( i .e . , 5 )  

It may be, however, that the work cannot be divided and balanced in fi ve stations—that six, or even seven, may be needed. For example, precedence relation- ships may interfere with assigning two tasks to the same workstation. This is why we referred to N

T as the theoretical number of workstations needed. If more worksta-

tions are actually needed than the theoretical number, the production line will be less effi cient. The effi ciency of the line with N

A actual stations may be computed from

Effi ciency   5   output

______ input   5   total task time ______________________

( N A stations )   3   cycle time  

1.9 _______ 5   3   0.4  

5   95 percent if the line can be balanced with 5 stations

1.9 ________ 6   3   0.4  

5   79 percent if 6 stations are required

In the formula for effi ciency, input is represented by the amount of work required to produce one unit, and output is represented by the amount of work that actually goes into producing one unit.

Now that the problem has been formulated, we can attempt to balance the line by assigning tasks to stations. We begin by assuming that all workers can do any of the tasks and check back on this later. There are many heuristic rules for deciding which task to assign to a station next. We will use the LOT rule; select the task with the longest operation time next. The general procedure for line balancing is:

• Construct a list of the tasks whose predecessor tasks have already been completed.

• Consider each of these tasks, one at a time, in LOT order and place them within the station.

• As a task is tentatively placed in a station, add new follower tasks to the list. • Consider adding to the station any tasks in this list whose time fi ts within the

remaining time for that station.

• Continue in this manner until as little idle time as possible remains for the station.

We will now demonstrate this procedure with reference to Longform, using the information in Table 2.1 and Figure 2.2 . The fi rst tasks to consider are those with no preceding tasks. Thus, task a , taking 0.2 of the 0.4 minute available, is assigned to station 1. This, then, makes tasks b (0.37 minute), c (0.21 minute), and

c02.indd 57c02.indd 57 25/09/12 4:20 PM25/09/12 4:20 PM

58 C h a p t e r 2 : P r o c e s s P l a n n i n g a n d D e s i g n

d (0.18 minute) eligible for assignment. Trying the longest fi rst, b , then c , and last d , we fi nd that only d can be assigned to station 1 without exceeding the 0.4-minute cycle time; thus, station 1 will include tasks a and d . Since only 0.02 minute remains unassigned in station 1 and no task is that short, we then consider assignments to station 2.

Only b and c are eligible for assignment (since e requires that c be completed fi rst), and b (0.37 minute) will clearly require a station by itself; b is, therefore, assigned to station 2. Only c is now eligible for assignment, since f requires that both e and b be completed and e is not yet completed. But when we assign c (0.21 minute) to station 3, task e (0.19 minute) becomes available and can also be just accommodated in station 3. Task f (0.39 minute), the next eligible task, requires its own station; this leaves g (0.36 minute) to station 5. These assignments are illustrated in Figure 2.3 and Table 2.2 .

T A B L E 2 .2 • S ta t ion Task A s s ignment s Station Time Available Eligible Tasks Task Assigned Idle Time

1 0.40 a a

0.20 b, c, d d

0.02 b, c none will fi t 0.02

2 0.40 b, c b

0.03 c c will not fi t 0.03

3 0.40 c c

0.19 e e 0.00

4 0.40 f f

0.01 g g will not fi t 0.01

5 0.40 g g 0.04

Figure 2.3 Station assignments.

0.36

Station 3

Station 1

Station 5Station 4

0.37

Station 2

0.20 0.190.21 0.39

0.18

c e f g

d

b

a

c02.indd 58c02.indd 58 25/09/12 4:20 PM25/09/12 4:20 PM

59F o r m s o f T r a n s f o r m a t i o n S y s t e m s

We now check the feasibility of these assignments. In many cases, several aspects must be considered in this check (as discussed later), but here our only concern is that the administrative assistant does not do task f and that the supervisor does not do task g (or, we hope, much of a through e ). As it happens, task f is a station by itself, so there is no problem.

As we saw, short tasks are often combined to reach the cycle time. However, long tasks may have to be split up to meet the cycle time requirements. If a task cannot be split, we can “clone” the station as many times as needed to effectively reduce its cycle time, with each station alternating in its output to match, in essence, the required cycle time.

Job Shop The job shop gets its name because unique jobs must be produced. In this form of transformation system, each output, or each small batch of outputs, is processed dif- ferently. Therefore, the fl ow of work through the facility tends to be intermittent.

The general characteristics of a job shop are grouping of staff and equipment according to function; a large variety of inputs; a considerable amount of transport of staff, materials, or recipients; and large variations in system fl ow times (the time it takes for a complete “job”). In general, each output takes a different route through the organization, requires different operations, uses different inputs, and takes a dif- ferent amount of time.

This transformation system is common when the outputs differ signifi cantly in form, structure, materials, or processing required. For example, an organization that has a wide variety of outputs or does custom work (e.g., custom guitars) would probably be a job shop. Specifi c examples of product and service organizations of this form are tailor shops, general offi ces, machine shops, public parks, hospitals, universities, auto- mobile repair shops, criminal justice systems, and department stores. By and large, the job shop is especially appropriate for service organizations because services are often customized, and hence each service requires different operations.

Clearly, the effi cient management of a job shop is a diffi cult task, since every out- put must be treated differently. In addition, the resources available for processing are limited. Furthermore, not only is it management ’s task to ensure the performance of the proper functions of each output, where considerations of quality and deadlines may vary, but management must also be sure that the available resources (staff, equipment, materials, supplies, capital) are being effi ciently utilized. Often there is a diffi cult tradeoff between effi ciency and fl exibility of operations. Job-based proc- esses tend to emphasize fl exibility over effi ciency.

Figure 2.4 represents the fl ow through a job shop. This facility might be a library, an auto repair shop, or an offi ce. Each particular “job” travels from one area to another, and so on, according to its unique routing, until it is fully processed. Temporary in-process storage may occur between various operations while jobs are waiting for subsequent processing (standing in line for the coffee machine).

Advantages of the Job Shop

The widespread use of the job shop form is due to its many advantages. The job shop is usually selected to provide the organization with the fl exibility needed to

c02.indd 59c02.indd 59 25/09/12 4:20 PM25/09/12 4:20 PM

60 C h a p t e r 2 : P r o c e s s P l a n n i n g a n d D e s i g n

respond to individual, small-volume demands (or even custom demands). The abil- ity to produce a wide variety of outputs at reasonable cost is thus the primary advan- tage of this form. General-purpose equipment is used, and this is in greater demand and is usually available from more suppliers at a lower price than special-purpose equipment. In addition, used equipment is more likely to be available, further reduc- ing the necessary investment. There is a larger base of experience with general- purpose equipment; therefore, problems with installation and maintenance are more predictable, and replacement parts are more widely available. Last, because general-purpose equipment is easier to modify or use elsewhere and disposal is much easier, the expense of obsolescence is minimized.

Because of the functional arrangement of the equipment, there are also other advantages. Resources for a function requiring special staff, materials, or facilities (e.g., painting or audiovisual equipment) may be centralized at the location of that function, and the organization can thus save expense through high utilization rates. Distracting or dangerous equipment, supplies, or activities may also be segregated from other operations in facilities that are soundproof, airtight, explosion-proof, and so forth.

One advantage to the staff is that with more highly skilled work involving con- stantly varying jobs, responsibility and pride in one ’s work are increased, and bore- dom is reduced. Other advantages to the staff are that concentrations of experience and expertise are available and morale increases when people with similar skills work together in centralized locations (all market researchers together). Because all workers who perform similar activities are grouped together, each worker has the opportunity to learn from others, and the workers can easily collaborate to solve diffi cult problems. Furthermore, because the pace of the work is not dictated by a moving “line,” incentive arrangements may be set up. Last, because no line exists that must forever keep moving, the entire set of organizational operations

Figure 2.4 A generalized job shop operation.

Dept. A Dept. B

Dept. E

Dept. G

Dept. C

Shipping

Dept. D Dept. F Receiving

Product B5 Product A63

Product B5

Product A63

c02.indd 60c02.indd 60 25/09/12 4:20 PM25/09/12 4:20 PM

61F o r m s o f T r a n s f o r m a t i o n S y s t e m s

does not halt whenever any one part of the operation stops working; other func- tional areas can continue operating, at least until in-process inventory is depleted. Also, other general-purpose resources can usually substitute for the nonfunction- ing resource: one machine for another, one staff member for another, one material for another.

Disadvantages of the Job Shop

The general-purpose equipment of job shops is usually slower than special-purpose equipment, resulting in higher variable (per-unit) costs. In addition, the cost of direct labor for the experienced staff necessary to operate general-purpose equip- ment further increases unit costs of production above what semiskilled or unskilled workers would require. The result, in terms of costs of the outputs, is that the vari- able costs of production are higher for the general-purpose than for the special- purpose equipment, facilities, and staff, but the initial cost of the equipment and facilities is signifi cantly less. For small-output volumes, the job shop results in a lower total cost. As volume of output increases, however, the high variable costs begin to outweigh the savings in initial investment. The result is that, for high- production volumes, the job shop is not the most economic approach (although its use may still be dictated by other considerations, as when particular equipment threatens workers ’ health or safety).

Inventories are also frequently a disadvantage in the job shop, especially in prod- uct organizations. Not only do many types of raw materials, parts, and supplies have to be kept for the wide variety of outputs anticipated, but in-process inventories , that is, jobs waiting for processing, typically become very large and thereby represent a sizable capital investment for the organization. It is not unusual for batches of parts in these environments to spend 90 to 95 percent of the time they are in the shop either waiting to be moved or waiting to be processed. Furthermore, because there are so many inventory items that must travel between operating departments in order to be processed, the cost of handling materials is also typically high. Because job routings between operations are not identical, inexpensive fi xed materials- handling mechanisms like conveyor belts cannot be used. Instead, larger and more costly equipment is used; therefore, corridors and aisles must be large enough to accommodate it. This necessitates allocating even more space, beyond the extra space needed to store additional inventories.

Finally, managerial control of the job shop is extremely diffi cult, as mentioned earlier. Because the output varies in terms of function, processing, quality, and tim- ing, the managerial tasks of routing, scheduling, cost accounting, and such become nearly impossible when demand for the output is high. Expediters must track down lost jobs and reorder priorities. In addition to watching the progress of individual jobs, management must continually strive to achieve the proper balance of materials, staff, and equipment; otherwise, highly expensive resources will sit idle while bot- tlenecks occur elsewhere.

Layout of the Job Shop

Because of its relative permanence, the layout of the operations is probably one of the most crucial elements affecting the effi ciency of a job shop. In general, the prob- lem of laying out operations in a job shop is quite complex. The diffi culty stems

c02.indd 61c02.indd 61 25/09/12 4:20 PM25/09/12 4:20 PM

62 C h a p t e r 2 : P r o c e s s P l a n n i n g a n d D e s i g n

from the variety of outputs and the constant changes in outputs that are characteris- tic of organizations with an intermittent transformation system. The optimal layout for the existing set of outputs may be relatively ineffi cient for the outputs to be pro- duced six months from now. This is particularly true of job shops where there is no proprietary product and only for-contract work is performed. One week such a shop might produce 1000 wheels, and the next week it might produce an 8000-gallon vat. Therefore, a job shop layout is based on the historically stable output pattern of the organization and expected changes in that pattern, rather than on current operations or outputs.

A variety of factors can be important in the interrelations among the operations of a job shop. If all the qualitative and quantitative factors can be analyzed and com- bined, the relative importance of locating each department close to or far from each of the other departments may be used to determine a layout. This approach is par- ticularly useful for service operations where movements of materials are not particu- larly signifi cant. To illustrate how this concept might be achieved in practice, we next present a simplifi ed example. Following this, we illustrate how a purely cost- based layout could be achieved.

Directly Specifi ed Closeness Preferences

As a simplifi ed example, consider Table 2.3 , where six departments have been ana- lyzed for the desirability of closeness to each other. Assume we are given the organization ’s closeness preferences, indicated by the letters A, E, I, O, U, and X, with the meanings given in the table. In general, the desirability of closeness decreases along the alphabet until U, which is “unimportant,” and then jumps to “undesirable” with X; there is no range of undesirability in this case, although there could be, of course.

One way of starting the layout process is simply to draw boxes representing the departments in the order given in the table and show closeness preferences on the arcs (line segments) joining them. Figure 2.5 a illustrates this for Table 2.3 . The next step is to shift the departments with A on their arcs nearer each other and those with X away from each other. When these have been shifted as much as possible, the E arcs, then the I arcs, and fi nally the O arcs will be considered for relocation, resulting in an improved layout, such as in Figure 2.5 b .

Department

Department 1 2 3 4 5 6

1 E A U U U

2 U I I U

3 U U A

4 I U

5 I

6

T A B L E 2 .3 • D i r ec t l y Spec i f i ed C lo senes s Pre fe rences*

* Note :

A 5 Absolutely O 5 Ordinary necessary closeness OK E 5 Especially U 5 Unimportant important X 5 Undesirable I 5 Important

c02.indd 62c02.indd 62 25/09/12 4:20 PM25/09/12 4:20 PM

63F o r m s o f T r a n s f o r m a t i o n S y s t e m s

Cost–Volume–Distance Model

In the cost–volume–distance (CVD) approach, the desirability of closeness is based on the total cost of moving materials or people between departments. Clearly, a layout can never be completely reduced to just one such objective, but where the cost of movement is signifi cant, this approach produces reasonable fi rst approximations. The objective is to minimize the costs of interrelations among operations by locating those operations that interrelate extensively close to one another. If we label one of the departments i and another department j , then the cost of moving materials between departments i and j depends on the distance between i and j , D

ij . In addi-

tion, the cost will usually depend on the amount or volume moving from i to j , such as trips, cases, volume, weight, or some other such measure, which we will denote by V

i j . Then, if the cost of the fl ow from i to j per-unit amount per-unit distance is C

i j , the

total cost of i relating with j is C i j V

i j D

ij . Note that C, V , and D may have different values

for different types of fl ows and that they need not have the same values from j to i as from i to j , since the fl ow in opposite directions may be of an entirely different nature. For example, information may be fl owing from i to j , following a certain paperwork path; but sheet steel may fl ow from j to i, following a lift truck or conveyor belt path.

Adding the fl ows from i to every one of N possible departments, we fi nd that the total cost of department i interrelating with all other departments is

∑ 

N

j   5 1  C

i j V

i j D

i j

(It is normally assumed that C ii V

ii D

ii 5 0, because the distance from i to itself is

zero.) Adding together the costs for all the departments results in the total cost.

T C   5   ∑ N  

i 5 1  ∑

N  

j 5 1 C i j V i j D i j

Our goal is to fi nd the layout that minimizes this total cost. This may be done by evaluating the cost of promising layouts or, as in the following simplifi ed example, by evaluating all possible layouts.

1 A

3 5

E

4

(a)

2

I I I

I A 6

4 I I

5 6

I

1

(b)

2

I A

E A 3

Figure 2.5 Closeness preferences layout: ( a ) Initial layout. ( b ) Final layout.

c02.indd 63c02.indd 63 25/09/12 4:20 PM25/09/12 4:20 PM

64 C h a p t e r 2 : P r o c e s s P l a n n i n g a n d D e s i g n

The section of a business school containing the administrative offi ces of the oper- ations management department is illustrated in Figure 2.6 . Each offi ce is approxi- mately 10 feet by 10 feet, so the walking distance ( D ) between adjacent offi ces (offi ces 1 and 2, and offi ces 2 and 3) is 10 feet, whereas the distance between diago- nal offi ces (offi ces 1 and 3) is approximately 15 feet.

T A B L E 2 .4 • Load Mat r ix , V ij ( t r ip s )

To

From 1

Assistant 2

Chair 3

Secretary

1 Assistant — 5 17

2 Chair 10 — 5

3 Secretary 13 25 —

The average number of interpersonal trips made each day is given in a travel or load matrix (Table 2.4 ). According to Table 2.4 , each day the assistant makes fi ve trips to the chairperson ’s offi ce and 17 trips to the secretary ’s offi ce. Thus, the assist- ant would travel 305 feet (10 feet 3 5 trips 1 15 feet 3 17 trips) each day.

Figure 2.6 Offi ce layout.

1 Junior

administrative assistant

2 Chairperson

Restrooms 3

Secretary

Assuming that the chairperson is paid approximately twice as much as the secretary and the junior administrative assistant, determine if the current arrangement is best (i.e., least costly) in terms of transit time and, if not, what arrangement would be better.

For convenience, the offi ces are numbered in Figure 2.6 . Before calculating total costs of all possible arrangements, some preliminary analysis is worthwhile. First, because of special utility connections, restrooms are usually not considered relocatable.

c02.indd 64c02.indd 64 25/09/12 4:20 PM25/09/12 4:20 PM

65F o r m s o f T r a n s f o r m a t i o n S y s t e m s

In addition, the relocation of the restrooms in this example would not achieve any result that could not be achieved by moving the other offi ces instead.

Second, many arrangements are mirror images of other arrangements and thus need not be evaluated, since their cost will be the same. For example, interchanging offi ces 1 and 3 will result in the same costs as the current layout. The essence of the problem, then, is to determine which offi ce should be located diagonally across from the restrooms . There are three alternatives: chairperson, assistant, or secretary.

Now, let us evaluate each of the three possibilities as the “diagonal offi ce”—fi rst the chairperson, then the assistant, and last the secretary. The costs will simply be denoted as 1 for the assistant and the secretary or 2 for the chairperson (who earns twice as much as the others). As noted, the V

i j “volumes” will be the number of trips from i to j

taken from the load matrix, and the distances will depend on who has the diagonal offi ce across from the restrooms. The calculations for each arrangement are shown here:

1. Chairperson: TC 5 1(5)10 1 1(17)15 1 2(10)10 1 2(5)10 1 1(13)15 1 1(25)10 5 1050

2. Assistant: TC 5 1(5)10 1 1(17)10 1 2(10)10 1 2(5)15 1 1(13)10 1 1(25)15 5 1075

3. Secretary: TC 5 1(5)15 1 1(17)10 1 2(10)15 1 2(5)10 1 1(13)10 1 1(25)10 5 1025 (lowest)

To better understand these calculations, consider the current arrangement in which the chair has the offi ce diagonal to the restrooms. In this case, the assistant must travel 305 feet each day, as was explained earlier. Each day the chairperson would have to travel 150 feet (10 feet 3 10 trips to the assistant) 1 (10 feet 3 5 trips to the secretary). Finally, the secretary would have to travel 445 feet each day (15 feet 3 13 trips to the assistant) 1 (10 feet 3 25 trips to the chair). Because the chairperson is paid twice as much as the secretary and assistant, we weight the chairperson ’s travel distance as twice that of the other two workers. Using this weighting scheme provides a total cost of the current offi ce arrangement of 1050: that is, 305 1 (2 3 150) 1 445. The best arrange- ment is to put the secretary in the offi ce diagonal to the restrooms for a relative cost of 1025. Again, if faced with an actual layout task, a computer package could be used.

DILBERT: © Scott Adams/Dist. by United Feature Syndicate, Inc.

c02.indd 65c02.indd 65 25/09/12 4:20 PM25/09/12 4:20 PM

66 C h a p t e r 2 : P r o c e s s P l a n n i n g a n d D e s i g n

Cellular Production Cellular production combines the advantages of the job shop and fl ow shop to obtain the high variety possible with the job form and the reduced costs and short response times available with the fl ow form. Figure 2.7 contrasts the job shop with cellular production for a manufacturing fi rm. The job shop in Figure 2.7 a has sepa- rate departments for welding, turning, heat treat, milling, and forming. This type of layout provides suffi cient fl exibility to produce a wide range of products simply by varying the sequence in which the products visit the fi ve processing departments. Also, fl exibility is enhanced, as machines are easily substituted for one another should a specifi ed machine be busy or nonoperational.

#1

#2

#3

Cell

Cell

Cell

F W

F

F

T

Job

(a)

(b)

Cell

Raw materials

Welding

Turning

FormingMilling

Raw materials HT

Heat treat

F F F

W

M

M

HT

T

Figure 2.7 Conversion of ( a ) a job shop layout into ( b ) a cellular layout for part families.

Figure 2.7 b shows a reorganization of the plant for cellular production. The cel- lular form is based on group technology , which seeks to achieve effi ciency by exploiting similarities inherent in parts. In production, this is accomplished by iden- tifying groups of parts that have similar processing requirements. Parts with similar processing requirements are called part families . Figure 2.8 provides an example of how a variety of parts can be organized into part families.

After the parts are divided into families, a cell is created that includes the human skills and all the equipment required to produce a family. Since the outputs are all similar, the equipment can be set up in one pattern to produce the entire family

c02.indd 66c02.indd 66 25/09/12 4:20 PM25/09/12 4:20 PM

67F o r m s o f T r a n s f o r m a t i o n S y s t e m s

and does not need to be set up again for another type of output (as is necessary in a job shop). Some cells consist of just one machine producing a complete prod- uct or service. Other cells may have as many as 50 people working with dozens of machines.

A facility using cells is generally organized on the basis of teams . That is, a team is completely responsible for conducting the work within its cell. The team members usually schedule and inspect the work themselves, once they know when it is due. Occasionally, work must be taken outside a cell for a special treatment or process that is unavailable within the cell, but these outside transfers are minimized as much as possible.

The families are derived from one of a number of different approaches. Sometimes the basis is the machines that are needed to produce the output, or the families may be based on the size of the equipment, the quality required, the skills needed, or any other overriding consideration. This is called the classifi cation stage. Items are clas- sifi ed into families—sometimes by simple inspection and other times by complex analysis of their routing requirements, production requirements, part geometry, and the like. It is generally not feasible to classify all the outputs into one of a limited number of families, so at some point all the miscellaneous outputs are placed in a “remainder” cell, which is operated as a minijob shop.

Unorganized parts

Parts organized by families

Formed partsGeometric partsTurned parts

Figure 2.8 Organization of miscellaneous parts into families.

c02.indd 67c02.indd 67 25/09/12 4:20 PM25/09/12 4:20 PM

68 C h a p t e r 2 : P r o c e s s P l a n n i n g a n d D e s i g n

Advantages of Cellular Production

Organizations adopt the cellular form to achieve many of the effi ciencies associated with products and services that are mass-produced using fl ow transformation sys- tems in less repetitive job shop environments. However, not all the advantages of a full fl ow shop or a full job shop can be obtained because not enough high-volume equipment can be purchased to obtain the economies of scale that fl ow shops enjoy. And because the equipment is dedicated to part families, some of the variety afforded by job shops is lost.

One of the most important advantages of the cellular form is reduced machine setup times. In the job shop, when a worker completes the processing of one batch, the machine is set up for the next batch. Because a wide variety of parts typically fl ow through each department in a job shop, the next batch of parts processed by the worker will likely be different from the one just completed. This means that the worker may have to spend several hours or more simply setting up and preparing the machine for the next batch of parts. In cellular production, machine setup times are minimized because each cell processes only parts that have similar (or identi cal) setup and processing requirements. It is extremely desirable to minimize machine setup times because setup time takes away from the amount of time machines can be used to produce the outputs.

Decreasing machine setup times provides several benefi ts. First, as we have just noted, when setup times decrease, the amount of time equipment is available to process parts increases. Second, increased capacity means that the company can produce at a given level with fewer machines. Reducing the number of machines used not only reduces the costs of equipment and maintenance but also reduces the amount of fl oor space needed. Third, shorter setup times make it more economical to produce smaller batches. For instance, if the setup time is four hours, it would not be effi cient to produce a small number of parts using a particular machine only to spend another four hours to set it up for the next batch. However, if the machine required only a few minutes of setup time, it might be practical to produce a few parts on the machine.

There are numerous benefi ts associated with producing parts in small batches. To begin with, producing small batches enhances an organization ’s fl exibility in respond- ing to changes in product mix. Also, reducing the size of batches leads to reductions in work-in-process inventory. Less inventory means that less space is needed to store it and less capital is tied up in it. Also, product lead times are shorter, and throughput times are faster due to overlapping the tasks. Shorter lead times and faster throughput facilitate more accurate forecasting, faster response to market changes, faster revenue generation, and perhaps the most important advantage of all— less time for engineers to change the output or customers to change (or cancel) the order!

Another major advantage of the cellular form is that parts are produced in one cell. Processing the parts in one cell simplifi es control of the shop fl oor. To illustrate this, compare the amount of effort required to coordinate the production activities in the job shop and the cellular layout shown in Figure 2.7 . Producing parts in a single cell also reduces the amount of labor and equipment needed to move materials because travel distances between successive operations are shorter. Additionally, producing the parts in one cell provides an opportunity to increase the workers ’ accountability, responsibility, and autonomy. Finally, reducing materials handling and increasing the workers ’ accountability typically translate into reduced defects. In

c02.indd 68c02.indd 68 25/09/12 4:20 PM25/09/12 4:20 PM

69F o r m s o f T r a n s f o r m a t i o n S y s t e m s

a job shop, it is diffi cult to hold the workers accountable for quality because the product is processed in several different departments and the workers in one depart- ment can always blame problems on another department.

A unique advantage of the cell form is that it maximizes the inherent benefi ts of the team approach. In a fl ow shop, there is little teamwork because the equipment does most of the work; the labor primarily involves oversight and maintenance. Job shops are organized by department, and this allows for some teamwork—but not in terms of specifi c jobs because everyone is working on a different job. In a cell, all the workers are totally responsible for completing every job. Thus, the effect is to enrich the work, provide challenges, encourage communication and teamwork, meet due dates, and maintain quality.

An additional advantage for manufacturers is the minimal cost required to move to cellular production. Although some cells may be highly automated, with expen- sive special-purpose equipment, it is not necessary to invest any additional capital in order to adopt the cellular form. It requires only the movement of equipment and labor into cells. Or, with even less trouble, though with some loss of effi ciency, the fi rm can simply designate certain pieces of equipment as dedicated to a single part family but not relocate them. The term used in this case is virtual cell or logical cell (also known as nominal ) because the equipment is not physically adjoining but is still reserved for production of only one part family.

Another form of cellular production is called a miniplant. Here, the cell not only does the manufacturing but also has its own industrial engineer, quality manager, accountant, marketing representative, and salesperson as well as almost all the other support services that a regular plant has. Only far-removed services, such as R&D and human resources, are not dedicated to the miniplant. The entire facility of the fi rm is thus broken down into a number of miniplants, each with its own general manager, production workers, and support services so that it can operate as an inde- pendent profi t center.

Disadvantages of Cellular Production

Some disadvantages of the cellular form are those of the fl ow shop and the job shop, but they are not as serious. As in a fl ow shop, if a piece of equipment should break down, it can stop production in the cell; but in a cell form—unlike a fl ow shop, where that might be the only piece of equipment in the facility—work might, if per- missible, temporarily be shifted to other cells to get a job out.

However, obtaining balance among the cells when demands for a product or service family keep changing is a lesser problem in both fl ow and job shops. Flow shops are relatively fi xed in capacity and produce a standard output, so there is no question of balance. Job shops simply draw from a pool of skilled labor for whatever job comes in. With cells, by contrast, if demand for a family dries up, it may be nec- essary to break up that cell and redistribute the equipment or re-form the families. In the short run, though, labor can generally be assigned to whatever cell needs it, including the remainder cell.

Of course, volumes are too small in cellular production to allow the purchase of the high-volume, effi cient equipment that fl ow shops use. The cellular form also does not allow for the extent of customization usually found in job shops, since the labor pool has largely been disbursed to independent cells (although the remainder cell may be able to do the work). Moreover, the fostering of specialized knowledge

c02.indd 69c02.indd 69 25/09/12 4:20 PM25/09/12 4:20 PM

70 C h a p t e r 2 : P r o c e s s P l a n n i n g a n d D e s i g n

associated with various operational activities is reduced because the workers who perform these activities are spread out and therefore have limited opportunities to collaborate.

Cellular Layout

Cellular production creates teams of workers and equipment to produce families of outputs. The workers are cross-trained so that they can operate any of the equip- ment in their cell, and they take full responsibility for the proper performance or result of the outputs. Whenever feasible, these outputs are fi nal products or services. At other times, particularly in manufacturing, the outputs are parts that go into a fi nal product. If the latter is the case, it is common to group the cells closely around the main production or assembly line so that they feed their output directly into the line as it is needed.

In some cases, a virtual cell is formed by identifying certain equipment and dedicat- ing it to the production of families of outputs, but without moving the equipment into an actual, physical cell. In that case, no “layout” analysis is required at all; the organiza- tion simply keeps the layout it had. The essence of the problem, then, is the identifi - cation of the output families and the equipment to dedicate to each of them.

It is more common for an organization to actually form physical cells. When physical cells are created, the layout of the cell may resemble a sort of minifl ow shop, a job shop, or a mix of these, depending on the situation. Thus, we will direct our attention here to the formation of the part or product families and their associ- ated equipment, leaving the issues of physical layout to be addressed in the discus- sions of the fl ow shop and job shop.

In practice, organizations often use the term cell to include a wide range of very different situations: a functional department consisting of identical machines, a sin- gle machine that automatically performs a variety of operations, or even a dedicated assembly line. Earlier, we also referred to the portion of a shop that is not associated with a specifi c part family as a cell: a remainder cell . Nevertheless, we do not con- sider all these groups to be part of what we are calling cellular production.

Organizations that formally plan their shop layouts typically choose to group their equipment on the basis of either the function it performs (i.e., job shops) or the processing requirements of a product or group of products (i.e., fl ow shops). As we discussed, the purpose of grouping equipment on the basis of its function is to maxi- mize fl exibility, whereas the purpose of grouping it on the basis of processing requirements is to maximize effi ciency.

Companies that adopt cellular manufacturing typically create a pilot cell initially to experiment with the cellular approach, and therefore most of the equipment in the shop remains in functional departments at this stage. As these fi rms gain experience with the cell and become convinced that it is benefi cial, they begin a phase of imple- menting additional cells. This can be referred to as the hybrid stage because as the shop is incrementally converted to cells, a signifi cant portion of the facilities are still arranged in functional departments. At some point, the formation of additional cells is terminated and the fi rm may or may not have the majority of its equipment arranged in cells. Often companies stop creating new cells when the volume of the remaining parts is insuffi cient to justify forming additional cells. To clarify the con- cept of a cellular layout based on product families and machine cells, we present a

c02.indd 70c02.indd 70 25/09/12 4:20 PM25/09/12 4:20 PM

71F o r m s o f T r a n s f o r m a t i o n S y s t e m s

detailed example based on one of the more common approaches to cell formation in the next subsection.

Methods of Cell Formation

There are a variety of ways to determine what outputs should constitute a family and be produced in the same cell. Sometimes a family is dictated by the size or weight of the output; for example, huge pieces of steel may require an overhead crane to lift them onto the machines for processing. Sometimes electronic parts have special requirements for quality, such as being produced in a “clean room” or being welded in an inert gas environment. Sometimes it is obvious what family a part belongs in simply by looking at it and seeing how it was made (i.e., by what machines).

Most commonly, some form of manual determination based on human judgment is used. One relatively simple approach involves taking photographs of a sample of the parts and then manually sorting these photographs into families based on the geometry, size, or other visual characteristics of the parts. Another approach is to sort the parts based on the drawing name.

A more sophisticated manual procedure is called production fl ow analysis (PFA). In this approach, families are determined by evaluating the resource requirements for producing the outputs. Outputs that have the same complete set of resource needs are grouped into a single family. It should then be possible to cluster a set of the necessary resources together in a cell to produce that family. However, this is not always the case because there may not be enough of all resources to place each one in each of the cells that needs it, or low levels of usage may not justify placing each resource in each cell. In these cases the resources can be shared between cells or additional resources acquired. For example, maternity wings at many hospitals are set up as cells having their own dedicated doctors, nurses, and even operating rooms. However, typically the amount of time that anesthesiologists are needed in the maternity wing does not justify dedicating anesthesiologists to the unit. Thus, the anesthesiologists split their time supporting several hospital units. At other times, even if there are suffi cient quantities of the resources to assign each to the appropri- ate cells, two or three such resources may be needed in one cell to handle its capac- ity requirements while half a resource or less is needed in another cell. These diffi culties are handled case by case.

The essence of PFA is to determine the resource–output matrix and then identify the outputs (parts or services) with common resource requirements. In manufactur- ing operations, the matrix is based on information contained in the part routings and is formed by listing all the outputs (parts, patients, services) across the top and all the resources (workers, machines, nurses) down the side. Then 1 ’s are written in the matrix wherever, say, a part uses a machine. For example, Table 2.5 shows a matrix with seven parts that together require six machines. The objective is to reorder the parts and machines so that “blocks” of 1 ’s that identify the cells are formed along the diagonal, as shown in Table 2.6 . Similar resource–output matrices could be developed for service organizations. For example, a hospital might identify type of treatment as the output (e.g., maternity, cardiac, oncology) and the resources as the equipment required (X-ray, respirator, defi brillator, heart monitor). Once the treatment–equipment matrix was developed, it could be reordered to identify the resources needed to set up dedicated treatment cells.

c02.indd 71c02.indd 71 25/09/12 4:20 PM25/09/12 4:20 PM

72 C h a p t e r 2 : P r o c e s s P l a n n i n g a n d D e s i g n

Note that it is acceptable for an output not to use every resource in a cell and for a resource not to process every output. However, no output should interact with a resource outside of its cell. Thus, in Table 2.6 , part 1 is listed as needing machine 3, but this is problematic. In this case, if we could duplicate machine 3, we could put it in both cell 1 and cell 2. Or we might consider putting machine 3 in cell 2 and send- ing component 1 to cell 2 after it is fi nished in cell 1 (but this violates our desire to produce cell-complete parts). Or we could remove part 1 from the families and put it in a remainder cell (if there are other components and machines not listed in Table 2.5 within the facility).

The general guidelines for reordering the matrix by PFA are as follows:

• Incompatible resources should be in separate cells. • Each output should be produced in only one cell. • Any investment in duplicate resources should be minimized. • The cells should be limited to a reasonable size. Another, less common method of cell formation is classifi cation and coding . With

classifi cation and coding, an alphanumeric code is assigned to each part on the basis of design characteristics, processing requirements, or both. Parts with similar codes can be identifi ed and grouped into families.

T A B L E 2 .5 • Or ig ina l Mach ine–Par t Mat r ix Parts

Machines 1 2 3 4 5 6 7

1 1 1

2 1 1 1 1 1 1

3 1 1 1 1

4 1

5

6 1 1

T A B L E 2 .6 • Reordered Mat r ix Parts

Machines 7 4 1 3 6 2 5

6 1 1

2 1 1 1 Cell 2

3 Cell 1 1 1 1

5 1 1

1 1 1

4 Cell 3

1

c02.indd 72c02.indd 72 25/09/12 4:20 PM25/09/12 4:20 PM

73S e l e c t i o n o f a T r a n s f o r m a t i o n S y s t e m

Project Operations Project operations are of large scale and fi nite duration; also, they are nonrepetitive, consisting of multiple, and often simultaneous, tasks that are highly interdependent. However, the primary characteristics of the tasks are their limited duration and, if the output is a physical product, their immobility during processing. Generally, staff, mate- rials, and equipment are brought to the output and located in a nearby staging area until needed. Projects have particularly limited lives. Resources are brought together for the duration of the project; some are consumed, and others, such as equipment and personnel, are deployed to other uses at the conclusion of the project. Typically, the output is unique (a dam, product development, a presidential campaign, a trial).

In designing a processing system, a number of considerations may indicate that the project form is appropriate. One of these is the rate of change in the organiza- tion ’s outputs. If one department must keep current on a number of markets that are rapidly changing, the typical organization would quickly fall behind its competition. The project form offers extremely short reaction times to environmental or internal changes and would thus be called for. In addition, if the tasks are for a limited dura- tion only, the project form is indicated. Finally, the project form is chosen when the output is of a very large scale with multiple, interdependent activities requiring close coordination. During the project, coordination is achieved through frequent meet- ings of the representatives of the various functional areas on the project team.

One of the advantages of the project form, as noted earlier, is its ability to perform under time and cost constraints. Therefore, if meeting a due date or staying within budget is crucial, the project form is most appropriate. However, a disadvantage of the project form, with its mixed personnel from different functional areas (an engi- neer, a scientist, a businessperson, a technician, etc.), is that it has less depth in any one technical area compared to functional organization by technical specialty. In that case, a number of specialists can be brought together to solve a problem. In addition, specialized resources (e.g., technical equipment) often cannot be justifi ed for a project because of its low utilization; hence, generalized resources must be used instead. The project form of transformation processes is discussed in more detail in Chapter 6.

S E L E C T I O N O F A T R A N S F O R M A T I O N S Y S T E M This section addresses the issue of selecting the appropriate transformation system, or mix of systems, to produce an output. From the preceding discussion, it should be clear that the fi ve transformation systems are somewhat simplifi ed extremes of what is likely to be observed in practice. Few organizations use one of the fi ve forms in a pure sense; most combine two or more forms in what we call a hybrid shop. For example, in manufacturing computer keyboards, some parts and subassemblies are produced in job shops or cells but then fed into a fl ow shop at the fi nal assembly line, where a batch of one model is produced. Then the line is modifi ed to produce a batch of another model. Even in “custom” work, jobs are often handled in groups of generally common items throughout most of their processing, leaving minor fi n- ishing details such as the fabric on a couch or the facade of a house to give the impression of customizing.

c02.indd 73c02.indd 73 25/09/12 4:20 PM25/09/12 4:20 PM

74 C h a p t e r 2 : P r o c e s s P l a n n i n g a n d D e s i g n

Although services typically take the form of a job shop, the emphasis has recently been on trying to mass-produce them (using cells or fl ow shops) so as to increase volume and reduce unit costs. Some examples are fast-food outlets, multiphasic medical screening, and group life insurance. Even with services, we often fi nd com- bined forms of process design: McDonald ’s prepares batches of Big Macs but will accept individual custom orders. Burger King uses a conveyor assembly line for its Whoppers but advertises its ability to customize its burgers to suit any taste.

The problem for the operations manager is to decide what processing form(s) is most appropriate for the organization, considering long-run effi ciency, effectiveness, lead time, capacity, quality, and fl exibility. Selection may be even more diffi cult because, as mentioned previously, it is possible to combine processing forms to attain effi ciency in some portions of the production process and fl exibility in other portions. It is clear that the tradeoffs must be well understood by the manager, and the expected benefi ts and costs must be well known.

Unfortunately, most plants do not have the luxury of time for completely reorgan- izing their processes. As a result, they often grow into a hodgepodge of machines and processes scattered somewhat randomly around the plant, barely resembling any of the above fi ve forms, even if they started out with one of them.

Considerations of Volume and Variety One of the most important factors in the design of a transformation system is estab- lishing the volume and variety of outputs the organization will produce. High vol- umes tend to indicate that highly automated mass production will be necessary. High variety, on the other hand, implies the use of skilled labor and general-purpose tools and facilities.

A related consideration here is whether the output will be make-to-stock or make- to-order. A make-to-stock item is produced in batches of some size that is economical (for the fi rm) and then stocked (in a warehouse, on shelves, etc.). As customers pur- chase them, the items are withdrawn from stock. A make-to-order item is usually produced in a batch of a size set by the customer (sometimes just one) and is deliv- ered to the customer upon its completion. Generally, make-to-stock items are pro- duced in large volumes with low variety, whereas make-to-order items are produced in low volumes with high variety. (Quite often, every item is different.) However, there are many variants of these two major categories, such as engineer-to-order , where a fi rm takes a customer ’s specifi cations for a product and actually designs the product and then makes it. (In the typical make-to-order process, the product is one that the fi rm already knows how to make and may even be listed in its sales catalog.) Another variant is called assemble-to-order , where the components of a set of products are stocked ahead of time (made-to-stock); when an order for one of the items in that set arrives, they are then quickly assembled to produce the fi nal desired product.

Organizations in the same industry may deliver their outputs using alternative approaches. For example, a bakery that designs and bakes custom wedding cakes exemplifi es engineer-to-order. A fi ve-star restaurant that prepares your steak to your specifi c doneness specifi cation exemplifi es make-to-order. A pizza restaurant that adds the toppings you request to your pizza illustrates an assemble-to-order opera- tion. Finally, a fast-food restaurant that prepares hamburgers ahead of time is employing a make-to-stock operation.

c02.indd 74c02.indd 74 25/09/12 4:20 PM25/09/12 4:20 PM

75S e l e c t i o n o f a T r a n s f o r m a t i o n S y s t e m

Clearly, many services will not normally be of a type that can be stocked, even if every service is identical (e.g., a physical examination). Also, exceptions to these generalizations are abundant. Automobiles, for example, are made to order but are produced in high volume and with high variety. (However, autos are really assem- bled to order, with the assembly components produced to stock and already available .) And general-purpose machine shops often produce high volumes of low- variety items for specifi c customers.

Figure 2.9 a , based on the product–process matrix developed by Hayes and Wheelwright (1979), illustrates these points as they relate to the various transforma- tion systems. The horizontal axis shows volume, as measured by the batch size, and the left vertical axis shows the variety of outputs. Organizations making a single unit of output that varies each time (such as dams and custom-built machines) use the project form or sometimes the job shop. Some services also fall into this region, as indicated by the upper left tip of the oval. Job shop and cellular systems, however, are mainly used when a considerable variety of outputs are required in relatively small batches. This is particularly characteristic of services. When the size of a batch increases signifi cantly, with a corresponding decrease in variety, then a fl ow shop is appropriate. Some services also fall into this category. Last, when all the output is the same and the batch is extremely large (or essentially infi nite, as in the ore, petro- chemical, and food and drink industries), the continuous process is appropriate. Very few services exist here.

Project

Job

Cell

Flow

None

None

One Few Many

Batch size 0� 50 100

% make-to-order (a) (b)

90% make-to-stock

None

Low

Much

High 0 50 100

O u

tp u

t va

ri et

y None

Low

Much

High

O u

tp u

t va

ri et

y

90% make-to-order

Continuous process

Figure 2.9 Effect of output characteristics on transformation systems—the product–process matrix.

Note that the standard, viable transformation forms lie on the diagonal of the product–process matrix. Operating at some point off this diagonal can be dangerous for the organization unless done carefully as a well-planned strategy. Essentially, no organizations operate in the upper right or lower left segments of this grid. The lower left does, however, represent manufacturing 200 years ago. If you wanted four identical dressers, say, for your four children, they were made one at a time by hand

c02.indd 75c02.indd 75 25/09/12 4:20 PM25/09/12 4:20 PM

76 C h a p t e r 2 : P r o c e s s P l a n n i n g a n d D e s i g n

(whether identical or all different, for that matter). Today, however, it is simply too expensive to produce items this way; if the items are all identical, they are made in a large batch and then sold separately. In some cases, it is almost impossible to buy a single unit of some items, such as common ten-penny nails—you have to buy a blister-pack of ten or so. Similarly, the upper right may represent manufacturing in the future, when advanced technology can turn out great masses of completely cus- tomized products as cheaply as standard items. Currently, however, we have trouble doing this (in spite of such popular concepts as “mass customization”). Some prod- ucts, however, lend themselves to approximations to this goal through such special- ized techniques as assembly-to-order: fast food, Internet-purchased computers, and so on. These fi rms have developed strategies, and the production techniques to accompany them, for successfully using “off-diagonal” transformation processes.

Note the overlap in the different forms . This means, for example, that on occasion some organizations will use a fl ow shop for outputs with smaller batches or larger variety, or both, than the outputs of organizations using a job shop. There are many possible reasons for this, including economic and historical factors. The organization may also simply be using an inappropriate transformation system. The point is that the categories are not rigid, and many variations do occur. Many organizations also use hybrids or combinations of systems, such as producing components to stock but assembling fi nished products to order, as in the auto industry.

Note in Figure 2.9 b the general breakdown of make-to-order and make-to-stock with output variety and size of batch. Project forms (high variety, unit batch size) are almost always make-to-order, and continuous processing forms (no variety, infi nite batch size) are almost always make-to-stock, though exceptions occasionally occur.

Product and Process Life Cycle In Chapter 1 we described the life cycle of an output: how long it takes to develop, bring to market, and catch on; how quickly it grows in popularity; how different ver- sions are developed for different market segments; how the output reaches market saturation; how price competition emerges. A similar life cycle occurs in the produc- tion system for an output. As a result, a project form of transformation system may be used for the development of a new output, may evolve into a job shop or cellular layout as a market develops, and fi nally may evolve into a fl ow shop as full stand- ardization and high volumes develop. (We assume here that a continuous process is not appropriate for the output.) We briefl y elaborate on this production life cycle.

In the R&D stage, many variations are investigated during the development of a product. As the output is being developed, prototypes are made in small volumes in a relatively ineffi cient, uncoordinated manner typically in a job shop. As demand grows and competitors enter the market, price competition begins and a cellular or fl ow sys- tem, with its high volume and low variable costs, becomes preferred. At the peak of the cycle, demand may increase to the point where such a system is justifi ed.

This progress is illustrated in Figure 2.10 , which presents a breakeven analysis for each of four transformation systems. The dark bold line illustrates the lowest-cost sys- tem for each stage of the life cycle. At the stage of project development and initiation (R&D and initial production), the cost of fi xed equipment is nil, and labor is the pre- dominant contributor to high variable costs. In the expansion stage, the job shop allows some tradeoff of equipment for labor with a corresponding reduction in variable unit

c02.indd 76c02.indd 76 25/09/12 4:20 PM25/09/12 4:20 PM

77S e l e c t i o n o f a T r a n s f o r m a t i o n S y s t e m

costs, thus leading, at these volumes, to a reduction in overall unit costs. Finally, at high volumes characterizing maturity, a nearly complete replacement of expensive labor with equipment is possible, using cellular form and the fl ow shop.

Be advised, however, that not all outputs can or should follow this sequence. The point is that the transformation system should evolve as the market and output evolve. But many organizations see their strength in operating a particular transfor- mation system, such as R&D or low-cost production of large volumes. If their out- puts evolve into another stage of the life cycle in which a different transformation system form is preferable, they drop the output (or license it to someone else) and switch to another output more appropriate to their strengths.

Failing to maintain this focus in the organization ’s production system can quickly result in a “white elephant”—a facility built to be effi cient at one task but being inef- fi ciently used for something else. This can also happen if the organization, in an attempt to please every customer, mixes the production of outputs that require dif- ferent transformation systems. Japanese plants are very carefully planned to main- tain one strong focus in each plant. If an output requiring a different process is to be produced, a new plant is acquired or built.

From the previous discussion it is clear that there is a close relationship between the design of a product or service and the design of the production system. Actually, the link is even closer than it seems. Figure 2.11 illustrates the relationship between the innovations throughout the life cycle of a product or service and inno- vations throughout the life cycle of its production system. At the left, when the prod- uct or service is introduced, innovations and changes in its design are frequent. At this point, the production system is more of the project or modeling/job shop form since the design is still changing (the number of product innovations is high). Toward the middle, the product design has largely stabilized, and cost competition is forcing

Use project form

Fixed equipment investment cost

Use job form

Use cell form

Volume

Use flow form

MaturityExpansionInitiation

C o

st

Variable cost

Cell Flow Best form

Job Project

Figure 2.10 Selection of transformation systems by stage of life cycle.

c02.indd 77c02.indd 77 25/09/12 4:20 PM25/09/12 4:20 PM

78 C h a p t e r 2 : P r o c e s s P l a n n i n g a n d D e s i g n

innovations in the production process, particularly the substitution of cellular or fl ow shop machinery for labor (the number of process innovations is high). At the right, this phenomenon has subsided and innovations in production methods are primarily the result of competitors ’ actions, government regulations, and other exter- nal factors.

Although not typically involved on the research side of such innovations in pro- duction methods (a laboratory engineering function), the operations manager is intimately involved in applying these developments in day-to-day production. The possible tradeoffs in such applications are many and complex. The new production system might be more expensive but might produce a higher-quality output (and thus, the repeat volume may be higher, or perhaps the price can be increased). Or the new production system might be more expensive and might produce a lower- quality output but be simpler and easier to maintain, resulting in a lower total cost and, ultimately, higher profi ts. Clearly, many considerations—labor, maintenance, quality, materials, capital investment, and so on—are involved in the successful application of research to operations.

Service Processes As with the design of transformation systems for products, the design of transforma- tion systems for services depends heavily on knowing exactly what characteristics of the service need to be emphasized: its explicit and implicit benefi ts, cost, time duration, location, and accessibility. Knowing the importance of each of these allows the designer to make the necessary tradeoffs in costs and benefi ts to offer an effec- tive yet reasonably priced service.

Unfortunately, service transformation systems are frequently implemented with lit- tle development or pretesting, which is also a major reason why so many of them fail.

Consider the extensive development and testing of the McDonald ’s fast-food pro- duction system, of airline reservations systems, and of many life insurance policies. Each of these also illustrates the many hours of training required to use equipment

Product innovations Process innovations

Time

N o

. o

f in

n o

va ti

o n

s

Figure 2.11 Product–process innovations over time.

c02.indd 78c02.indd 78 25/09/12 4:20 PM25/09/12 4:20 PM

79S e l e c t i o n o f a T r a n s f o r m a t i o n S y s t e m

and procedures properly and effi ciently. Yet most new service fi rms frequently fail to train their personnel adequately, again inviting failure.

In most cases, the various forms and layouts of manufacturing transformation processes apply equally well to services. Flow shops are seen in fast-food restau- rants, job shops are seen in banks and hospitals, and projects are seen in individual services such as salons and house construction. Chapter 8 includes an example of a service blueprint that is commonly used for process fl ow and capacity analysis pur- poses but may also be helpful when designing the service up front.

However, one important service element that is usually missing from manufactur- ing transformation design is extensive customer contact during delivery of the serv- ice. This presents both problems and opportunities. For one thing, the customer will often add new inputs to the delivery system or make new demands on it that were not anticipated when it was designed. In addition, customers do not arrive at smooth, even increments of time but instead tend to bunch up, as during lunch periods, and then complain when they have to wait for service. Furthermore, the customers ’ biased perception of the server, and the server ’s skills, can often infl uence their sat- isfaction with the quality of the service. Obviously, this can either be benefi cial or harmful, depending on the circumstances.

On the other hand, having the customer involved in the delivery of a service can also present opportunities to improve it. Since customers know their own needs best, it is wise to let them aid in the preparation or delivery of the service—as with automatic teller machines, salad bars, and pay-at-pump gas stations. In addition to improving the quality of the service, this can save the fi rm money by making it unnecessary to hire additional servers. However, the customer can also negligently— and quickly—ruin a machine or a tool and may even sue if injured by it, so the service fi rm must carefully consider how much self-service it is willing to let the customer perform.

Chase and Tansik (1983) devised a helpful way to view this customer contact when designing service delivery systems. Their suggestion is to evaluate whether the service is, in general, high contact or low contact, and what portions of the service, in particular, are each. The value of this analysis is that the service can be made both more effi cient and more effective by separating these two portions and designing them differently. For example, the high-contact portions of the service should be handled by workers who are skilled at social interaction, whereas the low-contact portion should employ more technical workers and take advantage of labor-saving equipment. For example, a bank might have a back offi ce, where checks are encoded, located apart from the front offi ce, where customers deposit them. In this back offi ce, equipment and effi ciency are the critical job elements, whereas in the front offi ce, interpersonal skills and friendliness are critical.

Whenever possible, the low-contact portion of a service should be decoupled from the high-contact portion so that it can be conducted with effi ciency, whereas the high-contact portion is conducted with grace and friendliness. Close analysis of the service tasks may also reveal opportunities for decreasing contact with the customer—through, for example, automated teller machines, phone service, self- service, or the Web, if this is appropriate—with a concomitant opportunity for improving both the effi ciency and level of service. In particular, enabling customers to use the Web to obtain service (e.g., obtain account information, place orders) offers them convenient access—24 hours per day, 365 days per year—and immediate attention (i.e., no longer being placed on hold for the next available representative).

c02.indd 79c02.indd 79 25/09/12 4:20 PM25/09/12 4:20 PM

80 C h a p t e r 2 : P r o c e s s P l a n n i n g a n d D e s i g n

Similarly, there may be some opportunities for increasing the amount of customer contact, such as phone or mail follow-ups after service, which should be exploited to improve the overall service and its image. The service provider should thoroughly investigate these opportunities.

Service Process Design

Like the product–process matrix for manufacturing, Schmenner (1986) has developed a similar matrix for services that not only classifi es four major and quite different types of services but gives some insights on how to design the best service system. The service matrix is shown in Figure 2.12 . Service systems are divided into those with high versus low contact intensity (similar to Chase and Tansik) and according to whether they are capital intensive or labor intensive. Schmenner names each of the quadrants with an easily understood identifi er that captures the essence of that quad- rant: service factory, service shop, mass service, and professional service.

Each of the quadrants represents a unique service transformation process, with unique managerial challenges and unique characteristics. Those services at the high- contact side of the matrix have low volumes with high customization and must attain their profi tability through high prices. Those on the other side with low contact and customization attain profi tability through high volumes. The investment axis identi- fi es whether the service provider puts its resources into expensive equipment or into labor. Thus, one axis is a combination of customer variety and volume (like the product–process matrix) and the other axis is based on the inputs needed to provide the service. Examples of typical services in each quadrant are given in the fi gure.

The matrix is also useful in identifying the managerial challenges for each of the quadrants. In the low-contact left side, the managerial challenge is making the serv- ice appear warm and friendly so as to attract high volumes. If the level of contact is high, the managerial challenge is to try to be optimally effi cient in using capital and

Figure 2.12 The service matrix.

Customer contact intensity

Low High

Capital Intensive

Service factory

Airlines Package/postal services Hotels Recreation

Service shop

Hospitals Cruise Line Repair services Expensive restaurants

Labor Intensive

Mass service

Sporting events School classes Retailing Fast food

Professional service

Legal services Physicians Interior decorators Tax preparers

c02.indd 80c02.indd 80 25/09/12 4:20 PM25/09/12 4:20 PM

81S e l e c t i o n o f a T r a n s f o r m a t i o n S y s t e m

labor resources while keeping prices high. If the service is equipment intensive, the challenge is to keep capital investment costs low. If instead the service is labor inten- sive, the challenge is to minimize wages and time spent on each customer.

The matrix is also useful in redesigning a service. For example, a fi rm may decide to move from one quadrant to another to better use its resources or environment. For example, a tax preparation service may start as a high-priced professional service but then move either toward a more automated service shop through computer prepara- tion of the forms or a less personalized mass service using less skilled tax preparers.

Servicescapes

Bitner (1992) uses the term servicescape to stress the importance of the physical environment of the service and its infl uence on both the customer and the server, as well as the effi ciency and the customer ’s perception of the service. The servicescape for a self-service operation is much more important to the successful delivery of a service than when the service is assisted or full service because the customer has to take cues from the layout or signage about how to serve him- or herself, such as at an ATM or cafeteria. When there is assistance from a server, such as at an airport kiosk or sit-down restaurant, the servicescape layout can be adjusted to be more effi cient from a processing point of view. And when there is full service, especially remote service (e.g., phoned orders, online tech support), the layout of the service facility can be even more structured so as to help the servers complete their tasks (since the layout does not affect the customers).

Bitner emphasizes three elements of servicescapes: their ambient conditions; their spatial layout and functionality; and the signs, symbols, and artifacts in the physical environment. The ambient conditions refer to the sights, sounds, scents, temperature, humidity, and other such conditions that affect the service. For example, a restaurant featuring live, soft piano music communicates a different (and probably more expen- sive) sensation to a customer than one where loud pop music is competing with loud noise from the bar, and table guests. Lighting, scents, and decor can play similar roles.

The spatial layout and functionality , as noted above, play important roles in minimizing the cost of providing the service (and sometimes maximizing the reve- nue) and maximizing the satisfaction of the customer. Again, whether the facility is a self-service or full-service one is an important consideration to the layout and func- tionality, such as with fast-food versus sit-down service restaurants. In terms of rev- enue maximizing, who hasn ’t tried to rush into a fancy department store to buy a shirt only to be obstructed by a 30-foot-wide cosmetics counter and then having to navigate through an array of jewelry counters?

Last, signs, symbols, and artifacts help tell customers what to expect of the service and their role in the process. This includes wall decorations, fl oor coverings, furniture, colors, room sizes, ceiling heights, and so on. Consider the heavy furniture, wood pan- eling, thick carpeting, wall art, and handsome receptionist in an expensive law offi ce versus the trash and recycle bins, and logo-imprinted shirts and aprons in a fast- food outlet.

Service Gaps

When designing services, it can be useful to inspect the service design and delivery for potential “gaps” between what the customer/client needs and what the service

c02.indd 81c02.indd 81 25/09/12 4:20 PM25/09/12 4:20 PM

82 C h a p t e r 2 : P r o c e s s P l a n n i n g a n d D e s i g n

provider is offering (Parasuraman et al. 1988). By identifying the possible gaps in the service process, a service provider can better control the quality, productivity, cost, and performance of its service offering, thereby resulting in greater profi t and mar- ket share. A gap analysis can also help identify service industries where better serv- ice might offer a competitive advantage.

Figure 2.13 illustrates the concept. Essentially, there is commonly a gap between what the customer/client actually needs and what is delivered that involves gaps throughout the selection, design, and delivery process. We start with gap 1, the rea- sonable difference between the ideal service that the customer actually needs and what the customer expects. This gap is often infl uenced heavily by advertising and other communications from the service provider. Gap 2 is the imbalance between what the customer expected and his or her perception of what was actually received. Gap 3 is the fi nal gap on the customer ’s side; it represents the difference between what was actually delivered and the customer ’s perception of that reality.

Perceived received service

Perceived delivered service

Expected service

Experience and knowledge

Ideal service

Need

Gap 8

Gap 1

Gap 2

Gap 3

Gap 10

Gap 7

Gap 9

Gap 6

Gap 5

Gap 4

Selected service

Perceived service need

Designed service

Communicated and advertised

service

Actual delivered service

Figure 2.13 Potential locations for service gaps.

c02.indd 82c02.indd 82 25/09/12 4:20 PM25/09/12 4:20 PM

83S e l e c t i o n o f a T r a n s f o r m a t i o n S y s t e m

The seven remaining gaps are all on the provider ’s side. Gap 4 is the mispercep- tion by the service provider of what the customer truly needs. Gap 5 is the difference between that misperception and what the provider chooses to offer (the selected service). Gap 6 is the discrepancy between the service that was selected and the service that was designed. Gap 7 concerns marketing and sales and is the disparity between the designed service and what these functions understand it to be. Continuing this path, gap 8 concerns the difference between what the provider is attempting to communicate to the customer and what the customer actually under- stands. Then, returning to the service delivery process, the last two gaps concern the contrast between what was designed and what was perceived as delivered (gap 9) or what was actually delivered (gap 10).

Clearly, with nine possible opportunities for the service provider to not meet the customer ’s expectations, not to mention the customer ’s needs (gap 1), there are a lot of ways to fail in the service provision process. It behooves all service (and product) providers to carefully examine each of these potential failure points in their own business to see if they can improve their service provision process, especially before someone else discovers the opportunity and moves to close the gaps.

Service Guarantees and Fail Safi ng

Service guarantees are increasingly common among service providers who have confi dence they can meet the customer ’s service expectations and who desire a com- petitive advantage in their industry. Package transportation companies were among the fi rst to use them, and since then have been adopted by hotels, restaurants, and others in those service businesses that have extensive contact with the public but a reputation for poor service.

There are four major elements of a service guarantee:

1. It must be meaningful to the customer in the sense that it in fact repays the customer for the failure of the service to meet his or her expectations. A guarantee with a trivial payoff that does not satisfy the customer will just increase the customer ’s dissatisfaction and negate the purpose of the guaran- tee program in the fi rst place.

2. The guarantee must be unconditional. Again, if there are “exceptions” that exclude the common reasons why the service might fail, the customer will only be more dissatisfi ed.

3. The guarantee must be easy to communicate and for the customer (and employees) to understand. If the guarantee is complex or complicated to explain, it will not attract customers to the service provider. And employees who are charged with making good on the guarantee also need to fully understand it and be able to execute the guarantee provisions.

4. The guarantee must be easy to “use” in the sense of immediately invoking it when a service failure occurs. If the customer has to return home, mail in a coupon, and wait for satisfaction, the guarantee program will not achieve its purpose.

The information technology fi eld is a leader in the development and use of for- malized service guarantees, referring to them as service-level agreements (SLAs).

c02.indd 83c02.indd 83 25/09/12 4:20 PM25/09/12 4:20 PM

84 C h a p t e r 2 : P r o c e s s P l a n n i n g a n d D e s i g n

1. When a line cannot be perfectly balanced, some people will have more work time than others in each cycle. What might be a solution for this situation?

2. A current sociological trend is to move away from paced lines. Yet increasing automation is pushing workers to match their work pace to that of machines and computers. How can both of these trends be happening at the same time?

3. if a job shop was being laid out in a third-world country, how might the procedure be different? What other factors might enter in that would not

exist in an industrialized country? Might the layout also differ among industrialized countries such as Europe and Japan? How about a fl ow shop?

4. In highly automated facilities, fi rms frequently increase the job responsibilities of the skilled work- ers who remain after automation has replaced the manual laborers, although there is less potential for applying their skills. Workers complain that they are under increased pressure to perform but have less control over the automated equipment. Is this

Here, an SLA is a written contract between an information technology provider and the user of the technology specifying the level of service that will be provided, typi- cally in measurable terms. In addition to using SLAs to specify the levels of service that will be provided by external organizations, it is also becoming increasingly common for internal information system departments to develop SLAs for the other departments they support within the enterprise. Representative SLA metrics for infor- mation technology providers include the percentage of uptime, help-desk response times, and the timely reporting of usage statistics.

One approach organizations use to help guarantee their service is a concept called fail safi ng (Chase and Stewart 1994), which anticipates where a service failure might occur and installs preventive measures. The service blueprint in Chapter 8, mentioned earlier, includes potential failure points in the service process that should be considered for fail safi ng. As an example of fail safi ng, fast-food playgrounds are ubiquitous these days, but children who are too large can be a danger to smaller children (or to themselves) when they play on the equipment; hence the “maximum height” signs at the playground entrance. And outpatient health clinics give vibrating beepers to patients who sign in so they aren ’t “lost” in the system.

But it is not only the customer who needs fail safi ng; the service providers also need their systems to be designed to force them into performing the service cor- rectly. A familiar example to both service providers and customers is computer screens that disallow entries in online forms that don ’t match the protocol or that reject the forms when required fi elds are inadvertently left empty. Another is McDonald ’s now-famous french fry scoop that picks up, straightens, and optimally sizes the amount of fries for the bag, all without human contact. Yet another is dan- gerous equipment, that can ’t be activated unless the operator ’s hands and body are sensed to be out of danger.

In large part, the emergence of the concepts of service guarantees and fail safi ng refl ects the tremendous growth and availability of services in our economy and the increasingly poor record of satisfactory service being provided by so many of these new services. Examples abound: telephone answering systems with unending menus that keep customers from reaching a real person who can fi x their problem, airline cancellations/delays/lost baggage, and so on. Although technology is often helpful in solving our problems, it can also multiply them.

E X P A N D Y O U R U N D E R S T A N D I N G

c02.indd 84c02.indd 84 25/09/12 4:20 PM25/09/12 4:20 PM

85A p p l y y o u r U n d e r s t a n d i n g

A P P L Y Y O U R U N D E R S T A N D I N G Parad i se S ta t e Un ive r s i t y

Paradise State University (PSU) is a medium-sized private university offering both undergradu- ate and graduate degrees. Students typically choose Paradise State because of its emphasis on high levels of interaction and relatively small classes. University policy prohibits classes with more than 75 students (unless special permission is obtained from the provost), and the target class size is 25 students. All courses are taught by tenure-track faculty members with appropri- ate terminal degrees. Faculty members teach two courses each semester.

The Business School at PSU offers only an MBA degree in one of six areas of concentra- tion: accounting, fi nance, general management, management information systems (MIS), mar- keting, and operations management (OM). The MBA program is a one-year (two-semester) lockstep program. Since the Business School does not offer undergraduate business courses, students entering the program are required to have completed all the undergraduate business prerequisites from an accredited university. The faculty is organized into six functional departments. The table below lists the number of faculty members in each department and the average number of students each year who choose a particular concentration. Students are not permitted to have double concentrations, and PSU does not offer minors at the graduate level.

Department Faculty Number of

Students per Year

Accounting 8 100 Finance 6 40 General management 7 70 MIS 10 150 Marketing 6 50 OM 10 30

The number of courses required by each concentration in each department is listed in the table below. For example, a student concentrating in accounting is required to take four accounting classes, one fi nance class, one management class, one MIS course, one marketing class, and two OM classes.

ethical on the part of the companies involved? What approach would be better?

5. Cellular production is often conducted in a U-shaped (horseshoe-shaped) cell, rather than the rectangular cells shown in Figure 2.7 b . What might be the advantages of this U shape?

6. What benefi ts would a virtual cell obtain, and not obtain, compared with a physical cell?

7. A number of fi rms are moving toward minifactories. What advantages might this offer over straight cel- lular production?

8. If effi ciency, variety, and so on are the important measures of the low-contact or no-contact portion

of a service, what are the important measures of the high-contact portion?

9. As the process life cycle changes with the product life cycle, should a fi rm change along with it or move into new products more appropriate to its existing process? What factors must be considered in this decision?

10. In Figure 2.9( a ) , showing the fi ve transformation systems, why don ’t fi rms operate in the regions marked “none”?

11. Identify the similarities in Figures 2.9( a ) and 2.12 . Identify the differences.

c02.indd 85c02.indd 85 25/09/12 4:20 PM25/09/12 4:20 PM

86 C h a p t e r 2 : P r o c e s s P l a n n i n g a n d D e s i g n

Number of Courses Taken in Respective Departments

Concentration Accounting Finance Management MIS Marketing OM

Accounting 4 1 1 1 1 2 Finance 1 4 1 1 1 2 General management 1 1 4 1 1 2

MIS 1 1 1 4 1 2 Marketing 1 1 1 1 4 2 OM 1 1 1 1 1 5

Questions

1. How many students must each department teach each semester? Given the target class size—25 students—are there enough faculty members?

2. Conceptually, how could the cellular production approach be applied to the Business School?

3. What would be the advantages and disadvantages of adopting a cellular approach at the Business School? As a student, would you prefer a functional organization or a cellular organization? As a faculty member, what would you prefer?

4. On the basis of the information given, develop a rough plan detailing how the Business School faculty might be assigned to cells.

X -Opo ly, Inc .

X-Opoly, Inc. was founded by two fi rst-year college students to produce a knockoff real estate board game similar to the popular Parker Brothers ’ game Monopoly ® . Initially, the partners started the company just to produce a board game based on popular local landmarks in their small college town, as a way to help pay for their college expenses. However, the game was a big success and because they enjoyed running their own business, they decided to pursue the business full time after graduation.

X-Opoly has grown rapidly over the last couple of years, designing and producing custom real estate trading games for universities, municipalities, chambers of commerce, and lately even some businesses. Orders range from a couple of hundred games to an occasional order for several thousand. This year X-Opoly expects to sell 50,000 units and projects that its sales will grow 25 percent annually for the next fi ve years.

X-Opoly ’s orders are either for a new game board that has not been produced before or repeat orders for a game that was previously produced. If the order is for a new game, the client fi rst meets with a graphic designer from X-Opoly ’s art department, and then the actual game board is designed. The design of the board can take anywhere from a few hours to several weeks, depending on how much the client has thought about the game before the meeting. All design work is done on personal computers.

After the design is approved by the client, a copy of the computer fi le containing the de- sign is transferred electronically to the printing department. Workers in the printing depart- ment load the fi le onto their own personal computers and print out the board design on special decals, 19.25 inches by 19.25 inches, using high-quality color inkjet printers. The side of the decal that is printed on is usually light gray, and the other side contains an adhesive that is covered by a removable backing.

The printing department is also responsible for printing the property cards, game cards, and money. The money is printed on colored paper using standard laser printers. Ten copies of a particular denomination are printed on each 8.5-inch by 11-inch piece of paper. The money is then moved to the cutting department, where it is cut into individual bills. The prop- erty cards and game cards are produced similarly, the major difference being that they are printed on material resembling posterboard.

c02.indd 86c02.indd 86 25/09/12 4:20 PM25/09/12 4:20 PM

87A p p l y y o u r U n d e r s t a n d i n g

In addition to cutting the money, game cards, and property cards, the cutting department also cuts the cardboard that serves as the substrate for the actual game board. The game board consists of two boards created by cutting a single 19-inch by 19.25-inch piece of cardboard in half, yielding two boards each measuring 19.25 inches by 9.5 inches. After being cut, game boards, money, and cards are stored in totes in a work-in-process area and deliv- ered to the appropriate station on the assembly line as needed.

Because of its explosive growth, X-Opoly ’s assembly line was never formally planned. It simply evolved into the 19 stations shown in the following table.

Station Number Task(s) Performed at Station

Time to Perform

Task

1

2

3

4

5

6

7

8

9

10

11

Get box bottom and place plastic money tray in box bottom. Take two dice from bin and place in box bottom in area not taken up by tray.

Count out 35 plastic houses and place in box bottom.

Count out 15 plastic hotels and place in box bottom.

Take one game piece from each of eight bins and place them in box bottom.

Take one property card from each of 28 bins. Place rubber band around prop- erty cards and place cards in box bottom.

Take one orange card from each of 15 bins. Place rubber band around cards and place cards in box bottom.

Take one yellow card from each of 15 bins. Take orange cards from box and remove rubber band. Place yellow cards on top of orange cards. Place rubber band around yellow and orange cards and place cards in box bottom.

Count out 25 $500 bills and attach to cardboard strip with rubber band. Place money in box bottom.

Count out 25 $100 bills. Take $500 bills from box bottom and remove rubber band. Place $100 bills on top of $500 bills. Attach rubber band around money and place in box bottom.

Count out 25 $50 bills. Take $500 and $100 bills from box bottom and remove rubber band. Place $50 bills on top. Attach rubber band around money and place in box bottom.

Count out 50 $20 bills. Take money in box and remove rubber band. Place $20 bills on top. Attach rubber band around money and place in box bottom.

10 seconds

35 seconds

15 seconds

15 seconds

40 seconds

20 seconds

35 seconds

30 seconds

40 seconds

40 seconds

55 seconds

12

13

14

15

16

17

18

19

Count out 40 $10 bills. Take money in box and remove rubber band. Place $10 bills on top. Attach rubber band around money and place in box bottom.

Count out 40 $5 bills. Take money in box and remove rubber band. Place $5 bills on top. Attach rubber band around money and place in box bottom.

Count out 40 $1 bills. Take money in box and remove rubber band. Place $1 bills on top. Attach rubber band around money and place in box bottom.

Take money and remove rubber band. Shrink-wrap money and place back in box bottom.

Take houses, hotels, dice, and game pieces and place in bag. Seal bag and place bag in box.

Place two cardboard game board halves in fi xture so that they are separated by ¼ in. Peel backing off of printed game board decal. Align decal over board halves and lower it down. Remove board from fi xture and fl ip it over. Attach solid blue backing decal. Flip game board over again and fold blue backing over front of game board, creating a ¼-in. border. Fold game board in half and place in box covering money tray, game pieces, and cards.

Place game instructions in box. Place box top on box bottom. Shrink wrap entire box.

Place completed box in carton.

45 seconds

45 seconds

45 seconds

20 seconds

30 seconds

90 seconds

30 seconds

10 seconds

c02.indd 87c02.indd 87 25/09/12 4:20 PM25/09/12 4:20 PM

88 C h a p t e r 2 : P r o c e s s P l a n n i n g a n d D e s i g n

E X E R C I S E S

Questions

1. What kind(s) of transformation system(s) does X-Opoly use? 2. What would be involved in switching the assembly line over from the production of one

game to the production of another? 3. What is the cycle time of the 19-station line? What is its effi ciency? 4. On the basis of the task descriptions, develop a precedence graph for the assembly tasks.

(Assume that tasks performed in the 19 stations cannot be further divided.) Using these precedence relationships, develop a list of recommendations for rebalancing the line in order to improve its performance.

5. What would be the impact on the line ’s effi ciency if your recommendations were implemented?

1. Given the following machine–part matrix, form cells using PFA.

Parts

Machines 1 2 3 4 5 6 7

1 1 1

2 1 1

3 1 1

4 1 1 1

5 1 1

2. a. Given the following load matrix of a DVD player factory in Nagoya, Japan, fi nd the best layout and its cost based on rectangular distances.

Department

Department 1 2 3 4 5 6

1 — 4 6 2 0 7

2 — 3 5 1 3

3 2 6 5

4 — 5 2

5 — 3

b. Resolve part a if the rates are ¥4 from odd to even departments, ¥5 from even to odd, ¥6 from odd to odd, and ¥7 from even to even.

3. An insurance offi ce in London is laid out as in the following table. The offi ce manager is considering switching departments 2 and 6 to reduce transport costs. Should this be done? [Use rectangular (rather than shortest) distances. Assume that offi ces are 10 feet on a side.] What is the difference in annual cost, assuming a 250-day work year?

1 2 3

4 5 6

D a i l y T r i p M a t r i x To

From 1 2 3 4 5 6

1 x 40 x x x 40

2 30 x 20 30 60 0

3 x 70 x x x 20

4 x 0 x x x 30

5 x 10 x x x 0

6 40 50 20 0 10 x

Trip Cost (to any department) per foot from

1 £0.02

2 £0.03

3 £0.01

4 £0.03

5 £0.02

6 £0.02

4. Lay out the offi ce in Exercise 3 again given the fol- lowing desired closeness ratings.

Department 1 2 3 4 5 6

1 2 3 4 5 6

I A X

X E O

O I X I

U O I E A

5. Demand for a certain subassembly in a toy manufac- turing facility at the North Pole is 96 items per eight- hour shift of elves. The following six tasks are re- quired to produce one subassembly.

c02.indd 88c02.indd 88 25/09/12 4:20 PM25/09/12 4:20 PM

89B i b l i o g r a p h y

Task Time Required (minutes) Predecessor Tasks

a 4 —

b 5 a

c 3 a

d 2 b

e 1 b,c

f 5 d,e

What is the required cycle time? Theoretically, how many stations will be required? Balance the line. What is the line ’s effi ciency?

6. An assembly line has the following tasks (times shown in minutes).

a. Six assemblies are required per hour. Balance the line.

b. What is the effi ciency of the line?

c. Rebalance the line if task e has a time of 1 minute instead of 3.

c 2

h 4

d 8

f 2

i 9

b 4

a 7

e 3

g 2

7. Given the following machine–part matrix, form cells using PFA. Parts

Machines 1 2 3 4 5

1 1 1

2 1

3 1 1

4 1 1

5 1 1

6 1

7 1

8 1

8. Kobenhavn Fine Products wishes to balance its line to meet a daily demand of 240 units. Kobenhavn works an eight-hour day on the following tasks:

Task Time (mins) Preceding Tasks

1 0.4 none

2 0.3 1

3 1.1 1

4 0.2 3

5 0.5 2

6 0.3 3

7 0.6 5

8 0.6 4, 6, 7

a. Find the cycle time, effi ciency, and minimum number of stations. Balance the line.

b. Rebalance the line if task 8 requires 0.7 minute.

B I B L I O G R A P H Y

Anupindi , R. , S. Chopra , S. D. Deshmukh , J. A. van Mieghem , and E. Zemel . Managing Business Process Flows . Upper Saddle River, NJ : Prentice Hall , 1999 .

Bitner , M. J. “ Servicescapes: The Impact of Physical Sur- roundings on Customers and Employees .” Journal of Marketing , 52 ( April 1992 ): 57 – 71 .

Bradley , T. “Fender Blenders.” In N. Goldwasser: 1988– 1989 Guitar Buyers’ Guide . Harris, 1988, 129 – 133 .

Bylinski , G. “ The Digital Factory .” Fortune (November 14, 1994 ): 92 – 107 .

Chase , R. B. , and S. Dasu . “ Want to Perfect Your Com- pany ’s Service? Use Behavioral Science .” Harvard Business Review ( June 2001 ): 78 – 84 .

Chase , R. B. , and D. M. Stewart . “ Make Your Service Fail-Safe .” Sloan Management Review (Spring 1994 ): 35 – 44 .

Chase , R. B. , and D. A. Tansik . “ The Customer Contact Model for Organization Design .” Management Science , 29 (September 1983 ): 1037 – 1050 .

c02.indd 89c02.indd 89 25/09/12 4:20 PM25/09/12 4:20 PM

90 C h a p t e r 2 : P r o c e s s P l a n n i n g a n d D e s i g n

Dennis , D. , and J. Meredith . “ An Empirical Analysis of Process Industry Transformation Systems .” Management Science , 46 (August 2000 ): 1085 – 1099 .

Ferras , L. “ Continuous Improvements in Electronics Manufacturing .” Production and Inventory Management Journal (Second Quarter 1994 ): 1 – 5 .

Firnstahl , T. W. “ My Employees Are My Service Guaran- tee .” Harvard Business Review ( July–August 1999 ): 28 – 32 .

Fitzsimmons , J. A. , and M. J. Fitzsimmons . Service Man- agement: Operations , Strategy, and Information Tech- nology , 5th ed. New York: Irwin/McGraw-Hill, 2006 .

Francis , R. L. , L. F. McGinnis Jr. , and J. A. White . Facility Layout and Location: An Analytical Approach , 2nd ed. Englewood Cliffs, NJ : Prentice-Hall , 1998 .

Hayes , R. H. , and S. C. Wheelwright . “ Link Manufactur- ing Process and Product Life Cycles .” Harvard Business Review ( January–February 1979 ): 133 – 140 .

Heskett , J. L. , W. E. Sasser Jr. , and L. A. Schlesinger . The Service Profi t Chain . New York : Free Press , 1997 .

Hyer , N. L., and K. A. Brown. “ Work Cells with Staying Power: Lessons for Process Complete Operations .” Cali- fornia Management Review , 46 : 1 ( 2003 ): 27 – 52 .

Meredith , J. R. , and S. J. Mantel Jr . Project Management: A Managerial Approach . Hoboken, NJ : Wiley , 2012 .

Metters , R. , K. King-Metters , and M. Pullman . Successful Service Operations Management . Mason, OH : South- Western , 2003 .

Parasuraman , A. , V. A. Zeithaml , and L. L. Berry . “ SERV- QUAL: A Multiple-Item Scale for Measuring Consumer Perceptions of Service Quality .” Journal of Retailing , 64 ( 1988 ): 12 – 40 .

Passariello , C. “Brand-New Bag: Louis Vuitton Tries Modern Methods on Old Factory Lines.” Wall Street Jour- nal , (October 9, 2006 ).

Passariello , C. “At Vuitton, Growth in Small Batches.” Wall Street Journal , ( June 11, 2011 ).

Schmenner , R. W. “ How Can Service Businesses Survive and Prosper? ” Sloan Management Review , 28 : 3 ( 1986 ): 21 – 32 .

Spear , S. , and H. K. Bowen . “ Decoding the DNA of the Toyota Production System .” Harvard Business Review (September–October 1999 ): 95 – 106 .

Verity , J. W. “A Company That ’s 100% Virtual.” Business Week (November 21, 1994 ): 85.

c02.indd 90c02.indd 90 25/09/12 4:20 PM25/09/12 4:20 PM

91

� Controlling Processes

C H A P T E R 3

ROLE OF OPERATIONS IN ORGANIZATIONS’ COMPETITIVENESS

Operations Strategy

Process Design and Control

Process Improvement

Process Execution

Ch. 1: Operations Strategy and Global

Competitiveness

Ch. 2: Process Planning and

Design

Ch. 3: Controlling Processes

Ch. 7: Supply Chain Management

Ch. 8: Capacity, Scheduling, and

Location Planning

Ch. 4: Process Improvement:

Six Sigma

Ch. 5: Process Improvement: Lean

Ch. 6: Managing Process

Improvement Projects

c03.indd 91c03.indd 91 25/09/12 10:43 AM25/09/12 10:43 AM

92

IN T R O D U C T I O N • The Columbus air mail processing center (AMC) is one of 77 AMCs operated by the

United States Postal Service (USPS). Each AMC is responsible for processing letters and packages that are to be distributed via air mail. A top priority of the USPS is out- standing customer service, which is assessed in part on the basis of the percent of letters and packages delivered on time. The USPS contracts with IBM Consulting Services to monitor its on-time delivery performance.

Each year the Columbus AMC processes in excess of 50 million fi rst-class letters that originate in central Ohio. Unfortunately, 8.7 percent of the letters processed there did not meet the established on-time delivery commitments. To address this problem, a Six Sigma project was initiated at the Columbus AMC with the objective of improving its on-time delivery performance. A key improvement made as a result of the Six Sigma project was the creation and documentation of a standardized process. In addition, improvements were incorporated into the new standardized process to “mistake-proof” it in order to further prevent late deliveries. As a result of the proc- ess improvements implemented, late deliveries were reduced by 14.3 percent. Furthermore, the improved process generated $15,000 in annual savings.

While making these types of process improvements is certainly important, it is equally important to carefully monitor and control the process to ensure that the improved level of performance is maintained. At the Columbus AMC, on-time deliv- ery performance was continuously monitored using a type of control chart called a

As the part of the organization that creates value for the customer, the operations area plays an important role in contributing to the organization ’s competitiveness. The way the operations will sup- port the organization ’s overall strategy and com- petitiveness is defi ned by its operations strategy, as discussed in Chaper 1. Once the operations strat- egy has been defi ned, the actual processes to exe- cute it must be planned and designed (discussed in Chapter 2). Once the organization ’s processes have been designed, they must be implemented and executed. The effective and effi cient execu- tion of processes is complicated by changes that occur both inside and outside the organization. Hence, every process must be monitored and con-

trolled to be sure it continues to achieve its objectives.

This chapter discusses the task of monitoring and control. It includes some discussion of the measures that will be monitored and ways to then exercise control to correct the process. We illus- trate the control process with the example of con- trolling quality through the use of quality control charts. Other topics include such well-popularized subjects as the balanced scorecard, strategy maps, ISO 9000 and ISO 14000, benchmarking, process capability, and service defections. After we com- plete our discussion of how to plan for process monitoring and control, the next two chapters will then delve into ways to improve these processes.

92

C H A P T E R I N P E R S P E C T I V E

c03.indd 92c03.indd 92 25/09/12 10:43 AM25/09/12 10:43 AM

93 I n t r o d u c t i o n

p -chart. The use of the p -chart enabled the Columbus AMC to track the proportion of late deliveries and discover when corrective measures were needed in a timely fashion (Franchetti 2008).

• Movistar is a mobile phone operator owned by the Telefónica Group, based in Argentina. In early 2008, Movistar raised concerns about the reliability of Telefónica ’s network service, citing over 400 service interruptions that resulted in lost revenue of $500,000. To address the concerns raised by Movistar, Telefónica initiated a Six Sigma project to better understand the root causes of the service interruptions. Over the course of the project, many process improvement tools were used to help better understand the problem, including process mapping, control charts, and box plots. In the end, the project resulted in dramatic reductions in service interruptions and in monthly savings of $300,000, allowing the project to recoup all its costs in six months. In refl ecting on the success of the project, managers cited the use of control charts incorporated into a scorecard that enabled the team to monitor the progress of the project and enhance communication among all stakeholders (Krzykowski 2011).

• North Shore–Long Island Jewish Health System (LIJ), located in Great Neck, New York, found that its “accessioning” registration process was out of control. Accessioning occurs when the information in the paper requisition for obtaining specimen samples is entered into the lab ’s information system and a label is gener- ated and placed on a sample tube. LIJ performs more than 3.5 million tests annually for all 18 of the system ’s hospitals as well as all the microbiology, molecular diagnos- tics, complex reference tests, long-term care facilities, clinical trials, and physician offi ces in the network. Accessioning errors had been measured for years and had been a chronic problem. However, when the rate of incomplete or inaccurate speci- mens reached 5 percent (i.e., 175,000!), LIJ decided to step in and exercise control. A multidisciplinary team was formed to investigate and correct the problem. Before the study began, the team assumed the handwriting of the physicians would be the cul- prit, but it was found that half the errors were due to incorrect entering of the patients’ Social Security numbers. This was due to the diffi culty in reading an addres- sograph label, which was then replaced with a bar-code reader. In addition, it was found that a small percentage of the staff made the majority of errors, indicating additional training was desirable. Further investigation showed that a lack of estab- lished best practices in the laboratory was another major source of errors, and dis- cussion with the staff led to the creation of a “lead accessioner” position to set standard practices for the lab. As a result of these changes, the lab increased its accuracy rate to 99 percent and its handling capacity by 43 percent, resulting in a positive fi nancial impact of about $339,000 a year (Riebling et al. 2004).

As these examples illustrate, monitoring and control are central to achieving the purposes for which processes and projects were initially designed. It doesn ’t matter if the process or project issue needing control is related to cost, quality, time, progress, or something else; to properly obtain the objectives we desire, we need to identify measures to track and, when needed, exercise control to bring results into agreement with our plans.

c03.indd 93c03.indd 93 25/09/12 10:43 AM25/09/12 10:43 AM

94 C h a p t e r 3 : C o n t r o l l i n g P r o c e s s e s

We begin the chapter with a general overview of the topics of monitoring and control. We then discuss the task of identifying appropriate measures to monitor, or track, as the processes get under way and describe a variety of tracking and control mechanisms that are available. Next, we illustrate one of the most fully developed control systems, statistical process control, using the control of quality as an exam- ple. Finally, we discuss a special case of control, that of service quality.

M O N I T O R I N G A N D C O N T R O L As we noted in Figure 1.1 back in Chapter 1 when we described the production sys- tem, we will want to monitor not only our processes but also our output—and even the environment—to make sure that our strategy, inputs, and transformation proc- esses are appropriate to achieve our goals. To do this, we fi rst need to identify the key factors to be monitored and controlled, which in turn depend on our goals for the production system.

The monitoring system is a direct connection between planning and control. But if it does not collect and report information on some signifi cant element of the produc- tion system, control can be faulty. Unfortunately, it is common to focus monitoring activities on data that are easily gathered—rather than important—or to concentrate on “objective” measures that are easily defended at the expense of softer, more subjective data that may be more valuable for control. When monitoring output performance, we should concentrate primarily on measuring various facets of output rather than inten- sity of activity. It is crucial to remember that effective managers are not primarily inter- ested in how hard their employees work—they are interested in the results achieved.

Although we will be monitoring the environment and our processes as well as our outputs, the measurement of output performance usually poses the most diffi - cult data-gathering problem. There is a strong tendency, particularly in service oper- ations, to let inputs serve as surrogate measures for output. If we have spent 50 percent of the budget (or of the scheduled time), we assume we have also com- pleted 50 percent of the work or reached 50 percent of our goal.

Given all this, performance criteria, standards, and data collection procedures must be established for each of the factors to be measured. However, more often than not, standards and criteria change because of factors that are not under the control of man- agement. For example, the market, our competitors, or government regulations may change. Standards may also be changed by the community as a response to some shift in public policy—witness the changes in the performance standards imposed on nuclear power installations or automotive exhaust systems. Shifts in the prime rate of interest or in unemployment levels often alter the standards that management must use in making decisions. The monitoring process is based on the criteria and stand- ards because they dictate, or at least constrain, the set of relevant measures.

Next, the information to be collected must be identifi ed. This may consist of accounting data, operating data, engineering test data, customer reactions, regula- tions, competitors’ prices, specifi cation changes, and the like. The fundamental problem is to determine precisely which of all the available data should be collected. It is worth repeating that the typical determinant for collecting data too often seems to be simply the ease with which they can be gathered. Again, the nature of the required data is dictated by the organization ’s objectives or goals.

c03.indd 94c03.indd 94 25/09/12 10:43 AM25/09/12 10:43 AM

95P r o c e s s M o n i t o r i n g

Perhaps the most common error made when monitoring data is to gather informa- tion that is clearly related to performance but has little or no probability of changing signifi cantly from one collection period to the next. Prior to its breakup, the American Telephone and Telegraph Company used to collect monthly statistics on a very large number of indicators of operating effi ciency. The extent of the collection was such that it fi lled a telephone-book-sized volume known as “Ma Bell ’s Green Book.” For a great many of the indicators, the likelihood of a signifi cant change from one month to the next was extremely small. When asked about the matter, one offi cial remarked that the mere collection of the data kept the operating companies “on their toes.” We feel that there are more positive and less expensive ways of motivating personnel. Certainly, “collect everything” is inappropriate as a monitoring policy.

Therefore, the fi rst task is to identify the objectives desired from the production system. Data must be identifi ed that measure achievement against these goals and mechanisms designed that will gather and store such data. Next, we must examine the purposes of the various processes and fi nd measures that provide us with insight into how these processes are performing. As an aside, if at least some of the data do not relate to the individual work level, no useful action is apt to be taken. In the end, it is the detailed work of the processes that must be altered if any aspect of perform- ance is to be changed. The fascinating book The Soul of a New Machine (Kidder 1981) reveals the crucial roles that organizational factors, interpersonal relationships, and managerial style play in determining process success.

There are a range of ways to determine what measures to monitor and then, if necessary, take action to control. We describe some of the more important ways in the next section.

P R O C E S S M O N I T O R I N G From the above discussion, it is clear that there are a wide variety of elements of the production system and environment that we may wish to monitor, but too much data can be worse than too little data since it can obscure the information that may be most important in the various measures we are watching. Most importantly, we need to monitor to make sure that we are effective in an overall competitive sense; the next section, Stages of Operational Effectiveness, discusses some measures for meeting this crucial objective. We next want to monitor our various processes, and we can make use of the balanced scorecard and strategy maps to guide us here. We also need to monitor the environment, including community standards, levels of risk, and other such elements. Since every situation faces unique environmental effects, we give only two examples of monitoring the environment, using ISO stand- ards to illustrate the international environment and FMEA to illustrate one approach to monitoring for risk.

Stages of Operational Effectiveness Wheelwright and Hayes (1985) suggest that organizations can progress through four stages of effectiveness in terms of the role their operations play in supporting and achieving the overall strategic objectives of a business ’s production system. As a

c03.indd 95c03.indd 95 25/09/12 10:43 AM25/09/12 10:43 AM

96 C h a p t e r 3 : C o n t r o l l i n g P r o c e s s e s

diagnostic tool, this framework helps determine the extent to which an organization is utilizing its operations to support and possibly attain a sustainable competitive advantage. As a prescriptive tool, the framework helps focus an organization on appropriate future courses of action because it is argued that stages cannot be skipped. Important managerial challenges are also identifi ed for each stage of effectiveness.

Organizations in stage 1 of the model are labeled internally neutral . These organ- izations tend to view operations as having little impact on the organization ’s com- petitive success. In fact, these organizations often consider the operations area as primarily a source of problems (e.g., quality problems, late shipments, too much capital tied up in inventory). Thus, believing that operations have little strategic importance, these organizations place an emphasis on minimizing the negative impact of operations.

Stage 2 is labeled externally neutral . As the name suggests, organizations at this stage attempt to match the operational practices of the industry. Thus, organizations in this stage still tend to view operations as having little strategic importance, but they at least attempt to follow standard industry practices and achieve operational parity with their competitors. Because these organizations follow industry practice, they tend to be more reactive than proactive in the operations area. Furthermore, operational investments and improvements tend to be tied to reducing costs.

Stage 3 is called internally supportive . In this stage of development, the organiza- tion expects its operations to support the overall business strategy and competitive position. In many cases this is stated as a formal operations strategy. Thus, opera- tional decisions are evaluated based on their consistency with and the extent to which they support the organization ’s overall mission. Internally supportive organi- zations tend to be more proactive in terms of identifying opportunities to support the organization ’s overall competitiveness. It is important to point out, however, that while stage 3 organizations expect operations to support the overall business strat- egy, operations is typically not involved in actually formulating it.

Stage 4 organizations depend on their operations to achieve a competitive advan- tage and are referred to as externally supportive . In effect, these organizations use core capabilities residing in the operations area to obtain a sustainable competitive advantage.

Because different parts of an organization may evolve at different rates, determin- ing an organization ’s stage of effectiveness may require making a judgment about where the balance of the organization is positioned. Thus, it is possible that some departments or areas of a stage 2 organization exhibit characteristics of a stage 3 organization. However, if the majority of the organization is most appropriately characterized as being in stage 2, then the organization should be categorized as being in stage 2.

Thus, evaluating an organization ’s evolution is based not on its most evolved area but rather on the balance of its organizational practices.

Using the defi nitions of the four operational effectiveness stages above, Dangayach and Deshmukh (2006) developed perceptual measures for each of these stages, some of which are given in Table 3.1 . By asking managers how closely their com- pany followed each of these policy measures (on a 1–5 scale), they were able to classify fi rms into one of the four stages and relate them to the fi rms’ overall per- formance. Thus, these measures would be excellent items for monitoring the evolv- ing competitive strength of manufacturing fi rms and taking control actions when the fi rms show competitive slippage.

c03.indd 96c03.indd 96 25/09/12 10:43 AM25/09/12 10:43 AM

97P r o c e s s M o n i t o r i n g

Balanced Scorecard The balanced scorecard approach (Kaplan and Norton 1996) is becoming increas- ingly recognized as a way to help organizations translate their strategy into appropri- ate performance measures in order to monitor their success. In the past, it was not uncommon for managers to rely primarily on fi nancial performance measures. However, when the inadequacies of these measures were discovered, managers often responded either by trying to improve them or by abandoning them in favor of operational performance measures such as cycle time and defect rates. Many organi- zations now realize that no single type of measure can provide insight into all the critical areas of the business. Thus, the purpose of the balanced scorecard is to develop a set of measures that provide a comprehensive view of the organization.

Organizations that have developed a balanced scorecard report numerous bene- fi ts, including the following:

• An effective way to clarify and gain consensus on the strategy • A mechanism for communicating the strategy throughout the entire

organization

• A mechanism for aligning departmental and personal goals to the strategy • A way to ensure that strategic objectives are linked to annual budgets • Timely feedback related to improving the strategy One problem with traditional performance measurement systems based primarily

on fi nancial measures is that they often encourage shortsighted decisions, such as reducing investments in product development, employee training, and information technology. The balanced scorecard approach corrects this problem by measuring per- formance in four major areas: (1) fi nancial performance, (2) customer performance, (3) internal business process performance, and (4) organizational learning and growth.

T A B L E 3 .1• Measure s fo r Opera t iona l E f fec t i venes s Stage Measures

Internally neutral The objective is to minimize operation s negative potential.

Firefi ghting is common.

Outside experts are called in for strategic decisions.

Operations is primarily reactive.

Externally neutral Industry practice is followed.

The aim is to achieve competitive parity.

Internally supportive Operations investments support the business strategy.

An operations strategy is formulated and pursued.

Externally supportive Operations is involved upfront in major strategic decisions.

The aim is to achieve a competitive advantage through operations.

The goal is to achieve competitive superiority.

c03.indd 97c03.indd 97 25/09/12 10:43 AM25/09/12 10:43 AM

98 C h a p t e r 3 : C o n t r o l l i n g P r o c e s s e s

The fi nancial performance measures included in the balanced scorecard are typically related to profi tability, such as return on equity, return on capital, and economic value added. Customer performance measures focus on customer satis- faction, customer retention, customer profi tability, market share, and customer acquisition. The internal business process dimension addresses the issue of what the organization must excel at to achieve its fi nancial and customer objectives. Examples of performance measures for internal business processes include quality, response time, cost, new-product launch time, and the ratio of processing time to total throughput time. Finally, the learning and growth dimension focuses on the infrastructure the organization must build to sustain its competitive advantage. Learning and growth performance measures include employee satisfaction, employee retention, worker productivity, and the availability of timely and accu- rate information.

The process of developing a balanced scorecard begins with top management translating the mission and strategy into specifi c customer and fi nancial objectives. Based on the customer and fi nancial objectives, related measures for the internal business processes are identifi ed. Finally, investments in employee training and information technology are linked to the customer, fi nancial, and internal business process objectives. Note that a properly constructed balanced scorecard contains an appropriate mix of outcome measures related to the actual results achieved and measures that drive future performance.

The balanced scorecard is based on the premise that a strategy is a set of hypoth- eses about cause-and-effect relationships that can be stated as if–then statements. For example, management of a department store might hypothesize that increasing the training that sales associates receive will lead to improved selling skills. These managers might further hypothesize that better selling skills will translate into higher commissions for the sales associates and will therefore result in less turnover. Happier and more experienced sales associates would likely lead to increased sales per store, which ultimately translates into an increase in return on investment. Since a properly developed balanced scorecard tells a story about the cause-and-effect relationships underlying the strategy, all measures included in the scorecard should be elements in the chain of cause-and-effect relationships.

One aspect of monitoring we have not yet addressed in detail is the role of our competitors in the environment. Obviously, being able to attain our mission and goals is not completely up to us, and our competitors have a strong impact on our success. Fortunately, there are again some well-formulated concepts for monitoring how our competitors are doing and whether or not they are threatening our success. One such approach is called “benchmarking,” essentially documenting the level of competence of either competitors or even noncompetitor “best-in-class” organiza- tions. This can show how our competitors are doing on those measures of impor- tance to our customers or clients as well as on other measures of interest to us, such as those included in the balanced scorecard.

But benchmarking can also be used simply to fi nd out how well some aspects of production can be done, known as “state-of-the-art,” “world-class,” or “best-in-class.” Industry leaders, often in a different industry from our own, can help us to improve our own operations even if our competitors are not at that level, giving us a potential competitive advantage. Hence, the topic of benchmarking will be described in more detail in Chapter 4.

c03.indd 98c03.indd 98 25/09/12 10:43 AM25/09/12 10:43 AM

99P r o c e s s M o n i t o r i n g

At the top of the strategy map the goal is specifi ed, which in our example is to improve the store ’s return on investment (ROI). Management has determined that the goal of improving the store ’s ROI can be accomplished by increasing revenue and/or improving the store ’s productivity. The remainder of the strategy map explic- itly shows the chain of cause-and-effect relationships management has hypothesized

The Strategy Map In extending their earlier work on the balanced scorecard, Kaplan and Norton (2000) proposed the development of strategy maps (see also Scholey 2005) as a way to illustrate and monitor the cause-and-effect relationships identifi ed through the devel- opment of a balanced scorecard. In particular, strategy maps provide organizations with a tool that helps them better monitor important details about their strategic business processes, thereby enhancing their employees’ understanding of the strat- egy interactions, which in turn facilitates implementing the business strategy.

Like the balanced scorecard, strategy maps address four perspectives: the fi nan- cial perspective, the customer perspective, the internal business process perspective, and the learning and growing perspective. A sample strategy map for a department store that wants to improve its performance is shown in Figure 3.1 .

Figure 3.1 Sample strategy map for a department store.

Increase sales/ft2

Improved selling skills

Friendly and courteous sales

associates

Provide training to sales

associates

Happier sales associates

Stronger relationships between customers

and associates

Increase inventory turns

Revenue growth strategy

Financial perspective

Customer perspective

Internal business process perspective

Learning and growth perspective

Productivity improvement strategy

Less turnover—more

experienced associates

Improve store performance (ROI)

c03.indd 99c03.indd 99 25/09/12 10:43 AM25/09/12 10:43 AM

100 C h a p t e r 3 : C o n t r o l l i n g P r o c e s s e s

about how the store ’s ROI can be improved. For example, it is believed that provid- ing the sales associates with additional training will lead to improved selling skills, which should then result in increased sales per square foot of retail space and hap- pier associates. Happier associates in turn should result in both friendly and courte- ous sales associates and less turnover among the associates. Ultimately, the strategy map hypothesizes that increased sales per square foot will help the store increase its revenue and its inventory turns, which will then lead to revenue growth and produc- tivity improvements. In the end, management should develop measures for each element in the strategy map and then track these measures to validate the hypothe- sized relationships proposed in the strategy map. In cases where the hypothesized relationships are not supported, new relationships should be hypothesized.

ISO 9000 and 14000 ISO 9000 was developed as a guideline for designing, manufacturing, selling, and servicing products; in a sense, it is a sort of checklist of good business practices. Thus, the intent of the ISO 9000 standard is that, if an organization selects a supplier that is ISO 9000–certifi ed, it has some assurance that the supplier follows accepted business practices in the areas specifi ed in the standard. However, one criticism of ISO 9000 is that because it does not require any specifi c actions, each organization determines how it can best meet the requirements of the standard.

ISO 9000 was developed by the International Organization for Standardization and fi rst issued in March 1987. A major revision to ISO 9000 was completed in December 2008; the new standard is commonly referred to as ISO 9000:2008. Since its introduction, ISO 9000 has become the most widely recognized standard in the world. To illustrate its importance, in 1993 the European Community required that companies in several industries become certifi ed as a condition of conducting busi- ness in Europe. In fact, over 630,000 organizations in 152 countries have imple- mented ISO 9000 and/or ISO 14000.

ISO 14000 is a series of standards covering environmental management systems, environmental auditing, evaluation of environmental performance, environmental labeling, and life-cycle assessment. Like ISO 9000, ISO 14001 (a subset of the ISO 14000 series) is a standard in which organizations can become certifi ed. The focus of ISO 14001 is on an organization ’s environmental management system. However, like ISO 9000, ISO 14001 does not prescribe specifi c standards for performance or levels of improvement. Rather, its intent is to help organizations improve their envi- ronmental performance through documentation control, operational control, control of records, training, statistical techniques, and corrective and preventive actions.

Clearly, this set of international standards for both production and environmental maintenance certifi cation will lead to a range of measures that should be considered for monitoring the proper functioning of the production system. But more focused measures specifi c to the production system at hand can also be developed, such as through Failure Mode and Effect Analysis (FMEA), described next.

Failure Mode and Effect Analysis (FMEA) FMEA was developed by the space program in the 1960s (Stamatis 2003) as a struc- tured approach to help identify and prioritize for close monitoring and control those

c03.indd 100c03.indd 100 25/09/12 10:43 AM25/09/12 10:43 AM

101P r o c e s s M o n i t o r i n g

elements of a system that might give rise to failure. It employs a scoring model approach set up in a series of six straightforward steps, as follows:

1. List the possible ways a production system might fail.

2. Evaluate the severity ( S ) of the consequences of each type of failure on a 10-point scale, where 1 is “no effect” and 10 is “very severe.”

3. For each cause of failure, estimate the likelihood ( L ) of its occurrence on a 10-point scale where 1 is “remote” and 10 is “almost certain.”

4. Estimate the ability to detect ( D ) a failure associated with each cause. Using a 10-point scale, 1 means detectability is almost certain using normal moni- toring/control systems and 10 means it is practically certain that failure will not be detected in time to avoid or mitigate it.

5. Find the Risk Priority Number ( R P N ) , where R P N   5   S   3   L   3   D .

6. Consider ways to reduce the S , L , and D for each cause of failure with a signifi cantly high RPN.

Table 3.2 illustrates the application of FMEA to a new concept for a fast-food restau- rant. Here we are primarily illustrating how to apply the approach, but in a real situa- tion our items of failure would be much more specifi c and narrow: a particular machine, a particularly diffi cult process, a unique government regulation, and other such items that clearly could result in missing our goals for the new concept. In the table, we have identifi ed training and marketing as the elements with the highest RPNs, and thus those we particularly want to monitor and control very carefully. We might invest additional time and effort in training our employees to offset the fi rst threat, but since L is low already, it might be more productive to fi nd ways to detect this inadequacy faster, such as surveying our customers to monitor their perceptions of our service. This might then reduce D to 3 instead of 5 and thereby mitigate the threat. In addition, we could include a question on the survey to help determine which mar- keting approaches are having the greatest results, reducing D from 8 to perhaps 5.

T A B L E 3 .2 • FMEA fo r a New Fas t -Food Concep t Potential Ways to Fall S L D RPN

Inadequate training 8 4 5 160

Weak marketing 6 3 8 144

Poor location 7 5 3 105

Defective concept 9 3 3 81

Local restaurant regulation change 3 5 8 120

Competitors’ reactions (e.g., price, ads) 4 6 4 96

It was noted above that more specifi c items might be included in the FMEA list, such as a particular machine or diffi cult process. One way to identify such items for inclu- sion in the FMEA list, or for monitoring in general, is by evaluating their “process capa- bility,” a topic we explain in detail in Chapter 4 when we talk about process

c03.indd 101c03.indd 101 25/09/12 10:43 AM25/09/12 10:43 AM

102 C h a p t e r 3 : C o n t r o l l i n g P r o c e s s e s

improvement. In essence, however, if the specifi cations for a machine (or process) are relatively tight compared to the natural variation in the machine, we have to carefully monitor the machine ’s process capability to satisfy the requirements of the production system. For example, in our fast-food example above, if the con- cept required a diffi cult and time-consuming step in the preparation of a particular menu item, we might fail because our customers got tired of waiting for their food. Hence, we would want to monitor the time this process required very carefully and take controlling steps if it started to exceed the required specifi cations of our production system.

P R O C E S S C O N T R O L Process control is the act of reducing differences between plan and reality for each process. Monitoring and comparing activities with the plan and then reporting these fi ndings is to no avail if actions are not taken when reality deviates signifi cantly from what was planned. In fact, the simple act of noting and reporting discrepancies may motivate the actions required to correct the deviations. When it does not, however, active control is needed to bring performance back in line with the plan. Control has the primary purpose of ensuring that the process is in compliance with its objectives. In large production systems particularly, early control is crucial since the longer we wait, the more diffi cult it is to correct the deviation.

Control is one of the manager ’s most diffi cult tasks, invariably involving both mechanical and human elements. In addition, control can be diffi cult because prob- lems are rarely clear-cut and hence the need for change and redirection may be fuzzy. Determining what to control raises further diffi culties—did someone take an incorrect action or is the system to blame, or perhaps simply Mother Nature?

A good control system should also possess some specifi c characteristics:

• The system should be fl exible. Where possible, it should be able to react to unforeseen changes in system performance.

• The system should be cost-effective, and the cost of control should never exceed the value of control. For example, bear in mind that control is not always less expensive than scrap.

• The system should be as simple as possible to operate. • The system must operate in a timely manner. Problems must be reported

while there is still time to do something about them.

• Sensors and monitors should be suffi ciently precise to control the process within limits that are truly functional for the organization.

• The control system should be easy to maintain. • The system should signal the manager if it goes out of order. • The system should be capable of being extended or otherwise altered. • Control systems should be fully documented when installed, and the docu-

mentation should include a complete training program in system operation.

• The system must operate in an ethical manner.

c03.indd 102c03.indd 102 25/09/12 10:43 AM25/09/12 10:43 AM

103P r o c e s s C o n t r o l

Next, we turn to a common tool for controlling processes, the control chart, and illustrate it with the case of controlling quality. However, it can also be used for con- trolling many other measures, such as scrap, turnover, revenues, progress (see, e.g., Meredith and Mantel 2009, pp. 490–493), costs, and so on.

Statistical Process Control One of management ’s most diffi cult decisions in quality control centers on whether an activity needs adjustment. This requires some form of inspection, and in quality control there are two major types of inspection, either (1) measuring something or (2) simply determining the existence of a characteristic .

1. Type 1, inspection by measuring, called inspection for variables , usually relates to weight, length, temperature, diameter, or some other variable that can be scaled .

2. Type 2, inspection by identifying a characteristic , called inspection of attri- butes, can also examine scaled variables but usually considers dichotomous variables such as right–wrong, acceptable–defective, black–white, on-time– late, and other such characteristics that either cannot be measured or do not need to be measured with any more precision than yes–no.

Walter A. Shewhart developed the concept of statistical control charts in the 1920s to distinguish between chance variation in a system that is still in control and variation caused by the system ’s being out of control, which he called assign- able variation . Should a process go out of control, that must fi rst be detected, then the assignable cause must be identifi ed, and fi nally the appropriate control action or adjustment must be performed. The control chart is used to detect when a proc- ess has gone out of control.

A repetitive operation will seldom produce items of exactly the same quality, size, and so on; rather, with each repetition the operation will generate variation around some average. This variation is particularly characteristic of a sampling process where random samples are taken and a sample mean is calculated. Because this variation usually has a large number of small, uncontrollable sources, the pattern of variability is often well described by a standard frequency distribution such as the normal distribution , shown plotted against the vertical scale in Figure 3.2 .

The succession of measures that results from the continued repetition of some process can thus be thought of as a population of numbers, normally distributed, with some mean and standard deviation. As long as the distribution remains the same, the process is considered to be in control and simply exhibiting chance varia- tion. One way to determine if the distribution is staying the same is to keep checking the mean of the distribution—if it changes to some other value, the operation may be considered out of control. The problem, however, is that it is too expensive for organizations to keep constantly checking operations. Therefore, samples of the out- put are checked instead.

In sampling the output for inspection, it is imperative that the sample fully repre- sent the population being checked; therefore, a rational subgroup of data should be used. But when checks are made only of sample averages, rather than 100 percent

c03.indd 103c03.indd 103 25/09/12 10:43 AM25/09/12 10:43 AM

104 C h a p t e r 3 : C o n t r o l l i n g P r o c e s s e s

of the output, there is always a chance of selecting a sample with an unusually high or low mean. The problem facing the operations manager is thus to decide what is too high or too low and therefore should be considered out of control. Also, the man- ager must consider the fact that the more samples eventually taken, the higher the likelihood of accidentally selecting a sample with too high (or too low) a mean when the process is actually still under control .

The values of the sample mean that are too high or low are called the upper control limit (UCL) and the lower control limit (LCL), respectively. These limits generally allow an approach to control that is known as management by exception because, theoretically, the manager need take no action unless a sample mean exceeds the control limits. The control limits most commonly used in organizations are plus and minus 3 standard deviations . We know from statistics that the chance that a sample mean will exceed 3 standard deviations, in either direction, due sim- ply to chance variation, is less than 0.3 percent (i.e., 3 times per 1,000 samples). Thus, the chance that a sample will fall above the UCL or below the LCL because of natural random causes is so small that this occurrence is strong evidence of assign- able variation. Figure 3.2 illustrates the use of control limits set at 3 standard devia- tions. Of course, using the higher limit values (3 or more) increases the risk of not detecting a process that is only slightly out of control.

An even better approach is to use control charts to predict when an out-of-control situation is likely to occur rather than waiting for a process to actually go out of con- trol. If only chance variation is present in the process, the points plotted on a control chart will not typically exhibit any pattern. On the other hand, if the points exhibit some systematic pattern, this is an indication that assignable variation may be present and corrective action should be taken.

The control chart, though originally developed for quality control in manufactur- ing, is applicable to all sorts of repetitive activities in any kind of organization. Thus, it can be used for services as well as products, for people or machines, for cost or quality, and so on. For example, the beginning of the chapter mentioned the USPS ’s use of control charts to monitor on-time delivery performance.

Figure 3.2 Control chart with the limits set at 3 standard deviations.

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

LCL

Mean

UCL

99.74% 0.13%

Sample number

Sa m

p le

m ea

n v

al u

e

c03.indd 104c03.indd 104 25/09/12 10:43 AM25/09/12 10:43 AM

105P r o c e s s C o n t r o l

For the control of variables —that is, measured characteristics —two control charts are required:

1. Chart of the sample means ( X – ) .

2. Chart of the range ( R ) of values in each sample (largest value in sample minus smallest value in sample).

It is important to use two control charts for variables because of the way in which control of process quality can be lost. To illustrate this phenomenon, we will use the data supplied in Table 3.3 , which correspond to the minutes needed to process a form at an insurance company. Three samples are taken each day: one in the morning, one near noon, and one just before closing. Each sample consists of three forms randomly selected from the staff working on these forms in the processing department.

T A B L E 3 .3 • Sample Data o f Proces s Times (minutes)

Sample Scenario 1 Scenario 2

1 4, 5, 6 5, 4, 6

2 6, 7, 8 3, 5, 7

3 7, 9, 8 8, 2, 5

Referring to scenario 1, we can easily determine that the average of sample 1 is 5 minutes and the range is 2 minutes ( X

– 1   5   5 ,   R

1   5   2 ) . Similarly, X

– 2   5   7 , R

2   5   2 ,

X – 3   5   8 , and R

3   5   2 . If we consider only the ranges of the samples, no problem is indi-

cated because all three samples have a range of 2 (assuming that a range of 2 min- utes is acceptable to management). On the other hand, the behavior of the process means shows evidence of a problem. Specifi cally, the process means (minutes) have increased throughout the day from an average of 5 minutes to an average of 8 min- utes. Thus, for the data listed in scenario 1, the sample ranges indicate acceptable process performance while the sample means indicate unacceptable process performance.

The sample statistics can be calculated in the same way for scenario 2: X – 1   5   5 ,

R 1   5   2 , X

– 2   5   5 , R

2   5   4 , X

– 3   5   5 , and R

3   5   6 . In contrast to scenario 1, the sample

means show acceptable performance while the sample ranges show possibly unac- ceptable performance. Thus, we see the necessity of monitoring both the mean and the variability of a process.

Figure 3.3 illustrates these two patterns of change in the distribution of process values more formally. These changes might be due to boredom, tool wear, improper training, the weather, fatigue, or any other such infl uence. In Figure 3.3 a the variabil- ity in the process remains the same, but the mean changes (scenario 1); this effect would be seen in the means ( X

– ) chart but not in the range ( R ) chart. In Figure 3.3 b

the mean remains the same, but the variability tends to increase (as in scenario 2 above); this would be seen in the range ( R ) chart but not the means ( X

– ) chart. In

terms of quality of the output, either type of change could result in lower quality,

c03.indd 105c03.indd 105 25/09/12 10:43 AM25/09/12 10:43 AM

106 C h a p t e r 3 : C o n t r o l l i n g P r o c e s s e s

depending on the situation. Regarding control limits, the lower control limit (LCL) for the means chart may be negative, depending on the variable being measured. For example, variables such as profi t and temperature can be negative, but variables such as time, length, diameter, and weight cannot. Since (by defi nition) the range can never be negative, if calculations indicate a negative LCL for the range chart, it should simply be set to zero.

As indicated earlier, control limits for the means chart are usually set at plus and minus 3 standard deviations. But if a range chart is also being used, these limits for the means chart can be found by using the average range, which is directly related to the standard deviation, in the following equations (where X

––   is the grand mean or

the average of the sample means):

U C L X –   5   X

––   1   A

2 R –  

L C L X –   5   X

––    2 A

2 R –  

Similarly, control limits for the range chart are found from

UCL R   5   D

4 R –  

LCL R   5   D

3 R –  

The factors A 2 , D

3 , and D

4 vary with the sample size and are tabulated in Table 3.4 .

Constructing Control Charts The best way to illustrate the construction of control charts is by example. Assume that a bank with 10 branches is interested in monitoring the age of the applications for home mortgages being processed at its branches. To maintain a continuing check on this measure of customer responsiveness, one could select branches at random each day and note the ages of the applications processed that day. To set up the control charts, initial samples need to be taken. These data will, if considered repre- sentative by management, be used to set standards (i.e., control limits) for future

Figure 3.3 Patterns of change in process distributions.

Process variable, X

F re

q u

en cy

April 3 April 4 April 5

Process variable, X

F re

q u

en cy

Oct. 30 Nov. 30 Dec. 30

(a) (b)

c03.indd 106c03.indd 106 25/09/12 10:43 AM25/09/12 10:43 AM

107P r o c e s s C o n t r o l

applications. For our example, we assume that a sample of n   5   4 of the 10 branches each day will give the best control for the trouble involved. The mean age and range in ages for the initial samples were entered into the spreadsheet shown in Table 3.5 . Note that each sample mean and sample range shown in Table 3.5 is based on data collected by randomly visiting four branches. The grand mean X

–– , and the average

range were also calculated (cells B23 and C23, respectively). The grand mean is then simply the average of all the daily means:

X ––   5   ∑ X

– ____ N

where N is 20 days of samples and the average range is

R –   5   ∑ R ___ N

The data in Table 3.5 can now be used to construct control charts that will indi- cate to management any sudden change, for better or worse, in the ages of the mort- gage applications. Both a chart of means, to check the age of the applications, and a chart of ranges, to check consistency among branches, should be used.

The grand mean and average range will give the centerline on these charts, respectively. The values of A

2 , D

3 , and D

4 are obtained from Table 3.4 for n   5   4 ,

resulting in the following control limits:

UCL X –   5   11   1   0.729 ( 14 )   5   21.206

LCL X –   5   11   2   0.729 ( 14 )   5   0.794

UCL R   5   2.282 ( 14 )   5   31.948

LCL R   5   0 ( 14 )   5   0

T A B L E 3 .4 • Cont ro l Char t Fac tor s to De te rmine Cont ro l L imi t s

Sample Size, n A 2 D

3 D

4

2 1.880 0 3.267

3 1.023 0 2.575

4 0.729 0 2.282

5 0.577 0 2.115

6 0.483 0 2.004

7 0.419 0.076 1.924

8 0.373 0.136 1.864

9 0.337 0.184 1.816

10 0.308 0.223 1.777

12 0.266 0.284 1.716

14 0.235 0.329 1.671

16 0.212 0.364 1.636

18 0.194 0.392 1.608

20 0.180 0.414 1.586

22 0.167 0.434 1.566

24 0.157 0.452 1.548

c03.indd 107c03.indd 107 25/09/12 10:43 AM25/09/12 10:43 AM

108 C h a p t e r 3 : C o n t r o l l i n g P r o c e s s e s

The control charts for this example were developed using a spreadsheet and are shown in Figures 3.4 and 3.5 . In addition, the data in Table 3.5 are graphed on the charts. As seen in Figure 3.4 , no pattern is apparent from the data; the points appear to fall randomly around the grand mean (centerline) and thus are considered by management to be representative.

The range chart, Figure 3.5 , again shows no apparent pattern and is also accept- able to management. Each day, as a new sample is taken, X

–   and R will be calculated

and plotted on the two charts. If either X –   or R is outside the LCL or UCL, manage-

ment must then undertake to fi nd the assignable cause for the variation. Control charts can also be used for controlling attributes of the output, the second

type of inspection we described earlier. The most common of these charts are the fraction-defective p-chart and the number-of-defects c chart . As with the range chart, the lower control limit for attribute charts can never be negative.

The fraction-defective p -chart can be used for any two-state ( dichotomous ) proc- ess such as heavy versus light, acceptable versus unacceptable, on-time versus late, or placed properly versus misplaced. The control chart for p is constructed in much

T A B L E 3 .5 • Mean and Range o f Age s o f Mor tgage App l i ca t ions

A B C

1 Sample Sample

2 Date Mean Range

3 June 1 10 18

4 June 2 13 13

5 June 3 11 15

6 June 4 14 14

7 June 5 9 14

8 June 6 11 10

9 June 7 8 15

10 June 8 12 17

11 June 9 13 9

12 June 10 10 16

13 June 11 13 12

14 June 12 12 14

15 June 13 8 13

16 June 14 11 15

17 June 15 11 11

18 June 16 9 14

19 June 17 10 13

20 June 18 9 19

21 June 19 12 14

22 June 20 14 14

23 Average 11 14

c03.indd 108c03.indd 108 25/09/12 10:43 AM25/09/12 10:43 AM

109P r o c e s s C o n t r o l

the same way as the control chart for X – . First, a large sample of historical data is

gathered, and the fraction (percent) having the characteristic in question (e.g., too light, defective, misplaced), p– , is computed on the entire set of data as a whole.

Large samples are usually taken because the fraction of interest is typically small and the number of items in the samples should be large enough to include some of

Figure 3.4 Mean mortgage application age.

1 5 10

Day of June

Sample means,

Grand mean

15 20

LCL

UCL

X

0.8

2

4

6

8

10

12

14

16

18

20

22

M ea

n m

o rt

ga ge

a p

p lic

at io

n a

ge , X

( d

ay s)

X

Figure 3.5 Range in mortgage application age.

1 5 10

Day of June

Sample ranges, R

Average range

R an

ge in

m o

rt ga

ge a

p p

lic at

io n

a ge

s, R

( d

ay s)

15 20 LCL

UCL

R

0

2

4

6

8

10

12

14

16

18

20

22

24

26

28

30

32

c03.indd 109c03.indd 109 25/09/12 10:43 AM25/09/12 10:43 AM

110 C h a p t e r 3 : C o n t r o l l i n g P r o c e s s e s

the defectives. For example, a fraction defective may be 3 percent or less. Therefore, a sample size of 33 would have to be taken (i.e., 1 / 0.03   5   33 ) to expect to include even one defective item. Since the control limits depend in part on the sample size used, it is best to use the same sample sizes for constructing the control charts and then collect additional data to monitor the process.

Since the fraction defective follows a binomial distribution ( bi means “two”: either an item is or it is not) rather than a normal distribution, the standard deviation may be calculated directly from p–   as

� p 5   �

_______

p– ( 1 2   p– )

_______ n

where n is the uniform sample size to be used for controlling quality. Although the fraction defective follows the binomial distribution, if np– and n ( 1   2   p– )   are both greater than 5, the normal distribution is a good approximation and the control lim- its of 3�

p will again represent 99.7 percent of the sample observations. Again, the

LCL cannot be negative. The number-of-defects c chart is used for a single situation in which any number

of incidents may occur, each with a small probability. Typical of such incidents are scratches on tables, fi re alarms in a city, typesetting errors in a newspaper, and the number of improper autoinsertions per printed circuit board. An average number of incidents, c- , is determined from past data by dividing the total number of incidents observed by the number of items inspected. The distribution of such incidents is known to follow the Poisson distribution with a standard deviation of

� c   5   �

__

c-  

Again, the normal distribution is used as an approximation to derive control limits with a minimum LCL of zero.

C O N T R O L L I N G S E R V I C E Q U A L I T Y For process control, strategy maps and control charts can also be used for control- ling the quality of services, assuming the right measures are being monitored. However, measuring the quality of the service portion of an output is often more diffi cult than measuring the quality of the facilitating good for a variety of reasons— including the service portion being abstract rather than concrete, transient rather than permanent, and psychological rather than physical. One way to cope with these diffi culties, as indicated earlier, is to use customer satisfaction surveys. For example, J.D. Power and Associates (www.jdpower.com) makes extensive use of customer satisfaction surveys to rate airlines, hotel chains, and rental car companies. For example, its ratings of airlines in 2011 were based on 13,500 fl ight evaluations supplied by a national survey of business and leisure travelers from July 2010 to April 2011. In the surveys, overall customer satisfaction was assessed on the basis of the airlines’ performance in seven areas: cost and fees, in-fl ight service, fl ight crew, aircraft, boarding/deplaning/baggage, check-in, and reservations. According to the travelers surveyed, the airlines made progress in improving overall traveler

c03.indd 110c03.indd 110 25/09/12 10:43 AM25/09/12 10:43 AM

111C o n t r o l l i n g S e r v i c e Q u a l i t y

satisfaction from 2010 to 2011. In 2011, Alaska Airlines received the highest ranking among the traditional network airlines for the fourth consecutive year, followed by Air Canada and Continental. In the low-cost segment, JetBlue had the highest rank- ing, followed by Southwest Airlines.

J. D. Power ranks hotel chains, wireless carriers, and rental car companies in a similar fashion. For example, in 2010, J. D. Power ranked the overall satisfaction for fi ve categories of hotels (luxury, upscale, economy, etc.) on the basis of 53,000 individual evaluations. According to J. D. Power, the fi ve most important amenities for hotel guests in 2010 were wireless Internet access, complimentary breakfasts, bedding and pillow choices, pillow-top mattresses, and free parking. In 2010, Ritz- Carlton was the highest-ranked luxury hotel and Omni Hotels was the highest ranked upscale hotel.

A common approach to improving the quality of services is to methodically train employees in standard procedures and to use equipment that reinforces this trai- ning. The ultimate example is McDonald ’s Hamburger University, where managers are intensively trained in the McDonald ’s system of food preparation and delivery. Not only is training intensive, but follow-up checkups are continuous, and incen- tives and rewards are given for continuing to pay attention to quality. Furthermore, the equipment is designed to reinforce the quality process taught to employees, and to discourage sloppy habits that lead to lesser quality.

Financial services can also benefi t from better quality. Several years ago, a major bank noticed that its requests for letters of credit were handled by nine different employees who conducted dozens of steps, a process that consumed four days. By retraining its employees so that each would be able to process a customer ’s request through all the steps, the bank was able to let each customer deal with only one employee, who could complete the process within a day. Now each time a letter of credit is ordered, the customer is placed back with the same employee. As a result, the department involved has been able to double its output of letters of credit using the services of 49 percent fewer employees.

By paying attention to the quality delivered to customers, American Express was able to cut the processing time for new credit applications from 22 days to 11 days, thereby more than doubling the revenue per employee in its credit card division. It had previously tracked errors and processing time internally but had ignored the impacts on the customer. When it began focusing on the customer, it suddenly found that speed in the credit department was often immaterial in shortening the customer ’s waiting time for credit approval because four more departments still had to process every new application.

Service Defections When a tangible product is produced, quality is often measured in terms of defects. In many services, the analogy to a product defect is a defecting customer—that is, a customer who takes his or her business elsewhere. Thus, service defections can be measured in a variety of ways, such as the percent of customers that do not renew their membership (health clubs), the percent of sales from new versus repeat cus- tomers (offi ce supply store), and the number of customers that cancel their service (wireless phone companies). Of course, the concept of a defecting customer is equally applicable to organizations that produce tangible outputs.

c03.indd 111c03.indd 111 25/09/12 10:43 AM25/09/12 10:43 AM

112 C h a p t e r 3 : C o n t r o l l i n g P r o c e s s e s

1. We often divide “control” into two categories: pre- ventive and corrective. Would you classify FMEA as a tool for monitoring, preventive control, or correc- tive control? Explain.

2. Does the balanced scorecard monitor all the ele- ments of the production system of Figure 1.1? If not, what does it miss? Does it include anything not in Figure 1.1? If so, why, and why isn ’t it in Figure 1.1?

3. Since service defections are analogous to defects in products, could they be controlled with control charts in the same way? Would they use charts based on inspection of variables or attributes? Give an example(s).

4. Under what kinds of circumstances might an organ- ization wish to use control limits of 2 standard devi- ations or even 1 standard deviation? What should it bear in mind when using these lower limits?

5. Why are two control charts not necessary in con- trolling for attributes? Might not the variability of the fraction defective or the number of defects also be going out of control?

6. It is generally not appropriate to apply control charts to the same data that were used to derive the mean and limits. Why? What are the two possible outcomes if this is done, how likely is each, and what are the appropriate interpretations?

7. In deriving the p -chart, why can the sample size vary? What must be remembered if the p -chart is applied to a different sample size each time?

8. Firms regularly employ a taster for drinkable food products. What is the purpose of this taster?

9. How is quality handled differently in service fi rms and product fi rms? Does quality mean something different in a service fi rm?

10. For many years, the balanced scorecard was seen as more appropriate for implementing strategy than for planning it. Now that the strategy map has been con- ceived, does this replace the balanced scorecard or does each do something different and, if so, which one is concerned with implementing strategy?

Organizations should monitor their defecting customers for a number of reasons. First, research suggests that long-time customers offer organizations a number of benefi ts. For example, the longer a customer has a relationship with an organization, the more likely that customer is to purchase additional products and services and the less price-sensitive he or she is. In addition, no advertising is necessary to get the business of long-term customers. In fact, long-term customers may actually be a source of free advertising for the company. One study published in the Harvard Business Review con- cluded that cutting defections in half more than doubles the average company ’s rate of growth. Likewise, improving customer retention rates by 5 percent can double profi ts.

Defections by customers can provide a variety of useful information. First, feed- back obtained from defecting customers can be used to identify areas that need improvement. Also, the feedback can be used to determine what can be done to win these customers back. Finally, monitoring for increases in the defection rate can be used as an early warning signal that control actions are immediately needed.

E X P A N D Y O U R U N D E R S T A N D I N G

A P P L Y Y O U R U N D E R S T A N D I N G Pa in t Tin t

Late last month, Jim Runnels, a sales representative for the Paint Tint Corporation, was called to the plant of Townhouse Paint Company, one of his largest accounts. The purchasing agent for Townhouse Paint was complaining that the tubes of paint tint it had received over the last couple of weeks were not within the specifi ed range of 4.9 to 5.1 ounces.

The off-weight tubes had not been detected by Townhouse ’s receiving clerks and had not been weighed or otherwise checked by their quality control staff. The problem arose

c03.indd 112c03.indd 112 25/09/12 10:43 AM25/09/12 10:43 AM

113A p p l y Y o u r U n d e r s t a n d i n g

when Townhouse began to use the tubes of tinting agent and found that the paint colors were not matching the specifi cations. The mixing charts used by the salespeople in Town- house ’s retail stores were based on 5-ounce tubes of tinting agent. Overfi lled or underfi lled tubes would result in improper paint mixes, and therefore in colors that did not meet cus- tomers’ expectations.

In consequence, Townhouse had to issue special instructions to all of its retail people that would allow them to compensate for the off-weight tubes. The Townhouse purchasing agent made it clear that a new supplier would be sought if this problem recurred. Paint Tint ’s qual- ity control department was immediately summoned to assist in determining the cause of the problem.

Paint Tint ’s quality manager, Ronald Wilson, speculated that the cause of the problem was with the second shift. To analyze the problem, he entered into a spreadsheet the data from all the previous samples taken over the last two months. As it turned out, 15 random samples had been taken over the two-month period for both the fi rst and the second shifts. Samples always consisted of 10 randomly selected tubes of paint tint. Also, separate sampling sched- ules were used for the fi rst and second shifts so that the second shift would not automati- cally assume that it would be subject to a random sample just because the fi rst shift had been earlier in the day.

After entering the sample weight data of the tubes into the following spreadsheet and cal- culating the sample means, Ronald was quite puzzled. There did not seem to be any notice- able difference in the average weights across the two shifts. Furthermore, although the lines were running at less than full capacity during the fi rst six samples, there still did not seem to be any change in either line after reaching full production.

A B C D E F G H I J K L M N O P

1 First Shift

2 Sample Number

3 Observation 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

4 1 4.90 5.05 4.96 4.92 4.96 5.03 4.99 5.00 5.02 5.03 5.01 4.95 5.02 4.96 5.06

5 2 5.03 5.04 4.96 5.00 5.00 4.99 5.03 5.01 5.05 4.90 4.94 4.95 4.95 4.97 4.97

6 3 5.00 5.00 4.92 5.05 5.03 4.98 5.01 4.95 5.00 4.95 5.00 5.06 5.00 4.93 5.00

7 4 5.03 5.11 5.01 5.03 4.98 4.99 5.02 5.01 5.01 5.01 5.00 5.02 4.98 5.01 5.00

8 5 5.02 4.94 4.98 5.01 5.00 4.98 5.01 4.99 5.03 5.01 4.96 4.94 5.04 5.00 5.03

9 6 4.92 5.02 5.00 5.02 5.02 5.01 4.99 4.98 5.00 4.94 4.98 4.99 5.02 5.04 5.08

10 7 5.04 5.03 4.98 5.02 5.00 4.99 5.06 4.96 5.01 4.98 5.01 4.97 4.99 4.98 4.97

11 8 4.92 5.00 5.00 4.96 5.01 5.01 5.05 5.00 4.97 4.98 4.97 4.97 5.05 5.08 4.98

12 9 4.95 4.95 4.94 5.02 4.95 4.98 4.97 4.94 5.07 5.00 5.00 4.96 5.02 4.94 5.00

13 10 5.02 4.99 5.08 4.94 5.00 4.95 5.04 4.98 5.02 5.01 4.98 5.02 5.06 5.02 4.97

14 Average 4.98 5.01 4.98 5.00 5.00 4.99 5.02 4.98 5.02 4.98 4.99 4.98 5.01 4.99 5.01

15

16

17 Second Shift

18 Sample Number

19 Observation 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

20 1 5.03 5.02 4.99 4.96 5.03 5.02 5.08 5.10 5.16 5.00 4.97 5.11 5.11 4.90 5.02

21 2 4.90 4.95 4.97 4.97 4.98 5.03 4.97 4.93 4.92 4.97 4.91 5.05 4.98 4.92 4.98

22 3 5.02 4.94 5.04 4.98 5.00 4.98 4.93 4.92 4.99 5.08 5.15 4.93 5.13 4.97 4.86

c03.indd 113c03.indd 113 25/09/12 10:43 AM25/09/12 10:43 AM

114 C h a p t e r 3 : C o n t r o l l i n g P r o c e s s e s

Questions

1. Can you identify any difference between the fi rst and second shifts that explains the weight problem? If so, when is this difference fi rst detectable?

Koa laTech , L td .

KoalaTech, Ltd., of Sydney, Australia, produces offi ce equipment for small businesses and home offi ces. Several months ago it launched its PFS 1000, a single unit that functions as a color printer, color scanner, color copier, and fax machine. The PFS 1000 won rave reviews for its functionality, affordable price, and innovative design. This, coupled with KoalaTech ’s repu- tation for producing highly reliable products, quickly led to a severe backlog. KoalaTech ’s plant simply could not keep up with demand.

Initially, KoalaTech ’s president, Nancy Samuelson, was extremely concerned about the backlog and put a great deal of pressure on the plant manager, George Johnson, to increase production. However, Nancy abruptly shifted gears when a new report indicated that returns and complaints for the PFS 1000 were running four times higher than the usual industry rate. Because KoalaTech ’s reputation was on the line, Nancy decided that the problem required immediate attention. She also decided that the quickest way to diagnose the problem and to avoid the usual mentality of “blaming it on the other department” would be to bring in an outside consultant with expertise in these matters.

Nancy hired Ken Cathey to investigate the problem. Nancy and Ken agreed that Ken should spend his fi rst week interviewing key personnel in an effort to learn as much about the problem as possible. Because of the urgency of the problem, Nancy promised Ken that he would have complete access to—and the cooperation of—all employees. Nancy would send out a memo immediately informing all employees that they were expected to cooperate and assist Ken in any way they could.

The next morning, Ken decided to begin his investigation by discussing the quality prob- lem with several of the production supervisors. He began with the supervisor of the fi nal as- sembly area, Todd Allision. Todd commented:

I received Nancy ’s memo yesterday and, frankly, the problem with the PFS 1000 does not surprise me. One of the problems we ’ve had in fi nal assembly is with the casing. Basically, the case is composed of a top and a bottom. The problem we are having is that these pieces rarely fi t together, so we typically have to force them together. I ’m sure this is adding a lot of extra stress on the cases. I haven ’t seen a breakdown on what the problems with quality are, but it wouldn ’t surprise me if one of the problems was cracked cases or cases that are coming apart. I should also mention that we never had this problem with our old supplier. However, when purchasing determined that we could save over $A1 per unit, we switched to a new supplier for the cases.

(Continued)

A B C D E F G H I J K L M N O P

23 4 4.98 5.05 5.02 5.00 4.97 5.06 4.84 4.93 5.00 5.07 4.96 5.15 5.15 4.92 4.94

24 5 5.01 4.95 5.02 5.02 4.98 5.04 5.07 5.03 4.98 4.94 4.91 4.98 5.10 5.04 4.93

25 6 4.99 4.99 4.99 5.03 5.00 5.04 4.95 4.96 4.99 4.96 5.07 4.88 5.12 5.03 4.97

26 7 4.99 4.97 5.00 4.98 4.99 4.99 4.93 4.86 5.01 5.13 5.15 4.74 5.01 4.91 5.05

27 8 5.02 5.00 5.00 4.96 4.98 4.98 4.99 5.08 5.07 4.93 4.95 4.90 4.93 4.95 4.97

28 9 5.01 5.00 5.05 5.02 5.03 4.97 4.82 4.96 4.93 4.96 4.91 5.03 5.04 4.98 5.03

29 10 4.97 4.99 4.95 5.03 5.00 4.99 5.05 5.14 5.03 4.91 5.11 5.04 5.03 5.08 4.92

30 Average 4.99 4.99 5.00 5.00 5.00 5.01 4.96 4.99 5.01 5.00 5.01 4.98 5.06 4.97 4.97

c03.indd 114c03.indd 114 25/09/12 10:43 AM25/09/12 10:43 AM

115A p p l y Y o u r U n d e r s t a n d i n g

The meeting with Todd lasted for about an hour and a half, and Ken decided that rather than meet with someone else, he would be better off reviewing the notes he had taken and fi lling in any gaps while the conversation was still fresh in his mind. Then he would break for lunch and meet with one or two additional people in the afternoon.

After returning from lunch, Ken stopped by to talk with Steve Morgan, the production supervisor for the printed circuit boards. Ken found Steve and an equipment operator staring at one of the auto-insertion machines used to place components such as integrated circuits, capacitors, and resistors on the printed circuit board before wave soldering. Upon arriving, Ken introduced himself to Steve and asked, “What ’s up?” Steve responded:

We are having an extremely diffi cult time making the printed circuit boards for the PFS 1000. The designers placed the components closer together than this generation of equipment was designed to handle. As a result, the leads of the components are constantly being bent. I doubt that more than 25 percent of the boards have all their components installed properly. As a result, we are spending a great deal of time inspecting all the boards and reworking the ones with prob- lems. Also, because of the huge backlog for these boards and the large number that must be reworked, we have been trying to operate the equipment 20 percent faster than its normal operating rate. This has caused the machine to break down much more frequently. I estimate that on a given eight-hour shift, the machine is down one to two hours.

In terms of your job—to determine the cause of the problems with quality—faulty circuit boards are very likely a key contributor. We are doing our best to fi nd and correct all the defects, but inspecting and reworking the boards is a very tedious process, and the employees are putting in a lot of extra hours. In addition, we are under enormous pressure to get the boards to fi nal assembly. My biggest regret is that I didn ’t have more input when they were building the prototypes of the PFS 1000. The prototypes are all built by highly trained technicians using primarily a manual process. Unfortunately, the prototypes are built only to give the engineers feedback on their designs. Had they shown some people in production the prototypes, we could have made suggestions on changes that would have made the design easier to produce.

Ken decided to end the day by talking to the plant manager, Harvey Michaels. Harvey was in complete agreement with Todd and Steve and discussed at length the enormous pressure he was under to get product out the door: “The bottom line is that no one cooperates. Purchasing changes suppliers to save a few bucks, and we end up with components that can ’t be used. Then our own engineers design products that we can ’t produce. We need to work together.”

On his second day, Ken decided to follow up on the information he had gathered the day before. He fi rst visited the director of purchasing, Marilyn Reagan. When asked about the problem of the cases that did not fi t together, Marilyn responded:

The fact of the matter is that switching suppliers for the cases saved $A1.04 per unit. That may not sound like a lot, but multiply that by the 125,000 units we are expecting to sell this year, and it turns out to be pretty signifi cant. Those guys in production think the world revolves around them. I am, however, sympathetic to their problems, and I plan on discussing the problem with the supplier the next time we meet. That should be some time next month.

After wrapping up the meeting with Marilyn, Ken decided he would next talk to the director of engineering. On the way, he recognized a person at a vending machine as the worker who had been standing next to Steve at the auto-insertion machine. Ken introduced himself and decided to talk with the worker for a few minutes. The worker introduced him- self as Jim and mentioned that he had been working in the shipping department until just two weeks ago. The operator before Jim had quit because of the pressure. Jim hadn ’t received any formal training in operating the new equipment, but he said that Steve tried to check on him a couple of times a day to see how things were going. Jim appreciated Steve ’s efforts, but the quality inspectors made him nervous and he felt that they were always look- ing over his shoulder.

c03.indd 115c03.indd 115 25/09/12 10:43 AM25/09/12 10:43 AM

116 C h a p t e r 3 : C o n t r o l l i n g P r o c e s s e s

Ken thanked Jim for his input and then headed off to meet with the director of engineering, Jack Carel. After introducing himself, Ken took a seat in front of Jack ’s desk. Jack began:

So you are here to investigate our little quality snafu. The pressure that we are under here in engineering is the need to shrink things down. Two years ago fax machines, printers, scanners, and copiers were all separate pieces of equipment. Now, with the introduction of the PFS 1000, all this functionality is included in one piece of equipment not much larger than the original printer. That means design tolerances are going to be a lot tighter and the product is going to be more diffi cult to manufacture. But the fact of the matter is that manufacturing is going to have to get its act together if we are going to survive. The engineering department did its job. We designed a state-of-the-art piece of offi ce equipment, and the prototypes we built proved that the design works. It ’s now up to the manufacturing guys to fi gure out how to produce it. We have done all that we can and should be expected to do.

To end his second day, Ken decided to meet with the director of quality assurance, Debbie Lynn. Debbie commented:

My biggest challenge as director of quality assurance is trying to convince the rest of the organization of the importance quality plays. Sure, everyone gives lip service to the importance of quality, but as the end of the month approaches, getting the product out the door is always the highest priority. Also, while I am offi cially held accountable for quality, I have no formal authority over the production workers. The quality inspectors that report to me do little more than inspect product and tag it if it doesn ’t meet the specifi cations so that it is sent to the rework area. In all honesty, I am quite optimistic about Nancy ’s current concern for quality and very much welcome the opportunity to work closely with you to improve KoalaTech ’s quality initiatives.

Questions

1. Which departments at KoalaTech have the most impact on quality? What role should each department play in helping KoalaTech improve overall quality?

2. What recommendations would you make to Nancy concerning KoalaTech ’s problem with quality? What role should the quality assurance department play?

E X E R C I S E S

1. The city government is planning a career fair week- end in two months and wishes to use an FMEA table to identify risk elements to monitor and deter- mine possible actions to take in controlling the risks. They believe that there are 4 major risks to the fair:

1. Insuffi cient employer turnout—the severity if this happens they rate as an 8, the likelihood as a 4, and the detectability as a 5.

2. Insuffi cient job seeker turnout—for this they rate the severity as a 6, likelihood as 7, and detectability as a 5.

3. Inclement weather—severity 4, likelihood 5, detectability 4.

4. Economy—severity 7, likelihood 5, detectability 6.

Determine the most important risk elements and suggest ways to control them.

2. Pick two different processes on which to run an FMEA analysis, such as running an end-of-semester party and passing an upcoming advanced fi nance course. Identify the major risks for success with each one and give each a severity, likelihood, and detectability rating to calculate the RPN for each risk. How would you monitor each of these risk elements for impending failure and what steps might you take to control them?

3. Top management of the Royal Scottish Bank monitors the volume of activity at 38 branch banks with control charts. If deposit volume (or any of perhaps a dozen other volume indicators) at a branch falls below the LCL, there is apparently some problem with the branch ’s market share. If, on the other hand, the vol- ume exceeds the UCL, it means the branch should be considered for expansion or that a new branch might be opened in an adjacent neighborhood.

c03.indd 116c03.indd 116 25/09/12 10:43 AM25/09/12 10:43 AM

117E x e r c i s e s

Given the 10-day samples for each of the six months below, prepare an X

– chart for the deposit vol-

ume (in hundreds of thousands of pounds) for the Kil- marnock branch. Use control limits of 63�. The aver- age range of the six samples was found to be £85,260.

Month Average of 10 Days of Deposits (£100,000)

June 0.93 July 1.05 August 1.21 September 0.91 October 0.89 November 1.13

4. Using the following weekly demand data for a new soft drink, determine the upper and lower control limits that can be used in recognizing a change in demand patterns. Use   6 3 � control limits.

Week Demand (6-packs)

1 3500 2 4100 3 3750 4 4300 5 4000 6 3650

5. A control chart has a mean of 50 and 2� control lim- its of 40 and 60. The following data are plotted on the chart: 38, 55, 58, 42, 64, 49, 51, 58, 61, 46, 44, 50. Should action be taken?

6. Given the following data, construct a 3� range con- trol chart.

Day of Sample Sample Values

Saturday 22, 19, 20 Sunday 21, 20, 17 Monday 16, 17, 18 Tuesday 20, 16, 21 Wednesday 23, 20, 20 Thursday 19, 16, 21

a. If Friday ’s results are 15, 14, and 21, is the proc- ess in control?

b. Construct a 3� means control chart and deter- mine if the process is still in control on Friday.

7. a. Using the following data, prepare a p -chart for the control of picking accuracy in a wholesale food warehouse. Sample size is expected to be 100 cases.

Day Number of

Cases Picked Number of

Incorrect Picks

1 4700 38 2 5100 49 3 3800 27 4 4100 31 5 4500 42 6 5200 48

b. Determine if days 7, 8, and 9 are under control.

Day Number of Cases

Picked Number of

Incorrect Picks

7 100 1

8 100 2

9 100 4

8. A new machine for making nails produced 25 defec- tive nails on Monday, 36 on Tuesday, 17 on Wednes- day, and 47 on Thursday. Construct an X

– chart,

p -chart, and c -chart based on the results for Monday through Wednesday and determine if Thursday ’s pro- duction was in control. The machine produces 1 mil- lion or so nails a day. Which is the proper chart to use?

9. Construct a p -chart using 2� limits based on the results of 20 samples of size 400 in the following table.

Sample Number Number of Defects

1 2 2 0 3 8 4 5 5 8 6 4 7 4 8 2 9 9

10 2 11 3 12 0 13 5 14 6 15 7 16 1 17 5 18 8 19 2 20 1

c03.indd 117c03.indd 117 25/09/12 10:43 AM25/09/12 10:43 AM

118 C h a p t e r 3 : C o n t r o l l i n g P r o c e s s e s

B I B L I O G R A P H Y

10. Twenty samples of 100 were taken, with the follow- ing number of defectives: 8, 5, 3, 9, 4, 5, 8, 5, 3, 6, 4, 3, 5, 6, 2, 5, 0, 3, 4, 2. Construct a 3 � p -chart.

11. Sheets of Styrofoam are being inspected for fl aws. The fi rst day ’s results from a new machine that pro- duced fi ve sheets are 17, 28, 9, 21, 14. Design a control chart for future production.

Alwan , L. Statistical Process Analysis . New York : Irwin/ McGraw-Hill , 2000 .

Besterfi eld , D. H. Quality Control , 6th ed. Upper Saddle River, NJ : Prentice Hall , 2001 .

Dangayach , G. S. , and S. G. Deshmukh . “ An Exploratory Study of Manufacturing Strategy Practices of Machinery Manufacturing Companies in India .” OMEGA: The Inter- national Journal of Management Science , 34 ( 2006 ): 254 – 273 .

Evans , J. R. , and W. M. Lindsay . The Management and Control of Quality . Cincinnati : South-Western , 1999 .

Fitzsimmons , J. A. , and M. J. Fitzsimmons . Service Man- agement for Competitive Advantage , 5th ed. New York : McGraw-Hill , 2006 .

Franchetti, M. “Six Sigma Receives USPS Stamp of Approval.” Six Sigma Formum Magazine (February 2008): 15–17.

Goetsch , D. L. , and S. B. Davis . Quality Management: Introduction to Total Quality Management for Produc- tion, Processing, and Services , 3rd ed. Upper Saddle River, NJ : Prentice Hall , 2000 .

Gupta, V. K., and D. J. Graham. “A Customer Driven Quality Improvement and Management Project at Dia- mond Offshore Drilling.” Project Management Journal (September 1997): 22–28.

Harvey , J. “ Service Quality: A Tutorial .” Journal of Oper- ations Management , 16 ( 1998 ): 583 – 597 .

Ibbs, C. W., and Y.-H. Kwak, “Benchmarking Project Management Organizations.” PM Network (February 1998): 49–53.

Jones, T. O., and W. E. Sasser, Jr. “Why Satisfi ed Custom- ers Defect.” Harvard Business Review (November– December 1995): 89–99.

Juran , J. M. , and F. M. Gryna Jr . Quality Planning and Analysis , 3rd ed. New York : McGraw-Hill , 1993 .

Kaplan , R. S. , and D. P. Norton . The Balanced Score- Card . Boston : Harvard Business School Press , 1996 .

Kaplan, R. S., and D. P. Norton. “Having Trouble with Your Strategy? Then Map It.” Harvard Business Review (September–October 2000): 167–176.

Kidder , T. The Soul of a New Machine . Boston : Little, Brown , 1981 .

Krzykowski, B. “Walking the Talk.” Quality Progress ( June 2011): 41–44.

Ledolter , J. , and C. Burrill . Statistical Quality Control: Strategies and Tools for Continual Improvement . New York : Wiley , 1999 .

Livingston, J. L., and R. Ronen. “Motivation and Manage- ment Control Systems.” Decision Sciences (April 1975): 360–375.

Meredith , J. R. , and S. J. Mantel Jr. Project Management: A Managerial Process , 7th ed. Hoboken, NJ : Wiley , 2009 .

McGrath, R. G., and I. MacMillan. “Discovery-Driven Planning.” Harvard Business Review ( July–August 1995): 44–54.

Montgomery , D. C. Introduction to Statistical Quality Control , 4th ed. New York : Wiley , 2001 .

Nixon, T. R. “Project Management at the New Parliament House, Canberra,” Project Management Journal (September 1987): 40–68.

Ott , E. , E. G. Schilling , and D. Neubauer . Process Quality Control . New York : McGraw-Hill , 2000 .

Reichheld, F. F. “Learning from Customer Defections.” Harvard Business Review (March–April 1996): 56–69.

Riebling, N. B., S. Condon, and D. Gopen. “Toward Error Free Lab Work.” Six Sigma Forum Magazine (November 2004): 23–29.

Schoderbek , C. G. , P. P. Schoderbek , and A. G. Kefalas . Management Systems , 4th ed. Homewood, IL : Irwin , 1989 .

Scholey , C. “ A Step-by-Step Guide to Measuring, Manag- ing, and Communicating the Plan .” Journal of Business Strategy , 26 ( 3 ) ( 2005 ): 12 – 19 .

Stamatis , D. H. Failure Mode and Effect Analysis: FMEA from Theory to Execution , 2nd ed. Milwaukee : ASQ Quality Press , 2003 .

Summers , D. Quality , 2nd ed. Upper Saddle River, NJ : Prentice Hall , 2000 .

Thamhain, H. J. “Best Practices for Controlling Technology- Based Projects.” Project Management Journal (December 1996): 37–48.

Wheelwright, S. C., and R. H. Hayes. “Competing Through Manufacturing.” Harvard Business Review ( January– February 1985): 99–109.

c03.indd 118c03.indd 118 25/09/12 10:43 AM25/09/12 10:43 AM

119

� Process Improvement:

Minimizing Variation Through

Six Sigma

C H A P T E R 4

ROLE OF OPERATIONS IN ORGANIZATIONS’ COMPETITIVENESS

Operations Strategy

Process Design and Control

Process Improvement

Process Execution

Ch. 1: Operations Strategy and Global

Competitiveness

Ch. 2: Process Planning and

Design

Ch. 3: Controlling Processes

Ch. 7: Supply Chain Management

Ch. 8: Capacity, Scheduling, and

Location Planning

Ch. 4: Process Improvement:

Six Sigma

Ch. 5: Process Improvement: Lean

Ch. 6: Managing Process

Improvement Projects

c04.indd 119c04.indd 119 25/09/12 10:50 AM25/09/12 10:50 AM

120

I N T R O D U C T I O N • Hewitt Associates provides human resource (HR) outsourcing services for client

organizations. HR outsourcing involves having a service provider take over an organi- zation ’s traditional HR functions, such as payroll processing, benefi ts administration, and employee recruiting. In effect, Hewitt becomes the HR department for its HR out- sourcing clients (see the BPO, Inc. case in the Cases section at the end of the book for more background information on the business process outsourcing industry).

Customer service (CS) representatives are the Hewitt employees on the frontline delivering the HR outsourcing service to its clients. These CS reps need substantial training in the proprietary information systems developed to process the data pro- vided by the client organizations as well as to provide them with the detailed knowl- edge about the benefi t plans Hewitt administers for its clients.

A key challenge Hewitt faced was that the annual turnover of its CS reps was close to 100 percent. Each time a CS rep left, Hewitt incurred signifi cant costs associ- ated with the separation process, hiring costs for the replacement, training costs, and costs related to lost productivity.

To address the turnover challenge, a team initiated a Six Sigma project using the traditional DMAIC methodology and multiple Six Sigma tools, including project char- ters, cause-and-effect diagrams, and prioritization matrices. At the outset of the project, the team determined that it fi rst needed to quantify the costs of CS rep turnover using hard data in order to get senior management buy-in for investing in initiatives to reduce the turnover rate. Based on the team ’s analysis, it was estimated

The operations strategy defi nes the way an organi- zation ’s operations will support the organization ’s overall competitiveness. Once the operations strat- egy has been defi ned, the value-creating processes must be designed (Chapter 2) and controlled (Chapter 3). While controlling the processes, it is often determined that there are opportunities to improve the process. Thus, the focus of this chap- ter is on the redesign and continuous improve- ment of business processes in support of the overall business strategy. To put our discussion in perspective, we begin with an overview of three alternative approaches for process improvement. We then turn our attention to the fi rst process improvement strategy, Business Process Design.

This is then followed by a detailed discussion of the second process improvement strategy, Six Sigma. Next, each phase in Six Sigma ’s DMAIC approach is discussed in more detail, including illustrating the use of representative Six Sigma tools in each phase. The chapter concludes with a discussion of Six Sigma in practice. Here we discuss the various roles associ- ated with Six Sigma, becoming certifi ed, and the need for organizations to customize their approach to Six Sigma training and implementation. In the next chapter we continue our discussion of process improvement strategies and address the third proc- ess improvement strategy, namely lean. The trend toward integrating Six Sigma and lean will also be discussed in the next chapter.

120

C H A P T E R I N P E R S P E C T I V E

c04.indd 120c04.indd 120 25/09/12 10:50 AM25/09/12 10:50 AM

that CS rep turnover was costing Hewitt $14.5 million annually in lost productivity and another $1.5 million for training and recruiting costs.

To analyze the situation further, Hewitt used regression analysis to identify the key factors that impacted employee turnover. The regression analysis identifi ed a number of potential factors, including providing employees with opportunities for growth, pay, recognition, work activities, and management. After studying the results, the team decided to experiment with one potential factor, base pay. To inves- tigate the impact base pay had on retention, the team selected a group of CS reps to pilot-test providing them with market adjustments to their base pay. Making these adjustments cost Hewitt $600,000. The result was that the retention rate of the CS reps who received the market adjustment to their base pay increased to 96 percent, which translated into not losing 80 profi cient CS reps. The team further estimated that the lower turnover rate translated into a savings of $1.9 million, providing a 217 percent return on investment (Leatherbury 2008).

• Error prone and ineffi cient, Bank of America (B of A) was paying a price in terms of both money and customer dissatisfaction. As one example, on a ten-point scale, only 40 percent of B of A ’s customers rated their experience with the company at the top, that is, a nine or ten. Internally, hundreds of thousands of defects were being created per million opportunities.

Ken Lewis, the company ’s CEO, decided in 2001 that a change in strategy was needed. This entailed a shift from fueling growth by mergers and acquisitions toward more organic growth based on retaining and deepening the relationship with existing customers. Thus, in 2001 the company embarked on its quality journey. Based on his belief that the company needed a more disciplined and comprehensive approach to process improvement, Lewis turned to Six Sigma. A new senior manager reporting directly to Lewis was hired to oversee quality and productivity.

Being a fi nancial services organization, it was to be expected that many in the organization would be skeptical of the applicability of an approach that was devel- oped for factories. One way CEO Lewis addressed this concern was by being among the fi rst to personally complete a Green Belt project and further requiring all of his direct reports to complete projects as well. Each of these projects was a success, pro- viding benefi ts such as improved customer satisfaction with problem resolution, sig- nifi cantly reduced travel expenses, and increased employee retention.

If you ask executives at B of A, they will tell you Six Sigma is not a fad but the way we conduct business. To get its Six Sigma initiative off the ground, B of A recruited Black Belts and Master Black Belts from leading Six Sigma organizations such as General Electric, Motorola, and Honeywell. By 2004, B of A estimated that there were over 100 open senior leadership positions requiring a background in Six Sigma. One senior executive at B of A had gone so far as to speculate that Black Belt certifi cation would become a mandatory qualifi cation for leadership roles at B of A. By early 2004, B of A had trained more than 10,000 employees in the use of Six Sigma tools to support the DMAIC methodology. But perhaps the most compelling statistics relate to the overall benefi ts B of A has received as a result of its Six Sigma initiatives. In particular, B of A estimates that it obtained benefi ts in excess of

I n t r o d u c t i o n 121

c04.indd 121c04.indd 121 25/09/12 10:50 AM25/09/12 10:50 AM

122 C h a p t e r 4 : P r o c e s s I m p r o v e m e n t : M i n i m i z i n g V a r i a t i o n T h r o u g h S i x S i g m a

$2 billion in less than three years while at the same time increasing its customer delight by 25 percent ( Jones 2004).

• In 2002, the nuclear medicine department of Southside Hospital, a not-for-profi t community hospital located in Bayshore, New York, was receiving numerous com- plaints regarding the turnaround times for stress tests. The turnaround time for a stress test is measured as the elapsed time from when the stress test was ordered by a physician until the results were signed off by a radiologist in the nuclear medicine department. Delays in receiving the results from stress tests impacted the timeliness of treating the patients, which in turn could affect the length of time a patient was required to stay in the hospital. To address the problem associated with excessive turnaround times, hospital administrators decided to test Six Sigma ’s defi ne, measure, analyze, improve, and control (DMAIC) approach to assess its applicability to health care operations. To execute the project, a team consisting of one Black Belt and three Green Belts was created. In the course of completing the project, the team uti- lized many traditional Six Sigma tools, including voice of the customer, “critical to quality” trees, process mapping, stakeholder analysis, defects per million opportuni- ties, cause-and-effect diagrams, regression analysis, and Pareto analysis. In the end, the team was able to reduce the turnaround times for stress tests by over 50 percent, from 68 hours to 32 hours (the standard deviation was also reduced from 32 hours to 9 hours). In addition, the team was able to increase the process sigma level (dis- cussed later in this chapter) from less than 0.1 to 2. Finally, the project resulted in an overall increase in capacity with no additional cost. In fact, costs actually decreased by $34,000 due to savings in salaries. In the end, hospital administrators acknowl- edged the extent to which such a data-driven approach enhanced the ultimate suc- cess of the project (Godin et al., 2004).

• One of the tasks performed by TRW ’s corporate law department is the registration of trademarks. The company estimates that it costs an average of $1200 (not including processing costs) to renew a trademark worldwide. To evaluate the trademark renewal process, a team utilized many traditional Six Sigma tools, including voice of the customer, determining the critical to quality characteristics, logistic regression analysis, and value-added process mapping to evaluate the process. One fi nding from the project was that in numerous cases trademarks were being renewed more out of a sense of history and nostalgia as opposed to providing value to the busi- ness. In the end, the project produced hard savings of $1.8 million by eliminating the renewal of entire classes of trademarks. Numerous process improvements in the trademark renewal process were also identifi ed, producing additional soft savings. Finally, by clearly defi ning defects in the trademark renewal process, the project team was able to establish a baseline process sigma level of 2.18, which can be used to assess the impact of future process improvements (Das et al., 2004).

These examples illustrate that the Six Sigma process improvement methodology is applicable to a wide variety of organizations and disciplines. The examples are also representative of signifi cant gains organizations achieve in improving their processes through the use of Six Sigma. These improvements translate directly into helping

c04.indd 122c04.indd 122 25/09/12 10:50 AM25/09/12 10:50 AM

123I n t r o d u c t i o n

As our discussion illustrates, Six Sigma is among the most timely topics in business today. Furthermore, Six Sigma is equally applicable to organizations that produce a tangible output (e.g., Cummins and Sun Microsystems) or deliver an intangible serv- ice (e.g., Air Canada and American Express) as well as to organizations that exist to make a profi t (e.g., B of A, American Standard, and Tyco International) or are non- profi t organizations such as Southside Hospital.

organizations become more competitive. As a result of the success organizations have had with their Six Sigma initiatives, it has become a particularly timely topic in business. This has further resulted in people with a background in Six Sigma being in high demand. In fact, at 3M, 25 percent of the 1000 employees who completed Six Sigma training were promoted two levels or more! As further support, a search in March 2012 at Monster.com using the keyword “Six Sigma” yielded over 1000 hits across virtually all industries, including manufacturing, consulting, technology, fi nan- cial services, insurance, health care, and retail.

As further evidence of the value industry is placing on individuals with Six Sigma experience, consider the following quote taken from the letter to shareholders in GE ’s 2000 annual report, whose co-authors included chairman and CEO Jack Welch and president and chairman-elect Jeffrey Immelt (p. 6):

It is reasonable to guess that the next CEO of this Company, decades down the road, is probably a Six Sigma Black Belt or Master Black Belt somewhere in GE right now, or on the verge of being offered—as all our early-career (3–5 years) top 20% perform- ers will be—a two- to three-year Black Belt assignment. The generic nature of a Black Belt assignment, in addition to its rigorous process discipline and relentless customer focus, makes Six Sigma the perfect training for growing 21st century GE leadership.

A question that naturally arises is: What is driving industry ’s interest in Six Sigma? Perhaps the primary reason for the current popularity of Six Sigma is that it works, as exemplifi ed by the signifi cant benefi ts several high-profi le organizations have reported from their Six Sigma initiatives. To illustrate this, Table 4.1 summarizes the fi nancial benefi ts obtained by organizations across a variety of industries. The table also provides summary information related to the number of employees who were trained at various Six Sigma levels.

T A B L E 4 .1 • Example s o f S ix S igma Tra in ing and Bene f i t s

Company Time Period

Number of Master Black Belts

Trained

Number of Black Belts

Trained

Number of Green Belts

Trained

Monetary Benefi ts from

Six Sigma ($M)

Air Canada 2002–2005 11 51 1200 $450

American Express 2002 $200

American Standard 2000–2004 44 673 4302 $170

Cummins 2000–2005 65 500 $1000

Merrill Lynch 2001–2005 20 406 874

Sun Microsystems 2000–2005 6 122 207 $1170

Tyco International 2002–2005 263 870 $800

c04.indd 123c04.indd 123 25/09/12 10:50 AM25/09/12 10:50 AM

124 C h a p t e r 4 : P r o c e s s I m p r o v e m e n t : M i n i m i z i n g V a r i a t i o n T h r o u g h S i x S i g m a

To put our discussion into perspective, we begin with a brief overview of alterna- tive approaches to process improvement. Based on this overview, we then turn our attention to the fi rst process improvement strategy, Business Process Design. This is then followed by a detailed discussion of the second process improvement approach—Six Sigma ’s DMAIC improvement process. Following this overview of the DMAIC approach, each of its phases is discussed in more detail and representative Six Sigma tools are overviewed. The chapter concludes with a discussion of Six Sigma in practice. In the next chapter, we continue our discussion of process improvement and address a third process improvement approach, namely, the lean approach to process improvement. The trend toward integrating Six Sigma and lean will also be discussed in the next chapter.

A P P R O A C H E S F O R P R O C E S S I M P R O V E M E N T The appropriate process improvement strategy to employ depends on the nature of the challenge to be addressed. Figure 4.1 provides a roadmap for selecting the appropriate process improvement strategy. As is shown in the fi gure, Business Process Design or Design for Six Sigma is the appropriate process improvement strategy to employ in situations when it is determined that the process is fundamen- tally fl awed or when a brand-new process must be created. In cases where the proc- ess is fundamentally fl awed, it is best to start with a clean slate and redesign the process from scratch. In cases where the process is not fundamentally fl awed but there are opportunities to improve it, we must consider the nature of the problem to determine the appropriate process improvement strategy. In cases where it is deter- mined that there is too much variation in the process, the Six Sigma methodology is appropriate. In cases where it is determined that the effi ciency of the process needs to be improved, lean is the appropriate process improvement strategy. Of course, it is also common for a process to suffer from both too much variation and waste, in which case a combination of Six Sigma and lean tools can be applied to the process improvement initiative. This is often referred to as Lean Sigma. We now turn our attention to discussing Business Process Design and Six Sigma in more detail. Lean is the topic of the next chapter.

Figure 4.1 Alternative process design and improvement strategies.

Is the process

fundamentally flawed or being

designed?

What is the nature of the

problem?

No

Yes

Business Process Design or Design for

Six Sigma Six Sigma (DMAIC)

Too much variation

Too much waste

Lean

c04.indd 124c04.indd 124 25/09/12 10:50 AM25/09/12 10:50 AM

125B u s i n e s s P r o c e s s D e s i g n ( R e e n g i n e e r i n g )

B U S I N E S S P R O C E S S D E S I G N ( R E E N G I N E E R I N G ) Business Process Design (BPD) is the appropriate strategy for processes that require improvements beyond what can be done via incremental enhancements or in situa- tions where a new process that does not currently exist must be developed. BPD is often needed when there is a major advance in technology and/or a major shift in customer requirements. It may also be needed in cases where a process has not been improved over a long period of time. BPD is perhaps most commonly referred to as reengineering , but a wide variety of other names are also frequently used, such as business process reengineering, business process engineering, business process innovation, and business process design (BPD) or redesign . To compound the confu- sion, often managers incorrectly use terms such as downsizing and restructuring interchangeably with BPD.

To help put BPD in perspective, consider that the roots of the functional organiza- tion date back to the late 1700s, when Adam Smith proposed his concept of the divi- sion of labor in An Inquiry into the Nature and Causes of the Wealth of Nations (1776). Referring to the 17 operations required to produce a pin, Smith argued that assigning one task to each of 17 workers would be more effi cient and would pro- duce more pins than having 17 workers each autonomously perform all 17 tasks.

Although there have been dramatic advances in technology and signifi cant shifts in customer requirements since Smith fi rst proposed the division of labor concept, it is only recently that organizations have begun to challenge the concept and look for better ways to organize and integrate work. Indeed, if you were to compare how companies are organized today and how they were organized 20 or 30 years ago, you would fi nd that little has changed in their organizational structures. This is true despite technological advances such as personal computers, fax machines, smart phones, laser printers, the Web, spreadsheets, word processors, client–server com- puting, e-mail, tablet computers, and Wi-Fi, to name a few.

Initially, when these technologies were fi rst adopted by organizations, the dra- matic improvements in performance that were expected did not materialize. One popular explanation for this is that organizations were not taking advantage of the capabilities the new technologies offered. Rather, companies were simply using technology to speed up and automate existing practices. Clearly, if an activity or a set of activities is not effective to begin with, performing it faster and with less human intervention does not automatically make it effective.

For instance, one major fi nancial institution reported that more than 90 steps were required for an offi ce worker to get offi ce supplies. These steps mostly involved fi ll- ing out forms and getting the required signatures. Given the capabilities of informa- tion technology, it is certainly true that these steps could be automated and speeded up. For example, an information system could be developed to generate all the forms automatically and then automatically e-mail them to the appropriate person for authorization. However, is automating all these steps the best solution? Might it not make more sense to eliminate most of them? Consider that even if the forms are gen- erated and dispatched faster, valuable managerial time is still being used to examine and approve these requests every time an employee needs a pad of paper or ball- point pen. Indeed, when the cost of the controls is weighed against the benefi ts, it might be much more effective to give employees access to the supply cabinet to

c04.indd 125c04.indd 125 25/09/12 10:50 AM25/09/12 10:50 AM

126 C h a p t e r 4 : P r o c e s s I m p r o v e m e n t : M i n i m i z i n g V a r i a t i o n T h r o u g h S i x S i g m a

retrieve their own supplies as needed. Dr. Michael Hammer uses the term paving cow paths to describe organizations that simply implement a new technology without considering the capabilities it offers to perform work in entirely new and better ways.

There are several themes that generally underlie BPD. First, BPD ’s primary objec- tive is improved customer service. With BPD, the goal is not to improve the effi - ciency or effectiveness of a process simply for the sake of doing so. Rather, all improvements must ultimately translate into benefi ts the customer cares about. This brings us to a second theme associated with BPD, a concern with making quantum improvements in performance rather than small, incremental improvements. Reducing errors by 10 percent would generally be considered an incremental improvement. Reducing errors by 75 percent or more is a quantum improvement.

A third important theme of BPD is the central role of technology. When many of the new information technologies were initially adopted by companies, the expected improvements in organizational effi ciency and effectiveness often did not material- ize. On closer examination, it was discovered that many companies were adapting new technology to fi t current business practices rather than attempting to take advantage of the capabilities offered by the technology to perform activities in per- haps entirely different and better ways. The early 1990s marked the beginning of the reengineering movement—companies started to consider the capabilities that tech- nology offered in relationship to the way work was performed and organized.

Michael Hammer and Steven Stanton, in The Reengineering Revolution (1995), defi ne reengineering as “the fundamental rethinking and radical redesign of busi- ness processes to bring about dramatic improvements in performance” (p. 3). The keywords radical, redesign, process , and dramatic are particularly important to understanding the concept of reengineering or BPD. The word radical is used to signify that the purpose of BPD is to profoundly change the way work is performed, not to make superfi cial changes. It has to do with understanding the foundation upon which work is based and eliminating old ways that no longer make sense. In other words, it refers to reinventing the way work is performed and organized, not simply improving it. Radically changing work is often best accomplished by starting with a clean slate and making no assumptions about how work activities are performed.

The second keyword, redesign , denotes the fact that BPD is concerned with the design of work. Typically people think of design as being primarily applicable to products. However, the way work is accomplished can also be designed. In fact, Hammer and Stanton point out that having intelligent, capable, well-trained, moti- vated employees is of little value if work is badly designed to begin with.

The third keyword is process . Although all organizations perform processes, it was not until recently that they began organizing work on the basis of these proc- esses. Partly as a result of total quality management (TQM, discussed in Chapter 1), companies began to focus more on meeting customers’ needs. As they did this, they soon realized that customers are not particularly interested in the individual activities that are performed to create a product or service. Rather, they are more concerned about the fi nal result of these activities. Of course, because companies were not organized on the basis of their processes, they were not managed on the basis of processes either. Therefore, no one was assigned responsibility for the entire process that created the results of interest to the customer. Using the scenario of product design, a typical company would have departmental managers to oversee market research, manufacturing, and customer service. However, there was no manager responsible for ensuring that the results of all these activities were meeting customers’

c04.indd 126c04.indd 126 25/09/12 10:50 AM25/09/12 10:50 AM

127B u s i n e s s P r o c e s s D e s i g n ( R e e n g i n e e r i n g )

requirements. We use the term process-centered to refer to companies that have organized their work activities on the basis of specifi c value-creating processes.

The last keyword is dramatic . BPD is concerned with making quantum improve- ments in performance, not small or incremental improvements. Thus, BPD focuses on achieving breakthroughs in performance. A company that lowers its lead time by 10 percent from the previous year does not exemplify a dramatic improvement. On the other hand, a company that reduces its lead time from three weeks to three days does.

To illustrate these concepts, consider the experiences of IBM Credit Corporation. IBM Credit is in the business of fi nancing purchases of IBM offi ce equipment. Numerous companies—including General Motors, Ford, Chrysler, and General Electric—are in the lending business. These companies have found that operating fi nancial units can be extremely profi table in addition to offering customers a higher level of service.

Originally, IBM Credit was organized into functional departments. The steps involved in processing a credit request are shown in Figure 4.2 . The process began when an IBM sales rep closed a deal and the customer wanted to fi nance the pur- chase through IBM Credit. In this case the sales rep would relay the pertinent infor- mation to one of 14 order loggers at IBM Credit. The order loggers sat in a conference room and manually wrote down on pieces of paper the information supplied by the sales reps. Periodically during the day, the pieces of paper were carted upstairs to the credit department. Employees in the credit department entered the pertinent information into a computer to check the borrower ’s creditworthiness. The results of this check were then recorded on another piece of paper.

Next, the documents would be transferred to the business practices department. This department would modify the standard loan covenant in response to specifi c requests by customers. The business practices department used its own computer system. After being processed in the business practices department, the documents

Figure 4.2 Processing credit requests at IBM credit.

Credit department

Business practices

department

Order logged

Pricer

Administrator

c04.indd 127c04.indd 127 25/09/12 10:50 AM25/09/12 10:50 AM

128 C h a p t e r 4 : P r o c e s s I m p r o v e m e n t : M i n i m i z i n g V a r i a t i o n T h r o u g h S i x S i g m a

were transported to the pricing department, where pricers entered the data into a program running on a PC to determine the appropriate interest rate. Finally, the entire dossier was transported to an administrator, who converted all the informa- tion into a “quote letter.” The quote letter was then sent by Federal Express to the fi eld sales rep.

The sales reps were extremely dissatisfi ed with this process. First of all, the entire process took an average of six days and sometimes as long as two weeks. What salesperson wants to give his or her customers two weeks to think over a purchase? On top of this, when a sales rep called to check on the status of a customer ’s credit request, often the request could not even be located.

As a result of complaints from the sales reps, a manager at IBM Credit decided to investigate the problem. The fi rst thing this manager wanted to determine was how much work time actually went into processing a credit request. To determine this, the manager employed the strategy of “becoming the part”—or in this case, “becom- ing the loan request”—by walking an actual request through the entire process. First, he recorded the time it took to log an actual order. Then he took the order that was just called in and personally carried it to the credit department. Arriving at the credit department, he selected a worker at random and told the worker to stop what he or she was currently working on and perform the credit check. After repeating this in the other departments, the manager determined that the actual processing time of a credit request was about 90 minutes. Thus, out of an average of six days, each appli- cation was being processed for only about 90 minutes, indicating a signifi cant oppor- tunity for improvement.

IBM Credit ’s approach to improving this process was to combine all these activi- ties into one job called a deal structurer . Thus, one worker handled all the activities required to process a credit request, from logging the information to writing the quote letter. As a result of using deal structurers, turnaround times were reduced to an average of four hours. Furthermore, with a small reduction in head count, the number of deals processed by IBM Credit increased 100 times (not 100 percent). Do these results qualify as dramatic?

Given these results, you may wonder why IBM Credit had ever adopted a func- tional organizational structure in the fi rst place. To answer this, let ’s put ourselves in the shoes of a manager at IBM Credit. Suppose we were asked to develop an organi- zation to process credit requests. One requirement that might occur to us is that the process should be able to handle any possible type of credit request. Given this requirement, if you look again at Figure 4.2 , you will see that IBM Credit ’s original functional arrangement accomplishes this objective. For example, no matter how diffi cult checking a particular borrower ’s creditworthiness might be, the process could handle it because everyone in the credit department was a highly trained spe- cialist. The same is true of all the other departments. However, another important question is: How often will this specialized knowledge be needed? In other words, what percent of the credit requests are relatively routine and what percent require deep, specialized knowledge? As IBM found out, the vast majority of credit requests could be handled relatively routinely.

Another explanation for why IBM Credit originally created a functional organiza- tion relates to the technology that was available at the time. A key ingredient that allowed IBM Credit to move to the deal-structurer model was advances in technol- ogy. For example, spreadsheets, databases, and other decision support tools were adopted so that the deal structurers could quickly check interest rates, access

c04.indd 128c04.indd 128 25/09/12 10:50 AM25/09/12 10:50 AM

129S i x S i g m a a n d t h e D M A I C I m p r o v e m e n t P r o c e s s

standard clauses, and check the creditworthiness of the borrowers. In effect, the new technology allowed the deal structurers, who had only general knowledge, to func- tion as though they had the specialized knowledge of an expert in a particular discipline.

S I X S I G M A A N D T H E D M A I C I M P R O V E M E N T P R O C E S S The Six Sigma concept was developed by Bill Smith, a senior engineer at Motorola, in 1986 as a way to standardize the way defects were tallied. As you probably already know, sigma is the Greek symbol used in statistics to refer to standard devia- tion, which is a measure of variation. Adding “six” to “sigma” combines a measure of process performance ( sigma ) with the goal of nearly perfect quality ( six ). More spe- cifi cally, to some the term Six Sigma literally translates into making no more than 3.4 mistakes (defects) per 1 million opportunities to make a mistake (defect).

While Six Sigma ’s original defi nition of 3.4 defects per million opportunities is a rather narrow measure of quality, Six Sigma itself has evolved and now encompasses a broad methodology for designing and improving business processes. In fact, many organizations (e.g., B of A and GE) view Six Sigma as an integral part of their overall business strategy. In the popular book The Six Sigma Way (Pande et al., 2000), Six Sigma is defi ned as:

a comprehensive and fl exible system for achieving, sustaining and maximizing business success. Six Sigma is uniquely driven by close understanding of customer needs, disciplined use of facts, data, and statistical analysis, and diligent attention to managing, improving, and reinventing business processes. (p. xi)

At Motorola, Six Sigma is defi ned as “a business improvement process that focuses an organization on customer requirements, process alignment, analytical rigor, and timely execution.” 1 While numerous additional defi nitions of Six Sigma could be cited, common themes tend to emerge across the range of suggested defi nitions, including rigorous (often statistical) analysis, customer focus, data-driven analysis, and improvement of overall business performance. Likewise, a number of benefi ts are commonly associated with Six Sigma initiatives, including increased profi tability, improved quality, improved employee morale, lower costs, higher productivity, mar- ket share growth, improved levels of customer retention and satisfaction, and shorter lead times. Interestingly, Motorola became the fi rst company to win the Malcolm Baldrige National Quality Award in 1988. Furthermore, Motorola estimated that as of 2006 it had saved in excess of $17 billion as a result of its Six Sigma initiatives.

Arguably, one reason for the success of Six Sigma programs where others have failed is that Six Sigma provides a structured, logical, and disciplined approach to problem solving. More specifi cally, as shown in Figure 4.3 , Six Sigma projects gener- ally follow a well-defi ned process consisting of fi ve phases. The phases are d efi ne, m easure, a nalyze, i mprove, and c ontrol, which are collectively referred to as

1 Retrieved September 20, 2004, from www.motorola.com/content/0,2409-4904,00.html.

c04.indd 129c04.indd 129 25/09/12 10:50 AM25/09/12 10:50 AM

130 C h a p t e r 4 : P r o c e s s I m p r o v e m e n t : M i n i m i z i n g V a r i a t i o n T h r o u g h S i x S i g m a

DMAIC (pronounced dey-MAY-ihk). As the names of the phases suggest, the DMAIC improvement process can be thought of as an adaptation of the scientifi c method to process improvement.

Before discussing the DMAIC phases in more detail, a couple of comments are in order. First, as shown in Figure 4.3 , the phases in a DMAIC project often serve as project milestones and can thus be used as tollgates to the next phase in the project. In particular, the progress and outcomes associated with the project are evaluated at the end of each phase to assess the merits of permitting the project to move on to the next phase. The extent to which organizational resources will continue to be allocated to the project is typically assessed at these milestones as well.

Figure 4.3 The Six Sigma DMAIC approach for process improvement.

• goals for process improvement • customer requirements • project scope • the problem/opportunity

Define

• Identify appropriate performance measures • Collect data • Evaluate current process performance

• Develop and test theories related to root causes of problems • Identify cause-and-effect relationships

• Develop, evaluate, and implement solutions to reduce gap between desired process performance and current performance

• Monitor process to sustain improved performance • Ensure that problems do not resurface

Define

Measure

Analyze

Improve

Control

c04.indd 130c04.indd 130 25/09/12 10:50 AM25/09/12 10:50 AM

131S i x S i g m a a n d t h e D M A I C I m p r o v e m e n t P r o c e s s

Second, there are large numbers of standard Six Sigma tools and methodologies that are used at various phases in a DMAIC project. Table 4.2 summarizes some fre- quently used Six Sigma tools and methodologies and lists the DMAIC phases where these tools/methodologies are most commonly used. Before turning our attention to discussing each DMAIC phase in more detail, we fi rst provide a detailed example of an actual Six Sigma project.

T A B L E 4 .2 • Common Too l s and Me thodo log ie s in the S ix S igma Too lk i t Six Sigma Tool/Methodology DMAIC Phase(s) Most Commonly Used in

Affi nity diagram D, A

Benchmarking D, M

Brainstorming A, I

Business case D

Cause-and-effect diagrams M, A

Control charts M, A, I, C

Critical to quality tree D

Data collection forms M, A, I, C

Data mining M

Design for Six Sigma (DFSS) An entire collection of tools/methodologies that can be used across all phases

Design of experiments (DOE) A, I

Defects per million opportunities (DPMO) M

Failure modes and effects analysis (FMEA) M, I, C

Gantt chart Tool used to manage entire DMAIC project

Kano model D, M

Lean tools An entire collection of tools/methodologies that can be used across all phases

Measurement systems analysis (gage R&R) M

Nominal group technique D, M

Pareto analysis D, M, A, I

Process capability M, A, I

Process maps D, M, A, I, C

Process sigma M, I

Project charter D

Quality function deployment (QFD) D, M

Regression A

Rolled throughput yield (RTY) D, M, A

Simulation A, I

SIPOC D

Stakeholder analysis D, I

Theory of constraints (TOC) One of the lean tools

Voice of customer (VOC) D

c04.indd 131c04.indd 131 25/09/12 10:50 AM25/09/12 10:50 AM

132 C h a p t e r 4 : P r o c e s s I m p r o v e m e n t : M i n i m i z i n g V a r i a t i o n T h r o u g h S i x S i g m a

E X A M P L E S I X S I G M A P R O J E C T With increasing patient volumes, the Northshore University Hospital in Manhasset, New York, initiated a Six Sigma project in 2004 to reduce the bed assignment turna- round time. The bed turnaround time is the elapsed time from when the discharge instructions are given to a patient to the time the admission nurse is notifi ed that a clean bed is ready. Within six months, the Six Sigma team was able to reduce the average bed turnaround time by over 2 hours.

The Six Sigma team began the defi ne phase by developing a process map that identifi ed all the steps in the process from the time the patient received the discharge instructions to the time the admission nurse was notifi ed that a clean bed was ready. The team also defi ned the admissions nurses as the customer of the process and surveyed these nurses to collect voice of the customer data. Based on the voice of the customer data collected, the team established a target bed turnaround time of 2 hours and an upper specifi cation limit of 2.5 hours.

In the measure phase, the team developed the operational defi nition of a defect as any case where more than 2.5 hours were required to turn around a bed. The team also conducted a measurement systems analysis to verify the effectiveness of its measurement system and hence the data collected to study the process. In addi- tion, the team calculated the defects per million opportunities (DPMO) of the cur- rent process. The team discovered that the average bed turnaround time was over 3.75 hours and that the DPMO was 672,725. Finally, the team created a cause-and- effect diagram to identify the variables that infl uenced the bed turnaround times.

In the analyze phase, the team performed hypothesis tests and used analysis of vari- ance to identify variables that were statistically signifi cant. As the team analyzed the data, they discovered communication and technical problems in two key steps. To address these problems, the team developed four recommendations in the improve phase. For example, one recommendation related to revising the discharge assessment sticker to include additional information. Another recommendation was to better utilize the admis- sion RN beepers so they would be immediately notifi ed of when a clean bed was ready.

In the control phase, control charts were created to monitor bed turnaround times. The control charts helped ensure that the improvements made to the process continued and also served as an early detection system should performance begin to deteriorate.

This example illustrates another characteristic of Six Sigma projects. Namely, with Six Sigma, we let the story naturally unfold as we objectively analyze the data. This means that, as diffi cult as it is to do at times, we resist the tendency to try to solve the problem as soon as it is defi ned. Rather, we engage in a process of progressively gain- ing additional insight into the root causes of the problem by sequentially answering new questions as they arise. Thus, as the insights we gain from investigating a particu- lar issue often naturally lead to new questions, we utilize the tools in the Six Sigma toolkit to help fi nd answers to these new questions. Through this process of applying the Six Sigma methodology and tools, we ultimately gain a clear understanding of the root causes of the problem, which in turn well positions us to address it.

It is also worth pointing out that often the hardest part of solving a problem is sim- ply fi guring out where to start. The DMAIC approach not only provides a disciplined approach for solving problems but also adds structure to what otherwise might appear to be an unstructured problem. Of course, it is possible that what we learn in one phase of a Six Sigma project requires that we revisit an earlier phase. We now turn our attention to discussing each of the phases of a Six Sigma project in more detail.

c04.indd 132c04.indd 132 25/09/12 10:50 AM25/09/12 10:50 AM

133T h e D e f i n e P h a s e

T H E D E F I N E P H A S E The defi ne phase of a DMAIC project focuses on clearly specifying the problem or opportunity, determining the goals for the process improvement project, and identi- fying the scope of the project. Identifying the customers and their requirements is also critical, given that the overarching goal for all Six Sigma projects is improving the organization ’s ability to meet the needs of its customers. In this section we over- view two tools commonly used in the defi ne phase of a DMAIC project: benchmark- ing and quality function deployment (QFD).

Benchmarking In conjunction with their efforts to improve their products and processes, many organizations are engaging in an activity called benchmarking . Essentially, bench- marking involves comparing an organization ’s processes with the best practices to be found. Benchmarking is used for a variety of purposes, including the following:

• Comparing an organization ’s processes with the best organization ’s processes. When used in this way, benchmarking activities should not be restricted to other organizations in the same industry. Rather, the companies that are best in the world at performing a particular activity, regardless of industry , should be studied. For example, Xerox used L.L. Bean to benchmark the order fulfi llment process.

• Comparing an organization ’s products and services with those of other organizations.

• Identifying the best practices to emulate. • Projecting trends in order to be able to respond proactively to future chal-

lenges and opportunities.

Benchmarking generally involves three steps. The fi rst step is concerned with pre- paring for the benchmarking study. In this phase it is important to get the support of senior management and its input on what should be benchmarked. Problem areas, activities related to serving the customer better, and activities related to the mission of the organization are all appropriate candidates for inclusion in the benchmarking study.

The second phase of benchmarking consists of collecting data. There are two general sources of benchmarking data. One source is published data . These are often available from universities, fi nancial fi lings (e.g., 10k reports), consultants, periodicals, trade journals, and books. The other source of data is original research conducted by the organization itself. If this approach is employed, a list of organiza- tions to benchmark might include companies that have recently received quality awards or other business awards, are top-rated by industry analysts, have been the subject of recent business articles, or have a track record of superior fi nancial performance. Once the companies have been identifi ed, data can be collected in a variety of ways, including interviews, site visits, and surveys.

The third and fi nal phase of benchmarking involves using what has been learned to improve organizational performance. Once the second phase has been com- pleted, identifi ed gaps in performance can be used to set challenging but realistic goals (often called stretch goals ). Also, the results of the benchmarking study can be used to overcome and eliminate complacency within the organization.

c04.indd 133c04.indd 133 25/09/12 10:50 AM25/09/12 10:50 AM

134 C h a p t e r 4 : P r o c e s s I m p r o v e m e n t : M i n i m i z i n g V a r i a t i o n T h r o u g h S i x S i g m a

Quality Function Deployment Arguably, two key drivers of an organization ’s long-term competitive success are the extent to which its new products or services meet customers’ needs and having the organizational capabilities to develop and deliver such new products and serv- ices. Clearly, no amount of clever advertising and no degree of production effi ciency will entice customers to continue to purchase products or services that do not meet their needs. Likewise, it serves no purpose for an organization to design new products or services that it does not have the capability to produce or deliver. To illustrate, it would make little sense for a local phone company to market a new service that offers voice over the Internet (VoIP) calling if the fi rm did not have the infrastructure to deliver this type of service. Even if the phone company was able to work out the bugs for such calling, it could still take years to install the hardware that is necessary to deliver this type of service on a large scale. Of course, the desire to offer new products and services can serve as the impetus for acquiring additional process capabilities; however, organizations typically seek to develop new products and services that capitalize on their existing capabilities.

Quality function deployment (QFD) is a powerful tool for helping translate customer requirements into process capabilities. In effect, the use of QFD ensures that newly designed or improved products and services satisfy market requirements and are ulti- mately producible by the fi rm. As Figure 4.4 illustrates, the QFD methodology utilizes a series of tables to maintain links among customer requirements, technical requirements, component requirements, process requirements, and ultimately specifi c process activi- ties. Because of their shape, these tables are often referred to as the houses of quality . Before discussing the contents of a house of quality in detail, we fi rst broadly overview the QFD process and discuss the links between the four houses of quality.

Broad Overview of QFD

QFD begins by using voice of the customer (VOC) data to specify the customer requirements in the rows of the fi rst house, the output planning matrix, shown in Figure 4.4 . The name VOC stems from the fact that the customer ’s own language is used to capture these requirements. As examples, a sample of mountain bike riders might offer responses regarding their preferences for a new bike such as “the bike should shift effortlessly,” “there should be no bob on climbs,” “the bike should climb,

Figure 4.4 Quality function deployment process.

Process deployment

matrix

Process activities

P ro

ce ss

re q

u ir

em en

ts

Process planning matrix

Process requirements

C o

m p

o n

en t

ch ar

ac te

ri st

ic s

Output specification

matrix

Component characteristics

Te ch

n ic

al re

q u

ir em

en ts

Output planning matrix

Technical requirements

V o

ic e

o f

cu st

o m

er

c04.indd 134c04.indd 134 25/09/12 10:50 AM25/09/12 10:50 AM

135T h e D e f i n e P h a s e

descend, and handle great,” “the bike should suck up the bumps,” and “I like a bike that is well balanced with a low center of gravity.”

Next, based on the customer requirements listed in the rows, the technical require- ments for the product or service are generated and listed in the columns of this house. While the customer requirements are expressed in the customer ’s own language, the technical requirements are often expressed in a more specialized language such as that used by engineers. Thus, technical requirements for a product might be expressed in terms of dimensions, weights, performance, tensile strength, and compression.

Once the rows and columns are generated for the output planning matrix, the relationship matrix in the middle of the house is completed. The cells in the relation- ship matrix correspond to the intersection of a particular customer requirement and technical requirement. In each cell the strength of the relationship between the cor- responding customer requirement and technical requirement is evaluated. An impor- tant use of the relationship matrix is to ensure that each customer requirement is addressed by one or more technical requirements. Likewise, the relationship matrix can be used to ensure that designers do not add technical requirements to the prod- uct or service that do not address specifi c customer requirements. Designers who do not specifi cally consider customers’ requirements run the risk of adding a number of “bells and whistles” that customers may not be interested in. In these cases the designers are simply adding to the cost of the fi nal product or service without pro- portionally increasing its value. For example, it might be an interesting challenge for an engineer to design a fi ve-speed motor for garage doors. However, since custom- ers would most likely operate it at only its fastest speed, adding the extra controls for additional speeds would not add value for the typical customer.

In the next house of quality, referred to as the output specifi cation matrix in Figure 4.4 , the technical requirements that were the columns in the previous house of quality now become the rows. Thus, in the output specifi cation matrix the task now becomes generating a list of the elemental or component characteristics for the product or service that will become the columns in this house of quality. Then, once the columns are specifi ed, the relationship matrix is completed to ensure that (1) each technical requirement is addressed by one or more component pieces of the product or service and (2) component requirements that are not related to specifi c technical requirements are not added.

In the third house shown in Figure 4.4 , the process planning matrix, component requirements listed in the columns of the previous house now become the rows of the new house. Next, process requirements are generated based on the component requirements listed in the rows and entered in the columns of the new house. Following this, the relationship matrix is completed for this house to ensure that each component requirement is matched to specifi c process requirements and that each process requirement is linked to specifi c component requirements. The former result ensures that the organization has the capability to produce or deliver the components; the latter result ensures that the organization does not attempt to develop process capabilities that are not related to the component requirements.

Finally, in the process deployment matrix, the process requirements from the columns of the previous house become the rows and specifi c process activities are generated for each process requirement. Then, in a similar fashion to the other houses, the relationship matrix is completed and checks are performed to ensure that all process requirements are addressed by one or more process activities and that no process activities are added that do not address at least one process requirement.

c04.indd 135c04.indd 135 25/09/12 10:50 AM25/09/12 10:50 AM

136 C h a p t e r 4 : P r o c e s s I m p r o v e m e n t : M i n i m i z i n g V a r i a t i o n T h r o u g h S i x S i g m a

Thus, we see that QFD provides a logical and straightforward approach for ensur- ing that designs for new or improved products and services meet customers’ require- ments and are ultimately producible. This is accomplished by fi rst translating the voice of the customer into the technical language of engineers and other specialists.

Next, these technical requirements are translated into specifi c requirements for the components of the new product or service. The component requirements are subsequently translated into specifi c process requirements, which in turn are trans- lated into specifi c process activities. In the end, however, it is the customer require- ments that drive the entire QFD process.

House of Quality Details

With this general overview of QFD and its four houses of quality, we now turn our attention to the specifi c information listed in each house of quality. A summary of the general structure of a house of quality is shown in Figure 4.5 . Constructing a house of quality begins by listing what it is we are trying to accomplish in the rows at the far left of the house. In the fi rst house, these “whats,” as they are generally called, are the customer requirements or the voice of the customer. After specifying what it is we would like to accomplish, the next task is to think about how to meet these require- ments. Thus, the “hows,” as they are called, are listed in the columns of the house. In the case of the fi rst house of quality, the hows are the technical requirements of the product or service given the customer requirements listed in the rows. The intersec- tion of each what (row) and how (column) is a cell in the relationship matrix. In this

Figure 4.5 The house of quality.

Hows

Interrelationships between Hows

Target values

Competitive evaluation

Importance weights

Relationship matrix

Whats Importance rating

Competitive evaluation

c04.indd 136c04.indd 136 25/09/12 10:50 AM25/09/12 10:50 AM

137T h e D e f i n e P h a s e

relationship matrix, the strength of the relationship between each what and how is evaluated. Likewise, the roof of the house corresponds to a triangular correlation table where the correlations between the technical requirements are assessed. This assessment helps identify those technical requirements that are synergistic with each other and that confl ict with each other and therefore where a tradeoff may exist. At the far right of the house, customer importance ratings and the results of a competi- tive evaluation are summarized. Finally, at the bottom of the house, target values, a competitive evaluation, and importance weights are summarized for each how.

To illustrate the QFD process, consider a fast-food restaurant chain that is inter- ested in improving its offerings. Figure 4.6 provides the completed output planning matrix for the chain. At the far left of the house, the voice of the customer data are listed, including customer statements such as “food that tastes good” and “get what I ordered.” Based on these customer requirements, a list of technical requirements was generated and listed in the columns. In the relationship matrix, the relationship between each customer requirement and technical requirement was assessed. Thus, we see that that there is a strong relationship between the taste of the food and the

1 2 3

Competitive evaluation

4 51 2 3

Importance ratings

Technical Requirements Fo

o d

c o

rr ec

t te

m p

er at

u re

4 5

95

87

88

4

A ve

ra ge

m ea

l p ri

ce <

$ 4

85

90

87

2

C o

rr ec

t o

rd er

98

86

84

12

Fo o

d d

el iv

er ed

in <

1 m

in

90

78

83

10

C le

an B

R a

n d

e at

in g

ar ea

s

95

77

75

4

C o

u rt

eo u

s st

af f

95

82

87

10

Li m

it f

o o

d s

h ig

h in

f at

a n

d c

ar b

s

92

70

80

10

P ro

p er

ly t

ra in

ed f

o o

d p

re p

ar er

s

95

80

82

15

U se

q u

al it

y in

gr ed

ie n

ts

85

70

70

18

U se

f re

sh in

gr ed

ie n

ts

Food that tastes good

Voice of Customer

Strong relationship

Food that is reasonably nutritious

Friendly employees

Clean restaurant

Short wait for food

Get what I ordered

Reasonable price

Moderate relationship

Weak relationship

90

70

75

15

Strong positive Positive Negative Strong negative

Target values

Competitive evaluation:

Competitor A

Competitor B

Importance weights

Competitor A

Competitor B

Us

Figure 4.6 Example output planning matrix for fast-food restaurant chain.

c04.indd 137c04.indd 137 25/09/12 10:50 AM25/09/12 10:50 AM

138 C h a p t e r 4 : P r o c e s s I m p r o v e m e n t : M i n i m i z i n g V a r i a t i o n T h r o u g h S i x S i g m a

use of fresh ingredients, while there is only a moderate relationship between the taste of the food and the time it takes to make and deliver it to the customer. In the roof of the house, the correlations between each of the technical requirements are evaluated and listed. In our example we note that the requirements of “fresh ingre- dients” and “quality ingredients” are consistent with each other, while a tradeoff exists between “limiting the fat and carbohydrate content of the food” and “keeping meal price down.” At the far right, customer importance ratings and a competitive evaluation are summarized. Thus we see that good taste, nutrition, and accuracy of the order are the most important aspects of the service to the customers surveyed. On these three dimensions we see that the organization in question has the best- tasting food, has the second most nutritional offerings, and has the highest order accuracy. At the bottom of the house, target values and a competitive evaluation are listed for each how. For example, in terms of using fresh ingredients, the chain in question has set a goal of achieving a score of 90 on this dimension and estimates that competitors A and B score 70 and 75, respectively, on this dimension. Finally, the importance weight or priority for each how is listed at the very bottom of the house. This can help in making tradeoffs when confl icts are discovered in the roof of the house. In the present case the most important technical requirement is the use of quality ingredients.

The other houses of quality are completed in a similar fashion. The process begins with the columns of the current house becoming the rows of the subsequent house. Then columns and the other information are determined for the new house.

One key advantage of QFD is that it is a visual tool. Through the use of QFD, a fi rm can analyze its outputs in terms of customers’ desires, compare its outputs with competitors’ outputs, determine what it takes to better meet each customer ’s require- ments, and fi gure out how to do it. In addition, it provides a means of linking these customer requirements through the entire planning process, ending with the specifi - cation of detailed process activities. A number of fi rms such as Toyota and Hewlett- Packard have adopted QFD and found that it cut their product development time by one-third to one-half and their costs by up to 60 percent (while improving quality).

T H E M E A S U R E P H A S E Typically the measure phase begins with the identifi cation of the key process per- formance metrics. Correctly choosing process performance metrics is critical in order to have an accurate picture of how the process is actually performing in terms of meeting customer requirements. Unfortunately, it is not uncommon for managers to select performance metrics based on their ease of measurement and/or the availa- bility of data, not on their ability to provide insights into how the process is meeting customer requirements. For example, some organizations use machine utilization to assess the performance of their manufacturing processes. In reality, machine utiliza- tion has at best an indirect relationship to what really matters to customers—shorter lead times, higher quality, the percent of orders shipped on time, and so on. As a service example, consider a call center. In this case, performance measures such as the percent of calls answered by the third ring, the percent of calls processed with- out having to be escalated, and the average hold time are all better indicators of how well the process is performing than, for example, labor utilization.

c04.indd 138c04.indd 138 25/09/12 10:50 AM25/09/12 10:50 AM

139T h e M e a s u r e P h a s e

To understand the implications associated with a shift in the process mean, con- sider a restaurant that buys frozen hamburger patties with a target weight of 4 ounces. The supplier has provided data that the average weight of its hamburger patties is 4.0 ounces and the process standard deviation is 0.1 ounce. According to Motorola ’s assumption, the average weight of a hamburger could change over time. More specifi - cally, as shown in Figure 4.8 , and based on Motorola ’s assumption of a potential shift in the process mean, the average weight of a hamburger could drop to as low as 3.85 ounces or increase to 4.15 ounces. Note that we are referring to the average weight of

Once the key process performance metrics have been specifi ed, related process and customer data are collected. One early use of these data is to evaluate the proc- ess ’s current performance, which can then be used as a baseline to evaluate the benefi ts of potential process improvements that are identifi ed later in the project.

As shown in Table 4.2 , there are a variety of tools in the Six Sigma toolkit that are useful during the measure phase. We begin our discussion of the measure phase with two commonly used process performance measures, namely, defects per mil- lion opportunities (DPMO) and process sigma . We then conclude our discussion of the measure phase with a brief overview of measurement systems analysis.

Defects per Million Opportunities Earlier it was noted that a literal interpretation of Six Sigma is 3.4 defects per million opportunities (DPMO). This may have caused some confusion for more statistically inclined readers, which we will now attempt to reconcile. To reconcile this difference, we need to discuss an important assumption Motorola made when it originally devel- oped the Six Sigma concept. Specifi cally, Motorola assumed that the mean of a proc- ess can shift (or drift) over time by as much as 1.5 standard deviations, as is illustrated in Figure 4.7 . In Figure 4.7 , the bold normal curve in the middle corresponds to the process mean when it is perfectly aligned with the target value, while the other two normal curves correspond to shifts in the process mean of 1.5 standard deviations both up and down. Note also that it is assumed that the shifts in the process mean do not affect the process standard deviation or customer requirements.

Figure 4.7 Motorola ’s assumption that the process mean can shift by as much as 1.5 standard deviations.

+s–s–2s

–1.5s +1.5s

–3s

Process mean shifts down 1.5 sigma

Process mean shifts up 1.5 sigma

–4s–5s–6s T

Customer requirements

+2s +3s +4s +5s +6s

c04.indd 139c04.indd 139 25/09/12 10:50 AM25/09/12 10:50 AM

140 C h a p t e r 4 : P r o c e s s I m p r o v e m e n t : M i n i m i z i n g V a r i a t i o n T h r o u g h S i x S i g m a

all hamburgers produced by the process at a particular point in time, not to the weight of individual hamburgers. Also, recall that we assume that the process standard devia- tion and restaurant ’s requirements are not affected by a shift in the process mean.

To see the impact that results from a shift in the process mean, let ’s assume that the restaurant ’s requirements for hamburgers are right at plus or minus 6 standard devia- tions from the target weight of 4 ounces, or at six sigma, as shown in Figure 4.8 . In this case the restaurant would consider a hamburger acceptable as long as it weighed between 3.4 and 4.6 ounces. It can be easily verifi ed that the area in one tail beyond 6 standard deviations is 0.0000001 percent, which yields a combined area of 0.0000002 percent in both tails. Multiplying this by 1 million yields 0.002 DPMO, which is con- siderably less than the 3.4 DPMO commonly associated with Six Sigma.

The reason for this difference concerns a shift in the process mean. For example, what happens when the process mean shifts up by 1.5 standard deviations such that the hamburgers have an average weight of 4.15 ounces, while the process standard deviation and the restaurant ’s requirements stay the same? Referring to Figure 4.7 , we observe that now the upper tail beyond the customers’ requirements would be 4.5 standard deviations above the new process mean, while the lower tail would be 7.5 standard deviations below the new process mean. For the normal distribution, the area in the tail beyond 4.5 standard deviations is 0.00034 percent, which works out to 3.4 DPMO, while the area beyond 7.5 standard deviations is 0.0000000000032 percent, which is less than 1 in a trillion and for practical purposes is zero. Therefore, we observe that when the process mean shifts by 1.5 standard deviations, the com- bined area in the tails is approximately 0.00034 percent, which is equivalent to 3.4 DPMO. Because the normal distribution is symmetrical, the same results are obtained in cases where the process mean shifts down by 1.5 standard deviations. The only difference is that the areas in the two tails are reversed. Finally, note that the 3.4 DPMO represents the worst possible performance because it corresponds to the largest shift in the process mean. Smaller shifts in the process mean yield lower DPMO values.

An important advantage of using DPMO as a measure of process performance is that it provides a standard measure of process performance. As such, it provides a mechanism for comparing the performance across a range of processes that other- wise would be diffi cult to compare. In fact, as we now illustrate, DPMO makes such

4.13.93.8

–1.5s +1.5s

3.73.63.53.4 4.0

Customer requirements

4.2 4.3 4.4 4.5 4.6

Figure 4.8 Shift in distribution of hamburger weights (after cooking) at a restaurant.

c04.indd 140c04.indd 140 25/09/12 10:50 AM25/09/12 10:50 AM

141T h e M e a s u r e P h a s e

comparisons possible across varying processes by incorporating an adjustment for the complexity of each process.

To illustrate the calculation of the DPMO and how it adjusts for process complex- ity, consider a bank that processes two types of loans. Process A is used to process relatively simple loans, such as for a car, and consists of fi ve steps; process B is used to process more complex loans, such as mortgages, and requires the completion of 25 steps. Let ’s further assume that of the last 10,000 loans processed by each process, a total of 100 errors were made in each process. In this case the number of defects per unit ( D P U )   is the same for both methods and is calculated as follows (note that here a loan represents a unit):

DPU   5   number of defects ________________ number of units

  5   100 ______ 10 , 000    5   0.01

This result suggests that each process is averaging 0.01 error (or defect) per loan, or 1 error per 100 loans. Is it then reasonable to conclude that both processes are performing at the same level? Because this comparison has not accounted for the differences between the methods in terms of their complexity, the answer is no. Thinking about this situation intuitively, we would generally expect the number of errors or defects to increase as the complexity of the process increases. Unfortunately, the DPU measure does not refl ect this logic.

To account for the differences in the complexity of the processes, an adjustment is needed. Up to this point, we have counted each loan (unit) processed as repre- senting one opportunity for a defect. In reality, there are typically multiple opportu- nities to create a defect (error). To illustrate, Figure 4.9 displays 33 specifi c defects organized into seven categories associated with staying at a hotel. As this fi gure illus- trates, for virtually all products and services there are numerous opportunities for introducing defects or making errors. Thus, rather than treating each unit or cus- tomer as a single opportunity for a defect, an alternative approach is to develop a list of all the opportunities for creating a defect for a given product or service. Then the number of defects per opportunity ( D P O )   can be calculated as follows:

DPO   5   number of defects _________________________________________________ number of units ( customers )   3   number of opportunities

An important issue that must be addressed in using the DPO measure relates to developing the list of opportunities. In particular, it is possible to make it appear that performance is better than it actually is by padding the list with additional opportu- nities. To illustrate, in Figure 4.9 there are 33 specifi c opportunities for defects listed in seven categories. Let ’s assume that a survey of 100 customers revealed 200 occur- rences of the 33 items listed in the fi gure. Based on this, if we consider each of these 33 items as a valid opportunity for a defect, then the DPO works out to be

200 ________ 100 3 33   5   0.06

Alternatively, if we consider each stage in the service delivery process (i.e., the seven categories listed in Figure 4.9 ) as an opportunity for a defect, then the DPO increases to

200 _______ 100 3 7 5   0.29

c04.indd 141c04.indd 141 25/09/12 10:50 AM25/09/12 10:50 AM

142 C h a p t e r 4 : P r o c e s s I m p r o v e m e n t : M i n i m i z i n g V a r i a t i o n T h r o u g h S i x S i g m a

Thus, we see that increasing the number of opportunities considered can make the performance look better. Along these lines, then, a less than honest manager or sup- plier could infl ate the list of opportunities for defects by including some opportuni- ties that in reality never occur. Besides being unethical, pursuing this strategy greatly undermines the value of using DPO as a standardized measure of process perform- ance. In the end, there are no fi rm rules for what to include and what not to include. As a general rule of thumb, however, it is suggested that only those defects that are meaningful to the customer be included. One strategy for determining the list of opportunities is to simply treat each stage as representing one opportunity for a defect. Based on this approach, there would be seven opportunities for a defect per

Figure 4.9 Defect opportunities associated with a stay at a hotel.

Hotel reservation Name entered incorrectly Wrong date of arrival entered Wrong departure date entered Error entering credit card number or expiration date Wrong address entered Incorrect number of people staying in room entered Wrong room reserved (e.g., smoking versus nonsmoking, number of beds) Incorrect number of baby cribs reserved Wrong room rate entered

Check-in Lost reservation Excessive wait Defective or wrong room key Desk staff not courteous No baggage carts available

Room cleaning Dirty shower Dirty linens Dirty sink Carpet not vacuumed Trash cans not emptied

Room supplies No clean towels No toilet paper No shampoo/hand soap

TV Cable out No remote control/remote control defective

Room service Late food order Missing items Billed incorrectly Food not prepared properly Food is cold

Checkout Incorrect charge for room service Incorrect telephone charges Excessive wait for desk clerk Excessive wait for bell captain

• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •

c04.indd 142c04.indd 142 25/09/12 10:50 AM25/09/12 10:50 AM

143T h e M e a s u r e P h a s e

hotel customer (of course, some of these defects could be repeated over a multiple-day stay). Thus, a hotel customer whose reservation was lost and who experienced an excessive wait for check-in would count as one check-in defect as opposed to two defects.

Based on this logic and returning to our original objective of comparing loan processes A and B, we determine that process A has 5 opportunities to create a defect while process B has 25. Then, based on the data collected indicating that 100 errors were made in both processes out of the last 10,000 loans processed, we can calculate the DPO for both processes as follows:

DPO A   5   100 __________

10 , 000 3 5 5   0.002

DPO B   5   100 ___________

10 , 000 3 25 5   0.0004

In contrast to our earlier analysis based on the DPU measure, we now observe a signifi cant difference in the performance of the two processes. This difference is now observable because we have adjusted the performance measure to account for the complexity of the processes. In particular, we see that process A produces an average of 0.002 defect per opportunity, while process B produces an average of only 0.0004 defect per opportunity. Because it is somewhat cumbersome to deal with such small numbers, it is common to multiply the DPO measure by 1 million to yield the DPMO measure discussed earlier. In this case, the DPMOs for processes A and B are 2000 and 400, respectively. Thus, it is expected that process A would make 2000 mistakes per 1 million opportunities to make a mistake, while Process B would make only 400 mistakes per million opportunities.

Process Sigma Regarding the term Six Sigma, it was noted earlier that sigma corresponds to a measure of process performance. Having clarifi ed why Six Sigma translates into 3.4 DPMO positions us to now examine how sigma itself can be used to measure the performance of a process. We now know that when customers consider an output to be of acceptable quality as long as it is within plus or minus 6 standard devia- tions from the target value and that the process mean can shift by as much as 1.5 standard deviations, the process itself will produce 3.4 DPMO. In effect, then, the inherent variability of the process itself relative to the customer requirements pro- vides a measure of the capability of the process to meet the customers’ require- ments. In other words, one way to measure the performance of a process is to calculate the number of standard deviations the customer requirements are from the process mean or target value. According to this measure, called process sigma , a higher value corresponds to higher process performance. In the examples at the beginning of the chapter, it was noted that Southside Hospital was able to increase its process sigma from 0.1 to 2 while TRW ’s legal department established a baseline process sigma level of 2.18.

To see why a higher process sigma corresponds to higher process performance, refer to Figure 4.10 . In particular, two levels of process sigma are shown in the fi gure.

c04.indd 143c04.indd 143 25/09/12 10:50 AM25/09/12 10:50 AM

144 C h a p t e r 4 : P r o c e s s I m p r o v e m e n t : M i n i m i z i n g V a r i a t i o n T h r o u g h S i x S i g m a

The normal curve drawn with the thin line corresponds to a process where the cus- tomer requirements are at plus or minus 3 of the process ’s standard deviations (3 sigma) from the target value (T) and the normal curve drawn with the thick, bold line corresponds to a process where customer requirements are at plus or minus 6 of the process ’s standard deviations (6 sigma) from the target value. From this fi gure we observe that the 3 sigma process exhibits more variation and therefore has more area in its tails beyond the customer requirements. This greater amount of area in the tails corresponds to a higher probability of producing an outcome that is not accept- able to the customer.

Table 4.3 examines the impact on process performance for a range of process sigma values for two levels of process drift. Referring back to Figure 4.10 , we see from the middle column in Table 4.3 , which assumes the process mean can shift by as much as 1.5 standard deviations, that a 3 sigma process produces 66,811 DPMO

Figure 4.10 Comparison of three sigma process and Six Sigma process.

Three sigma

process Six Sigma process

T

Customer requirements

T A B L E 4 .3 • DPMO for A l t e rna t i ve Proce s s S igma Leve l s

Process Sigma

DPMO (Based on Process Mean Shifting by Up to 1.5 Standard

Deviations)

DPMO (Based on Process Mean Shifting by Up to 1.0

Standard Deviation)

1.0 697,672 522,751

1.5 501,350 314,747

2.0 308,770 160,005

2.5 158,687 67,040

3.0 66,811 22,782

3.5 22,750 6,213

4.0 6,210 1,350

4.5 1,350 233

5.0 233 32

5.5 32 3.4

6.0 3.4 0.3

c04.indd 144c04.indd 144 25/09/12 10:50 AM25/09/12 10:50 AM

145T h e M e a s u r e P h a s e

while a 6 sigma process produces 3.4 DPMO. Note that we are defi ning a defect rather broadly here in terms of the entire unit of output being either defective or not defective.

The last column of Table 4.3 examines the impact of process stability on process performance. Specifi cally, while the DPMO values listed in the middle column of the table are based on Motorola ’s original assumption of shifts in the process mean of as much as 1.5 standard deviations, the last column is based on a process shift of no more than 1 standard deviation. As can be seen in comparing these two columns, the number of defects produced by a process can be signifi cantly reduced by increas- ing the stability of the process. For example, at 3 sigma, a process produces 66,811 DPMO when the process mean shifts by as much as 1.5 standard deviations. If the process ’s stability is increased so that the process mean shifts by no more than 1 standard deviation, the DPMO drops by 66 percent to 22,782 DPMO. Based on the results shown in the table, we point out that another way to achieve the target value of 3.4 DPMO is to operate a process at 5.5 sigma while ensuring that the process mean shifts by no more than 1 standard deviation.

As another example, 3.4 DPMO would be achieved in cases where the process operated at 5 sigma and the process mean shifted by no more than 0.5 standard deviation. This discussion illustrates that there are three drivers of process sigma: the actual customer requirements, the variation in the process as measured by the process standard deviation, and the stability of the process as measured by how much it can shift over time. Thus, the sigma of a process can be increased by widening the cus- tomer requirements surrounding the target value, reducing the variation of the proc- ess, and/or increasing the stability of the process.

Measurement Systems Analysis Whenever we deal with data that were collected via measurement, measures of vari- ation like the standard deviation and variance may not accurately refl ect the true variation in the sample or population of interest. In particular, measurement errors may introduce another source of variation. For example, measuring a person ’s blood pressure manually requires both good eyesight and good hearing. It also requires that the dial gauges be properly recalibrated over time.

As an example, consider the 10 systolic blood pressure values shown in Table 4.4 for a random sample of male diabetic patients. In particular, we note that the aver- age systolic blood pressure for this sample of patients was 130.5 with a standard deviation of 21 and variance of 442.9. As you may recall from an earlier statistics course, the purpose for measures like the standard deviation and variance is to pro- vide a sense of how much variation or dispersion there is across the population of interest or, in the present case, how much variation there is across male diabetics’ systolic blood pressure.

Figure 4.11 illustrates that the measurement system contributes to the total vari- ance that is actually calculated. More specifi cally, we note that the observed variation can be broken down into two major components: the actual variation in the process and the variation introduced by the measurement system itself. Based on this insight and referring back to our blood pressure example, we observe that the total calcu- lated variance of 442.9 is the result of the actual differences in the patients’ blood pressure (process variation) and perhaps errors made in measuring the patients’

c04.indd 145c04.indd 145 25/09/12 10:50 AM25/09/12 10:50 AM

146 C h a p t e r 4 : P r o c e s s I m p r o v e m e n t : M i n i m i z i n g V a r i a t i o n T h r o u g h S i x S i g m a

blood pressure (measurement system variation). Mathematically this can be expressed as follows:

� T 2   5   �

p 2   1   �

m 2

where:

� T 2 5 the total observed or calculated variation

� p 2 5 the actual variation inherent in the process and commonly referred to as the part-to-part variation (or, in our example, patient-to-patient variation)

� m 2 5 variation introduced by the measurement system

T A B L E 4 .4 • Sy s to l i c B lood Pre s sure Va lues fo r Sample o f Ma le D ia be t i c Pa t i en t s

Patient Systolic Blood Pressure

S. Jones 123

K. Smith 106

T. Carter 136

F. Lance 145

J. Porter 153

L. Davis 157

H. Johnson 101

R. Jones 124

G. Scott 152

B. Regan 108

Average 130.5

Std. dev. 21.0

Variance 442.9

Figure 4.11 Components of total process variation.

Reproducibility (σ )20

Repeatability (σ )2e

Measurement system variation

(σ )2m

Process variation (σ )2p

Total variation (σ = 442.9)2T

c04.indd 146c04.indd 146 25/09/12 10:50 AM25/09/12 10:50 AM

147T h e M e a s u r e P h a s e

Recall from basic statistics that calculating the total variation arising from multiple sources requires summing the variances from each source, not their standard devia- tions. Therefore, the total observed standard deviation, �

T , is calculated as

� T   5   �

________ �

p 2 1 �

m 2

The variance introduced by the measurement system can be further broken down into two sources: repeatability and reproducibility. Repeatability corresponds to the ability of the same person doing the measuring to get consistent measurement results when measuring a given item. Ideally, a person would obtain the same measurement results when the same item is measured using the same measurement instrument at different points in time. In our blood pressure example, repeatability corresponds to the ability of a given technician to get the same blood pressure reading for a given patient (assuming the patient ’s blood pressure has not changed) using the same sphygmomanometer. In contrast, reproducibility corresponds to the consistency in the measurement readings when different people measure a particular item using the same measurement instrument. Ideally, different people would obtain the same measurement result when the same item is measured using a common measurement instrument. In the blood pressure example, reproducibility corresponds to the extent to which different technicians get the same blood pressure reading when they take the blood pressure of a given patient (again assuming the patient ’s blood pressure has not changed) using a common sphygmomanometer. Mathematically, the varia- tion introduced by the measurement system can be expressed as

� m 2   5   �

e 2   1   �

o 2

where:

� e 2 5 repeatability, or the variation introduced when the same person measures the same item at different points in time using the same measuring instrument and obtains different results across the trials

� o 2 5 reproducibility, or the variation introduced when different people measure the same item using the same measuring instrument

The measurement standard deviation is calculated as:

� m   5   �

_______ �

e 2   1   �

o 2

To assess the variation introduced by the measurement system, a Measurement Systems Analysis study (or a Gage R&R, for repeatability and reproducibility study as it is commonly called) is conducted. The purpose of a Measurement Systems Analysis study is to assess what percent of the observed variation is being introduced by the measurement system itself and what percent represents the actual underlying varia- tion in the process. The smaller the percentage of variation introduced by the meas- urement system, the better.

Performing a Measurement Systems Analysis requires having two or more work- ers take repeated measurements of multiple test units with known standard values. For example, a Measurement Systems Analysis for taking blood pressure could be done by selecting two or more nurses and having the nurses rotate across multiple

c04.indd 147c04.indd 147 25/09/12 10:50 AM25/09/12 10:50 AM

148 C h a p t e r 4 : P r o c e s s I m p r o v e m e n t : M i n i m i z i n g V a r i a t i o n T h r o u g h S i x S i g m a

patients, taking a patient ’s blood pressure one time and then moving on to the next patient. Once a nurse fi nished taking the blood pressure of all patients, the process would be repeated one or more times so that each nurse took each patient ’s blood pressure two or more times. Prior to the study, the patients could be required to lie down in order to stabilize their blood pressure and remain lying down throughout the study in order to maintain a constant blood pressure. Furthermore, each patient ’s blood pressure would need to be taken both at the beginning and the end of the study using a digital blood pressure monitor to establish the patient ’s true blood pressure and ensure that the patient ’s blood pressure did not change during the study. If one or more of the patients’ blood pressures changed during the study, the study would need to be repeated for these patients. Once the data are collected from a Measurement Systems Analysis study, they are analyzed and the total observed variation is partitioned into its constituent parts using analysis of variance (ANOVA) or other similar techniques.

Both the repeatability and reproducibility of a measurement system relate to the impact the measurement system has on the observed variation. In addition to con- sidering the impact the measurement system has on the variation of observed val- ues, it is also important to consider the impact the measurement system has on the mean (or location) of the observed values. To access the impact the measurement system has on the mean of observed values, three additional measurement system metrics are employed:

1. Bias represents the difference between the average of a number of observa- tions and the true value. For example, assume a nurse takes the systolic blood pressure of a patient three times and observes the following blood pressures: 126, 128, and 127. Further, assume that it was known at the time the patient ’s blood pressure was taken that the patient had an actual blood pressure of 125. In this case we can average the three nurse readings and calculate an average blood pressure value of 127. Based on this, the bias would be 2 (i.e., 127−125), suggesting that the nurse tends to overestimate a patient ’s blood pressure by 2. Thus, bias is a measure of the tendency of the measurement system to under- or overestimate the measurement value of interest, perhaps due to a defective instrument, such as the sphygmomanom- eter. While errors may be made in taking individual measurements, ideally these errors should cancel out over time and the average should be close to the true value, which would in turn yield a bias of approximately zero.

2. Linearity of a measurement system corresponds to the accuracy of the measurement system across the entire range of possible entities to be meas- ured. Ideally, a measurement system ’s accuracy should not be impacted by an entity ’s position in the range of possible values. Thus, a blood pressure measuring system that is more accurate for people with a blood pressure of 115 than for people with a blood pressure of 155 does not possess the char- acteristic of linearity.

3. Stability of a measurement system corresponds to the ability of the meas- urement system to get consistent results over time. For example, assuming a patient ’s blood pressure remains constant over time, the results of taking the patient ’s blood pressure at his or her annual physical should yield similar results.

c04.indd 148c04.indd 148 25/09/12 10:50 AM25/09/12 10:50 AM

149T h e A n a l y z e P h a s e

T H E A N A L Y Z E P H A S E Having fi rst defi ned the problem/opportunity, the customer, and the goals for the Six Sigma project and then subsequently considered appropriate performance meas- ures, collected the relevant data, and evaluated the process ’s current performance, we are now ready to begin the analyze phase. In this phase of the project, our objec- tive is to utilize the data that have been collected to develop and test theories related to the root causes of existing gaps between the process ’s current and desired per- formance. Ultimately, our goal is to identify key cause-and-effect relationships that can be leveraged to improve the overall performance of the process.

Referring to Table 4.2 again, there are a number of tools in the Six Sigma toolkit that are useful in the analyze phase. In this section we will overview three of these tools: (1) brainstorming, (2) cause-and-effect diagrams, and (3) process capability analysis. Because the design of experiments is equally applicable to the improve phase, we defer our discussion of it until the next section.

Brainstorming Brainstorming is among the most, if not the most, widely used techniques in busi- ness to stimulate and foster creativity. It is widely used to facilitate the identifi cation of ways to improve business processes. Brainstorming was originally developed by Alex Osborn, an advertising executive, in the 1950s. The basis for brainstorming was Osborn ’s belief that while on the one hand there can be a synergistic effect associ- ated with having people work in teams (i.e., two heads are better than one), the team ’s overall creativity and effectiveness is often limited by a tendency to prema- turely evaluate ideas as they are being generated. In an effort to capitalize on the strengths of working in teams while at the same time eliminating the drawbacks, Osborn developed the brainstorming approach, which includes the following four guidelines:

1. Do not criticize ideas during the brainstorming session.

2. Express all ideas no matter how radical, bizarre, unconventional, ridiculous, or impractical they may seem.

3. Generate as many ideas as possible.

4. Combine, extend, and/or improve on one another ’s ideas.

As you can see, brainstorming focuses more on the quantity of ideas generated rather than the quality. This is intentional, the point being that there will be ample opportunity to critically evaluate the ideas after the brainstorming session has ended. Therefore, the temptation to criticize or judge ideas early in the process should be avoided so as to not stifl e the creativity of the participants during the brainstorming session.

It is interesting to note that despite the wide acceptance and use of brainstorming in industry, much of the research in the area questions its effectiveness. As one example, Diehl and Stroebe (1987) compared both the quantity and quality of ideas generated by people working in teams and individually. In this study, they found

c04.indd 149c04.indd 149 25/09/12 10:50 AM25/09/12 10:50 AM

150 C h a p t e r 4 : P r o c e s s I m p r o v e m e n t : M i n i m i z i n g V a r i a t i o n T h r o u g h S i x S i g m a

that the teams generated an average number of 28 ideas, while the same number of people working individually generated an average of 74.5 ideas. Furthermore, when the ideas were evaluated by experts, only 8.9 percent of the ideas generated by the teams on average were considered “good ideas” compared with 12.7 percent of the ideas generated by the people working independently. Based on these results and the results of other studies, Professor Leigh Thompson (2003) identifi ed four threats to team creativity:

1. Social loafi ng. When working in teams, people may feel they will not get credit for their ideas and therefore may not work as hard in groups com- pared with the amount of effort they would invest if working individually.

2. Conformity. When working in teams, people may be overly conservative with what they are willing to share with the team because of concerns they may have about the reaction of others.

3. Production blocking. There are physical limitations that can restrict the productivity of a team. For example, only one person can speak at a time. Likewise, people cannot listen and concentrate on what others are saying and simultaneously generate their own new ideas. When working alone, there are likely to be fewer distractions interrupting a person ’s train of thought.

4. Downward norm setting. Research on teams suggests that individuals working in a team environment tend to match the productivity of the least productive team member.

Fortunately, in addition to identifying these threats to team creativity, Professor Thompson also identifi ed a number of specifi c actions that can be used to mitigate the threats and actually enhance team creativity. These actions include the following:

• Create diversifi ed teams. Teams consisting of members with a variety of dif- ferent skills, experiences, training, and so on will position the team to view a problem from multiple perspectives.

• Use analogical reasoning. With analogical reasoning, concepts from one dis- cipline or area are applied to other areas. For example, Dr. Eliyahu Goldratt originally developed the theory of constraints (discussed in the next chapter) as a way to improve the effi ciency of a factory. More recently, his theory has been extended and applied to the fi eld of project management.

• Use brainwriting. Brainwriting involves having the participants in a brain- storming session take periodic breaks to write down their own ideas silently. A key benefi t of brainwriting is that it can greatly eliminate production blocking.

• Use the nominal group technique. The nominal group technique (NGT) is often included as part of the Six Sigma toolkit, as shown in Table 4.3 . With the NGT, team members fi rst work independently for perhaps fi ve to ten min- utes generating a list of ideas. The team members then share their ideas with the team, often in a round-robin fashion, and the ideas are listed. After all the ideas are listed, the team moves on to discussing, clarifying, and extending them. At the end of the discussion, team members individually rank-order the

c04.indd 150c04.indd 150 25/09/12 10:50 AM25/09/12 10:50 AM

151T h e A n a l y z e P h a s e

ideas. Depending on the number of ideas, the team may rank-order each idea or, alternatively, select their fi ve or ten favorite ideas and then rank-order this subset. The team then evaluates the results, perhaps considering the scores of the ideas or the frequency with which an idea was selected.

• Record team ideas. Recording ideas can help a team utilize its time together more effectively by eliminating repetitive discussions and ensuring that ideas are not forgotten.

• Use trained facilitators to run the brainstorming session. As experts in brainstorming, trained facilitators can ensure that the rules are followed and that the discussion stays on task.

• Set high standards. In some cases, a team ’s lack of performance may be the result of misunderstandings of what is expected or even possible. For exam- ple, feedback regarding how many ideas the team has generated in compari- son to the number of ideas other teams have generated after similar durations may help increase the quantity of ideas generated.

• Change the composition of the team. Research supports the fact that periodi- cally replacing team members helps increase both the quantity of ideas gen- erated by the team and the number of different types of ideas generated.

• Use electronic brainstorming. With electronic brainstorming, team members are seated in a room with individual computer workstations for each team member. The team members work individually and enter their ideas into the computer. A computer screen located at the front of the room anonymously displays all ideas generated. Thus, the participants are able to build off of one another ’s ideas without the constraint that only one person can speak at a time and without having to listen to someone else while trying to think independently.

• Make the workplace a playground. Creativity can be fostered by making sim- ple changes to the work environment. While the possibilities are endless, the common denominator is to make the environment fun. Some ideas include placing toys that foster creativity at each seat (e.g., Play-Doh, Legos, building blocks), painting a conference room in nontraditional colors, and changing the name of the room from “the conference room” to “the innovation zone.”

Cause-and-Effect Diagrams Cause-and-effect diagrams are another widely used Six Sigma tool. In fact, develop- ing a cause-and-effect diagram often goes hand-in-hand with brainstorming. For example, a cause-and-effect diagram provides an effective way to organize the ideas that are generated in a brainstorming session addressing the causes of a particular problem. Alternatively, a brainstorming session may be held to develop the cause- and-effect diagram.

As an example, Figure 4.12 provides a simplifi ed version of the cause-and-effect diagram developed at the West Babylon School District in Long Island, New York. In particular, there was a common perception by the teachers that insuffi cient time was being spent covering the curriculum. To help better understand the problem and analyze it, a cause-and-effect diagram was developed.

Creating cause-and-effect diagrams is a fairly straightforward process. First, a box summarizing the problem is drawn at the far right of the workspace, and a horizontal

c04.indd 151c04.indd 151 25/09/12 10:50 AM25/09/12 10:50 AM

152 C h a p t e r 4 : P r o c e s s I m p r o v e m e n t : M i n i m i z i n g V a r i a t i o n T h r o u g h S i x S i g m a

line with an arrow terminating at the box is added. Next, the major causes of the problem are identifi ed and connected to the original horizontal line. Referring to Figure 4.12 , four major causes were identifi ed regarding the problem associated with a lack of teaching time in the fi fth grade: (1) scheduling, (2) staffi ng, (3) no priority given to classroom instruction time, and (4) state mandates. The process of creating a cause-and-effect diagram continues by attempting to break down each major cause into more detailed causes and then possibly breaking down these detailed causes even further. For example, according to Figure 4.12 , we observe that issues related to shared staffi ng and lack of funding contribute to the staffi ng problem.

Not only can cause-and-effect diagrams be developed quickly, they provide an intuitive approach for better understanding problems. Furthermore, important insights are often obtained through the process of creating the diagram, as well as from the diagram itself. However, it should also be noted that, while the cause-and- effect diagram is well structured, the process of creating one is usually not. It is typi- cal to bounce around from the detailed analysis of a particular cause to adding one or more additional major causes. Furthermore, as additional ideas are generated, it may be decided to eliminate, move, combine, and/or rename the major causes or the more detailed causes. Finally, note that because of its appearance, cause-and- effect diagrams are often referred to as fi shbone diagrams.

Process Capability Analysis With the advent of total quality management programs and their emphasis on “mak- ing it right the fi rst time,” organizations are becoming increasingly concerned with the ability of their production processes and service delivery processes to meet cus- tomer requirements. Process capability analysis allows an organization to measure the extent to which its processes can meet its customer requirements or the design

Figure 4.12 Fishbone diagram to analyze the problem of insuffi cent time being spent covering the curriculum. Source: Adapated from R. Manley and J. Manley. “Sharing the Wealth: TQM Spreads from Business to Education.” Quality Progress ( June 1996), pp. 51–55.

Shared staffing

Additional activities

Pull-out programs

Lack of funding

Lack of coordination and communication

Special education

Reading

Math

Staffing Scheduling

State mandates

Lack of teaching time

in grade 5

No priority for classroom

instruction time

c04.indd 152c04.indd 152 25/09/12 10:50 AM25/09/12 10:50 AM

153T h e A n a l y z e P h a s e

specifi cations for the product or service. As shown in Figure 4.13 , process capability depends on the following:

1. Location of the process mean

2. Natural variability inherent in the process

3. Stability of the process

4. Product ’s design requirements

In Figure 4.13 a the natural variation inherent in the process and the product ’s design specifi cations are well matched, resulting in a production system that is con- sistently capable of meeting the design requirements. However, in Figure 4.13 b the natural variation in the production process is greater than the product ’s design requirements. This will lead to the production of a large amount of product that does not meet the requirements: the production process simply does not have the necessary ability. Options in this situation include improving the production process, relaxing the design requirements, or producing a large quantity of product that is unfi t for use.

In Figure 4.13 c the situation is reversed: the product has wider design specifi ca- tions than the natural variation inherent in the production system. In this case the production process is easily able to meet the design specifi cations, and the organiza- tion may choose to investigate a more economical production process in order to lower costs. Finally, although the widths of design specifi cations and process varia- tion are equal in Figure 4.13 d , their means are out of sync. Thus, this process will produce a fair amount of output above the upper specifi cation limit (USL). In this situation the solution would be to shift the process mean to the left so that it is better aligned with the design specifi cations.

More formally, the relationship between the natural variation in the production system and the product ’s design specifi cations can be quantifi ed using a process capability index . The process capability index ( C

p )   is typically defi ned as the ratio

of the width of the product ’s design specifi cation to 6 standard deviations of the

Figure 4.13 Natural variation in a production system versus product design specifi cations.

Design specification Design specification

Design specification Design specification

(a)

(c)

Natural variation in process

Natural variation in process

(b)

(d )

Natural variation in process

Natural variation in process

c04.indd 153c04.indd 153 25/09/12 10:50 AM25/09/12 10:50 AM

154 C h a p t e r 4 : P r o c e s s I m p r o v e m e n t : M i n i m i z i n g V a r i a t i o n T h r o u g h S i x S i g m a

production system. Six standard deviations for the production process is used because 3 standard deviations above and below the production process ’s mean will include 99.7 percent of the possible production outcomes, assuming that the output of the production system can be approximated with a normal distribution. Mathematically, the process capability index is calculated as

C p   5  

product ’s design specifi cation range _________________________________________

6 standard deviations of the production system   5 USL �   LSL __________

6 �

Design specification range

LSL USL

Process mean

+ 3σ

(a) Cp = 1.6

– 3σ

LSL USL

LSL USL

Process mean

Process mean

+ 3σ

(b) Cp = 0.8

(c) Cp = 1.0

– 3σ

+ 3σ– 3σ

Design specification range

Design specification range

Figure 4.14 Effect of production system variability on process capability index. (a) C

p   5   1.6 ; (b) C

p   5   0.8 ; (c) C

p   5   1.0 .

c04.indd 154c04.indd 154 25/09/12 10:50 AM25/09/12 10:50 AM

155T h e I m p r o v e P h a s e

where LSL and USL are a product ’s lower and upper design specifi cation limits, respectively, and � is the standard deviation of the production system.

According to this index, a C p of less than 1 indicates that a particular process is

not capable of consistently meeting design specifi cations; a C p greater than 1 indi-

cates that the production process is capable of consistently meeting the require- ments. As a rule of thumb, many organizations desire a C

p index of at least 1.5.

Achieving Six Sigma quality with no more than 3.4 defective parts per million pro- vides a C

p index of 12 � / 6 �   5   2.0 (assuming the process mean can shift by as much

as 1.5 standard deviations). Figure 4.14 illustrates the effect that changes in the natural variation of the pro-

duction system have on the C p index for fi xed product design specifi cations. In

Figure 4.14 a the natural variation in the process is much less than the product ’s design specifi cation range, yielding a C

p index greater than 1. In contrast, in

Figure 4.14 b the natural variation in the process is larger than the product ’s design specifi cations, yielding a C

p index less than 1. Finally, in Figure 4.14 c the natural

process variation and the design specifi cations are equal, yielding a C p index

equal to 1. One limitation of the process capability index is that it only compares the magni-

tudes of the product ’s design specifi cation range and the process ’s natural variation. It does not consider the degree to which these ranges are actually aligned. For example, the situations shown in Figure 4.13 a and Figure 4.13 d both yield a C

p

index of 1. However, as was pointed out earlier, a considerable amount of defective product would be produced in the situation shown in Figure 4.13 d , owing to the lack of alignment between the design specifi cations and the process mean. The most common way to evaluate the extent to which the process mean is centered within the product ’s design specifi cation range is to calculate a one-sided process capability index C

p k , as follows:

C p u

  5   USL –  process mean

__________________ 3 �

C p l   5  

process mean   –   LSL __________________ 3 �

C p k   5   min ( c

p l ,   c

p u )  

Typically, the data collected to construct the control charts (see Chapter 3) for the process is used to calculate the process standard deviation used in the capability index formulas.

T H E I M P R O V E P H A S E Having defi ned the problem, measured the process ’s current performance, and ana- lyzed the process, we are now in a position to identify and test options for improv- ing the process. In the remainder of this section, our focus will be on the use of design of experiments (DOE) as a process improvement tool.

c04.indd 155c04.indd 155 25/09/12 10:50 AM25/09/12 10:50 AM

156 C h a p t e r 4 : P r o c e s s I m p r o v e m e n t : M i n i m i z i n g V a r i a t i o n T h r o u g h S i x S i g m a

Design of Experiments Perhaps the most common approach to analyzing problems is to investigate one fac- tor at a time (aka OFAT or 1FAT). Unfortunately, the one factor at a time approach suffers from several important shortcomings. To illustrate these shortcomings, con- sider the operation of a fi nancial institution ’s call center for its credit cards.

A representative performance measure would be the time it takes the call center ’s customer service reps (CSRs) to process an incoming call. Factors or variables that might initially be identifi ed as having an impact on the time to process a call include the nature of the call, the time of day, and the CSR who handles the call.

The fi rst shortcoming with OFAT is that it is not typically possible to test one factor at a time and hold all the other factors constant. For example, assume that processing time data were collected for two CSRs over some period of time and it was determined that CSR A averaged fi ve minutes per call while CSR B averaged seven minutes per call. Based on these data, can we conclude that CSR A is more effi cient than CSR B? The answer is no because the impact of other variables has not been accounted for. In this case it could be that CSR B ’s calls were of a more diffi cult nature to handle. Or perhaps the data for CSR A were collected around lunch time when there was a high volume of calls and numerous callers on hold, while the data from CSR B were collected early in the morning when the call volume was lower and there were virtu- ally no callers on hold. The point is that when one variable is studied at a time and the values of other variables are not controlled or otherwise accounted for, it is diffi cult, if not impossible, to draw valid conclusions about the impact of a single variable.

One approach to overcoming the shortcomings associated with the OFAT approach is to use design of experiments (DOE) techniques. DOE techniques utilize the princi- ples of statistics to design experiments to investigate multiple process variables simul- taneously. With DOE techniques, multiple factors are varied and therefore studied simultaneously, and repeated measurements are typically taken for each combination of factor-level settings.

Some major considerations associated with DOE include the following:

• Determining which factors to include in the experiment. Interviewing subject-matter experts (SMEs) is one way to identify relevant factors. Work in the previous phases of the Six Sigma project may also provide important insights into relevant factors. Along these lines, cause-and-effect diagrams are often particularly helpful for identifying relevant factors.

• Specifying the levels for each factor. Once the factors are identifi ed, the levels of each factor must be specifi ed. For example, referring to the Southside Hospital example from the beginning of the chapter, Table 4.5 summarizes an experiment that investigates four factors that are hypothesized to have an impact on the lead time for stress tests. For the fi rst factor, method used to order the stress test, two levels have been specifi ed—using either a fax machine or the Web to order the test. Thus, the study will investigate the impact these two alternative methods for ordering stress tests has on the over- all stress test lead time. Notice that the factor related to the method used to educate the patients about the stress test has three levels, while each of the other factors has two levels.

• Determining how much data to collect. In the experimental design listed in Table 4.5 there are a total of 24 treatment combinations (2 levels of the

c04.indd 156c04.indd 156 25/09/12 10:50 AM25/09/12 10:50 AM

157T h e I m p r o v e P h a s e

method used to order the stress test 3 2 levels of the patient scheduling method 3 3 levels of the patient education method 3 2 levels of the dictation technology). Therefore, 24 patients are needed in order to obtain one obser- vation for each possible treatment combination. Of course, to have confi - dence in the results of a study, it is necessary to collect more than one observation. In DOE terminology, we refer to multiple observations for a treatment combination as replication. If historical data are available, prelimi- nary calculations may be performed to determine how many replications are needed in order to obtain a specifi ed level of statistical confi dence. In other cases, limitations such as the time available to complete the study, personnel, or money may dictate the number of replications possible.

• Determining the type of experimental design. Fundamentally, there are two types of experimental designs. A full factorial experiment corresponds to a study where data are collected for all possible treatment combinations. A frac- tional factorial experiment corresponds to a study where data are collected for only a subset of all possible treatment combinations. Fractional factorial experiments are used when the number of treatment combinations is so large that it is not practical to collect data for each treatment combination. For example, a study with seven factors, each with three levels, would have 2187 treatment combinations (3 7 ). In many cases, investigating 2187 treatment com- binations is not practical. Fortunately, in these cases fractional factorial designs can be developed that reduce the number of treatment combinations while at the same time still providing the most relevant information. In effect, DOE techniques utilize statistical principles to maximize the amount of infor- mation that can be obtained from a given number of treatment combinations. Because the calculations are quite complex, the design of fractional factorial experiments is typically done with the aid of specialized statistics packages or published experimental design catalogs.

Taguchi Methods

Among the more popular approaches used to design experiments are Taguchi meth- ods, named after Genichi Taguchi. According to Taguchi, most of the quality of products and services is determined at the design stage, and therefore the produc- tion system can affect quality only slightly. Taguchi focused on this fact to develop

T A B L E 4 .5 • Repre sen ta t i ve Fac tor s and The i r Leve l s fo r a S t r e s s Te s t S tudy

Factor Levels

Method used to order stress test Fax, Web

Method used to schedule patient appointments

Fixed time appointments; patients given a time window

Method used to educate patients about stress test

Information sheet; phone call from nurse; in-person meeting with nurse

Dictation technology Tape recorder and transcriber; speech recognition

c04.indd 157c04.indd 157 25/09/12 10:50 AM25/09/12 10:50 AM

158 C h a p t e r 4 : P r o c e s s I m p r o v e m e n t : M i n i m i z i n g V a r i a t i o n T h r o u g h S i x S i g m a

an approach to designing quality into outputs. Rather than trying to constantly con- trol equipment and workers to stay within specifi cations—sizes, fi nishes, times—he has devised a procedure for statistical testing to determine the best combination of product and process design to make the output relatively independent of normal fl uctuations in the production system. To do this, statistical experimentation is con- ducted to determine what product and process designs produce outputs with the highest uniformity at lowest cost.

T H E C O N T R O L P H A S E As the Six Sigma project nears completion, the focus in the fi nal phase shifts to the development of procedures to again monitor the process. Here our purpose is to ensure that the process ’s new higher level of performance is maintained and that previous problems do not resurface. We discussed the use of control charts—the most commonly used control tool—in Chapter 3 and will describe the use of “earned value” to control projects in Chapter 6.

S I X S I G M A I N P R A C T I C E To conclude our introduction to Six Sigma, we now turn our attention to issues related to employing Six Sigma to improve business processes and performance. Here our focus will be on the various roles played in Six Sigma initiatives, becoming certifi ed in Six Sigma, and the need for each organization to customize its Six Sigma program to its unique needs.

Six Sigma Roles One aspect that differentiates Six Sigma from earlier process improvement programs, including total quality management and reengineering, is that with Six Sigma, spe- cifi c roles and titles for these roles have been defi ned and generally accepted. The central roles to Six Sigma include the following:

• Master Black Belts. Master Black Belts combine an advanced knowledge of the Six Sigma toolkit with a deep understanding of the business. The primary roles of Master Black Belts are to develop and execute Six Sigma training programs and to work with senior management to ensure that Six Sigma initiatives are being best leveraged to help the organization achieve its strategic goals.

• Black Belts. The two primary roles of Black Belts who have a solid back- ground in the Six Sigma toolkit, are conducting Six Sigma training and lead- ing Six Sigma improvement projects. Both Black Belt and Master Black Belt positions tend to be full-time ones.

• Green Belts. Green belts have broad knowledge of the Six Sigma toolkit but not nearly as much depth in the tools as Black Belts and Master Black Belts. The majority of the work of Six Sigma projects is typically completed by

c04.indd 158c04.indd 158 25/09/12 10:50 AM25/09/12 10:50 AM

159S i x S i g m a i n P r a c t i c e

Green Belts under the guidance and direction of Black Belts and, on occa- sion, Master Black Belts. Green Belts usually split their time between their work on Six Sigma projects and other work responsibilities.

• Yellow Belts. Although Yellow Belts are not as common as Master Black Belts, Black Belts, and Green Belts, some organizations have added this rank as a designation for those employees who have completed Six Sigma awareness training.

In addition to these central roles, there are also a number of important supporting roles, including the following:

• Champions/Sponsors. Champions are senior managers who support and pro- mote Six Sigma projects. The employees working on a Six Sigma project rely on the champion ’s senior position within the organization to help them obtain the resources needed to successfully complete the project as well as to remove hurdles that might otherwise derail the successful completion of the project.

• Process owners. Although process owners are the managers with end-to-end responsibility for a particular business process, typically they do not direct Six Sigma projects. Rather, they are best viewed as being the customers of Six Sigma projects.

Becoming Certifi ed In addition to having well-defi ned roles, another aspect that differentiates Six Sigma from earlier programs is that accompanying each of the central Six Sigma roles is a cer- tifi cation process. Along these lines, it is a common practice for organizations to make a distinction between employees who are Six Sigma trained at a certain level and those who are certifi ed at the level. In these organizations, becoming certifi ed at a given level entails meeting additional requirements beyond receiving the training, such as passing an examination and/or successfully completing one or more Six Sigma projects.

Generally speaking, there are four alternative ways of obtaining Black Belt certifi - cation. Perhaps the most common approach is for employees to be trained and certi- fi ed internally by their current employer. In fact, based on their success with Six Sigma, some organizations actually open their training programs to people outside their organizations. One notable example is Motorola University.

A second approach for obtaining certifi cation is through numerous consulting organizations that offer both training and certifi cation programs. These organizations can be easily found by doing a Web search using a search string such as “Six Sigma certifi cation.” Third, a number of universities have begun offering training and in some cases certifi cation programs. Several universities even offer this training and certifi cation through online distance education programs. Finally, individuals can obtain certifi cation through professional societies, perhaps most commonly through ASQ (www.asq.org).

Given the wide range of options for becoming certifi ed, it is somewhat surprising to observe the extent to which these varied certifi cation programs are standardized in terms of both duration and content. For example, the standard for Black Belt training is a four-month program during which students receive one week of formal in-class training each month and use the time between training sessions to work on a Black Belt project.

c04.indd 159c04.indd 159 25/09/12 10:50 AM25/09/12 10:50 AM

160 C h a p t e r 4 : P r o c e s s I m p r o v e m e n t : M i n i m i z i n g V a r i a t i o n T h r o u g h S i x S i g m a

A P P L Y Y O U R U N D E R S T A N D I N G

Three Do t Four Cap i ta l Management

John Galt was recently promoted to Senior Vice President of Consumer Lending at Three Dot Four Capital Management. Three Dot Four is a large fi nancial services organization ranked among the top 20 fi nancial institutions in terms of total assets. A key consideration in John ’s

The Need to Customize Six Sigma Programs Although there is a fair amount of consensus related to the roles, body of knowledge, and training practices surrounding Six Sigma programs, organizations that have suc- ceeded with their Six Sigma programs also recognize the need to tailor their Six Sigma approach to their unique needs. ScottishPower, an electric and gas provider to more than 5 million customers in the United States and United Kingdom, serves as an excel- lent example of this. In May 2001, ScottishPower brought in an external consultant to begin training its fi rst wave of 20 full-time Black Belts. Because of the limited data that were available and the relatively generic nature of the external consultant ’s training program, ScottishPower emphasized the use of the more simple tools in its early Six Sigma projects, including process mapping, Pareto analysis, cause-and-effect diagrams, and stakeholder analysis. Six months after the fi rst wave of Black Belt training, a sec- ond group of 20 employees was selected for full-time Black Belt training. Based on their experience from the fi rst round, ScottishPower identifi ed the need to better use statistical analysis to identify the root causes of problems. Thus the second wave of training, which was performed by ScottishPower ’s own Black Belts who had been trained in the fi rst wave, emphasized statistical techniques such as the use of t - tests , ANOVA, correlation, regression, and DOE. Based on a desire to gain additional value from the data that were being collected, ScottishPower began emphasizing additional techniques in subsequent waves, including chi-square tests, nonparametric tests such as Mann-Whitney and Kruskall-Wallis, and Box-Cox transformations. ScottishPower ’s experience and success with Six Sigma highlights the importance of customizing a Six Sigma program to the organization ’s unique needs and adopting the training as the organization becomes more sophisticated in its ability to both collect and analyze data.

E X P A N D Y O U R U N D E R S T A N D I N G

1. Contrast Six Sigma and Business Process Design.

2. Is there any relationship between process sigma, DPMO, and process capability?

3. Measurement Systems Analysis focuses primarily on the variation introduced into the measurement system by human operators. Can you think of other sources of variation introduced by the measure- ment system beyond the human operators?

4. Is the DMAIC approach more applicable to projects focusing on incremental change or radical change? Why?

5. Are there any limitations you see associated with QFD? Benchmarking?

6. What phase of DMAIC do you imagine is the most important phase? What phase do you imagine takes the longest?

7. In the example given in the discussion of DPMO, it was stated that the restaurant ’s requirements for the hamburger patties that it purchased from its sup- plier should be between 3.4 and 4.6 ounces. Why would the restaurant be concerned about hamburg- ers weighing too much?

c04.indd 160c04.indd 160 25/09/12 10:50 AM25/09/12 10:50 AM

161A p p l y Y o u r U n d e r s t a n d i n g

promotion was his past success leading the bank ’s quality and productivity group. In particu- lar, there appears to be a signifi cant amount of dissatisfaction among both the bank ’s custom- ers and the bank ’s loan offi cers with its online mortgage application process.

The online mortgage application process is initiated when a customer clicks on the Apply for a Home Mortgage Now link on Three Dot Four ’s homepage. This link takes the user to an Instruction page that overviews the application process and provides a checklist of the information the applicant will be asked to supply on subsequent Web pages. Clicking on the Continue button at the bottom of the Instruction page takes the user to the fi rst of four Web pages, each containing a Web-based form to collect the required data.

The fi rst Web page, Personal Information , solicits information regarding the applicant and requires entering information in 33 fi elds. Information collected on this page includes the applicant ’s and co-applicant ’s names, full addresses, previous addresses, dates of birth, Social Security numbers, phone numbers, e-mail addresses, and so on. To continue on to the next page, the user selects the Continue button located at the bottom of the Personal Informa- tion page. When the Continue button is selected on a given page, a check is made to ensure that none of the required fi elds has been left blank. If the validation check is passed, the next Web page in the application process is displayed. In cases where the validation check fails, the blank fi elds are highlighted and the user is asked to enter the information in these fi elds.

Once information has been entered for all required fi elds on the Personal Information page, the second page— Property, Loan, and Expenses —is displayed. This page is used to collect information about the property the loan will be used to purchase, the type of loan the applicant desires, and information about the applicant ’s monthly expenses. In total, the Prop- erty, Loan, and Expenses page contains 10 data fi elds.

Once the information has been entered for all required fi elds on the Property, Loan, and Expenses page, the Employment page is displayed. This page captures information about the applicant ’s and co-applicant ’s employment history, including salary and other income in- formation. The Employment page contains 16 user fi elds. Finally, the last Web page in the application process captures information about the applicant ’s Assets and Liabilities . In particular, the user is asked to supply information about checking accounts, savings accounts, credit card accounts, investment accounts, car loans, and so on. In total, this page contains 22 data fi elds.

When the applicant clicks on the Submit Application button at the bottom of the Assets and Liabilities page, a fi nal validation check is performed and the information is transferred to one of the bank ’s loan offi cers. The loan offi cers subsequently print out the information and then add the application to their backlog of other in-process applications. To even out the work across the loan offi cers, all loan offi cers process loan applications submitted via the Web, as well as applications received via the mail and applications completed at one of the bank ’s branch offi ces.

Initially, John identifi ed two areas in need of improvement: the fairness of loan approval decisions and the accuracy of the information in loan applications submitted online. In terms of the fairness of loan approval decisions, over the last couple of years the company has re- ceived numerous complaints from applicants questioning the organization ’s fairness in mak- ing loan approval decisions. To begin understanding this problem, John initiated a study in which 25 loan applications were randomly selected. These loan applications were then evalu- ated by a panel of three experts to determine whether the loan should be approved or re- jected. Next, three loan offi cers were selected and asked to evaluate each of the 25 loans two times. The data collected from this study is summarized in Table 1 .

To investigate the issue related to the accuracy of information in online mortgage applica- tions, John formed a process improvement team. The team began by collecting data on the total number of hits each page in the Web application process received as well as the number of times the page was actually completed during the month of January. In addition, the team performed a detailed audit of all the information that was submitted during January and tal- lied the number of fi elds that contained errors across all submitted information. A summary of the team ’s preliminary results is given in Table 2 .

c04.indd 161c04.indd 161 25/09/12 10:50 AM25/09/12 10:50 AM

162 C h a p t e r 4 : P r o c e s s I m p r o v e m e n t : M i n i m i z i n g V a r i a t i o n T h r o u g h S i x S i g m a

T A B L E 1 • Summar y o f Loan Approva l Fa i rnes s S tudy Loan Offi cer 1 Loan Offi cer 2 Loan Offi cer 3 Loan Expert Panel 01/01/2005 02/01/2005 01/01/2005 02/01/2005 01/01/2005 02/01/2005 1 A A A A A A A

2 R R R R R R A

3 A A A A A A A

4 A R R R R R R

5 R R R R R R R

6 R R R A A R R

7 A A A A R A A

8 R R R R R R R

9 A A R R R A R

10 R R R R R R R

11 R R R R R R R

12 A A A A A A A

13 A A A A R A A

14 R R R R R R R

15 R R R A R R R

16 A A A A A A A

17 A A A R A A A

18 A A A A A A A

19 R R R R R R R

20 R R R R A R R

21 R R A R A R A

22 A A A A A A A

23 A R R R R R R

24 A A R A R R R

25 A A A A A R A

A   5   Loan approved R   5   Loan not approved

T A B L E 2 • On l ine Mor tgage App l i ca t ion Submis s ions , J anuar y 2005 Web Page Number of Hits Number Submitted Number of Errors

Personal information 108,571 68,400 45,144

Property, loan, and expense information

68,400 62,928 22,025

Employment information 62,928 59,781 28,695

Asset and liability information

59,781 52,009 51,489

c04.indd 162c04.indd 162 25/09/12 10:50 AM25/09/12 10:50 AM

163A p p l y Y o u r U n d e r s t a n d i n g

Questions

1. What is the DPMO for the loan applications submitted via the Web? 2. What could be done to improve the DPMO? 3. Approximately, what is the process sigma of the loan application process? 4. Develop intuitive measures of repeatability and reproducibility for the loan approval

process. What do the results of this analysis tell you about the fairness of the loan approval process?

5. Regarding the fairness of the loan approval process, what recommendations would you make?

Va l l ey County Med ica l C l in i c

Valley County operates a walk-in medical clinic (VCMC) to meet the nonacute medical needs of its approximately 15,000 citizens. Patients arriving at the clinic are served on a fi rst-come, fi rst-served basis.

As part of a new total quality management program, VCMC conducted an in-depth, four- month study of its current operations. A key component of the study was a survey, distributed to all county citizens. The purpose of the survey was to identify and prioritize areas most in need of improvement. An impressive 44 percent of the surveys were returned and deemed usable. Follow-up analysis indicated that the people who responded were representative of the population served by the clinic. After the results were tabulated, it was determined that the walk-in medical clinic was located near the bottom of the rankings, indicating a great deal of dissatisfaction with the clinic. Preliminary analysis of the respondents’ comments indicated that people were reasonably satisfi ed with the treatment they received at the clinic but were very dissatisfi ed with the amount of time they had to wait to see a caregiver.

Upon arriving at the clinic, patients receive a form from the receptionist requesting basic biographical information and the nature of the medical condition for which treatment is being sought. Completing the form typically requires two to three minutes. After the form is re- turned to the receptionist, it is time-stamped and placed in a tray. Clerks collect the forms and retrieve the corresponding patients’ fi les from the basement. The forms typically remain in the tray for about fi ve minutes before being picked up, and it takes the clerk approximately 12 minutes to retrieve the fi les. After a patient ’s fi le is retrieved, the form describing the medical problem is attached to it with a paper clip, and it is placed in a stack with other fi les. The stack of fi les is ordered according to the time stamps on the forms.

When the nurse practitioners fi nish with their current patient, they select the next fi le from the stack and escort that patient to one of the treatment rooms. On average, fi les remain in the stack for ten minutes, but this varies considerably depending on the time of day and the day of the week. On Monday mornings, for example, it is common for fi les to remain in the stack for 30 minutes or more.

Once in the treatment room, the nurse practitioner reads over the form describing the pa- tient ’s ailment. Next, the nurse discusses the problem with the patient while taking some standard measurements such as blood pressure and temperature. The nurse practitioner then makes a rough diagnosis, based on the measurements and symptoms, to determine if the ail- ment is one of the 20 that state law permits nurse practitioners to treat. If the condition is treatable by the nurse practitioner, a more thorough diagnosis is undertaken and treatment is prescribed. It typically takes about fi ve minutes for the nurse practitioners to make the rough diagnosis and another 20 minutes to complete the detailed diagnosis and discuss the treat- ment with the patient. If the condition (as roughly diagnosed) is not treatable by the nurse practitioner, the patient ’s fi le is placed in the stack for the on-duty MD. Because of the higher cost of MDs versus nurse practitioners, there is typically only one MD on duty at any time. Thus, patients wait an average of 25 minutes for the MD. On the other hand, because of their greater training and skill, the MDs are able to diagnose and treat the patients in 15 minutes,

c04.indd 163c04.indd 163 25/09/12 10:50 AM25/09/12 10:50 AM

164 C h a p t e r 4 : P r o c e s s I m p r o v e m e n t : M i n i m i z i n g V a r i a t i o n T h r o u g h S i x S i g m a

despite the fact that they deal with the more diffi cult and less routine cases. Incidentally, an expert system for nurse practitioners is being tested at another clinic that—if shown to be effective—would initially double the number of ailments treatable by nurse practitioners and over time would probably increase the list even more as the tool continued to be improved.

Questions

1. Develop a process map for the medical clinic that shows the times of the various activities. Is the patients’ dissatisfaction with the clinic justifi ed?

2. What do you imagine are the patients’ key requirements for the clinic? 3. What assumptions are being made about the way work is performed and treatment

administered at the clinic? 4. Redesign the process of treating patients at the clinic, using technologies you are familiar

with, to better meet the patients’ needs as listed in Question 2.

E X E R C I S E S

1. A call center has determined fi ve types of defects can occur in processing customer calls: the cus- tomer spends too long on hold, the customer is given the wrong information, the customer rep han- dles the call in an unprofessional way, the customer is transferred to the wrong destination, and the cus- tomer is disconnected. A total of 468 calls were sub- ject to a quality audit last month, and the results obtained from the audit are summarized in the list below. What is the DPMO for the call center?

Number of Defects/Call Frequency

1 73

2 13

3 3

4 1

5 0

2. Over the last quarter, 742 shots were administered at a walk-in clinic. To be treated properly, patients must be given the correct dosage of the correct medica- tion. During the quarter in question, it was deter- mined that one patient received the incorrect dosage of the correct medication, another patient received the wrong medication, and a third patient received both the wrong medication and the wrong dosage given her age and weight. What is the DPMO and process sigma level for the clinic?

3. A silk screening company prints 6000 decals per month. A random sample of 150 decals is taken

every week and inspected based on four character- istics. The data for the last four weeks are summa- rized in the table below. Assuming the data in the table are representative of the process, what is the DPMO and process sigma level for the silk screening process?

Decal Characteristic Number of Defects

Observed

Color accuracy 10

Image alignment 7

Color consistency 8

Image sharpness 3

4. A hospital made 225 medication errors last year. Of these errors, 30 percent were errors with the pre- scription while 70 percent were errors made while dispensing the medication. The hospital admitted 8465 patients last year, and the patients received an average of 4.8 prescriptions per hospital stay. Medications are dispensed at the time they are needed, and each medication is dispensed four times per day on average. The average patient stay at the hospital is 3.5 days. Compute the DPMO and process sigma level for the patient medicine proc- ess. What assumptions, if any, were needed to cal- culate the DPMO?

5. In the chapter it was noted that when the process mean can shift by as much as 1.5 standard devia- tions, a C

p of 2.0 is needed in order to achieve 3.4

defective parts per million. What C p is needed in

c04.indd 164c04.indd 164 25/09/12 10:50 AM25/09/12 10:50 AM

165B i b l i o g r a p h y

B I B L I O G R A P H Y

order to achieve the same 3.4 defective parts per million assuming the process is perfectly stable and its mean does not shift?

6. Customers of Dough Boy Ltd. have specifi ed that pizza crusts they order should be 28–32 centimeters in diameter. Sample data recently collected indicate that Dough Boy ’s crusts average 30 centimeters in diameter, with a standard deviation of 1.1 centime- ters. Is Dough Boy ’s pizza crust production system

capable of meeting its customers’ requirements? If not, what options does Dough Boy have to rectify this situation?

7. Design specifi cations for a bottled product are that it should contain 350–363 milliliters. Sample data indicate that the bottles contain an average of 355 milliliters, with a standard deviation of 2 milliliters. Is the fi lling operation capable of meeting the design specifi cations? Why or why not?

Alwan , L. Statistical Process Analysis . New York : Irwin/ McGraw-Hill , 2000 .

Arndt, M. “Quality Isn ’t Just for Widgets; Six Sigma, the Quality-Control and Cost-Cutting Power Tool, Is Proving Its Worth on the Service Side.” Business Week ( July 22, 2002): 72.

Breyfogel , F. W. , III. Implementing Six Sigma , 2nd ed. New York : Wiley , 2003 .

Das, R., S. Colello, and H. Davidson. “Six Sigma in Cor- porate Law.” Six Sigma Forum Magazine (November 2004): 30–36.

Diehl , M. , and W. Stroebe . “ Productivity Loss in Brain- Storming Groups: Towards a Solution of a Riddle .” Journal of Personality and Social Psychology , 53 ( 1987 ): 497 – 509 .

Godin, E., D. Raven, C. Sweetapple, and F. R. Del Giu- dice. “Faster Test Results.” Quality Progress ( January 2004): 33–39.

Hammer , M. Beyond Reengineering . New York : Harper Business , 1996 .

Hammer , M. , and S. A. Stanton . The Reengineering Revo- lution . New York : Harper Business , 1995 .

Harrington , H. J. , and J. S. Harrington . High Perform- ance Benchmarking . New York : McGraw-Hill , 1996 .

Jones, M. H., Jr. “Six Sigma … at a Bank?” Six Sigma Forum Magazine (February 2004): 13–17.

Leatherbury, J. “Talent Show: A Process Improvement Approach to Employee Retention Yields Impressive

Business Results.” Quality Progress (November 2008): 48–55.

Pande , P. S. , R. P. Neuman , and R. R. Cavanagh . The Six Sigma Way . New York : McGraw-Hill , 2000 .

Peace , G. S. Taguchi Methods: A Hands-On Approach . Reading, MA : Addison-Wesley , 1993 .

Pellicone, A., and M. Martocci. “Faster Turnaround Time.” Quality Progress (March 2006): 31–36.

Pyzdek , T. The Six Sigma Handbook , rev. ed. New York : McGraw-Hill , 2003 .

Rath & Strong, Inc. Lexington, MA: Rath & Strong’s Six Sigma Pocket Guide . Rath & Strong, Inc., 2000.

Rohleder , T. R. , and E. A. Silver . “ A Tutorial on Business Process Improvement .” Journal of Operations Manage- ment , 15 (May 1997 ): 139 – 154 .

Sester, D. “Motorola: A Tradition of Quality.” Quality (October 2001): 30–34.

Shafer , S. M. “ Karate in Business School? This Is Not Your Father ’s Black Belt .” Quality Management Journal , 12 ( 2 ) ( 2005 ): 47 – 56 .

Thompson , L. “ Improving the Creativity of Organizational Work Groups .” Academy of Management Executive , 17 ( 2003 ): 96 – 111 .

Wortman , B. , W. R. Richardson , G. Gee , M. Williams , T. Pearson , F. Bensley , J. P. Patel , J. DeSimone , and D. R. Carlson . CSSBB Primer . West Terre Haute, IN : Quality Council of Indiana , 2001 .

c04.indd 165c04.indd 165 25/09/12 10:50 AM25/09/12 10:50 AM

c04.indd 166c04.indd 166 25/09/12 10:50 AM25/09/12 10:50 AM

167

� Process Improvement: Reducing

Waste Through Lean

C H A P T E R 5

ROLE OF OPERATIONS IN ORGANIZATIONS’ COMPETITIVENESS

Operations Strategy

Process Design and Control

Process Improvement

Process Execution

Ch. 1: Operations Strategy and Global

Competitiveness

Ch. 2: Process Planning and

Design

Ch. 3: Controlling Processes

Ch. 7: Supply Chain Management

Ch. 8: Capacity, Scheduling, and

Location Planning

Ch. 4: Process Improvement:

Six Sigma

Ch. 5: Process Improvement: Lean

Ch. 6: Managing Process

Improvement Projects

c05.indd 167c05.indd 167 25/09/12 9:05 AM25/09/12 9:05 AM

168

I N T R O D U C T I O N • It was not uncommon for chemotherapy patients at Virginia Mason Medical Center, a

350-bed hospital located in downtown Seattle, to spend an entire day receiving their weekly chemotherapy treatment. To illustrate the process, after arriving at 8:00 A.M. and checking in on the fi rst fl oor, the patient would be asked to go to the laboratory for blood testing located on the sixth fl oor. After having the blood drawn, the patient would then wait for the results to be sent to the oncologist and then eventually meet with the oncologist on the second fl oor. If things progressed smoothly, the patient would begin receiving the intravenous chemotherapy treatment by noon in an open and noisy room that was shared with six other patients.

To improve this process, Virginia Mason has turned to the concepts of lean pio- neered by Toyota shortly after World War II. Virginia Mason ’s overarching goal was to improve the patient experience while at the same time increasing the overall effi - ciency of the process. Using lean concepts, Virginia Mason completely redesigned the process for chemotherapy patients so that everything fl ows to the patient as opposed to the patient fl owing through the process. For example, instead of being located on separate fl oors, the labs and doctors’ offi ces are now adjacent to private patient treatment rooms. Furthermore, each private treatment room has a fl at-screen TV, a computer, nursing supplies, and toilet facilities. A dedicated pharmacy was also added to the cancer treatment center, thereby eliminating delays for patients of up to two hours. Other improvements have reduced the preparation time for chemotherapy treatments from three hours to less than one hour. Across the entire medical center, hospital administrators estimate that its lean initiatives have resulted in savings of $6 million in capital spending, freed up 13,000 square feet, reduced inventory costs by $360,000, and reduced the distance hospital staff walk each day by 34 miles. In addi- tion to these tangible results, the hospital achieved a number of other

Based on their unique operations strategy, organiza- tions design value-creating processes to achieve their strategy (Chapter 2). Frequently opportunities are identifi ed to improve these processes either by com- pletely redesigning the process through Business Process Design or reducing the variation inherent in the process through Six Sigma, as described in Chapter 4. In this chapter, we discuss another approach for process improvement that seeks to minimize waste and maximize value.

More specifi cally, “lean management” has taken on the aura of a global competitive philosophy

because so many fi rms that embrace it have been so successful: Toyota, Deere, and numerous oth- ers. We fi rst address the history and philosophy of lean and then make a comparison between tradi- tional production systems and lean enterprises. Following this, we continue with a discussion of fi ve lean principles: (1) specify value from the cus- tomer ’s point of view, (2) identify the value stream, (3) make value fl ow, (4) have the customer pull value, and (5) pursue perfection. The chapter con- cludes with a discussion of the benefi ts associated with lean.

168

C H A P T E R I N P E R S P E C T I V E

c05.indd 168c05.indd 168 25/09/12 9:05 AM25/09/12 9:05 AM

169I n t r o d u c t i o n

benefi ts, including improved patient satisfaction, shorter bill collection times, and lower infection rates. To achieve these benefi ts, the hospital spent approximately $1.5 million, primarily for consultants, travel, and training (Connolly 2005).

• Although Xerox often gets more publicity for its fi nancial diffi culties, it does have a long track record in the area of quality management. Xerox ’s journey began in the early 1980s when it established its Leadership Through Quality Initiative, which focused on improving business processes in order to improve customer satisfaction, quality, and productivity. Fast forward to the late 1990s and we see Six Sigma and lean being adopted by Xerox ’s manufacturing and supply chain functions. While lim- ited in scope, the lean and Six Sigma programs help Xerox improve its operating effi ciency and effectiveness. Perhaps in part due to the success of these limited ini- tiatives, in mid-2002 Xerox ’s leadership decided to integrate its lean and Six Sigma programs across the entire enterprise, naming the initiative Xerox Lean Six Sigma. To support this initiative, Xerox kicked off an intense Black Belt training program in January 2003 that included employees from all functional areas. By August 2004, 400 Black Belts had been trained, 2500 employees had completed or were in the process of completing Green Belt training, 2000 leaders had completed a two-day workshop, and 10,000 employees had completed Yellow Belt awareness training. Furthermore, a total of 700 Lean Six Sigma projects have been completed across all areas of Xerox, including product design, supply chain, marketing and sales, customer service, and strategy deployment. Xerox estimates that it achieved an initial $6 million return in 2003 based on a $14 million investment in Lean Six Sigma and expects even bigger gains in the years ahead (Fornari and Maszle 2004).

• Honeywell International, a diversifi ed technology company with 2004 sales in excess of $25 billion, is another company that has successfully integrated its Six Sigma initiatives with its lean initiatives. In particular, Honeywell has combined Six Sigma ’s traditional emphasis on variation reduction with lean ’s emphasis on waste reduction to create its Six Sigma Plus program.

Honeywell competes in four major industry segments: (1) aerospace, (2) automa- tion and control solutions, (3) specialty materials, and (4) transportation systems. The business unit for each of these industry segments is headed by a group president. Reporting to each group president is a Six Sigma Plus Leader (SSPL) who is responsi- ble for developing the strategic plans and deploying these plans for Six Sigma Plus initiatives within the group. Reporting to the SSPL is a team composed of Master Black Belts (MBBs), Lean Masters (LMs), Black Belts (BBs), and Lean Experts (LEs). MBBs and LMs, also referred to as Honeywell Masters (HMs), work on projects that have more than a $1 million fi nancial impact and are also responsible for training, mentoring, and certifying BBs and LEs. BBs and LEs work on projects that have a fi nancial impact in the range of $200,000 to $250,000 and train, mentor, and certify Green Belts. In 2002, two HM training waves, fi ve LE waves, 15 BB waves, and 15 BB for leaders training waves, each with 20 to 30 employees, were conducted in North America and additional waves were conducted in Europe and Asia. The HM waves consist of fi ve weeks of training, while the BB and LE waves consist of four weeks of training. The BB for leaders training supplements the core BB training with additional lean topics.

c05.indd 169c05.indd 169 25/09/12 9:05 AM25/09/12 9:05 AM

170 C h a p t e r 5 : P r o c e s s I m p r o v e m e n t : R e d u c i n g W a s t e T h r o u g h L e a n

In addition to the SSPLs in each business unit, Honeywell has a Vice President of Six Sigma who reports directly to the CEO. This VP chairs the Six Sigma Plus Executive Council, which in turn oversees all training and ensures a common curriculum is used companywide. In 2002, Honeywell reported productivity improvement gains of $1.2 billion. Honeywell ’s 2004 Annual Report announced the introduction of the Honeywell Operating System (HOS). The HOS is based on the Toyota Production System and will be used to provide a roadmap to further integrate Six Sigma and lean tools (Hill and Kearney 2003).

• The hospital patient discharge process is often associated with substantial patient dis- satisfaction. However, in addition to frustrating patients, delays in the discharge proc- ess can create problems in other areas of the hospital such as admitting and the emergency department, as these areas must wait for the rooms vacated by discharged patients. To address the ineffi ciencies often associated with the patient discharge process, Valley Baptist Hospital in Harlingen, Texas, utilized lean, Six Sigma, and change management techniques. One specifi c goal of this project was to substantially reduce the time from when a patient discharge order was entered into the computer until the time the patient was transported from the room.

The process improvement team began by mapping the current patient discharge process. In mapping the process, the team discovered that there was little consistency across the nurses in terms of their approaches to discharging patients. Further analysis of the process map was undertaken to identify the activities in the discharge process that were not adding value or, in this case, helping discharge patients faster. In the end, all activities were classifi ed as either value-added, non-value-added, or value enabler. The team further enhanced the process map to show rework loops, communication fl ows among the staff, and physical movements. Key performance metrics were also added to the process map, that highlighted a substantial amount of non-value-added time.

As the team further embellished and analyzed the process map, it was able to identify several primary drivers of waste in the patient discharge process. For exam- ple, the team discovered that in 21 percent of cases, nurses required clarifi cation from a doctor before the discharge order could be entered into the computer. The need to clarify an order added an average of 33 minutes to the discharge process. As another example, the team discovered that in some cases the primary nurse took the patient ’s vital signs, while in other cases a second nurse took the vital signs, which were then reported to the primary nurse. Having the primary nurse take the vital signs himself or herself reduced the elapsed time by an average of 64 minutes.

Based on these insights and others, the process improvement team developed a new standard operating procedure consisting of six steps for the patient discharge process. After adopting the new process, the mean time to discharge a patient was reduced by 74 percent, from 185 minutes to 48 minutes. Furthermore, the standard deviation of discharge times also decreased by 71 percent, from 128.7 minutes to 37.2 minutes. Finally, the percentage of discharged patients who vacated their rooms in 45 minutes or less increased from 6.9 percent to 61.7 percent (DeBusk and Rangel 2005).

c05.indd 170c05.indd 170 25/09/12 9:05 AM25/09/12 9:05 AM

171H i s t o r y a n d P h i l o s o p h y o f L e a n

As these examples illustrate, lean is a philosophy that seeks to eliminate all types of waste, whether it be excessive delays in treating patients, excessive lead times, carrying excessive levels of inventory, workers or parts traveling excessive distances, spending too much time setting up equipment, unneeded space, reworking defec- tive products, clarifying patient orders, idle facilities, and scrap. The examples also illustrate several other important themes associated with lean. First, since waste can be thought of as those activities and outcomes that do not add value for the cus- tomer, a strong customer orientation is central to lean. This was illustrated in the Virginia Mason example, where the chemotherapy process was redesigned so that the treatment process fl owed to the patient as opposed to the patient fl owing through the process. In fact, Virginia Mason made additional changes not mentioned in the example to improve the customer experience, including allocating the best rooms with windows to patients, adding a waterfall and meditation room to help ease patient stress, and adding an Internet café. A second theme relates to the large payoff that can be achieved through lean initiatives. This is exemplifi ed by Honeywell ’s $1.2 billion in productivity gains; Virginia Mason ’s cost, space, and travel savings; and Valley Baptist ’s 74 percent reduction in patient discharge time. Finally, a third theme that emerges from the examples is the trend for organizations to merge their Six Sigma (discussed in Chapter 4) programs with their lean pro- grams. This trend refl ects the complementary nature of these two programs: Six Sigma ’s focus on variation reduction and lean ’s focus on eliminating waste. Clearly, a process with little variation but lots of waste would not be desirable and vice versa.

In essence, the goal of lean is to accomplish more with fewer resources: fewer workers and less inventory, space, equipment, time, scrap, and so on. To accomplish this goal, Womack and Jones, in their book Lean Thinking (2003), identify fi ve lean principles:

1. Specify value from the customer ’s point of view.

2. Identify the value stream, the complete set of activities required to create the output valued by the customer.

3. Make value fl ow through the value stream by eliminating non-value-added activities and streamlining the remaining value-added steps.

4. Have the customer pull value through the value stream.

5. Pursue perfection.

Lean cannot be reduced to a “formula,” and therefore every fi rm must apply the philosophy differently. In the remainder of this chapter, we discuss these fi ve lean principles in more detail. However, before doing so, we begin our discussion with an overview of the lean philosophy to put it in proper context.

H I S T O R Y A N D P H I L O S O P H Y O F L E A N Lean production (also known as synchronous manufacturing or simply lean ) is the name given to the Toyota Production System. Toyota began developing its approach to manufacturing shortly after World War II. The Toyota system is known for its mini- mal use of resources and elimination of all forms of waste, including time.

c05.indd 171c05.indd 171 25/09/12 9:05 AM25/09/12 9:05 AM

172 C h a p t e r 5 : P r o c e s s I m p r o v e m e n t : R e d u c i n g W a s t e T h r o u g h L e a n

Thus, for example, just-in-time ( JIT) is a substantial portion of the Toyota system. Similarly, lean production is an integral element of supply chain management as it is currently envisioned. As such, it requires identifying and eliminating all forms of non-value-added activities throughout the entire supply chain. Teams of multiskilled workers are employed at all levels of the organization to root out ineffi ciency and waste. To understand why lean was developed, it is important to understand a little about the history and culture of Japan.

Japan is a small country with minimal resources and a large population. Thus, the Japanese have always been careful not to waste resources, including space (espe- cially land) as well as time, and labor. Waste is abhorrent because the country has so little space and so few natural resources to begin with. Therefore, the Japanese have been motivated to maximize the gain or yield from the few resources available. It has also been necessary for them to maintain their respect for each other in order to work and live together smoothly and effectively in such a densely populated space. As a result, their work systems tend to be based on three primary tenets:

1. Minimizing waste in all forms

2. Continually improving processes and systems

3. Maintaining respect for all workers

During production, the Japanese studiously avoid waste of materials, space, time, and labor. They therefore pay signifi cant attention to identifying and correcting problems that could potentially lead to such waste. Moreover, operations and proce- dures are constantly being improved and fi ne-tuned to increase productivity and yield, further eliminating waste. Equal respect is paid to all workers, and the trap- pings of status are minimized so that respect among all can be maintained.

Although low cost and consistent quality are important goals when a fi rm adopts lean, many other benefi ts also have accrued in those fi rms where it has been imple- mented. Examples include reduced inventories of all types (and thus less need for the space they require), greater productivity among both labor and staff, shorter lead times, improved processes, increased equipment productivity and utilization, better quality, fewer errors, and higher morale among the workforce and managers. Because of its broad nature and wide range of benefi ts, lean has become for many companies a major element in their competitive strategy, as the Xerox example at the beginning of the chapter illustrated.

The second tenet of Japanese work systems is continuous improvement, which corresponds to the lean principle of pursuing perfection. Accordingly, lean is not considered simply a one-time event to streamline the transformation system from a sloppy, wasteful form to an effi cient, competitive form. Rather, it is an ongoing jour- ney that seeks to make continuing improvements throughout the system to keep the fi rm competitive and profi table in the future.

Perhaps the most important of the three tenets is the third, maintaining respect for all workers. Unfortunately, U.S. industry seems to be moving more slowly in this direction, and U.S. fi rms seem far behind the Japanese in obtaining respect and loy- alty from their workers. This is probably because these fi rms and industries do not show respect for and loyalty to their employees in the fi rst place.

Initially, in the early 1980s, the Japanese approach to production was greeted with a great deal of ambivalence in the United States. Typical of the sentiment at this time was, “It will never work here.” However, this view abruptly changed when a

c05.indd 172c05.indd 172 25/09/12 9:05 AM25/09/12 9:05 AM

173T r a d i t i o n a l S y s t e m s C o m p a r e d w i t h L e a n

number of domestic companies, such as Hewlett-Packard and Harley-Davidson, began demonstrating the signifi cant benefi ts of JIT, an important component of lean.

Next, we describe the most common characteristics of lean systems and compare them with the more traditional systems.

T R A D I T I O N A L S Y S T E M S C O M P A R E D W I T H L E A N Table 5.1 presents a dozen characteristics of lean systems that tend to distinguish them from the more traditional systems historically used in U.S. industry. These

T A B L E 5 .1 • Compar i son o f Trad i t iona l Sy s t ems and Lean Characteristic Traditional Lean

Priorities Accept all orders Many options

Limited market Few options Low cost, high quality

Product/Service design Customized outputs Design from scratch

Standardized outputs Incremental design Simplify, design for manufacturing

Capacity Highly utilized Infl exible

Moderately utilized Flexible

Transformation system Job shop Flow shops, cellular manufacturing

Layout Large space Materials-handling equipment

Small space Close, manual transfer

Workforce Narrow skills Specialized Individualized Competitive attitude Change by edict Easy pace Status: symbols, pay, privilege

Broad skills Flexible Work teams Cooperative attitude Change by consensus Hard pace No status differentials

Scheduling Long setups Long runs

Quick changeovers Mixed-model runs

Inventories Large WIP buffers Stores, cribs, stockrooms

Small WIP buffers Floor stock

Suppliers Many Competitive Deliveries to central receiving area Independent forecasts

Few or single-sourced Cooperative, network Deliveries directly to assembly line Shared forecasts

Planning and control Planning-oriented Complex Computerized

Control-oriented Simple Visual

Quality Via inspection Critical points Acceptance sampling

At the source Continuous Statistical process control

Maintenance Corrective By experts Run equipment fast Run one shift

Preventive By operator Run equipment slowly Run 24 hours

c05.indd 173c05.indd 173 25/09/12 9:05 AM25/09/12 9:05 AM

174 C h a p t e r 5 : P r o c e s s I m p r o v e m e n t : R e d u c i n g W a s t e T h r o u g h L e a n

characteristics range from philosophy and culture to standard operating procedures. Several of the contrasts summarized in Table 5.1 are elaborated on in the remainder of this section.

Priorities Traditionally, most fi rms want to accept all customer orders or at least provide a large number of options from which customers may order. However, this confuses the pro- duction task, increases the chance of errors, and increases costs. With lean, the target market is usually limited and the options are also limited. A wise lean fi rm knows which customers it does not want.

Thus, we see that right from the start the overall priorities of lean fi rms are differ- ent from those of the traditional fi rm. This perspective is refl ected in the approach lean fi rms take to each of the other characteristics as well. In one sense, their “strat- egy” for competing is different from that of the traditional fi rm, and this strategy permeates their production system.

Product/Service Design In line with the priorities, engineering in the lean fi rm designs standard outputs and incrementally improves each design. The parts and subassemblies that make up each output are also standardized; over time, they are further simplifi ed and improved. More traditionally, engineers attempt to design custom outputs to satisfy unique customers, starting from scratch each time and designing new parts and sub- assemblies. The reason for the new parts and subassemblies is often that the engi- neers change and do not know what their predecessors have already designed. Yet even if the same engineers are doing the design work, they often design new parts when a previously designed, tested, and proven part would do—because they can- not afford the time to fi nd the previous design.

Furthermore, designers in lean organizations usually include considerations about the manufacturability of the part or product. This is called design for manufactur- ability (DFM) or design for assembly (DFA). Too often, the traditional fi rm whips up an engineering design as quickly as it can (since it has had to start from scratch) and then passes the design on to manufacturing without giving a thought to how it can be made (sometimes it cannot). With this approach, poor quality and high costs often result and cannot be improved on the shop fl oor, since they were designed in from the start. If the product or part absolutely cannot be made, or perhaps cannot be assembled, then the design is sent back to engineering to modify, taking more time and costing more in engineering hours.

Capacity In terms of capacity, traditional fi rms tend to design extra capacities of all kinds into the system just in case a problem arises and they are needed. These capacities may consist of extra equipment, overtime, partial shifts, and, frequently, large work-in- process (WIP) inventories. All of them cost extra money to acquire and maintain, which eventually increases the cost of the product.

c05.indd 174c05.indd 174 25/09/12 9:05 AM25/09/12 9:05 AM

175T r a d i t i o n a l S y s t e m s C o m p a r e d w i t h L e a n

In lean organizations, excess capacities are kept to a minimum to avoid inherent waste, particularly the WIP inventories, as will be discussed in more detail later. In place of the excess capacities, tighter control is exerted over the production system so that conditions do not arise where signifi cant additional capacity is needed in the fi rst place.

Layout The traditional method of layout follows the job shop approach of using widely spread-out equipment with space for stockrooms, tool cribs, and work-in-process inventories between the equipment. To handle and move all this inventory, auto- mated or semiautomated equipment such as conveyors, carousels, and forklifts is also required, which takes even more space.

With lean, equipment is moved as close together as possible so that parts can be actually handed from one worker or machine to the next. The use of cells and fl ow lines permits the production of parts in small lots with minimal work-in-process and material-moving equipment. The cells are often U-shaped so that one worker can easily access all the machines without moving very far, and fi nished products will exit at the same point where raw materials enter the cell.

It is not unusual for the work fl ows in a traditional job shop to look like a plate of spaghetti when traced on a diagram of the shop. In fact, creating such a diagram in which the physical fl ows of the parts are mapped onto the shop fl oor is referred to as a spaghetti chart and is a commonly used lean tool. In particular, spaghetti charts can be used to identify excessive travel distances, backtracking, and other sources of waste. Based on the insights gained from creating and analyzing a spaghetti chart, ideas for shortening work fl ows and making them more direct, with fewer major part-family fl ow streams, can be identifi ed and implemented. In service-oriented and transactional processes, spaghetti charts can be created by having a person assume the role of the part and actually walking through the process as would a patient or invoice.

Workforce A key element of lean is the role of the workforce as a means of uncovering and solving problems. Rather than considering the workers as the traditional cogs in the great plant machine, each with its own tasks, skills, and narrow responsibilities, lean strives for a broadly skilled, fl exible worker who will look for and solve production problems wherever they appear.

In the traditional shop, much of the employees’ time is nonworking time: looking for parts, moving materials, setting up machines, getting instructions, and so on. Thus, when actually working, the employees tend to work fast, producing parts at a rapid pace whether or not the parts are needed. (This, of course, results in errors, scrap, and machine breakdowns, which again provide a reason to stop working.) The outcome is a stop-and-go situation that, overall, results in a relatively ineffi cient, ineffective pace for most workers.

Conversely, with lean, the workers produce only when the next worker is ready. The pace is steady and fast, although never frantic. In spite of the built-in rule that workers should be idle if work is not needed, the focus on smooth fl ows, short

c05.indd 175c05.indd 175 25/09/12 9:05 AM25/09/12 9:05 AM

176 C h a p t e r 5 : P r o c e s s I m p r o v e m e n t : R e d u c i n g W a s t e T h r o u g h L e a n

setups, and other such simplifi cations means that workers are rarely idle. (Of course, if they are idle, that is an immediate signal to the system designers that work is not progressing smoothly through the plant and adjustments need to be made.)

Inventories In Japan, inventory is seen as an evil in itself. It is a resource sitting idle, wasting money. But, more important, inventory tends to hide problems. In the traditional plant, inventories are used to buffer operations so that problems at one stage don ’t affect the next stage. However, inventories also hide problems, such as defective parts, until the inventory is needed and then is found to be defective. For example, in a plant with lots of work-in-process inventory, a worker who discovers a batch of defective parts can simply put them aside and work on something else. By the time the worker returns to the defective batch, if ever, so much time has elapsed since the batch was processed upstream that the cause of the problem is unlikely to be discovered and corrected so as to prevent a recurrence. In contrast, in an environment where there is little or no buffer inventory, a worker who discovers a defective batch has no choice but to work on the batch. Furthermore, the worker can then notify upstream operations of the problem so they can correct it and ensure that it does not occur in the future.

The Japanese liken inventory, and the money it represents, to the water in a lake. They see problems as boulders and obstacles under the water, as shown in Figure 5.1 . To expose the problems, they reduce the inventories, also shown in Figure 5.1 , and then solve the problems. Then they lower the inventory more, exposing more prob- lems, and solve those, too. They continue this until all the problems are solved and the inventory investment is practically gone. The result is a greatly improved and smoother production system.

In the traditional plant, almost the opposite happens. Because managers know that their plant produces, say, 15 percent defective products, they produce 15 percent extra, which goes into inventory. That ’s the wrong way to handle the problem—they should fi x the problem in the fi rst place, not cover it up with expensive inventory.

All types of inventories are considered liabilities: work-in-process, raw materials, fi nished goods, component parts, and so on. By eliminating storage space, not only do we save space, but we also disallow inventories where defectives can be hidden until no one knows who made them. And by eliminating queues of work waiting for machines, we facilitate automatic inspection by workers of hand-passed parts,

$ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $

$ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $

$ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $

$ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $$ $ $ $ $ $

$ $ $ $

$

$

$

Figure 5.1 Lowering inventory investment to expose problems.

c05.indd 176c05.indd 176 25/09/12 9:05 AM25/09/12 9:05 AM

177T r a d i t i o n a l S y s t e m s C o m p a r e d w i t h L e a n

thereby identifying problems when they begin rather than after 1000 units have been made incorrectly.

If the space saved when operations are moved closer to each other—frequently 33 percent of the original space—is immediately used for something else, then inventory can ’t be dumped there. This facilitates simultaneously reducing the lead time, smoothing the workload, and reducing the inventory.

Last, with minimal or no inventory, control of materials is much easier and less expensive. Parts don ’t get lost, don ’t have to be moved, don ’t have to be labeled, and don ’t have to be held in computer memory or inventory records. Basically, discipline and quality are much improved and cost is reduced, simultaneously.

In Chapter 8, the economic order quantity (EOQ) will be presented as the optimal order quantity, given the tradeoff between inventory carrying cost and setup or ordering cost. The EOQ model also demonstrates the relationship between setup cost and average inventory levels: order quantities, and consequently average inven- tory levels, increase as setup time and cost increase. Knowing that the EOQ mini- mized total costs, managers in the United States simply plugged values into the EOQ formula to determine optimal order quantities. However, use of the EOQ model assumes that its inputs are fi xed. In contrast to their American counterparts, manag- ers in Japan did not assume that these inputs were fi xed. In fact, they invested sig- nifi cant amounts of time and other resources in fi nding ways to reduce equipment setup times. These efforts led to substantial reductions in setup times and therefore in setup costs, and ultimately to much smaller batch sizes, which became the basis of the JIT system.

As discussed in Chapter 2, one approach to reducing setup times is to adopt cel- lular manufacturing. Another approach, if the equipment is available and utilization rates are not a problem, is to use multiple machines that have already been set up for the new task. Alternatively, some of the more advanced and automated equip- ment will automatically reset itself. In the remaining cases, the setup task can be made much more effi cient through a number of techniques that have been largely identifi ed and cataloged by the Japanese. Some of these are described next.

One lean tool used to reduce setup times is SMED, which stands for single minute exchange of die. Developed by Shigeo Shingo, SMED is an important component of the Toyota Production System. While SMED literally translates to a single minute for practical purposes, the goal is to reduce setup times to under 10 minutes (i.e., a sin- gle digit, not a single minute). An excellent example of SMED is provided by the CMI factory that is jointly operated by GM and Suzuki, where machine setup times were reduced from 36 hours to 6 minutes.

A key element of SMED is distinguishing between internal setup time, which requires that the machine be turned off, and external setup time, which can be con- ducted while the machine is still working on the previous part. First, a major effort is directed toward converting internal to external setup time, which is easier to reduce. This is largely done by identifying all the previously internal setup tasks that can either be conducted just as easily as external setup work or, with some changes in the operation, be done externally. Then the external task times are reduced by such techniques as staging dies, using duplicate fi xtures, employing shuttles, and install- ing roller supports. Last, internal time is reduced by such creative approaches as using hinged bolts, folding brackets, guide pins, or Lazy Susans.

Once the setup times are reduced, the fi rm gains not only in smoother work fl ows and shorter lead times but also in fl exibility to any changes in production schedules

c05.indd 177c05.indd 177 25/09/12 9:05 AM25/09/12 9:05 AM

178 C h a p t e r 5 : P r o c e s s I m p r o v e m e n t : R e d u c i n g W a s t e T h r o u g h L e a n

stemming from accidents, unexpected breakages, customers’ problems, and so on. Clearly, this fl exibility is immensely valuable.

Suppliers Traditional practice has been to treat suppliers as adversaries and play them off against each other. Multiple sourcing purportedly keeps prices down and ensures a wide supply of parts. However, multiple sourcing also means that no supplier is get- ting an important fraction of the order; thus, there is no incentive to work with the fi rm to meet specifi cations for quality and delivery.

With lean, the desire is for frequent, smooth deliveries of small lots, with the sup- plier considered part of the team. As part of the team, the supplier is even expected to help plan and design the purchased parts to be supplied. Schedules must be closely coordinated, and many small deliveries are expected every day. Thus, it is in the supplier ’s interest to locate a plant or warehouse close to the customer. Clearly, then, the supplier must have a large enough order to make this trouble worthwhile; thus, single-sourcing for 100 percent of the requirements is common. But with such large orders, the customer can expect the supplier to become more effi cient in pro- ducing the larger quantities of items, so quantity discounts become available. Moreover, having just one source is also more convenient for a fi rm that must interact and coordinate closely with the supplier. Companies that develop single-sourcing rela- tionships recognize the mutual dependency of the supplier–customer relationship. Specifi cally, for the customer to prosper in the marketplace, the supplier must supply high-quality items in the right quantities on time. On the other hand, the more success- ful the customer, the more business is generated for the supplier.

Perhaps equally signifi cant, there is no incoming inspection of the materials to check their quality—all parts must be of specifi ed quality and guaranteed by the supplier. Again, this requires a cooperative rather than an adversarial approach, with the supplier working with the team. Many lean fi rms are now establishing a list of “certifi ed” suppliers that they can count on to deliver perfect quality and thus become members of their production teams. In fact, many organizations implementing such programs will purchase products only from suppliers that pass their certifi cation criteria. Often, companies that set up certifi cation programs work with their suppli- ers to help them become certifi ed.

Single-sourcing also has some disadvantages, however. The largest, of course, is the risk of being totally dependent on one supplier. If the supplier, perhaps through no fault on its part, cannot deliver as needed, the fi rm is stuck. With the minimal buffers typical of a lean organization, this could mean expensive idled production and large shortages. There is also some question about the supplier ’s incentive to become more creative in terms of producing higher-quality or less expensive parts, because it already has the single-source contract. Yet the Japanese constantly pres- sure their suppliers to continue reducing prices, expecting that, at the least, the effect of increased learning with higher volumes will result in lower prices.

Planning and Control In the traditional fi rm, planning is the focus, and it is typically complex and compu- terized. Materials Requirements Planning (MRP, discussed in Chapter 7) is a good

c05.indd 178c05.indd 178 25/09/12 9:05 AM25/09/12 9:05 AM

179T r a d i t i o n a l S y s t e m s C o m p a r e d w i t h L e a n

example of the level of planning and analysis that goes into the traditional production system. Unfortunately, plans often go astray, but since the fi rm is focused on planning rather than control, the result is to try to improve planning the next time, and this, in turn, results in ever more complex plans. Thus, these fi rms spend most of their time planning and replanning and very little time actually executing the plans.

In the lean approach, the focus is on control. Thus, procedures are kept simple and visual and made as routine as possible. Rather than planning and forecasting for an uncertain future, the fi rm attempts to respond to what actually happens in real time with fl exible, quick operations. Some planning is certainly conducted, but to be even more effective and effi cient in responding to actual events, the planning is directed to simple expectations and improvements in the control system.

But is there any way to combine the advantages of the lean JIT approach and MRP? Yes, there is a way. It consists of using MRP to pull the long-lead-time items and pur- chases into the shop, and it then employs JIT once the parts and raw materials have entered the shop. Dover ’s OPW Division used this approach, for example. It employed MRP ’s explosion and lead-time offsetting to identify and order the external parts and raw materials and used JIT ’s procedures to run a smooth, effi cient plant once the parts and materials arrived. In other cases MRP is used as a planning tool for order releases and fi nal assembly schedules, while JIT is used to execute and implement the plan.

Quality The traditional approach to quality is to inspect the goods at critical points in the production system to weed out bad items and correct the system. At the least, fi nal inspection on a sample should be conducted before a lot is sent to a customer. If too many defectives are found, the entire lot is inspected and the bad items are replaced with good ones. Scrap rates are tracked so that the fi rm knows how many to initiate through the production system in order to yield the number of good items desired by the customer.

With lean, the goal is zero defects and perfect quality. A number of approaches are used for this purpose, as described in Chapters 1 and 3. But the most important elements are the workers themselves—who check the parts as they hand them to the next worker—and the small lot sizes produced, as described earlier. If a part is bad, it is caught at the time of production, and the error in the production system is corrected immediately.

Maintenance In the traditional approach to production, maintenance has been what is termed cor- rective maintenance , although preventive maintenance is also common. Corrective maintenance is repairing a machine when it breaks down, whereas preventive main- tenance is conducting maintenance before the machine is expected to fail or at regu- lar intervals. Corrective maintenance is more acceptable in the traditional fi rm, because there are queues of material sitting in front of the machines to be worked on so that production can continue undisturbed, at least until the queues are gone.

But in the lean enterprise, if a machine breaks down, it will eventually stop all the following downstream equipment for lack of work. (It will almost immediately stop

c05.indd 179c05.indd 179 25/09/12 9:05 AM25/09/12 9:05 AM

180 C h a p t e r 5 : P r o c e s s I m p r o v e m e n t : R e d u c i n g W a s t e T h r o u g h L e a n

all upstream equipment as well, through the pull system.) Therefore, in lean organi- zations, the maintenance function assumes greater responsibility and has greater visibility. To refl ect its expanded role, lean organizations refer to the maintenance function as total productive maintenance (TPM). One key aspect of TPM is that instead of employing a “crew” of experts who do nothing but repair broken equip- ment, the lean enterprise relies much more heavily on the operator for many of the maintenance tasks, especially simple preventive maintenance. We will return to the topic of TPM later in the chapter.

With our comparison of traditional and lean organizations complete, we now turn our attention to the fi ve principles of lean and discuss each in more detail. This dis- cussion will include an overview of representative tools and methodologies com- monly used to support each principle.

S P E C I F Y V A L U E At the heart of lean is the concept of value. While producers and service providers seek to create value for their customers, it is important to recognize that value is ultimately defi ned by the customer. Thus, one way to defi ne value is to consider what and how much a customer is willing to pay for a particular product or service. Of course, related to how much a customer is willing to pay for a product or service is the strength of the customer ’s desires and needs, and the variety of options available to satisfy these needs.

Alternatively, another common defi nition of value is that it is the opposite of waste, or muda . Waste can be defi ned as those activities that consume resources but from the customer ’s perspective create no value. From this perspective, waste is often classifi ed into one of the following seven categories:

1. Overproduction. Overproduction means creating more of an output than is needed at a particular point in time. Producing more than is needed creates the need for additional space to store the surplus, requires purchas- ing more raw materials than were needed, and often has a detrimental effect on profi t margins as the surplus may need to be disposed of at dis- tressed prices.

2. Inventory. Inventory takes a variety of forms, including raw materials, work- in-process, and fi nished goods. It requires space for its storage, leading to lease and utility expenses. Furthermore, the inventory must be insured, han- dled, fi nanced, and tracked, further increasing the cost of holding it. However, despite all these efforts, some portion of inventory will tend to get damaged, some may become obsolete, and some may even be stolen. Unfortunately, most, if not all, of the work related to maintaining inventory is not value added in the eyes of the customer.

3. Waiting. Waiting relates to delays or events that prevent a worker from per- forming his or her work. A worker with nothing to work on because of a delay in an upstream activity, a worker who is idle because a piece of equip- ment broke down, or a worker who is idle while waiting for a piece of equipment to be set up all exemplify the waste of waiting.

c05.indd 180c05.indd 180 25/09/12 9:05 AM25/09/12 9:05 AM

181I d e n t i f y t h e V a l u e S t r e a m

4. Unnecessary transport. Any time a worker or a part must be moved, it is considered waste. One goal of lean is to seek ways to reduce the distance people or work must travel, as was illustrated by the Virginia Mason Medical Center example at the beginning of the chapter.

5. Unnecessary processing. Unnecessary processing relates to extra steps in a process. Examples of unnecessary processing include removing burrs from machined parts, reworking defective parts, and entering the same informa- tion into multiple databases. Also, from the lean perspective, inspections are generally considered unnecessary processing.

6. Unnecessary human motions. Using the human body effi ciently and effec- tively is vital not only to the health of the workers, but also to the productiv- ity of the organization. Time and motion studies as well as ergonomic studies are used to help design work environments that increase the effi ciency, safety, and effectiveness of workers.

7. Defects. Parts that must be reworked, or in more extreme cases scrapped, represent the fi nal category of waste. Having to perform rework requires repeating steps that were already performed, while scrapping parts results in extra material and processing charges with no corresponding output to offset these charges. The key to providing outputs that are valued by the customer is developing a solid understanding of customer needs. Establishing the voice of the customer, perhaps through a quality function deployment initia- tive (see Chapter 4), is one approach commonly used to help identify and better understand customer needs.

Based on a better understanding of how the customer defi nes value, the next logical task is to defi ne a target cost. Generally speaking, the low-cost producer in an industry has more options available to it than other organizations in the industry. For example, the low-cost producer has the option of matching its competitors’ prices and thereby maintaining a higher profi t margin. Alternatively, the low-cost provider can offer its products and services at a lower price than the competition in an effort to increase its market share.

I D E N T I F Y T H E V A L U E S T R E A M Once value has been defi ned from the customer ’s perspective and a target cost established, the next step is to identify the set of activities or value stream required to create the customer-valued output. Broadly speaking, the value stream includes all activities (value-added and non-value-added) from the creation of the raw materi- als to the fi nal delivery of the output to the end consumer. Within the organization, the value stream includes the design of the output; continues through the operations function, where raw materials are transformed into fi nished goods; and ends with the delivery of its output to the customer. However, it should also be pointed out that a properly crafted value stream map should transcend organizational bounda- ries. Thus, a complete value stream map would include an organization ’s suppliers, the suppliers to its suppliers, and any distributors, retailers, and so on between the organization and the end consumer.

c05.indd 181c05.indd 181 25/09/12 9:05 AM25/09/12 9:05 AM

182 C h a p t e r 5 : P r o c e s s I m p r o v e m e n t : R e d u c i n g W a s t e T h r o u g h L e a n

A value stream map displays the fl ow of materials and information with suppliers, shown on the left of the diagram, and customers on the right. In studying the value stream map shown in Figure 5.2 , we see that our contract manufacturer gets a shipment of steel coils from its supplier, National Steel, Inc., every Monday. Also shown in the value stream map are the quantities of inventory held at each stage of the value stream and the details of the operational activities at each stage in the process. Inventory is represented by triangles and processing activities by rectangles in the value stream map. In the case of our contract manufacturer, a 15-day supply

The activities within a value stream map are often broadly categorized as follows:

• Value-added (e.g., patient diagnosis) • Non-value-added but necessary (e.g., requiring patients to sign a HIPAA form) • Non-value-added and not necessary (e.g., waiting for the doctor) The challenge associated with value-added activities is to identify ways to per-

form these activities in such a way that more value is created and/or fewer resources are consumed. Likewise, the challenge for both types of non-value-added activities is to identify opportunities to eliminate them or perhaps transform the activity into something that is valued by the customer.

An example value stream map for a service fi rm, in this case a contract manufac- turer of metal cases for servers, is shown in Figure 5.2 . In this example, the contract manufacturer provides the service of fabricating the metal cases for the servers that their customer, Allied Computer, Inc. (shown in the upper right-hand corner of the fi gure), assembles for its customers. The most frequently used value stream symbols are summarized in Table 5.2 .

Monday

National Steel, Inc. Weekly

Fax Steel Coils

Coils 15 Days

1000 Tops 1000 Bottoms

100 Tops 100 Bottoms

160 Tops 160 Bottoms

10 hours16 hours16 hours16 hours15 days 5 seconds 3 seconds 5 minutes 3 minutes 8 minutes

Cutting Stamping Welding Drilling Assembly

I I I I

Production Control

Weekly Production Schedule

Weekly Fax Allied

Computer, Inc.ERP

Monday

C/T = 5 s

C/O = 30 m

Uptime = 87%

C/T = 3 s

C/O = 1.5 h

Uptime = 85%

C/T = 5 m

C/O = 10 m

Uptime = 90%

C/T = 3 m

C/O = 20 m

Uptime = 90%

C/T = 10 m

C/O = 0

Uptime = 100%

Production Lead Time

Value Added Time

= 22.3 days

= 16.1 min.

50 Tops 50 Bottoms

I

Figure 5.2 As-Is value stream map for metal case contract manufacturer. Source: Adapted from www.mamtc.com/lean/building_vsm.asp

c05.indd 182c05.indd 182 25/09/12 9:05 AM25/09/12 9:05 AM

183I d e n t i f y t h e V a l u e S t r e a m

T A B L E 5 .2 • Commonly Used Va lue S t r eam Symbol s Value Stream Map Symbol Description

Customer/Supplier When in upper left represents a supplier. When in upper right represents a customer. Supplier or customer name entered inside symbol.

Used to represent shipments from a supplier or to a customer. The frequency of the shipment is often entered inside the symbol.

Block arrows used to show the movement of raw materials and finished goods.

Used to show inventory between stages in the process. The amount of inventory and a description of what is being stored is often entered below the symbol.

This symbol represents a process, operation, machine, or department that material flows through.

Data Boxes are used with other symbols to provide additional information. They most frequently are used with Process symbols. Information frequently captured about a process includes its cycle time (C/T), changeover time (C/O), uptime, available capacity, batch size, and scrap rate.

A timeline is often placed at the bottom of the value stream map to show value added (VA) and non-value-added (NVA) time.

The Production Control symbol is used to capture how production is scheduled and controlled.

External Shipment

Shipments

Inventory

Process

Data Box

Timeline

Production Control

Use

Frequency

Customer/ Supplier

Process

Production Control

C/T =

C/O =

Avail =

NVA NVA

VA VA VA

(Continued)

c05.indd 183c05.indd 183 25/09/12 9:05 AM25/09/12 9:05 AM

184 C h a p t e r 5 : P r o c e s s I m p r o v e m e n t : R e d u c i n g W a s t e T h r o u g h L e a n

of coils is maintained in front of the cutting operation. Also, additional details about the processing steps are included in Data Boxes near each operational activity. Information often captured in Data Boxes includes the cycle time 1 of the activity, how long it takes to change over the equipment, the capacity of the processing stage, and other relevant details about the activity or operation. We also see that a weekly pro- duction schedule is generated by an ERP system and that each stage in the process works at its own pace and pushes its product on to the next operation. We contrast push and pull systems later in this chapter. Along the bottom of the value stream map

Manual Information A straight thin arrow is used to show the flow of information that is conveyed manually such as memos, reports, and meetings. The frequency with which the information is conveyed can also be added.

A wiggle arrow represents information that is conveyed electronically such as via the Web or faxes. The frequency with which the information is conveyed can also be added.

This symbol is used to document specific process improvement projects that are expected to be executed.

This symbol represents the production of part families in cells.

This symbol is used when the output of one process stage is pushed to the next stage in the process.

Production kanbans are used to trigger production. Withdrawal kanbans are used to authorize the material movement to down- stream processes.

A supermarket is a small amount of inventory that is stored at the point of usage.

Electronic Information

Kaizen Blitz

Workcell

Push Arrow

Production and Withdrawal Kanbans

Supermarket

P W

1Note that usage of the term cycle time has a slightly different meaning compared to its usage in the context of assembly line balancing. Here our usage of cycle time refers to the amount of time it takes the particular stage to complete its operation, which may vary across the different processing stages. In assembly line balancing, cycle time refers to the amount of time each station has to complete its activities where the cycle time is constant across all processing stations.

T A B L E 5 .2 • Commonly Used Va lue S t r eam Symbol s (Cont ’d)

c05.indd 184c05.indd 184 25/09/12 9:05 AM25/09/12 9:05 AM

185I d e n t i f y t h e V a l u e S t r e a m

is a timeline that tracks value-added and non-value-added time. In our example, we see that out of the total lead time of just over 22 days, only 16.1 minutes are consid- ered to be adding value. Finally, at the far right, we observe that the contract manu- facturer makes weekly shipments on Monday to its customer, Allied Computer, Inc. Also note that information from the contract manufacturer ’s ERP system is used to generate orders for raw materials, which are faxed to its supplier, National Steel, weekly. Likewise, the contract manufacturer receives a weekly fax from its customer, Allied Computer, which is input into its ERP system.

The value stream map shown in Figure 5.2 is referred to as the As-Is value stream map since it describes the current value stream. After the As-Is map is created, it is carefully studied to identify opportunities for improving the process and a To-Be value stream map is crafted. Following this, a transition plan from the As-Is process to the To-Be process is developed.

Figure 5.3 provides the To-Be value stream map for the contract manufacturer. The To-Be map calls for the following process improvements:

• Reducing the quantity of steel coils held in inventory from 15 day’s, worth to 1 day ’s worth, made possible in part by getting daily shipments from National Steel as opposed to weekly shipments

Figure 5.3 To-Be value stream map for metal case contract manufacturer. Source: Adapted from www.mamtc.com/lean/building_vsm.asp

Daily

Daily

National Steel, Inc.

Steel Coils

Daily Orders

Production Control

Kanban

Daily Orders Allied

Computer, Inc.

I

Coils 1 day’s

300 Tops 300 Bottoms

20 Tops 20 Bottoms

Cutting Stamping

2 hours30 minutes1 day 5 seconds 3 seconds 6 minutes

Convert to Cells

Fabrication & Assembly Cell

Convert to Kanban

C/T = 5 s

C/O = 30 m

Uptime = 87%

C/T = 3 s

C/O = 30 m

Uptime = 85%

C/T = 18 min

C/O = 10 min

Uptime = 95%

Production Lead Time

= 10.6 hours

= 6.1 min

Value Added Time

Change over

P

W W

P

Hourly Production Schedule

c05.indd 185c05.indd 185 25/09/12 9:05 AM25/09/12 9:05 AM

186 C h a p t e r 5 : P r o c e s s I m p r o v e m e n t : R e d u c i n g W a s t e T h r o u g h L e a n

• Reductions in work-in-process inventory levels made possible by eliminating the old push system and adopting a kanban pull system (discussed later in the chapter)

• Kaizen events (discussed later in the chapter) to reduce the changeover time of the stamping machine, convert to the kanban system, and convert from the functional layout to a cellular layout (as discussed in Chapter 2)

• Communicating electronically and daily with its suppliers and customers • Reducing the production lead time from 22 days to 10.6 days

M A K E V A L U E F L O W Erratic fl ows in one part of the value stream often become magnifi ed in other parts of the system, not only farther down the stream but, because of scheduling, farther up the line as well. This is due to the formation of queues in the production system, the batching of parts for processing on machines, the lot-sizing rules we use to initi- ate production, and many other similar policies. These disruptions to the smooth fl ow of goods are costly to the production system and waste time, materials, and human energy. Thus, having identifi ed the value stream, the next step is to transform it from the traditional batch and wait approach to one where the fl ow is continuous. This is accomplished by eliminating non-value-added activities and streamlining the remaining value-added steps. In fact, many lean organizations make the analogy that goods should “fl ow like water.” A key aspect to achieving such a smooth fl ow is to master-schedule small lots of fi nal products.

Another obstacle to smooth fl ows is the traditional functional organization struc- ture. In the functional organization, work is organized based on the similarity of the work performed. Thus, you have accounting departments, marketing departments, radiology units, quality assurance departments, and so on. The problem with organ- izing work on the basis of the type of work performed is that work must then be handed off from department to department. Such hand-offs inevitably create delays in the process and introduce opportunities for making errors. Therefore, lean organi- zations have a bias toward organizing work based on the value-creating process the work supports, as opposed to organizing work functionally.

It should also be pointed out that early production or delivery is just as inappro- priate as late delivery. The goal is perfect adherence to schedule—without this, erratic fl ows are introduced throughout the value stream. With continuous, smooth fl ows of parts come continuous, level fl ows of work, so there are no peak demands on workers, machines, or other resources. Then, once adequate capacity has been attained, it will always be suffi cient.

Continuous Flow Manufacturing An important tenet to making value fl ow in lean enterprises is continuous fl ow manufacturing (CFM). According to this tenet, work should fl ow through the process without interruption, one unit at a time, based on the customer ’s demand rate. Thus, once the processing of a unit has begun, the work should continue uninterrupted

c05.indd 186c05.indd 186 25/09/12 9:05 AM25/09/12 9:05 AM

187M a k e V a l u e F l o w

until the unit is completed. This is refl ected by the phrase, “Don ’t let the parts touch the fl oor.” To accomplish this, delays associated with setting up equipment, moving work between departments, storing work because a needed resource is unavailable, equipment breakdowns, and so on must be eliminated.

To synchronize the fl ow of work with the customer ’s demand rate, the takt time is calculated (the same as the cycle time, as noted in Chapter 2). The term takt time , German for the baton used by orchestra conductors, was coined by Toyota and translates the customer demand rate into time. In effect, the takt time defi nes the rhythm or pace at which work must be completed at in order to meet the customer demand rate. More specifi cally, takt time is calculated as:

Takt time   5   available work time _______________________ customer required volume

To illustrate the concept of takt time, consider an insurance company that oper- ates nine hours per day processing claims. Assume that the employees get two 15-minute breaks and one hour for lunch. Further assume that the company receives 6000 claims per month and that there are 20 working days per month. In this case, the takt time would be calculated as:

Takt time   5   540 min   2   30 min   2   60 min _________________________ 6000 _____ 20

  5   450 ____ 300   5   1.5 minutes / claim

In this case, the insurance company must process a claim every 1.5 minutes. But sup- pose the processing of an application requires 15 minutes of work. Then, in this case, 10 employees working in parallel would be needed. In other words, processing 10 applica- tions every 15 minutes is equivalent to processing one application every 1.5 minutes.

Converting to Mixed-Model Assembly and Sequencing Another approach for enhancing the fl ow of work is mixed-model assembly and sequencing. With mixed-model assembly, items are produced smoothly throughout the day rather than in large batches of one item, followed by long shutdowns and setups and then by another large batch of another item. Let us demonstrate with an example.

Suppose three different models are being produced in a plant that operates two shifts, and the monthly demands are as given in Table 5.3 . Dividing the monthly demand by 20 working days per month and then again by two shifts per day gives the daily production requirements per shift. A common divisor of the required pro- duction per shift of 20 A ’s, 15 B ’s, and 10 C ’s is 5. Using 5 as the common divisor means that we would produce fi ve batches of each of these models each shift.

T A B L E 5 .3 • M ixed-Mode l A s sembly Cyc l e Model Monthly Demand Required/Shift Units/Cycle

A 800 800/(20 3 2) 5 20 4

B 600 15 3

C 400 10 2

Total 1800 45 9

c05.indd 187c05.indd 187 25/09/12 9:05 AM25/09/12 9:05 AM

188 C h a p t e r 5 : P r o c e s s I m p r o v e m e n t : R e d u c i n g W a s t e T h r o u g h L e a n

Dividing the required production per shift of each model by fi ve batches indicates that on each production cycle, 4 units of A, 3 units of B, and 2 units of C will be produced. Assuming two 15-minute breaks per 8-hour shift (480 minutes), the pro- duction rate must be 45 units per 450-minute shift ( 480   2   15   2   15   5   450 ) , or 10 minutes per unit (450 minutes per shift/45 units per shift). Because one cycle consists of 9 units (4 A ’s, 3 B ’s, and 2 C ’s), the entire cycle will take 90 minutes. Thus, each production cycle of 4 A ’s, 3 B ’s, and 2 C ’s will be repeated fi ve times each shift to produce the required 45 units.

One possible production cycle would be to produce the three models in batches using a sequence such as A-A-A-A-B-B-B-C-C. Alternatively, to smooth the production of the nine units throughout the production cycle, a sequence such as A-B-A-B-C-A-B- A-C might be used. Clearly, numerous other sequences are also possible. With daily production of all models, no erratic changes are introduced into the plant through customer demand, because some of every product is always available. When models are produced in traditional batches (such as producing 1000 A ’s, then 750 B ’s, followed by 500 C ’s), one or more batches may well be depleted before the other batches are fi nished. This then necessitates putting a “rush” order through the plant (in order not to lose a customer for the models that are out of stock), disrupting ongoing work, and adding to the cost of all products—not to mention the frustration involved.

The Theory of Constraints The theory of constraints (Goldratt 1990) offers a systematic way to view and ana- lyze process fl ows. Key aspects of the theory of constraints (TOC) include identify- ing the bottlenecks in the process and balancing the work fl ows in the system. Other names for the same concept are drum-buffer-rope (DBR), goal system, constraint management, and synchronous manufacturing. TOC is often compared to kanban (discussed later in this chapter) and MRP (Chapter 7) as another way to plan produc- tion and schedule operations. Studies comparing these systems seem to show that each has different strengths—MRP to generate time-phased requirements, TOC to plan medium-time-horizon bottleneck facilities, and JIT to maximize throughput— and manufacturers should employ a combination of the three.

The theory of constraints was originally implemented through a proprietary pack- age primarily used in the make-to-order and automotive industries called optimized production technology (OPT), which is based on an alternative approach to capacity planning. The basic procedure is fi rst to identify bottleneck workstations in the shop, schedule them to keep them fully utilized, and then schedule the non- bottleneck workstations to keep the bottlenecks busy so that they are never waiting for work. The following ten guidelines capture the essence of the theory:

1. Flows rather than capacities should be balanced throughout the shop. The objec- tive is to move material quickly and smoothly through the production system, not to balance capacities or utilization of equipment or human resources.

2. Fluctuations in a tightly connected, sequence-dependent system add to each other rather than averaging out.

3. Utilization of a non-bottleneck is determined by other constraints in the sys- tem, such as bottlenecks . Non-bottleneck resources do not restrict the amount of output that a production system can create. Thus, these resources should

c05.indd 188c05.indd 188 25/09/12 9:05 AM25/09/12 9:05 AM

189M a k e V a l u e F l o w

be managed to support the operations of those resources (i.e., the bottle- necks) that do constrain the amount of output. Clearly, operating a nonbot- tleneck resource at a higher rate of output than the bottleneck resource does nothing to increase the output produced by the entire production system.

4. Utilizing a workstation (producing when material is not yet needed) is not the same as activation . Traditionally, managers have not made a distinction between “using” a resource and “activating” it. However, according to the theory of constraints, a resource is considered utilized only if it is helping the entire system create more output. If a machine is independently producing more output than the rest of the system, the time the machine is operated to produce outputs over and above what the overall system is pro- ducing is considered activation, not utilization.

5. An hour lost at a bottleneck is an hour lost for the whole shop . Since the bottleneck resource limits the amount of output the entire system can create, time when this resource is not producing output is a loss to the entire system that cannot be made up. Lost time at a bottleneck resource can result because of downtime for maintenance or because the resource was starved for work. For example, if a hair stylist is idle for an hour because no custom- ers arrive, this hour of lost haircuts cannot be made up, even if twice as many customers as usual arrive in the next hour.

6. An hour saved at a nonbottleneck is a mirage . Since nonbottlenecks have plenty of capacity and do not limit the output of the production system, sav- ing time at these resources does not increase total output. The implication for managers is that time-saving improvements to the system should be directed at bottleneck resources.

7. Bottlenecks govern shop throughput and work-in-process inventories .

8. The transfer batch need not be the same size as the process batch . The size of the process batch is the size of the batch produced each time a job is run. Often, this size is determined by trading off various costs, as is done with the economic order quantity (EOQ) model discussed in Chapter 7, Supp. B. On the other hand, the size of the transfer batch is the size of the batch of parts moved from one work center to another work center. Clearly, parts can be moved in smaller batches than the process batch. Indeed, consider- able reductions in batch fl ow times can often be obtained by using a transfer batch that is smaller than the process batch. For example, assume that a manufacturer produces a part in batches of 10. This part requires three oper- ations, each performed on a different machine. The operation time is 5 min- utes per part per operation. Figure 5.4 a demonstrates the effect on fl ow time when a process batch of 10 units is reduced to a transfer batch of one unit. Specifi cally, in Figure 5.4 a the transfer batch is the same size as the process batch, and a fl ow time of 150 minutes results. In Figure 5.4 b , the one-unit transfer batch reduces fl ow time to 60 minutes. The reason for long fl ow time with a large transfer batch is that in any batch, the fi rst part must always wait for all the other parts to complete their processing before it is started on the next machine. In Figure 5.4 a , the fi rst part in the batch has to wait 45 minutes for the other nine parts. When the transfer batch is reduced to one unit, the parts in the batch do not have to wait for the other parts in the process batch.

c05.indd 189c05.indd 189 25/09/12 9:05 AM25/09/12 9:05 AM

190 C h a p t e r 5 : P r o c e s s I m p r o v e m e n t : R e d u c i n g W a s t e T h r o u g h L e a n

9. The size of the process batch should be variable, not fi xed . Because the eco- nomics of different resources can vary, the process batch does not need to be the same size at all stages of production. For example, consider an item that is produced on an injection molding machine and then visits a trimming department. Because the time and cost to set up injection molding equip- ment are likely to be very different from the time and cost to set up the trim- ming equipment, there is no reason why the batch size should be the same at each of these stages. Thus, batch size at each stage should be determined by the specifi c economics of that stage.

10. A shop schedule should be set by examining all the shop constraints simulta- neously . Traditionally, schedules are determined sequentially. First, the batch size is determined. Next, lead times are calculated and priorities set. Finally, schedules are adjusted on the basis of capacity constraints. The theory of constraints advocates considering all constraints simultaneously in develop- ing schedules (this is why it is also referred to as constraint management). The theory also argues that lead times are the result of the schedules and therefore cannot be determined beforehand.

The critical aspect of these guidelines is the focus on bottleneck workstations, not overloading the workstations, and splitting batches in order to move items along to

Figure 5.4 Transfer batch size and its effects on fl ow time. ( a ) Transfer batch size equals process batch size. ( b ) Transfer batch size equals one part.

(b)

(a)

Time 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100105110115120125130135140145150

Opn 1 P 1

P 2

P 3

P 4

P 5

P 6

P 7

P 8

P 9

P 10

Opn 2 P 1

P 2

P 3

P 4

P 5

P 6

P 7

P 8

P 9

P 10

Opn 3 P 1

P 2

P 3

P 4

P 5

P 6

P 7

P 8

P 9

P 10

Time 5 10 15 20 25 30 35 40 45 50 55 60

Opn 1 P 1

P 2

P 3

P 4

P 5

P 6

P 7

P 8

P 9

P 10

Opn 2 P 1

P 2

P 3

P 4

P 5

P 6

P 7

P 8

P 9

P 10

Opn 3 P 1

P 2

P 3

P 4

P 5

P 6

P 7

P 8

P 9

P 10

c05.indd 190c05.indd 190 25/09/12 9:05 AM25/09/12 9:05 AM

191P u l l V a l u e T h r o u g h t h e V a l u e S t r e a m

the next workstation when desirable. A fi ve-step process is recommended for imple- menting the theory of constraints:

1. Identify the system ’s constraint(s). Usually the process fl ow diagram will help identify the constraints, but the ultimate constraint may in fact be sales repre- sentatives’ time, capital available for investment, mandated policies such as a single shift, or even demand in the marketplace.

2. Exploit the constraint. Find ways to maximize the return per unit of the con- straint. An example here would be to use the scarce resource to produce as much of the highest-profi t item as possible.

3. Subordinate all else to the constraint. The objective here is to make sure the constraint is always productive and that something else isn ’t drawing resources away from the constraint. For example, perhaps inventories should be built in front of a scarce machine or worker.

4. Elevate the constraint. Again, fi nd ways to make the constraint as productive as possible, such as extra maintenance; saving time on the constraint by using other, perhaps less effi cient machines more intensively; or even obtain- ing more of the constraint.

5. If the constraint is no longer a bottleneck, fi nd the next constraint and repeat the steps . Once a bottleneck has been eliminated, something else becomes the bottleneck—perhaps another machine or storage facility, or perhaps the demand in the marketplace.

P U L L V A L U E T H R O U G H T H E V A L U E S T R E A M In the traditional fi rm, long lead times are often thought to allow more time to make decisions and get work performed. But in the lean enterprise, short lead times mean easier, more accurate forecasting and planning. Moreover, a way to capitalize on the increasing strategic importance of fast response to the customer is to minimize all the lead times. If lead times are reduced, there is less time for things to go awry, to get lost, or to be changed. For example, it is quite common for an order placed two months ago to be changed every three weeks until it is delivered: change an option, change the quantity ordered, and so on. However, if the delivery time is one week or less, customers can place the order when they know exactly what they need and can therefore delay ordering until the week before they need it.

As opposed to the MRP approach of “pushing” materials through a plant, lean enter- prises rely on pull systems whereby actual customer demand drives the production process. Push systems are planning-based systems that determine when workstations will probably need parts if everything goes according to plan. However, operations rarely go according to plan; as a result, materials may be either too late or too early. To safeguard against being too late and to make sure that people always have enough work to keep busy, safety stocks are used, even with MRP; these may not even be needed, but they further increase the stocks of materials in the plant. Thus, in a push system we see workers always busy making items and lots of material in the plant.

In comparison, a pull system is a control-based system that signals the requirement for parts as they are needed in reality. The result is that workers may occasionally

c05.indd 191c05.indd 191 25/09/12 9:05 AM25/09/12 9:05 AM

192 C h a p t e r 5 : P r o c e s s I m p r o v e m e n t : R e d u c i n g W a s t e T h r o u g h L e a n

In a pull system, the worker at machine A would produce only in response to requests for more materials made by the worker at machine B. Furthermore, the worker at machine B is authorized to make additional product only to replenish product that is used to meet actual customer demand. If there is no customer demand, machine B will sit idle. And if machine B sits idle, machine A will be idle. In this way, the production of the entire operation is matched to actual demand.

The signals used in a pull system to authorize production may be of various kinds. Dover Corporation ’s OPW Division makes gasoline nozzles for gas pumps and uses wire bins as signals. Each bin holds 500 nozzles, and two are used at any time. Raw material is taken out of one bin until it is empty, and then material is drawn from the second bin. A bin collector constantly scouts the plant, looking for empty bins, and returns them to the stockroom, where they are refi lled and returned to the workstations. In this manner, no more than two bins’ worth of material (1000 units) is ever in process.

Hewlett-Packard uses yellow tape to make squares about 1 foot on a side as the signals for its assembly lines. One square lies between every two workers. When workers fi nish an item, they draw the next unit to work on from the square between

(and sometimes frequently) be idle because more materials are not needed. This keeps material from being produced when it is not needed (waste). The appearance of a plant using a pull system is quiet and slow, with minimal material around.

To further contrast the differences between push and pull systems, consider the production system shown in Figure 5.5 . The system consists of one machine of type A and one machine of type B. Machine A has the capacity to produce 75 units per day, and machine B has the capacity to produce 50 units per day. All products are fi rst produced on machine A and then processed on machine B. Daily demand for the organization is 50 units.

In a push system, each work center would work as fast as it could and push the product on to work centers downstream, regardless of whether they needed addi- tional materials. In Figure 5.5 , after the fi rst day of operation, machine A would pro- duce 75 units, machine B would process 50 of the 75 units it received from machine A, and 25 units would be added to work-in-process inventory. Each day the system operates in this fashion, 25 more units will be added to the work-in-process inventory in front of machine B. This might seem irrational to you, but the only way for inven- tory not to be built up is for machine A to produce less than it is capable of produc- ing. In this example, we could idle machine A 33 percent of the day and produce and transport only 50 units to machine B. However, if you were the plant manager and you noticed that the worker assigned to machine A was working only 67 percent of the time, what would you think? You might think the worker was goofi ng off and order him or her to run the machine. Of course, doing this only increases the amount of money tied up in inventory and does nothing to increase the amount of product completed and shipped to the customer.

Machine A Machine B

75 units/day 50 units/day

Demand

50 units/day

Figure 5.5 Sequential production system with two machines.

c05.indd 192c05.indd 192 25/09/12 9:05 AM25/09/12 9:05 AM

193P u l l V a l u e T h r o u g h t h e V a l u e S t r e a m

them and the previous worker. When the square is empty, this is the signal that another item is needed from the previous worker. Thus, there are never more than two items in process per worker.

These two examples are actually modifi cations of Toyota ’s original JIT system. Toyota ’s materials management system is known as kanban , which means “card” in Japanese. The idea behind this system is to authorize materials for production only if there is a need for them. Through the use of kanban authorization cards, production is “pulled” through the system, instead of pushed out before it is needed and then stored. Thus, the Master Production Schedule (MPS) authorizes fi nal assembly, which in turn authorizes subassembly production, which in its turn authorizes parts assem- bly, and so on. If production stops at some point in the system, immediately all down- stream production also stops, and soon thereafter all upstream production as well.

Typically, two cards are used—a withdrawal kanban and a production kanban. The cards are very simple, showing only the part number and name, the work cent- ers involved, a storage location, and the container capacity. The approach is illus- trated in Figure 5.6 .

Assume that work fl ows from work center (WC) 4 to WC5, and containers are used to transport the output from WC4 to WC5, where they are used as inputs. When WC5 sees that it will be needing more input parts, it takes an empty container and a withdrawal kanban back to WC4. There it leaves the empty container and locates a full one, which has a production kanban with it. WC5 ’s withdrawal kanban authorizes it to remove the full container and put the production kanban in a rack at WC4, thereby authorizing the production of another container of parts. Back at WC5, the withdrawal kanban is placed back in its rack. WC4 cannot initiate production and fi ll an empty container until it has a production kanban on the rack authorizing additional production. Thus, with- drawal kanbans authorize the acquisition of additional materials from a supplying work center and production kanbans authorize a work center to make additional product.

The advantage of such a system is its simplicity. Being entirely visual in nature, it facilitates smooth production fl ow, quality inspection, minimization of inventory, and clear control of the production system.

Figure 5.6 Kanban process.

Inputs Inputs

Rack

WC5WC4

Outputs Outputs

Empty container

Withdrawal kanban

Full container

Withdrawal kanban

Production kanban

Kanban holding

rack

P

W

W

W

P

P W

W

c05.indd 193c05.indd 193 25/09/12 9:05 AM25/09/12 9:05 AM

194 C h a p t e r 5 : P r o c e s s I m p r o v e m e n t : R e d u c i n g W a s t e T h r o u g h L e a n

Kanban/JIT in Services Of course, many services, especially pure services, have no choice but to provide their service exactly when it is demanded. For example, a hair stylist cannot build up inven- tories of cuts and styles before the actual customers arrive. Now JIT is being adopted in other services that use materials rather extensively. For example, professors can choose materials from a wide variety of sources and let a “just-in-time” publisher com- pile the material into a custom-made book as quickly and cheaply as a standard book. Supermarkets replenish their shelves on a JIT basis as customers withdraw purchases. And everyone is familiar with fast-turnaround operations such as cleaners, automobile oil changes, photo processing, and eyeglass lenses, not to mention fast food.

Many, if not most, of the techniques used in manufacturing to become lean are equally applicable to services such as close supplier ties (food spoils), maintaining a fl exible workforce (customization), and using reservation systems and off-peak pric- ing to keep level loads on the system. In addition, the general advantages that manu- facturers accrue through defect-free operations, fl exible layouts, minimal inventories, preventive maintenance, advanced technologies, standardized work methods, and other such approaches provide equal advantages to service organizations—and in some cases greater advantages.

P U R S U E P E R F E C T I O N At the risk of stating the obvious, competition is a moving target. By the same token, opportunities to improve processes never end. Therefore, it is common for lean enterprises to focus less on meeting the immediate challenges posed by the compe- tition and to focus more on the long-term goal of achieving perfection. In the remain- der of this section, we overview fi ve commonly used tools that lean organizations turn to in their pursuit of perfection: 5S, the visual factory, kaizen, poka yoke, and total productive maintenance.

5S A widely used approach for increasing the effi ciency of individual work activities is 5S. The approach consists of the following fi ve steps:

1. Sort. Distinguish what work must be performed to complete a task from what does not need to be done. Eliminate the unnecessary steps.

2. Straighten (Set in order). A common phrase in industrial engineering is “a place for everything and everything in its place.”

3. Scrub (Shine). Maintain a workplace that is clean and free of clutter.

4. Systemize. Develop and implement standardized procedures for maintaining an orderly work environment.

5. Standardize (Sustain). Make the previous four steps a habit.

The Visual Factory With little or no slack to absorb disruptions, successful execution in a lean environment requires that workers and decision makers be constantly up to date with the conditions

c05.indd 194c05.indd 194 25/09/12 9:05 AM25/09/12 9:05 AM

195P u r s u e P e r f e c t i o n

in the work environment. One way lean organizations accomplish this is through an approach called the visual factory. The objectives of the visual factory are to help make problems visible, help employees stay up to date on current operating conditions, and communicate process improvement goals. With the visual factory, problems can be made visible through the use of charts displayed throughout the workplace that plot trends related to quality, on-time delivery performance, safety, machine downtime, pro- ductivity, and so on. Likewise, visual factories make use of production and schedule boards to help employees stay up to date on current conditions. It should be noted that the concept of a visual factory is equally applicable to services. For example, a call center one of the authors visited had a board that displayed updated information on the percent of calls that were answered within the desired time frame.

Kaizen The Japanese word kaizen literally translates as “continuous improvement.” The lean journey in the pursuit of perfection requires a continuous series of incremental improvements. In some cases, a continuous improvement initiative may take a year or longer to implement. However, recently a short-term approach to continuous improvement, called the kaizen blitz (aka kaizen workshops, kaizen events), is becoming increasingly popular. In a kaizen blitz, a cross-functional team com- pletes a continuous improvement project in under a week. Often the kaizen blitz begins with a day or two of formal training in lean concepts. The training is then followed by the team completing a continuous improvement project. The project requires the team to collect any needed data, analyze the data, and then immediately implement the proposed improvements. Typical goals for a kaizen blitz include one or more of the following: reducing the amount of fl oor space needed, increasing process fl exibility, improving work fl ows, improving quality, enhancing the safety of the working environment, and reducing or eliminating non-value-added activities.

Poka Yoke The goal of poka yoke is to mistake-proof work activities in a way that prevents errors from being committed in the fi rst place. Examples of poka yoke include sup- plementing electronic forms with computer code or scripts that check the validity of information entered into fi elds as it is being entered, designing a machine that requires the operator to press two buttons simultaneously to cycle the machine so that neither hand can be caught in the machine when it is operating, and placing parts in kits based on their assembly sequence.

Total Productive Maintenance A key driver of waste and therefore an important component of lean is the effective use of equipment. In particular, equipment impacts waste in a number of ways, including the following:

• Breakdowns. When a piece of equipment fails, it is no longer creating valued outputs, which can lead to customer dissatisfaction as well as economic repercussions for the fi rm. Furthermore, workers may be made idle during the breakdown, further adding to the fi rm ’s cost without corresponding increases in sales.

c05.indd 195c05.indd 195 25/09/12 9:05 AM25/09/12 9:05 AM

196 C h a p t e r 5 : P r o c e s s I m p r o v e m e n t : R e d u c i n g W a s t e T h r o u g h L e a n

• Setups. As with breakdowns, a piece of equipment undergoing a setup or changeover is not creating valued outputs. Moreover, during the setup, work- ers are being paid to make the changeover; if a specialized group performs the setup, the machine operators may be idled during the setup period.

• Stoppages. At times, production on a piece of equipment may need to be halted because its output is unacceptable.

• Reduced speed. Another potential loss occurs when a piece of equipment is operated at a lower production rate than the rate at which it was designed to operate at.

• Yields. Yield relates to the percent of the total output produced that is accept- able. Lower yields correspond to greater amounts of waste in the form of scrap and rework.

Total productive maintenance (TPM) focuses broadly on the cost of equipment over its entire life cycle and encompasses a variety of tools and techniques to improve equipment maintenance practices as well as to help prevent and predict equipment failures. Key components of a TPM program include the following:

• Identifying ways to maximize equipment effectiveness • Developing a productive maintenance system for maintaining equipment over

its entire life cycle

• Coordinating the work of engineering, operations, and maintenance employees

• Giving employees the responsibility to maintain the equipment they operate

A measure of overall equipment effectiveness is calculated as the product of equip- ment availability, equipment effi ciency, and the rate of quality output. Equipment availability represents the percent of time a piece of equipment is available to pro- duce output. Equipment effi ciency is a function of the theoretical cycle time, the actual cycle time, actual processing time, and equipment operating time. The rate of quality output corresponds to the yield on the piece of equipment. An overall effec- tiveness rating of 85 percent is considered excellent.

B E N E F I T S O F L E A N In summary, it appears that lean is not one of the annual fads of American manage- ment but rather a philosophy for effi ciently and effectively using the resources an organization already has at its disposal. As such, it will not disappear from the scene, though its tenets are increasingly being merged with other programs such as Six Sigma and supply chain management. And in spite of concerns about the timely physical transportation of goods, a major challenge for the future will be the effec- tive utilization of the many information technologies available to managers, such as the Internet. In too many cases, organizations are relying too much on internal fore- casts rather than using the Internet and other information and communications tech- nologies to access their customers’ real-time production schedules. In the future, lean organizations will increasingly make use of satellite tracking systems, wireless

c05.indd 196c05.indd 196 25/09/12 9:05 AM25/09/12 9:05 AM

197L e a n S i x S i g m a

communication, scanning technology, global positioning systems, two-dimensional bar codes and RFID tags, and paperless documentation across the entire value chain.

As we have seen, lean offers a variety of possible benefi ts: reduced inventories and space, faster response to customers due to shorter lead times, less scrap, higher qual- ity, increased communication and teamwork, and greater emphasis on identifying and solving problems. In general, there are fi ve primary types of benefi ts: (1) cost savings, (2) revenue increases, (3) investment savings, (4) workforce improvements, and (5) uncovering problems.

1. Cost savings. Costs are saved in a number of ways: inventory reductions, reduced scrap, fewer defects, fewer changes due to both customers and engineering, less space, decreased labor hours, less rework, reduced over- head, and other such effects.

2. Revenue increases. Revenues are increased primarily through better service and quality to the customer. Short lead times and faster response to customers’ needs result in better margins and higher sales. In addition, rev- enues will be coming in faster on newer products and services.

3. Investment savings. Investment is saved through three primary effects. First, less space (about a third) is needed for the same capacity. Second, inventory is reduced to the point that turns run about 50 to 100 a year (compared with 3 or 4, historically). Third, the volume of work produced in the same facility is signifi cantly increased, frequently by as much as 100 percent.

4. Workforce improvements. The employees of lean fi rms are much more satis- fi ed with their work. They prefer the teamwork it demands, and they like the fact that fewer problems arise. They are also better trained for the fl exibility and skills needed with lean (e.g., problem solving, maintenance), and they enjoy the growth they experience in their jobs. All this translates into better, more productive work.

5. Uncovering problems. One of the unexpected benefi ts is the greater visibility of problems that lean allows if management is willing to capitalize on the opportunity to fi x these problems. In trying to speed up a process, all types of diffi culties are uncovered and most of them are various forms of waste, so not only is response time improved but cost is also.

L E A N S I X S I G M A Earlier in the chapter it was mentioned that a current trend is for organizations to merge their lean and Six Sigma initiatives. Xerox calls its merged program Lean Six Sigma, while Honeywell refers to its program as Six Sigma Plus. Merging Six Sigma and lean makes sense because being competitive in today ’s environment requires processes that are both effi cient and consistent. An organization that focuses only on the elimination of waste through lean and does not also address the consistency of its processes will be missing signifi cant opportunities to enhance the effectiveness of its processes. Likewise, focusing only on reducing the variation inherent in an inconsistent process and not simultaneously addressing ways to make it more effi - cient also limits the overall effectiveness of the process.

c05.indd 197c05.indd 197 25/09/12 9:05 AM25/09/12 9:05 AM

198 C h a p t e r 5 : P r o c e s s I m p r o v e m e n t : R e d u c i n g W a s t e T h r o u g h L e a n

1. In your opinion, does it make more sense for an organization to merge its lean and Six Sigma pro- grams or keep them separate?

2. Describe how trying to please every customer turns into a “trap” for traditional production. Aren ’t cus- tomization and multiple options the way of the future, particularly for differing national tastes and preferences?

3. The Japanese say that “a defect is a treasure.” What way do they mean, and how does this relate to lean?

4. How smooth is a production fl ow where every item requires a setup? Wouldn ’t fl ows be smoother with long runs where no setups were required for days?

5. Does the theory of constraints apply to services as well as to products?

6. One JIT consultant suggests that managers imple- ment JIT by just removing inventories from the fl oor. What is likely to happen if they do this? What would the Japanese do?

7. With single-sourcing, how does the fi rm protect itself from price gouging? From strikes or interrup- tions to supply?

8. How might lean apply to a service like an airline? A retailer? A university?

9. American managers hate to see high-paid workers sitting idle, even maintenance employees. What is the alternative?

However, beyond simply addressing different aspects of process effectiveness, Six Sigma and lean have proven to be excellent complements to each other. For exam- ple, most of the tools associated with lean are more applicable to the improve and control phases of the Six Sigma DMAIC approach. Thus, merging lean with Six Sigma provides lean practitioners with the disciplined and structured DMAIC approach and a richer set of tools, particularly those associated with the defi ne, measure, and analyze phases. Likewise, the lean tools nicely complement the tradi- tional Six Sigma tools and methodologies.

A project completed at one of Honeywell ’s European chemical plants provides an excellent illustration of the complementary nature of lean and Six Sigma. The impe- tus for the project was the fact that the operation was losing almost $1 million per year. To turn around the operation, it was determined that capacity needed to be doubled while prices needed to be reduced by 50 percent.

The project team began by analyzing a detailed process map and categorizing the steps in the process as value-adding or non-value-adding activities. Through this analysis, the team also discovered that each stage in the process was a bottleneck. Using the detailed process map, the team determined whether a quality requirement or an issue related to process fl ow was the cause of the bottleneck. As it turned out, one bottleneck was the result of a quality issue, and the bottlenecks at the other four stages were caused by issues related to process fl ow.

To resolve the bottleneck caused by quality issues, the team relied on traditional Six Sigma tools, including Measurement Systems Analysis, process mapping, FMEA, cause-and-effect matrices, and design of experiments (see Chapter 4 for a discussion of these tools). To address the bottlenecks that resulted from process fl ow issues, lean tools were used to eliminate the non-value-adding steps such as unnecessary product movements and material handling. By utilizing both lean and Six Sigma tools, the team was able to meet the goal of doubling capacity and reducing costs by 50 percent. And the profi t of the operation increased from losing almost $1 million annually to generating a profi t of $3.4 million annually. Of course, the similarity between the $3.4 million in annual profi t is just a coincidence with Six Sigma ’s goal of achieving no more than 3.4 defects per million opportunities. Or is it?

E X P A N D Y O U R U N D E R S T A N D I N G

c05.indd 198c05.indd 198 25/09/12 9:05 AM25/09/12 9:05 AM

199A p p l y Y o u r U n d e r s t a n d i n g

A P P L Y Y O U R U N D E R S T A N D I N G

A IRCO , Inc .

AIRCO, Inc. offers a range of aviation services to airlines. Its primary services are repairing and overhauling planes and their interiors for its client airlines, which range in size from small regional carriers to large international carriers. As an extension to its seat repairing services, AIRCO recently expanded into the design and assembly of airplane seats.

Since it is new to the business, a typical seat order for AIRCO is for one plane, consisting of anywhere between 100 to 350 seats. Its more established competitors often receive seat orders for multiple planes. The process of fi lling each seat order is complicated by the fact that many variations of seats go into a single plane, based on how the seats are grouped to- gether in the rows; whether the seat is an aisle seat, window seat, or middle seat; whether the seat will be placed in an exit row; and what class the seat will be placed in. There is addi- tional variation across orders because different airlines require the use of different fabrics and cushion materials, require different types of electronics installed in the seats, and even have different dimensions of their seats.

AIRCO ’s seat design and assembly business completed the upfi t of its South Carolina fac- tory a couple of months ago. The facility houses both a group of design engineers and three assembly lines. In addition to designing the seats to the airlines’ specifi cations, the design engineers also develop detailed instructions for the assembly workers, explaining the steps to follow in assembling the seats.

Currently, each of the three assembly lines consists of 12 stations. The takt time of the as- sembly lines is 30 minutes. Unfortunately, as is often the case for the start-up of production operations, the assembly line is experiencing some signifi cant challenges. For example, virtu- ally no seats pass fi nal inspection at the end of the line and are therefore routed to a “penalty box,” where they wait to be reworked. Regarding the quality problems, Christine Chadwick, the supervisor of the penalty box, commented:

While I haven ’t done a detailed analysis of why so many seats fail fi nal inspection, my experi- ence tells me that by far the number-one reason for the seats ending up here is because they are missing parts. Sometimes we also see seats where the parts were installed incorrectly or where a part was damaged when it was assembled.

Parts are delivered to the assembly line stations in carts. Each cart contains the parts for one seat, taken from the “supermarket.” Supermarkets are locations on the shop fl oor where a small amount of inventory is stored to support production activities. While the assembly worker is assembling a seat, a second cart is being fi lled with the parts for the next seat. Two parts pickers support each assembly line. These parts pickers travel up and down the assem- bly line; when they fi nd an empty cart, they take it to the supermarket, where they pick the parts that are needed at the station.

To replenish a cart, the parts pickers refer to the laminated list of parts needed by the workstation. The parts pickers then walk around the supermarket to fi nd the needed parts. The supermarket was designed in a U-shape to facilitate replenishing the cart with the needed parts; however, the laminated list is rarely in the order in which the parts are stored, thus

10. The theory of constraints distinguishes between proc- ess batches and transfer batches. It also recommends that process batches vary according to the econom- ics of effi ciency at each stage of production. Considering the effects of order size and size of the preceding process batch, how should transfer batches be determined?

11. Consider a service you are familiar with. List exam- ples for each of the seven categories of waste for the service.

12. In identifying the value stream, why is it important to go beyond the boundaries of the organization of interest?

c05.indd 199c05.indd 199 25/09/12 9:05 AM25/09/12 9:05 AM

200 C h a p t e r 5 : P r o c e s s I m p r o v e m e n t : R e d u c i n g W a s t e T h r o u g h L e a n

requiring the parts pickers to backtrack through the supermarket. As needed parts are found, they are tossed into the cart except for smaller items like bolts and washers, which are placed in a box attached inside the cart. Immediately after taking parts from the supermarket shelves, the parts pickers scan the bar code label on the shelf with a bar code reader and enter the quantities taken in order to update the inventory level.

Each assembly line has its own dedicated supermarket. Frequently, however, parts pickers fi nd the inventory level depleted for some parts on their list; in these cases, they go to one of the other supermarkets to fi nd the needed parts. Each assembly line is dedicated to the seats for a specifi c plane order. All the parts needed for the entire seat order are delivered to the supermarket at once. Typically, it takes one to two weeks to assemble all the seats for a plane order. Occasionally, there are some bulky items that do not fi t in the supermarkets; in these cases, the parts pickers must travel a couple of hundred yards to retrieve these parts from bulk storage. Bulk storage also holds the parts that arrive before they are moved to the supermarkets.

The assembly workers refer to instructions displayed on monitors at their workstation to assemble the seats. After reading the instructions and looking at the diagrams, the assembly operator roots through the cart of parts to fi nd the next pieces to install. With the parts in hand, the operator next searches for the needed tools and then installs the parts. This process is repeated until the seat is completed, with the operation cycling between reading instruc- tions, fi nding parts, and fi nding tools. If the work is completed in less than the 30 minute takt time, the worker can take a short break. More typical, however, are cases where the assembly workers are waiting for their carts of parts so they can begin working on the next seat.

Questions

1. What types of waste are present at AIRCO ’s new South Carolina facility? 2. What lean tools do you see as being particularly applicable to AIRCO? Explain the

potential benefi ts these tools could provide to AIRCO.

J . Ga l t Lock L td .

J. Galt Lock Ltd., located in Sydney, Australia, produces a line of door locksets and hardware for the residential, light commercial, and retail markets. The company ’s single plant is just over 200,000 square feet and is organized into the following functional departments: screw ma- chines, presses, machining, maintenance, tool and dies, latches, plating, buffi ng, subassembly, and fi nal assembly. The company employs approximately 375 people, 290 of whom are hourly workers. The largest category of employees—assemblers—accounts for two-thirds of the workforce.

The company uses a proprietary planning and scheduling system that uses both an AS/400 minicomputer and spreadsheet analysis performed on a microcomputer to determine produc- tion and purchasing requirements. At any given time, there are 1500 to 3000 open work orders on the shop fl oor. The average lot size is 50,000 parts, but for some products the size is as high as 250,000 parts.

The planning system creates work orders for each part number in the bills of materials, which are delivered to the various departments. Department supervisors determine the order in which to process the jobs, since the system does not prioritize the work orders. A variety of scheduling methods are used throughout the plant, including kanbans, work orders, and expe- diters; however, the use of these different methods often creates problems. For example, one production manager commented that although a “kanban pull scheduling system is being used between subassembly and fi nal assembly, frequently the right card is not used at the right time, the correct quantity is not always produced, and there are no predetermined schedules and paths for the pickup and delivery of parts.” In fact, it was discovered that work orders were often being superseded by expediters and supervisors, large lag times existed between the

c05.indd 200c05.indd 200 25/09/12 9:05 AM25/09/12 9:05 AM

201E x e r c i s e s

E X E R C I S E S

decision to produce a batch and the start of actual production, and suppliers were not being included in the “information pipeline.” One production supervisor commented:

We routinely abort the plans generated by our formal planning system because we fi gure out other ways of pushing product. Although we use kanban systems in two areas of the plant, in reality everything here is a push system. Everything is based on inventory levels and/or incoming customer orders. We push not just the customer order but all the raw materials and everything that is associated with the product being assembled.

In an effort to improve its operations, Galt Lock hired a consulting company. The consultant determined that 36 percent of the fl oor space was being used to hold inventory, 25 percent was for work centers, 14 percent for aisles, 7 percent for offi ces, and 18 percent for non-value- adding activities. The production manager commented:

We have an entire department that is dedicated to inventory storage consisting of 10 to 11 aisles of parts. What is bad is that we have all these parts, and none of them are the right ones. Lots of parts, and we still can ’t build.

The consultant also determined that the upstream “supplying” work centers were often far from the downstream “using” work centers, material fl ows were discontinuous because the parts were picked up and set down numerous times, and workers and supervisors often spent a considerable amount of time hunting for parts. The production manager commented:

Work-in-process is everywhere. You can fi nd work-in-process at every one of the stations on the shop fl oor. It is extremely diffi cult to fi nd materials on the shop fl oor because of the tremendous amount of inventory on the shop fl oor. It is also very diffi cult to tell what state a customer order is in or the material necessary to make that customer order, because we have such long runs of components and subassemblies.

The plant manager commented:

My biggest concern is consistent delivery to customers. We just started monitoring on-time delivery performance, and it was the fi rst time that measurement had ever been used at this operation. We found out how poorly we are actually doing. It is a matter of routinely trying to chase things down in the factory that will complete customer orders. The challenge of more consistent delivery is compounded by the fact that we have to respond much faster. Our customers used to give us three to six weeks of lead time, but now the big retailers we are starting to deal with give us only two or three days. And if we don ’t get it out in that short period of time, we lose the customer.

Questions

1. Evaluate and critique the existing operation and the management of J. Galt Lock. 2. How applicable is JIT to a situation like this? Would converting from a functional layout

to a cellular layout facilitate the implementation of JIT? 3. Where could the principles of lean production be of value to J. Galt Lock?

1. The time between patient arrivals to the blood- drawing unit of a medical lab averages 2 minutes. The lab is staffed with two nurses who actually draw the patients’ blood. The nurses work from 9:00 A.M. to 5:30 P.M. and get two 10-minute breaks and a half-hour for lunch. What is the takt time for drawing patient blood?

2. Referring to Exercise 1, assume that additional anal- ysis was performed and it was determined that an average of 255 patients (with a standard deviation of 30) requiring blood work come to the lab each day. It was further determined that the duration of the nurses’ breaks ranged from 9 to 13.5 minutes, with all times in the range equally likely, and that

c05.indd 201c05.indd 201 25/09/12 9:05 AM25/09/12 9:05 AM

202 C h a p t e r 5 : P r o c e s s I m p r o v e m e n t : R e d u c i n g W a s t e T h r o u g h L e a n

the time taken for lunch ranged from 28 minutes to 34 minutes, again with all times in the range equally likely. Using Crystal Ball or another simulation package, develop a distribution for the takt time, assuming that the number of patients that arrive on a given day is normally distributed. What are the managerial implications of your analysis?

3. The high-speed copier of a printing and document services fi rm is available for 95 percent of the oper- ating hours. The copier is operated at a rate of three copies per second, although it was designed to make four copies per second. Data suggest that about 3 percent of the copied pages must be scrapped and recopied. What is the overall equip- ment effectiveness of the high-speed copier?

4. Referring to Exercise 3, assume that the copier avail- ability follows a triangular distribution and that sometimes the copier is available for as few as 2 hours in a 12-hour day, is typically available 10 hours a day, and occasionally is available for all 12 hours. Further assume that the rate at which the copier operates varies, based on the type of job, from two copies per second to four copies per second, with all rates in this range equally likely. Finally, assume that the scrap rate is normally distributed with a mean of 3 percent and standard deviation of 0.7 percent. Using Crystal Ball or another simulation package, develop the distribution for the copier ’s overall equipment effectiveness. What are the managerial implications of your analysis?

Connolly, C. “Hospital Takes Page from Toyota.” Wash- ington Post, January 3, 2005: www.msnbc.msn.com/ id/8079313/

DeBusk, C., and A. Rangel Jr. “Creating a Lean Six Sigma Hospital Process”: healthcare.isixsigma.com (May 30, 2005).

Fornari, A., and S. Maszle. “Lean Six Sigma Leads Xerox.” Six Sigma Forum Magazine (August 2004): 11–16.

George , M. Lean Six Sigma: Combining Six Sigma Qual- ity with Lean Production Speed . New York : McGraw- Hill , 2002 .

Goldratt , E. M. Theory of Constraints , 2nd rev. ed. Croton-on-Hudson, NY : North River Press , 1990 .

Hill, W. J., and W. Kearney. “The Honeywell Experience.” Six Sigma Forum Magazine (February 2003): 34–37.

Mascitelli, R. “Lean Thinking: It ’s About Effi cient Value Creation.” Target (Second Quarter 2000): 22–26.

Monden , Y. Toyota Production System: An Integrated Approach to Just-in-Time . Atlanta, GA : Institute of Indus- trial Engineers , 1998 .

Pyzdek , T. The Six Sigma Handbook . New York : McGraw-Hill , 2003 .

Schniederjans , M. Topics in Just-in-Time Management . Boston : Allyn & Bacon , 1993 .

Schonberger , R. World Class Manufacturing: The Next Decade: Building Power, Strength, and Value . New York : Free Press , 1996 .

Snee, R. D. “When Worlds Collide: Lean and Six Sigma.” Quality Progress (September 2005): 63–66.

Tonkin, L. “System Sensor ’s Lean Journey.” Target (Second Quarter 2002): 44–47.

White , R. E. , J. N. Pearson , and J. R. Wilson . “ JIT Manu- facturing: A Survey of Implementations in Small and Large U.S. Manufacturers .” Management Science , 45 ( January 1999 ): 1 – 15 .

Womack, J. “Lean Thinking for Process Management.” Paper presented at the Decision Sciences Institute Annual Meeting, November 22, 2004.

Womack , J. , and D. Jones . Lean Thinking: Banish Waste and Create Wealth in Your Corporation . New York : Simon & Schuster , 2003 .

Wortman , B. , W. R. Richardson , G. Gee , M. Williams , T. Pearson , F. Bensley , J. P. Patel , J. DeSimone , and D. R. Carlson . CSSBB Primer . West Terre Haute, IN : Quality Council of Indiana , 2001 .

Zipkin , P. H. Foundations of Inventory Management . New York : Irwin/McGraw-Hill , 2000 .

B I B L I O G R A P H Y

c05.indd 202c05.indd 202 25/09/12 9:05 AM25/09/12 9:05 AM

203

� Managing Process Improvement

Projects

C H A P T E R 6

ROLE OF OPERATIONS IN ORGANIZATIONS’ COMPETITIVENESS

Operations Strategy

Process Design and Control

Process Improvement

Process Execution

Ch. 1: Operations Strategy and Global

Competitiveness

Ch. 2: Process Planning and

Design

Ch. 3: Controlling Processes

Ch. 7: Supply Chain Management

Ch. 8: Capacity, Scheduling, and

Location Planning

Ch. 4: Process Improvement:

Six Sigma

Ch. 5: Process Improvement: Lean

Ch. 6: Managing Process

Improvement Projects

c06.indd 203c06.indd 203 25/09/12 10:48 AM25/09/12 10:48 AM

204

I N T R O D U C T I O N • Décor Cabinets, a custom cabinet maker in Canada, adopted the strategic goal of 100

percent on-time delivery of their cabinets to achieve long-term customer loyalty and create added value that enhances their profi tability. Having such a clear objective helped them assemble a project portfolio uniquely focused on their goal, although it also meant declining some seemingly profi table project ideas requested by custom- ers. However, if demand increased for the requested products, it could have had a serious negative impact on their delivery goals. It was diffi cult to resist pressure from all different areas of the company to support these kinds of projects: “You can easily lose focus,” the CEO admitted. “Sometimes, when ROI drives all decision making, you miss the bigger picture” (Gale, 2007 ).

• One day, Melvin Wilson was simply a marketing manager for little 1250-employee Mississippi Power in Gulfport, Mississippi. But the next day, after Hurricane Katrina hit New Orleans and Gulfport, he was the fi rm ’s “Director of Storm Logistics,” responsible for 11,000 repairmen with a goal of restoring power to 195,000 custom- ers within 12 days. Although Mississippi Power ’s primary storm center at headquar- ters was knocked out, they had a backup storm center 5 miles inland. However, when Wilson got there, the cars were fl oating in the parking lot, so he moved his group in charge to a third location, an old service offi ce without electricity or run- ning water. In spite of the phone lines being down, the group managed to get word of their needs to the outside world; within days, 11,000 repairmen from 24 states and Canada came to help. To support the 11,000 workers, the group needed housing, beds, food, clean water, showers, laundry, bulldozers, 5000 trucks, 140,000 gallons of fuel each day, 8000 tetanus shots, and hundreds of other such items. Directing such

In the last two chapters, we discussed improving our process to make it more competitive through tools like lean management to reduce waste and Six Sigma to reduce process variability. Such efforts are typically done through projects, either small, short-term, focused projects or, sometimes, major long-term projects for making changes in the organization ’s systems and procedures. In this chapter, we address the management of these projects. We use a process improvement project as an example, but projects are used in all kinds of organizations for every conceivable purpose. They range from simple combinations of tactical tasks to

strategic organizational change, and from setting up a party to putting a person on the moon.

The chapter begins with a discussion of the cru- cial topics of project selection, project planning, and organizing the project team. We then move on to an explanation of some project-scheduling techniques, showing some typical project manage- ment software printouts that are available to project managers. The chapter continues with a discussion of controlling project cost and perform- ance, primarily through the use of “earned value,” and then concludes with a brief description of Goldratt ’s “critical chain.”

204

C H A P T E R I N P E R S P E C T I V E

c06.indd 204c06.indd 204 25/09/12 10:48 AM25/09/12 10:48 AM

a massive project as the restoration of power was far beyond the experience of Mississippi Power ’s group, but they succeeded, and the power was restored to every customer who could handle it within 12 days (Cauchon 2005 ).

• In March 2003, the United Kingdom ’s Child Support Agency (CSA) started using their new $860 million software system for receiving and disbursing child support pay- ments. However, by the end of 2004, only about 12 percent of all applications had received payments, and even those took about three times longer to process than they were supposed to take. CSA thus threatened to scrap the entire system and withhold $2 million per month in service payments to the software vendor. The problem was thought to be due to both scope creep and the lack of a risk management strategy. The vendor claimed that the project was disrupted constantly by CSA ’s 2500 change requests, while CSA maintained there were only 50, but the contract did not include a scope management plan to help defi ne what constituted a scope change request. And the lack of a risk management strategy resulted in no contingency or fallback plans in case of trouble, so when project delays surfaced and inadequate training became apparent, there was no way to recover (Project Management Institute 2005).

• To speed passengers to Shanghai ’s new international airport, China built a magnetic levitation (maglev) train that runs every 10 minutes from Shanghai ’s business center to the Pudong International Airport. Reaching speeds over 300 miles an hour, it whisks people to the airport 20 miles away in less than 8 minutes. However, according to the vice director of the train company, “We are not lucky with ticket sales,” since the trains are virtually empty. The reason is that because to meet the project ’s time deadline and budget, the train station was located 6 miles outside the city center, requiring lengthy public transportation to get there. So in spite of the technical, budget, and timing suc- cess of the project, it failed to meet the needs of the passengers. China is currently investigating extending the line to the downtown area, but that will be a much more expensive and time-consuming project (Project Management Institute 2004a).

• Boston ’s “Big Dig” highway/tunnel project is considered one of the largest, most com- plex, and technologically challenging highway projects in U.S. history. In early 2003, the Big Dig, originally expected to cost less than $3 billion, was declared complete, after two decades and over $14 billion spent for planning and construction. This project was clearly one that offered little value to the city if it wasn ’t completed, so it continued far past what planners thought was a worthwhile investment, primarily because the federal government was paying 85 percent of its cost. With an estimated benefi t of $500 million per year in reduced congestion, pollution, accidents, fuel costs, and lateness, but a total investment cost of $14.6 billion (a 470 percent cost overrun), it is expected to take 78 years to pay its costs back. The overrun is attrib- uted to two major factors: (1) a major underestimate of the initial project scope, typi- cal of government projects, and (2) lack of control, particularly of costs and including confl icts of interest between the public and private sectors. One clear lesson from the project has been that unless the state and local governments are required to pay at least half the cost of these megaprojects, there won ’t be serious local deliberation of their pros and cons (Abrams 2003 ; Project Management Institute 2004b).

I n t r o d u c t i o n 205

c06.indd 205c06.indd 205 25/09/12 10:48 AM25/09/12 10:48 AM

206 C h a p t e r 6 : M a n a g i n g P r o c e s s I m p r o v e m e n t P r o j e c t s

From the above examples, we see the ubiquitous role of project management in all types of organizations and in all countries. And we see how some tools can help in the management and control of projects, as well as what can go wrong when such man- agement is missing. Project management is concerned with managing organizational activities that result in some particular, desired output. Although we have titled this chapter Managing Process Improvement Projects, it should be clear that this is only one use among many for project management. For example, in the traditional func- tional organization, a product development team with representatives from produc- tion, fi nance, marketing, and engineering can be assembled to ensure that new product designs simultaneously meet the requirements of each area. Ensuring that each area ’s requirements are being met as the new design is developed reduces the likelihood that costly changes will have to be made later in the process. The result is that new products can be developed faster and less expensively, thereby enhancing the fi rm ’s overall responsiveness. Perhaps a better product is developed as well, owing to the synergy of including a variety of different perspectives earlier in the design process.

In this chapter, we describe the many activities required in the successful man- agement of projects. We start with the defi nition of a project and why project management is different from managing functional activities. We then move into the project life-cycle activities, starting with planning the project, which includes an understanding of the role of the project in the organization ’s strategy. We next describe the two major types of project life cycles and why it is important to be able to tell which is applicable for the project at hand. This is then followed by a discus- sion about organizing the project team and the various techniques available to the project manager for planning the project activities. Following this, we discuss the major topics of scheduling the project and determining the probability of completing it by its due date. In the process, we describe the capabilities and outputs of some project management software packages. Moving along the project life cycle, we then address the topics of controlling the project ’s cost and performance. The chapter concludes with a discussion of the “critical chain” concept of project management.

D E F I N I N G A P R O J E C T Up to this point, you might not have realized that projects are actually a special type of process. As described in Chapter 2, the term process refers to a set of activities that, taken together, create something of value to customers. Typically, the term pro- cess is used to refer to a set of activities that are routinely repeated, such as process- ing insurance forms, handling customers ’ complaints, and assembling an MP3 player. The term project also refers to a set of activities that, taken together, produce a val- ued output. However, unlike a typical process, each project is unique and has a clear beginning and end. Therefore, projects are processes that are performed infrequently and ad hoc, with a clear specifi cation of the desired objective.

There are two other typical characteristics of projects that are less obvious. One is that there is a limited budget to attain the unique desired objective. The second is that the objective is extremely important to the organization. If neither of these two char- acteristics exists, it would be foolish to designate a special project team to accomplish the project, since, with an unlimited budget, it could be done by regular functional departmental employees doing their routine work.

c06.indd 206c06.indd 206 25/09/12 10:48 AM25/09/12 10:48 AM

207D e f i n i n g a P r o j e c t

In Chapter 2, the project form of the transformation process was briefl y described. The choice of the project form usually indicates the importance of the project objec- tive to the organization and to the many other stakeholders , such as the client, sub- contractors, consultants, the project team, the project management offi ce, the government (sometimes), the community, and possibly others. Thus, top-grade resources, including staff, are often made available for project operations. As a result, project organizations become professionalized and are often managed on that basis. That is, minimal supervision is exercised, administrative routine is minimized, and the project team professionals are charged with solving the problem and obtain- ing the required results (cost, performance, deadline). The project team is then given the privacy and freedom to decide how to solve the problem.

Projects frequently require different emphases during their life cycle. For exam- ple, technical performance may be crucial at the beginning, cost overruns in the middle, and on-time completion at the end. The fl exibility of making spur-of-the- moment changes in emphasis by trading off one criterion for another is basic to the project design form. This ability results from the close contact of the project manager with the technical staff—there are few, if any, “middle managers.”

Following are some examples of projects:

• Constructing highways, bridges, tunnels, and dams • Building ships, planes, and rockets, or a doghouse • Erecting skyscrapers, steel mills, homes, and processing plants • Locating and laying out amusement parks, camping grounds, and refuges • Organizing conferences, banquets, conventions, and weddings • Managing R&D projects such as the Manhattan Project (which developed the

atomic bomb)

• Running political campaigns, war operations, advertising campaigns, or fi re- fi ghting operations

• Chairing ad hoc task forces, overseeing planning for government agencies, or conducting corporate audits

• Converting from one computer system to another As may be noticed by reading this list, the number of project operations is grow-

ing in our economy, probably at about the same rate as services (which many of them are). Some of the reasons for this growth in project operations are as follows:

1. More sophisticated technology . An outgrowth of our information age, and its technology, has been increased public awareness of project operations (e.g., Project Apollo) and interest in using the project form to achieve society ’s goals (Project Head Start).

2. Better-educated citizens . People are more aware of the world around them and of techniques (such as project management) for achieving their objectives.

3. More leisure time . People have the time available to follow, and even partici- pate in, projects.

4. Increased accountability . Society as a whole has increased its emphasis on the attainment of objectives (affi rmative action, environmental protection,

c06.indd 207c06.indd 207 25/09/12 10:48 AM25/09/12 10:48 AM

208 C h a p t e r 6 : M a n a g i n g P r o c e s s I m p r o v e m e n t P r o j e c t s

better fuel economy) and the evaluation of activities leading toward those objectives.

5. Higher productivity . People and organizations are involved in more activities, and are more productive in those activities, than ever before.

6. Faster response to clients . Today ’s intense competition has escalated the importance of quick response to clients ’ needs, and projects are often more responsive and fl exible than bureaucracies or functionally organized fi rms.

7. Greater customization for clients . Intense competition has also increased the importance of better meeting the client ’s unique needs in terms of both the service and the facilitating good. The project form of organizing is more likely to meet this need.

In physical project operations, such as bridge construction, most of the produc- tion per se is completed elsewhere and brought to the project area at the proper time. As a result, a great many project activities are assembly operations. The project design form concentrates resources on the achievement of specifi c objectives, pri- marily through proper scheduling and control of activities, many of which are simul- taneous. Some of the scheduling considerations in project management are knowing what activities must be completed and in what order, how long they will take, when to increase and decrease the labor force, and when to order materials so that they will not arrive too early (thus requiring storage and being in the way) or too late (thus delaying the project). The control activities include anticipating what can and might go wrong, knowing what resources can be shifted among activities to keep the project on schedule, and so forth.

P L A N N I N G T H E P R O J E C T In this section, we focus in detail on the planning of projects. In the area of project management, planning is probably the single most important element in the success of the project, and considerable research has been done on the topic. We start with the role of the organization ’s many projects in achieving its strategy, known as the organization ’s project portfolio . This portfolio evolves over time, since projects have a fi nite life cycle, as discussed in the next subsection. Following this, we discuss the project team and its tie to the parent organization. Last, we describe the actual project planning tools.

The Project Portfolio The long-term purpose of projects in the organization is to ultimately achieve the organization ’s goals. This is accomplished through the project portfolio, also known as the organization ’s aggregate project plan . In making project selection decisions, it is vital to consider the interactions among various projects and to manage the projects as a set. This is in stark contrast to the common practice of simply setting a project budget and specifi ed return on investment (ROI) hurdle rate, then funding projects until either the budget or supply of acceptable projects is exhausted.

c06.indd 208c06.indd 208 25/09/12 10:48 AM25/09/12 10:48 AM

209P l a n n i n g t h e P r o j e c t

Organizations that fund all projects that meet their ROI criterion typically end up with signifi cantly more ongoing projects than they can competently manage, and thus their contribution to the organization ’s long-term goals can be lost. Because ROI is an insuffi cient selection criterion, the set of projects chosen may not consti- tute close to an optimal portfolio for achieving their purpose.

In an attempt to better tie the fi rm ’s product development projects to its strategic objectives, Professors Wheelwright and Clark (1992) of the Harvard Business School developed a framework for categorizing projects that they call the aggregate project plan. The purpose of the framework is to illustrate the distribution of all the organi- zation ’s product/service design projects across a variety of measures such as resource demands, innovativeness, product lines, time, and project type. It is important to point out, however, that it is typically not a single project that determines the organi- zation ’s long-run success, but rather the set of research projects pursued by the organization, or its project portfolio . Therefore, in making project selection decisions, it is vital to consider the interactions among various projects and to manage the projects as a set in order to achieve the organization ’s strategic objectives.

Using this framework, output development projects are categorized along two dimensions: (1) the extent of changes made to the output and (2) the degree of process change. Based on these two dimensions, projects can then be categorized into the following four categories, as shown in Figure 6.1 :

1. Derivative projects . Derivative projects seek to make incremental improve- ments in the output and/or process. Projects that seek to reduce the output ’s cost or make minor product line extensions exemplify these types of projects. Developing a stripped-down version of a notebook computer or adding a

Figure 6.1 The aggregate project plan.

R & D projects

Extensive process changes

Extensive product changes

Minor process changes

Minor product changes

Platform projects

Derivative projects

Breakthrough projects

c06.indd 209c06.indd 209 25/09/12 10:48 AM25/09/12 10:48 AM

210 C h a p t e r 6 : M a n a g i n g P r o c e s s I m p r o v e m e n t P r o j e c t s

new menu item at a fast-food restaurant would qualify as a derivative project. This category accounts for a large majority of all innovations.

2. Breakthrough projects . These projects are at the opposite end of the continuum from derivative projects and typically seek the development of a new generation of outputs. A computer that operates by voice recognition as opposed to a keyboard and mouse, or an entirely online grocery store are examples of breakthrough projects.

3. Platform projects . Platform projects fall between derivative and breakthrough projects. In general, the result of these projects is an output that can serve as the platform for an entire line of new outputs. A key difference between plat- form projects and breakthrough projects is that platform projects stick with existing technology. As an example, the development of an ultrathin netbook computer would qualify as a platform project. If this computer succeeds, it could serve as the basis for a number of derivative projects focusing on cost improvement and the development of other computer models with different features.

4. R&D projects . R&D projects entail working with basic technology to develop new knowledge. Depending on its focus, an R&D project might lead to breakthrough, platform, or derivative innovations.

Use of the aggregate project plan requires that all projects be identifi ed and plotted. The size of the points plotted for each project should be proportional to the amount of resources the project will require. In Figure 6.2 , we have used different shapes to indi- cate different types of projects. Internal projects are plotted using circles, while projects pursued as part of a strategic alliance with other fi rms are plotted using squares.

Figure 6.2 An example aggregate project plan.

Extensive process changes

Extensive product changes

Minor process changes

Minor product changes

c06.indd 210c06.indd 210 25/09/12 10:48 AM25/09/12 10:48 AM

211P l a n n i n g t h e P r o j e c t

There are a number of ways the aggregate project plan can be used. The identifi - cation of gaps in the types of projects being undertaken is probably most important. For example, are the types of projects undertaken too heavily skewed toward derivatives-type projects? This might indicate an inadequate consideration of the fi rm ’s long-run competitive position. Also, the aggregate project plan facilitates eval- uation of the resource commitments of the ongoing as well as proposed projects. Finally, this framework can serve as a model for employee development. New employees can be initially assigned to a team working on a derivative project. After gaining experience, employees can be assigned to a platform project, then assigned to manage a derivative project. As managerial skill accumulates, the employee will qualify for larger and more valuable projects. Of course, we must remember that the fundamental purpose of this entire process is to ensure that the set of projects accu- rately refl ects the organization ’s strategic goals and objectives.

As one example, Hewlett-Packard (HP) found itself in the common position of having more ongoing projects than it could effectively control, so it turned to the use of an aggregate project plan. By intensely scrutinizing all their projects and taking into consideration the gaps and excesses on their aggregate project plan, HP was able to prioritize the projects and concentrate on those that made the greatest con- tribution to their strategic goals for the least resource use. The initiative helped one HP organization systematically reduce 120 projects down to 30, and another organi- zation from 50 to 17, thereby increasing the chances of success for the most impor- tant projects (Englund and Graham 1999).

The Project Life Cycle It has been found, for example, that progress in a project is rarely uniform, but instead often follows one of two common forms, as shown in Figure 6.3 . In the stretched-S life-cycle form, illustrated in Figure 6.3 a , when the project is initiated, progress is slow as responsibilities are assigned and organization takes place. But the project gathers speed during the implementation stage, and much progress is made. As the end of the project draws near, the more diffi cult tasks that were post- poned earlier must now be completed, yet people are being drawn off the project and activity is “winding down,” so the end keeps slipping out of reach.

In the exponential form, illustrated in Figure 6.3 b , after the project is initiated there is continuous activity on numerous aspects of the project, but until all the elemental parts come together at the end, there is no fi nal output. This is typical of projects that require fi nal assembly of components to produce the whole (like a car) or goods (like a cake, which is only glop until it is baked in the oven). It is especially typical of offi ce and other such service work where the fi nal output is a life insurance policy, or ad piece, or perhaps even an MBA degree. Without that last signature, or piece of paper, or earned credit, there is virtually no product. (However, if a student is audit- ing courses with the goal of understanding rather than getting a degree, the progress toward their goal may indeed be linear with every day spent in class.)

The reason it is important to contrast these two forms, besides pointing out their difference in managerial needs, is that during the budgeting stage, if there is a fl at across-the-board budget cut of, say, 10 percent and the project is of the stretched-S form, then not being able to spend that last 10 percent of the budget is of no urgency, since probably 97 percent of the benefi ts will be achieved anyway. However, if the

c06.indd 211c06.indd 211 25/09/12 10:48 AM25/09/12 10:48 AM

212 C h a p t e r 6 : M a n a g i n g P r o c e s s I m p r o v e m e n t P r o j e c t s

project is of the exponential form, then missing the last 10 percent is catastrophic because this is where all the value is attained. Another perspective on the same issue is the effect of early termination of the project. Terminating the stretched-S form early will have negligible impact, but terminating the exponential form will be a complete disaster. It is imperative that the project manager and top management know which type of project they are working with before taking such actions.

Projects in the Organizational Structure One of the factors in the success of a project is making sure it is reporting to the proper level of management in the organization. In a functionally structured organization with the typical departments of marketing, human resources (HR), engineering, and so on, projects are frequently housed in the department or division that has a major interest in the project ’s success or has the special skills to ensure that it will be executed properly. More generic projects, such as a division-wide effort to speed up product development, might report to the division head or a vice president, and the project team is composed of members drawn from many of the individual departments.

Some organizations are structured by projects, called “projectized” organizations. Construction, engineering, legal, consulting, and auditing fi rms are frequently struc- tured this way; the line units are the particular contracts the fi rm is working on, while the administrative support groups, like HR and fi nance, are staff units that report higher up in the organization but render support to all the projects. If some of the contracts are very large and long-running, the project may have it’s own dedicated staff.

In some ways, the projectized organization is at the opposite end of the spectrum from the functional organization, and, as might be suspected, each has its own advan- tages and disadvantages. For example, functionally organized projects have the advantage of tremendous depth of knowledge in the department relating to the goals of the project. But projectized projects have the advantage of having the entire project team all directed to the same goal and reporting to a single project manager.

Figure 6.3 Two project life cycles. ( a ) Stretched-S. ( b ) Exponential.

Time (a)

Project initiation

Project implementation

Project termination

% p

ro je

ct c

o m

p le

ti o

n

Time (b)

Project initiation

Project implementation

Project termination

% p

ro je

ct c

o m

p le

ti o

n

c06.indd 212c06.indd 212 25/09/12 10:48 AM25/09/12 10:48 AM

213P l a n n i n g t h e P r o j e c t

In an attempt to obtain the benefi ts of each of these ways of organizing projects, some fi rms have adopted what is known as a “matrixed” structure, where projects draw their workers from each appropriate functional department (the columns of the matrix) but the workers are grouped together into independent projects (the rows). Thus, each worker has two bosses, a functional boss and a project boss. This arrange- ment often does gain some of the advantages of each of the two major organizational arrangements, but it also has some additional disadvantages, such as confl icts between the two bosses. Combinations of these forms are also common, such as the “weak” or functional matrix, and the “strong” or project matrix. Each way of organizing the project has its own advantages and disadvantages, and what works best depends largely on the circumstances of the organization and the reason it started a project.

Organizing the Project Team Regardless of the form of the project, a team will be required to run it. Some mem- bers of the team may be directly assigned to the project manager for the duration, while others may have only partial responsibilities for the project and still report to their functional superior. There are three types of team members who should report directly to the project manager (PM), however:

• Those who will be having a long-term relationship with the project • Those with whom the PM will need to communicate closely or continuously • Those with rare skills necessary to project success Yet even if these people report to the PM, it is still not common for the PM to have

the authority to reward these people with pay bonuses, extra vacation, or other such personnel matters—that authority normally still resides with the functional manager. Thus, there are not a lot of incentives the PM can give people for working hard on the project. The main ones are the fun and excitement of the challenge, and doing something that will be important to the organization.

With the pressures that tend to gravitate toward such important and high-profi le projects, it may be assumed that there is also a lot of opportunity for confl ict to arise. This is true, and not only between the PM and other organizational units, but even among members of the project team. According to Thamhain and Wilemon (1975), at project formation the main sources of confl ict are priorities and procedures. As the project gets under way, priorities and schedules become the main points of confl ict. During the main implementation stage, confl ict shifts to technical issues and sched- ules. But toward the end of the project, when timing is becoming crucial, only schedules are a source of confl ict. Knowing when to expect trouble, and what kinds, throughout the project can help the PM keep peace within the project team and facilitate smooth project progress.

Another major factor in the success of a project is selecting the proper project manager. The responsibilities of and special demands on a PM are extensive (e.g., acquiring resources, motivating the team, dealing with obstacles, making goal trade- offs, communicating across a range of stakeholders), but four major attributes stand out (Meredith and Mantel 2012):

1. Credibility. The PM needs both technical and administrative credibility, technical to handle the project ’s details and administrative to handle the range of important stakeholders.

c06.indd 213c06.indd 213 25/09/12 10:48 AM25/09/12 10:48 AM

214 C h a p t e r 6 : M a n a g i n g P r o c e s s I m p r o v e m e n t P r o j e c t s

2. Sensitivity. The PM must be sensitive to both politics and personalities. Projects are pressure-cookers and interpersonal confl icts can erupt without warning, not only between team members but also with higher administra- tion and outsiders. The sensitive PM will detect a potential confl ict and then confront and defuse it before it turns into a crisis.

3. Leadership, ethics, and managerial style. Clearly, the PM needs many competencies to run complex projects, but different types of projects need more emphasis on certain competencies than others. For example, the PM will require vastly different competencies in an HR versus a construction project.

4. Ability to handle stress. Projects are created because the normal function- ing of the organization isn ’t able to handle the special requirements of the project. Hence, stress is expected in every project, but tight schedules, insuf- fi cient resources, and diffi cult (or constantly changing) goals add substan- tially to a PM ’s stress level.

Some people enjoy the job of a project manager and others don ’t, so selecting a PM who has a good record on past projects is one indicator of the proper PM for a project. Another is whether the person shows interest in the profession of project management, such as by joining the Project Management Institute (PMI), knowing the project management body of knowledge (PMBOK), or becoming certifi ed as a Project Management Professional or holding one of the other certifi cations offered by PMI and other organizations. (For more on PMI, PMBOK, and certifi cation, refer to Table 1-1 in Meredith and Mantel 2012.)

Project Plans The initiation of a project should in most cases include the development of some level of project charter (also known as the project plan ), unless the project is highly routine. The elements that constitute the project plan and form the basis for more detailed planning of the budgets, schedules, work plan, and general management of the project are described below. It should be noted that the process of developing the project charter varies from organization to organization, but any project charter should contain some level of information regarding the following elements:

• Overview . This is a short summary of what the client expects from the project. It is directed to top management and contains a statement of the goals of the project, a brief explanation of their relationship to the fi rm ’s objectives, a description of the managerial structure that will be used for the project, and a list of the major milestones in the project schedule.

• Goals, or scope . This contains a more detailed statement of the general goals noted in the overview section and the specifi c requirements of the stakeholders, sponsor, and client that must be satisfactorily completed.

• Business case . This describes the justifi cation for the project in terms of the benefi ts to the project organization. It includes the expected profi ts and ROI (return on investment), of course, but also other gains such as experience, establishing a strong track record, working with an important client, and so on, which may be more important than profi ts.

c06.indd 214c06.indd 214 25/09/12 10:48 AM25/09/12 10:48 AM

215P l a n n i n g t h e P r o j e c t

• General approach . This describes both the managerial and the technical approaches to the work, such as whether the project is an extension of work done by the company for an earlier project and whether there are any deviations from routine procedure—for instance, the use of subcontractors for some parts of the work.

• Contractual aspects . This includes a complete list and description of all reporting requirements, customer-supplied resources, liaison arrangements, and so on, as well as the technical deliverables and their specifi cations, delivery schedules, and the procedures for changing any of the above.

• Schedule and milestones . This outlines the schedule and lists milestone events. Each task is listed, and the estimated time for each task should be obtained from those who will do the work. The project master schedule is constructed from these inputs.

• Resources . There are two primary aspects to be considered here. The fi rst is the project budget . Both capital and expense requirements are detailed by task. Second, cost monitoring and control procedures should be described.

• Personnel . This lists both who must be involved in the review and approval process as well as the time-phased personnel requirements of the project, that is, the team and other involved departments. If known, the name and author- ity level of the project manager should also be included here. Special skills, types of training needed, possible recruiting problems, legal or policy restric- tions on workforce composition, and any other special requirements, such as security clearances, should be noted. Time-phasing the personnel needs makes clear to management and other departments when the various types of contributors are needed and in what numbers.

• Risk management plan . This covers potential problems that could affect the project. One or more issues, such as subcontractor default, unexpected technical breakthroughs, strikes, hurricanes, new markets for the fi rm’s tech- nology, tight deadlines and budgets, and sudden moves by a competitor, are certain to occur—the only uncertainties are which, when, and their impact. Plans to deal with unfavorable (or favorable) contingencies should be devel- oped early in the project ’s life.

• Evaluation method . Every project should be evaluated against standards and by methods established at the project ’s inception. This includes a brief description of the procedures to be followed in monitoring, collecting, stor- ing, and evaluating the history of the project.

Almost by defi nition, a project is an attempt to meet specifi ed performance or “scope” requirements by a specifi c deadline within a limited budget. These objec- tives are generally illustrated in Figure 6.4 . The project plan described above lays out these three objectives in detail, but the task for the manager is to “make it happen.”

To achieve these three project objectives, one of the project manager ’s fi rst responsibilities is to defi ne all the tasks in as much detail as possible so that they can be scheduled and costed out and so that responsibility can be assigned. This set of task descriptions, called the work breakdown structure (WBS), provides the inputs for the project schedule (usually put into a format known as the project Gantt chart ) and the linear responsibility chart that depicts the tasks of those outside the project team but with responsibilities related to the project. The linear responsibility chart is

c06.indd 215c06.indd 215 25/09/12 10:48 AM25/09/12 10:48 AM

216 C h a p t e r 6 : M a n a g i n g P r o c e s s I m p r o v e m e n t P r o j e c t s

similar to a RACI matrix, which stands for the four main project responsibilities: responsible (for a task), approval, consult (or coordinate for support), and inform (notify). The main difference is that the letters R, A, C, and I are put in place of the symbols in a RACI matrix, and a linear responsibility chart may have other personnel involvement included as well, such as “initiate” or “supervise.”

A typical WBS and project schedule are illustrated in Figures 6.5 and 6.6 for a project installing robots on a manufacturing assembly line. Milestone, commitment,

Performance (“scope”)

Target

Time (“schedule”)

Cost

Required performance

Budget limit

Due date

Figure 6.4 Three project objectives. Reprinted with permission from J. Meredith and S. J. Mantel, Jr., Project Management: A Managerial Approach , 8th ed. New York: Wiley, 2012.

Figure 6.5 Work breakdown structure.

E. Install and

start-up

A. Determine

need

B. Solicit

quotations

C. Appropriation

request

Robot installation

D. Purchase resources

2. Tooling

1. Order

3. Write

2. $

1. Tooling

1. Contacts

2. Size, type

1. Benefit

3. Gages

1. Install

3. Run-off

2. Train

c06.indd 216c06.indd 216 25/09/12 10:48 AM25/09/12 10:48 AM

217P l a n n i n g t h e P r o j e c t

and completion points are shown, and actual progress is graphed. The last status update shows that the project is a month behind schedule. Figure 6.7 illustrates the linear responsibility chart or RACI matrix for the robot project.

The scheduling of project activities is highly complex because of (1) the number of activities required, (2) the precedence relationships among the activities, and (3) the limited time of the project. Project scheduling is similar to the scheduling discussed earlier in some ways but still differs signifi cantly. For example, the basic network approaches— program evaluation and review technique (PERT) and critical path method (CPM)—are based on variations of the Gantt chart. Figure 6.6 is a project Gantt chart but is inadequate for scheduling the multitude of subtasks that compose,

Figure 6.6 Project baseline schedule. Reprinted with permission from J. Meredith and S. J. Mantel, Jr., Project Management: A Managerial Approach , 8th ed. Hoboken, NJ: Wiley, 2012.

Install and start-up

Purchase machine,

tooling, and gages

Write appro- priation request

Solicit quotations

Determine need

Subproject Task 20X4

J F M A M J J A S O N D J F M A M J J A S O N D

20X5

A1

A2

B1

C1

C2

C3

D1

D2

D3

E1

E2

E3

Industrial engr. (IE)

Responsible Dept.

Project engr. (PE)

Project engr. (PE)

Project engr. (PE)

Tool design

Industrial engr. (IE)

Purchasing

Tool design

Quality control

Plant layout

Personnel

Manufacturing

Dependent Dept.

Industrial engr. (IE)

Finance, IE, purchasing

Tool design, finance, IE

Industrial engr. (IE)

Industrial engr. (IE)

Project engr. (PE)

Purchasing, tool making

Tool design, purchasing

Mill- wrights

Process engr.,

manufacturing

Quality control

Note: As of 31 Jan., 20X5 the project is one month behind schedule. This is due mainly to the delay in task C1, which was caused by the late completion of A2.

Find operations that benefit most

Approx. size and type needed

Contact suppliers & review quotes

Determine tooling costs

Determine labor savings

Actual writing

Order robot

Design and order or manufacture tooling

Specify needed gages and order of manufacturing

Install robot

Train employees

Run-off

Legend:

Project completion

Contractual commitment

Planned completion

Actual completion

Status date

Milestone planned

Milestone achieved

Planned progress

Actual progress

c06.indd 217c06.indd 217 25/09/12 10:48 AM25/09/12 10:48 AM

218 C h a p t e r 6 : M a n a g i n g P r o c e s s I m p r o v e m e n t P r o j e c t s

for example, task A1. That is, a project schedule has to handle an enormous number of different activities, which must be coordinated in such a way that the subsequent activities can take place and the entire project (job) can be completed by the due date.

The scheduling procedure for project operations must be able not only to identify and handle the variety of tasks that must be done, but also to handle their time sequencing. In addition, it must be able to integrate the performance and timing of all the tasks with the project as a whole so that control can be exercised, for exam- ple, by shifting resources from operations with slack (permissible slippage) to other operations whose delay might threaten the project ’s timely completion. The tasks involved in planning and scheduling project operations are as follows:

• Planning . Determining what must be done and which tasks must precede others

• Scheduling . Determining when the tasks must be completed; when they can and when they must be started; which tasks are critical to the timely comple- tion of the project; and which tasks have slack in their timing and how much

S C H E D U L I N G T H E P R O J E C T The project scheduling process is based on the activities that must be conducted to achieve the project ’s goals, the length of time each requires, and the order in which they must be completed. If a number of similar projects must be conducted, sometimes these activities can be structured generically to apply equally well to all the projects.

Responsibility

Project OfficeWBS

Subproject Task

A1 A

A

I A

I

I R

C I

C

C

R C

C

I

I

R

A

R

A

R

R

C

C

R

A2

B1

C1

C2

C3

Project Manager

Contract Admin.

Project Eng.

Industrial Eng.

Field Manager

Field Oper.

Determine need

Solicit quotations

Write approp. request.

Legend: Responsible

Consult Inform

"

"

"

"

"

"

A Approval

Figure 6.7 Linear responsibility chart or RACI matrix. Reprinted with permission from J. Meredith and S. J. Mantel, Jr., Project Management: A Managerial Approach , 8th ed. Hoboken, NJ: Wiley, 2012.

c06.indd 218c06.indd 218 25/09/12 10:48 AM25/09/12 10:48 AM

219S c h e d u l i n g t h e P r o j e c t

Two primary techniques have been developed to plan projects consisting of ordered activities: PERT and CPM. Although PERT and CPM originally had some dif- ferences in the way activities were determined and laid out, many current approaches to project scheduling minimize these differences and present an integrated view, as we will see here. It will be helpful to defi ne some terms fi rst.

• Activity . One of the project operations, or tasks; an activity requires resources and takes some amount of time to complete.

• Event . Completion of an activity, or series of activities, at a particular point in time.

• Network . The set of all project activities graphically interrelated through prec- edence relationships. In this text, boxes (called nodes ) represent activities and arrows between the boxes represent precedence. (This is typical of the CPM approach; in PERT the arrows represent activities.)

• Path . A series of connected activities from the start to the fi nish of the project. • Critical path . Any path that, if delayed, will delay the completion of the entire

project.

• Critical activities . The activities on the critical path or paths. We next illustrate the process of scheduling with a Six Sigma process improve-

ment project example. We use the DMAIC approach (see Chapter 4) to improve a bank ’s process for handling mortgage refi nancing applications.

Project Scheduling with Certain Activity Times: A Process Improvement Example The primary inputs to project planning are a list of the activities that must be com- pleted, the activity completion times (also called activity durations ), and precedence relationships among the activities (i.e., what activities must be completed before another activity can be started). In this section, we assume that activity completion times are known with certainty. Later, we relax this assumption and consider situa- tions in which activity completion times are uncertain.

Important outputs of project scheduling include the following:

• Graphical representation of the entire project, showing all precedence rela- tionships among the activities

• Time it will take to complete the project • Identifi cation of critical path or paths • Identifi cation of critical activities • Slack times for all activities and paths • Earliest and latest time each activity can be started • Earliest and latest time each activity can be completed

Project Completion and Critical Paths

Table 6.1 shows the activity times and precedence for the ten activities that must all be fi nished to complete a bank ’s process improvement project. According to the

c06.indd 219c06.indd 219 25/09/12 10:48 AM25/09/12 10:48 AM

220 C h a p t e r 6 : M a n a g i n g P r o c e s s I m p r o v e m e n t P r o j e c t s

table, activities A, B, and C can be started at any time. Activity D can be started once activity A is completed. Activities E, F, and G cannot be started until both activities B and C are fi nished, and so on. The network diagram for this project is shown in Figure 6.8 , in which ellipses show the start and end of the project, arrows represent the required precedence, and rectangular nodes represent activities A–J in Table 6.1 . The rectangular nodes list the activity by letter, followed by its expected time (in days). This way of depicting a project is known as activity-on-node (AON) and is typical of CPM; the PERT alternative, activity-on-arrow (AOA), is also common, how- ever (see Meredith and Mantel 2012 for examples).

T A B L E 6 .1 • Da ta fo r a Mor tgage Re f inanc ing Pro j ec t Activity Expected Time, t

e Preceding Activities

A: Identify all stakeholders 10 —

B: Develop the project charter 10 —

C: Uncover all relevant regulations 5 —

D: Set up project procedures 7 A

E: Determine total refi nancing time 5 B, C

F: Use accounting data for total cost 7 B, C

G: Interview to determine unknown risks 2 B, C

H: Redesign so as to reduce task times 5 C

I: Determine cost reductions of new design 8 G, H

J: Uncover any new constraints on design 4 D, E

A, 10 0, 10 0, 10

B, 10 0, 10 1, 11

0

Start

21

End

C, 5 0, 5 3, 8

D, 7 10, 17 10, 17 J, 4

17, 21 17, 21

Activity, te TES, TEF TLS, TLF

E, 5 10, 15 12, 17

F, 7 10, 17 14, 21

G, 2 10, 12 11, 13

H, 5 5, 10 8, 13

I, 8 12, 20 13, 21

Figure 6.8 Network diagram for process improvement project.

c06.indd 220c06.indd 220 25/09/12 10:48 AM25/09/12 10:48 AM

221S c h e d u l i n g t h e P r o j e c t

To determine the expected completion time of each of the nodes on the network and thus the entire project, early start times T

ES and early fi nish times T

EF are calculated

for each activity, as shown in Figure 6.8 . The values of T E S and T

E F are calculated mov-

ing left to right through the network. Thus, we begin with the Start node and work our way to the End node. To illustrate, the project starts at time zero (sometimes this is not the case), then activities A, B, and C can also be started as early as time zero, since none of them are preceded by another activity. Since activity A requires 10 days, if it is started at time zero ( T

E S ), it can be completed ( T

E F ) as early as day 10. Likewise,

if activity B is started at time zero, it can be completed as early as day 10 and activity C can be completed by day 5. Continuing on, since activity A can be fi nished as early as day 10, activity D can start as early as day 10 and, since it takes seven days, can fi nish as early as day 17. The same logic applies to activities E through H.

Now consider activity J. Activity J cannot be started until activities D and E are both completed. Since activity D can be fi nished as early as day 17 and E can be fi nished as early as day 15, activity J can be started only as early as day 17, the latest of its preceding activities (remember, J cannot start until both activities D and E are completed). Since J takes four days, then it can be fi nished as early as 21 days. The same logic applies to activity I. It cannot start until the latest of its predecessors is completed, 12 for G and 10 for H, and hence 12. Since it takes eight days, it can be completed by day 20. Now that T

E S and T

E F have been calculated for all the activi-

ties, we can determine the earliest time that the project can be completed. Since the project “End” cannot be completed until all its predecessors are completed, the earli- est it could be completed is day 21, based on activity J.

We can now fi nd the critical path and critical activities for the project. Since the End of the project depended on activity J, we bold the arrow from J to End. Similarly, J ’s start time depended on activity D, rather than E, so we bold the arrow from D to J. And D depended on the completion of A, so we bold that arrow, and then the arrow from the Start node to A, resulting in the critical path A-D-J and the critical activities A, D, and J.

Once T E S and T

E F have been calculated for each activity, the latest times each activ-

ity can be started and fi nished without delaying the completion of the project can be determined. In contrast to T

E S and T

E F , the latest start time ( T

L S ) and latest fi nish time

( T L F ) are calculated by moving backward through the network, from right to left.

These T L S and T

L F times are also shown in Figure 6.8 .

In calculating T E S and T

E F , we determined that the project could be completed by

day 21. If the project is to be completed by day 21, then activities J, F, and I can be completed as late as day 21 without delaying completion of the project. Thus, the latest fi nish time for activities J, F, and I is 21. Since activity J requires four days, it can start as late as 21 2 4, or 17, and still fi nish by day 21. Likewise, activity I can start as late as 21 2 8, or 13, and still fi nish by day 21, and activity F can start as late as 21   2   7   5   14 . Continuing on, since activity J can start as late as day 17, activity D can fi nish as late as day 17. Since activity D requires seven days, it can start as late as day 17   2   7   5   10 without delaying the entire project. Activities G and H are handled similarly, resulting in the late start and fi nish times for activities G and H.

Now let ’s look at activity B, which precedes activities E, F, and G. The latest it can fi nish is the earliest late start date of activities E, F, and G since if it doesn ’t start by then, that activity will be late and will delay the entire project. Since the late start dates of these three activities are 12 (for E), 14, and 11, activity B ’s latest fi nish date must be 11, because if it doesn ’t fi nish by 11, activity G can ’t start and will delay

c06.indd 221c06.indd 221 25/09/12 10:48 AM25/09/12 10:48 AM

222 C h a p t e r 6 : M a n a g i n g P r o c e s s I m p r o v e m e n t P r o j e c t s

activity I, which will delay the project. The latest dates for activity C are found in the same manner.

Note in Figure 6.8 that the latest dates for the critical activities (A, D, and J) are identical to the earliest dates. That is, the latest dates cannot be delayed from the earliest dates at all, or else the entire project will be delayed! This will always be the case since this represents the critical path of the project.

Slack Time

The times T E S , T

E F , T

L S , and T

L F can be used by the project manager to help plan and

develop schedules for the project. For example, if an activity requires a key resource or individual, its earliest and latest start times provide a window during which that resource can be acquired or assigned to the project. Alternatively, if an activity falls behind schedule, the latest completion time provides an indication of whether the slippage will delay the entire project or can simply be absorbed.

Notice in Figure 6.8 that for some activities, T E S is less than its T

L S and its T

E F is less

than its T L F . In these cases, the project manager can exercise some discretion in terms

of when the activity is started and when it is completed. The amount of fl exibility the project manager has in terms of starting and completing an activity is referred to as its slack (or fl oat ) and is calculated as

Activity slack   5   T L S   2   T

E S   5   T

L F   2   T

E F

All activities on the critical path have zero slack—that is, there is no room for delay in any activity on the critical path without delaying the entire project. But activities off the critical path may delay up to a point where further delay would delay the entire project. For example, activity H has a late start time of 8 and an early start time of 5, leaving three days of possible slack. If resources for activity H are sit- ting idle and could be used to expedite activity A, for example, the project manager may choose to do this, perhaps reducing the duration of activity A by one day and bringing the project in at day 20 instead of 21, for an early completion!

In addition to calculating slack times for individual activities, slack times can be calculated for entire paths. Since all paths must be fi nished to complete the project, the time to complete the project is the time to complete the path with the longest duration. Thus, the path with the longest duration is critical in the sense that any delay in completing it will delay the completion of the entire project. Path slacks are calculated as

Path slack   5   duration of critical path   2   path duration

If we consider path C-H-I, it has a duration of 18, so its path slack is 21   2   18   5   3 days, but path B-G-I has a path slack of only 1 day. Since activity I is on both paths, its slack is always the lesser of the two paths, one day in this case.

Before leaving the topic of slack, it is important to point out that the slack times computed for individual activities are not additive over a path. To illustrate, both activities C and H have slacks of three days, but if we use those three days for activ- ity C, starting it at day 3 instead of 0, there is then no slack for activities H or I. The point is that slack times for individual activities are computed on the assumption that only one particular activity is delayed.

c06.indd 222c06.indd 222 25/09/12 10:48 AM25/09/12 10:48 AM

223S c h e d u l i n g t h e P r o j e c t

Project Scheduling with Uncertain Activity Times The previous section discussed project planning in situations where the activity completion times were known with certainty before the project was actually started. In reality, however, project activity times are frequently not known with certainty beforehand. In these cases, project managers often develop three estimates for each activity: an optimistic time t

o , a pessimistic time t

p , and a most likely time t

m . The opti-

mistic time is the amount of time the project manager estimates it will take to complete the activity under ideal conditions; that is, only one time in a hundred would it take less time than this. The pessimistic time refers to how long the activity will take to complete under the worst-case scenario; again, there is only a 1 percent chance it would ever take longer than this. The most likely time is the project manager ’s best estimate of how long the activity will actually take to complete. In addition to these three time estimates, the precedence relationships among the activities are also needed as inputs to the project planning process.

The primary outputs of project planning when activity times are not known with certainty include the following:

• Graphical representation of the entire project, showing all precedence relationships among the activities

• Expected activity and path completion times • Variance of activity and path completion times • Probability that the project will be completed by a specifi ed time • That time corresponding to certain probability of the project being complete In Table 6.2 we present the three activity times that gave rise to the expected

times in Table 6.1 . Table 6.2 also includes the variance of the expected time, whose calculation, as well as the calculation of the expected time, we describe next.

Calculating Activity Durations

The estimates of the three activity times in Table 6.2 are based on the assumption that the activities are independent of one another. Therefore, an activity whose duration

T A B L E 6 .2 • S ix S igma Ac t i v i t y Times (days)

Project Activity Optimistic

Time t o

Most Likely Time t

m

Pessimistic Time t

p

Expected Time t e ,

and Variance � 2

A 5 11 11 10, 1

B 10 10 10 10, 0

C 2 5 8 5, 1

D 1 7 13 7, 4

E 4 4 10 5, 1

F 4 7 10 7, 1

G 2 2 2 2, 0

H 0 6 6 5, 1

I 2 8 14 8, 4

J 1 4 7 4, 1

c06.indd 223c06.indd 223 25/09/12 10:48 AM25/09/12 10:48 AM

224 C h a p t e r 6 : M a n a g i n g P r o c e s s I m p r o v e m e n t P r o j e c t s

is changed will not necessarily affect the duration of the other activities. Additionally, it is assumed that the difference between t

o and t

m need not be the same as the differ-

ence between t p and t

m . For example, a critical piece of equipment may be wearing

out. If it is working particularly well, this equipment can do a task in two hours that normally takes three hours; but if the equipment is performing poorly, the task may require 10 hours. Thus, we may see nonsymmetrical optimistic and pessimistic task times for project activities, as for activities E and H in Table 6.2 . Note also that for some activities, such as B, the durations are known with certainty.

The general form of nonsymmetrical or skewed distribution used in approximat- ing PERT activity times is called the beta distribution and has a mean (expected completion time t

e ) and a variance, or uncertainty in this time, � 2 , as given below.

The beta distribution is used because it is fl exible enough to allow one tail of the distribution to be longer than the other (more things will typically go worse than expected than will go better than expected in a project) and is thus a more appropri- ate distribution for activity completion times.

t e   5

t o 1 4 t

m 1 t

p _________ 6

� 2   5   � t p   2   t o _____ 6 � 2

The above equation for the expected completion time is simply a weighted aver- age of the three time estimates, with weights of 1, 4, and 1, and the denominator of 6 is, of course, the sum of the weights. The value of 6 in the estimate of the variance, however, comes from a different source, the assumption that the optimistic and pes- simistic times are each 3 standard deviations from the mean. This only applies, how- ever, to estimates made at the 99 percent certainty level. If a manager is reluctant to make estimates at that level and feels that a 95 percent, or 90 percent, level is easier to estimate, then the equations for the standard deviation change (the approxima- tion for the mean is still acceptable, however) to:

95 %   level :   �   5   ( t p   2   t

o ) / 3 . 3

90 %   level :   �   5   ( t p   2   t

o ) / 2 . 6

The results of these calculations (at the 99 percent level) are listed in Table 6.2 . The discussion of project management with known activity times included critical

paths, critical activities, and slack. These concepts are not quite as useful in situations where activity times are not known with certainty. Without knowing the activity times with certainty, any of the paths may have the potential to be the longest path. Furthermore, we will not know which of the paths will take longest to complete until the project is actually completed. And since we cannot determine with certainty before the start of the project which path will be critical, we cannot determine how much slack the other paths have. We can, however, use probability estimates and simu- lation to help us gain more confi dence, as described in the next two subsections.

Probabilities of Completion

When activity times are not known with certainty, we cannot determine how long it will actually take to complete the project. However, using the variance of each

c06.indd 224c06.indd 224 25/09/12 10:48 AM25/09/12 10:48 AM

225S c h e d u l i n g t h e P r o j e c t

activity (the variances in Table 6.2 ), we can compute the likelihood or probability of completing the project in a given time period, assuming that the activity durations are independent of each other. The distribution of a path ’s completion time will be approximately normally distributed if the path has a large number of activities on it. (Recall from the central limit theorem in statistics that this is true regardless of the distribution of the activities themselves, beta in our case.) For example, the mean time along path A-D-J was found to be 21 days. The variance is found by summing the variances of each of the activities on the path. In our example, this would be

V p a t h

  A - D - J

  5   � A 2   1   �

D 2   1   �

J 2

  5   1   1   4   1   1

  5   6

The probability of completing this path in, say, 23 days is then found by calculat- ing the standard normal deviate of the desired completion time less the expected completion time, and using the table of the standard normal probability distribution (inside rear cover) to fi nd the corresponding probability:

Z   5   desired completion time   2   expected completion time

_______________________________________________ � __ V

  5   23   2   21 ________ � __ 6

  5   0 . 818

which results in a probability of 79 percent (see Figure 6.9 ). This can also be found in Excel ® using the NORMDIST function with the syntax   5   NORMDIST (   D ,   t

e ,   � ,   TRUE ) ,

where D is the desired time of interest, 23 days in our case. Similarly, we can calculate that completion time by which we would be, say, 90 percent sure the project would be completed. From Appendix A, we fi nd the standard normal deviate corresponding to 90 percent as about 1.28, so 21   1   ( 1 . 28

� __

V )   5   24 . 14 days. Again, this could also be found in Excel ® from the NORMINV function with syntax   5   NORMINV ( probability , t

e ,   � ) ,

which in our case would be 5 NORMINV(.90,21,2.449) 5 24.14 days. So far, we have determined only that there is a 79 percent chance that path A-D-J

will be completed in 23 days or less. If we were interested in calculating the proba- bility that the entire project will be completed in 23 days, we would need to calcu- late the probability that all paths will be fi nished within 23 days. To calculate the

21

Time (days)

F re

q u

en cy

Area = 79%

23

V = 6

Figure 6.9 Probability distribution of path completion time.

c06.indd 225c06.indd 225 25/09/12 10:48 AM25/09/12 10:48 AM

226 C h a p t e r 6 : M a n a g i n g P r o c e s s I m p r o v e m e n t P r o j e c t s

Based on historical data, it has been determined that all the activity times are approximately normally distributed, with the means and standard deviations given in the nodes of Figure 6.10 following the letter of the activity. Inspection of the net- work diagram reveals three paths: A-C-F, B-D-F, and B-E.

To simulate the completion of this project using Crystal Ball ® (see www.oracle.com/ crystalball/index.html for details on Crystal Ball ® ), the spreadsheet in Figure 6.11 was developed. In the spreadsheet, completing the project is simulated by generat- ing random numbers for the six activities and then adding up the activity times that

probability that all paths will be fi nished in 23 days or less, we fi rst calculate the probability that each path will be fi nished in 23 days or less, as we just did for path A-D-J. Then we multiply these probabilities together to determine the probabil- ity that all paths will be completed by the specifi ed time. The reason we multiply these probabilities together is that we are assuming that path completion times are independent of one another. Of course, if the paths have activities in common, they are not truly independent of one another and a more complex analysis or simula- tion, illustrated next, is necessary.

To simplify the number of calculations required to compute the probability that a project will be completed by some specifi ed time, for practical purposes it is reasona- ble to include only those paths whose expected time plus 2.33 standard deviations is more than the specifi ed time. The reason for doing this is that if the sum of a path ’s expected time and 2.33 of its standard deviations is less than the specifi ed time, then the probability that this path will take longer than the specifi ed time is very small (i.e., less than 1 percent), and therefore we assume that the probability that it will be com- pleted by the specifi ed time is 100 percent. Finally, note that to calculate the probability that a project will take longer than some specifi ed time, we fi rst calculate the probabil- ity that it will take less than the specifi ed time and then subtract this value from 1.

Simulating Project Completion Times

When activity times are uncertain, it is usually not possible to know which path will be the critical path before the project is actually completed. In these situations, simu- lation analysis can provide some insights into the range and distribution of project completion times. To illustrate this, we use the network diagram in Figure 6.10 con- sisting of six activities labeled A through F.

Figure 6.10 Network for simulating.

Start

A, 32.1, 1.2 C, 22.2, 2.2

B, 24.6, 3.1

D, 26.1, 5.2

E, 34.4, 6.2

F, 34.5, 4.1

End

c06.indd 226c06.indd 226 25/09/12 10:48 AM25/09/12 10:48 AM

227S c h e d u l i n g t h e P r o j e c t

make up each path to determine how long the paths take to complete. The longest path determines the project completion time.

In the spreadsheet, randomly generated activity times from a normal distribution for each activity are generated in cells A3:F3 by defi ning these cells as assumption cells. For example, cell A3 was defi ned as an assumption cell with a normal distribu- tion and mean and standard deviation of 32.1 and 1.2, respectively. In column G the time to complete path A-C-F is calculated based on the activity times generated in cells A3:F3. For example, in cell G3, the formula 5 A3 1 C3 1 F3 was entered. In a similar fashion, cells H3 and I3 are used to calculate the time to complete paths B-D-F and B-E, respectively. Cell J3 keeps track of when the project is actually com- pleted on a given replication. Since the longest path determines the time when the project is completed, 5 MAX(G3:I3) was entered in cell J3.

The results of simulating the project are summarized in Figure 6.12 . The results indicate that, on the average, the project required 90.28 days to complete. Furthermore, across the 1000 replications of the project, the fastest project completion time was

Figure 6.11 Spreadsheet for simulating the network.

1 2 3 4 5 6 7 8 9 10 11 12 13

Activity A

A 32.1

Formulae: Cell G3 Cell H3 Cell I3 Cell J3

= A3 + C3 + F3 = B3 + D3 + F3 = B3 + E3 = MAX (G3:I3)

Activity B

B 24.6

Activity C

C 22.2

Activity D

D 26.1

Activity E

E 34.4

Activity F

F 34.5

Path G

ACF 88.8

Path H

BDF 85.2

Path I

BE 59

Completion J

Time 88.8

Assumption Cells Forecast Cell

Figure 6.12 Simulation results.

c06.indd 227c06.indd 227 25/09/12 10:48 AM25/09/12 10:48 AM

228 C h a p t e r 6 : M a n a g i n g P r o c e s s I m p r o v e m e n t P r o j e c t s

75.77 days and the longest was 109.77 days. The simulation package can also show the probabilities of completing the project before any given date, or after any given date, or even between any two dates.

Project Management Software Capabilities A wide range of project management software packages and capabilities are availa- ble, depending on the project need and the funds available. The main aspects to consider when selecting a package are the capabilities required and the time and money available to invest in a package. If the project is very large and complex, or if it interacts with a number of other projects that must also be managed with the software, then some of the more sophisticated packages are appropriate. However, not only do these cost more, they also take longer to learn and need greater compu- ter power to run. On the other hand, if the project is simpler, a less elaborate pack- age that is easier to learn and use may be the best choice.

WBS Name Duration Sch. start Sch. finish 4 11 18 25 1 8 15 22 29 5 12 19 25 5 12 19 26 2 9 16

December January February March April

1

2

3

4

4.1

4.2

4.3

5

6

7

7.1

7.2

7.3

7.4

8

9

10 Purchase order prepared

Purchase recommendation prepared

Check out references

Evaluate demos

Survey participants

Software loaded on system

Participants selected

Demo evaluation

Price evaluation

Demos received

Reference list

Prices gathered

Demos ordered

Vendor calls

Literature reviewed

Literature search

Software review begins 0d

2d

12d

10d

10d

1d

1d

1d

5d

40d

1d

1d

9d

30d

3d

5d

0d Mar 14 Mar 14

Mar 14

Mar 7

Mar 7

Mar 8

Dec 28 Dec 30

Jan 25

Jan 11 Jan 11

Jan 11 Jan 11

Jan 11

Jan 10 Jan 16

Jan 10

Jan 9

Jan 9

Jan 10

Dec 27

Dec 27

Dec 27

Dec 27

Dec 27

Dec 27

Dec 7 Dec 7

Dec 7 Dec 8

Dec 9 Dec 26

Jan 12 Jan 24

Project: software evaluation

Date: 1/20/94

Critical

Noncritical

Progress

Milestone

Summary

Rolled up

Figure 6.13 Microsoft Project ’s Gantt chart.

c06.indd 228c06.indd 228 25/09/12 10:48 AM25/09/12 10:48 AM

229S c h e d u l i n g t h e P r o j e c t

A yearly survey and analysis of such packages is conducted by the Project Management Institute. These surveys give details on the friendliness of each package, their capabilities (schedules, calendars, budgets, resource diagrams and tables, graph- ics, migration capabilities, report generation, tracking capability, etc.), their computer requirements, and their cost.

Probably the most commonly used package these days is Microsoft ’s Project. This package is fairly sophisticated for the cost and is extremely easy to learn and use. Examples of some of its report capabilities are given in Figures 6.13 , 6.14 , and 6.15 .

Goldratt ’s Critical Chain 1 In Critical Chain , Eliyahu Goldratt ( 1997 ) applies his theory of constraints (described in Chapter 5) to the fi eld of project management. In this theory, he primarily focuses on three phenomena that tend to bias the expected completion time of projects, based on the network techniques we described above, toward shorter times than occur in reality. These three phenomena are infl ated activity time estimates, activity time variability with path interdependencies, and resource dependence.

Figure 6.14 PERT chart generated by Microsoft Project.

12/9/98

Project approval

1

12/3/98

0d

12/3/98

Begin scheduling

4

12/3/98

0d

12/3/98

Scheduling complete

8

12/16/98

0d

12/16/98

Deliver video to client

14

12/4/99

0d

12/4/99 Schedule shoots

7

12/10/98

5d

12/16/98

Propose shoots

5

12/3/98

5d

12/9/98

Hire secretary

6

12/3/98

5d

Project: Producing a video tape Date: 12/3/98

ID

Scheduled start

Duration

Critical Milestone

Noncritical Summary

Subproject

MarkedScheduled finish

Script writing

2

12/3/98

14d

12/22/98

Schedule shoots

3

12/3/98

10d

12/16/98

Script approval

9

12/23/98

5d

12/29/98

Revise script

10

12/30/98

5d

1/5/99

Shooting

11

1/6/99

10d

1/19/99

Editing

12

1/20/99

7d

1/28/99

Final approval

13

1/31/99

5d

2/4/99

Name

1Adapted from S. J. Mantel, Jr., J. R. Meredith, S. M. Shafer, and M. M. Sutton. Project Management in Practice , 4th ed. Hoboken, NJ: Wiley, 2011.

c06.indd 229c06.indd 229 25/09/12 10:48 AM25/09/12 10:48 AM

230 C h a p t e r 6 : M a n a g i n g P r o c e s s I m p r o v e m e n t P r o j e c t s

Infl ated Activity Time Estimates

Assuming that project workers have a general desire to be recognized for good per- formance, what do you imagine they do when they are asked to provide time esti- mates for tasks they will be responsible for? Do you think they give an estimate that they believe has only a 50 percent chance of being met? Or, more likely, do you imagine they infl ate, or pad , their estimate to increase the likelihood of successfully completing the task on time? What would you do?

We suspect that if you are like most people, you would tend to somewhat infl ate your time estimate. Unfortunately, infl ated time estimates tend to create even more problems. First, infl ating the time estimate has no impact on the actual probability dis- tribution of completing the activity. Second, what do you imagine happens in cases where a project team member fi nishes early? More than likely, the team member believes that it is in his or her best interest to remain silent about completing activities in less than the allotted time so that future time estimates are not automatically dis- counted by management based on a track record of early task completions. Moreover, there are sometimes penalties for completing early, such as storage of materials.

Third, just as things tend to fi ll the available closet and storage space in your home, work tends to fi ll the available time. Thus, the scope of the task may be

Figure 6.15 Calendar of activities created by Microsoft Project.

Sun Mon Tue Wed Thu Fri Sat

321

1098764 5

171615141311 12

242322212018 19

313029282725 26

December

Software Evaluation

Software review beg...

Literature search, 2d Literature reviewed, 12d

Literature reviewed, 12d

Literature reviewed, 12d

Reference list, 1d

Prices gathered, 1d Check out references, 3d

Demos ordered, 10d

Vendor calls, 10dLiterature reviewed, 12d

c06.indd 230c06.indd 230 25/09/12 10:48 AM25/09/12 10:48 AM

231S c h e d u l i n g t h e P r o j e c t

expanded to fi ll the available time. Perhaps even more dangerous than the infl ated estimate becoming a self-fulfi lling prophecy is that, after receiving approval for a task based on an infl ated time estimate, workers may perceive that they now have plenty of time to complete the task and therefore delay starting the task . Goldratt refers to this as the student syndrome , likening it to the way students often delay writing a term paper until the last minute. The problem of delaying the start of a task is that obstacles are frequently not discovered until the task has been under way for some time. If the start of the task is delayed, the opportunity to effectively deal with these obstacles and complete the task on time is greatly diminished.

Activity Time Variability with Path Interdependencies

Another factor that tends to favorably bias the expected project completion time is the effect of variability in the activity times when there are multiple and interconnecting paths in a network. First consider a project with, say, ten activities all in a line (i.e., in series), each of the same expected duration and variability. It seems clear that if random events affect the activities, some will fi nish early and others late, but the general overall effect will be that the early completions will largely offset the late completions and the project will fi nish about when expected. However, suppose now that another project also has ten activities, but they are all in parallel, and all must be completed to complete the project. Since the project will not be done until every activity is completed, the slowest activity of the ten—that is, the one whose random events delay the activity the most—will be the one that determines when the project is actually completed.

Most projects are not like either of the above two examples but instead have many activities in series and many in parallel. As we saw above, the activities in series tend to cancel out their random effects, but this is not true of the parallel activities, which tend to delay the project. Eventually, all the interacting paths of activities throughout the network act like our parallel activities and have a delay- ing effect on the project due to their random variations. In particular, if there is another path(s) through the network that is close to the length of the critical path and has substantial variability, it is quite likely that this path, rather than the suppos- edly “critical path,” will determine when the project is completed.

Resource Dependence

Last, it frequently happens that some activities need the same (scarce) resource (per- haps a machine, or a particularly skilled person) at the same time. If this happens, then there is no alternative but for one activity to wait until the other activity has fi nished with the resource, unless, of course, the organization is willing to spend extra funds to acquire or rent another resource—but this will then negatively affect the budget. As a result, resource dependence within a project can also seriously delay a project beyond its expected completion time based on the critical path.

Goldratt ’s approach for addressing these three issues is based on elementary sta- tistics. It is easily shown that the amount of safety time needed to protect a particular path in a project is less than the sum of the safety times required to protect the indi- vidual activities making up the path. The same approach is commonly used in inven- tory management, where it can be shown that less safety stock is needed at a central warehouse to provide a certain service level to customers than the amount of safety stock that would be required to provide this same service level if carried at multiple distributed (e.g., retail) locations.

c06.indd 231c06.indd 231 25/09/12 10:48 AM25/09/12 10:48 AM

232 C h a p t e r 6 : M a n a g i n g P r o c e s s I m p r o v e m e n t P r o j e c t s

Clearly, activities on the critical chain should be given the highest priority. Likewise, to ensure that resources are available when needed, they should be con- tacted at the start of the project. It is also wise to keep these resources updated on the status of the project and to remind them periodically of when their input will be needed. Goldratt suggests reminding these resources two weeks before the start of their work, then three days prior to their start, and fi nally the day before they start. Since any delay of an activity on the critical chain can cause a delay of the entire project, it is important that a resource immediately switch to the task on the critical chain when needed.

Based on this intuition, Goldratt suggests reducing the amount of safety time added to individual tasks and using some fraction of that reduction as a safety buffer for the entire project, called the project buffer . The amount of time each task is reduced depends on how much of a reduction is needed to get project team mem- bers to change their behavior. For example, the allotted time for tasks should be reduced to the point that the student syndrome is eliminated. To motivate the project team members, Goldratt suggests using activity durations where in fact there is a high probability that the task will not be fi nished on time.

To address the need to consider both precedence relationships and resource dependencies, Goldratt proposes thinking in terms of the longest chain of consecu- tively dependent tasks where such dependencies can arise from a variety of sources, including precedence relationships among the tasks and resource dependencies. Goldratt coined the term critical chain to refer to the longest chain of consecutively dependent activities.

Based on this defi nition of the critical chain, there are two potential sources that can delay the completion of a project. In a similar fashion to the critical path concept, one source of delay is the tasks that make up the critical chain. The project buffer discussed earlier is used to protect against these delays (see Figure 6.16 ). But, as noted above, tasks external to the critical chain can also delay the completion of the project if these delays end up delaying one or more of the tasks on the critical chain. As shown in Figure 6.16 , safety time can be added to these paths as well to ensure that they do not delay tasks on the critical chain. The safety time added to chains other than the critical chain is called a feeding buffer , since these paths often feed into the critical chain. Thus, the objective of feeding buffers is to ensure that noncritical chains are completed so that they do not delay tasks on the critical chain.

Figure 6.16 Project and feeding buffers.

Critical chain

Project buffer

Feeding buffer

c06.indd 232c06.indd 232 25/09/12 10:48 AM25/09/12 10:48 AM

233C o n t r o l l i n g t h e P r o j e c t : E a r n e d V a l u e

These two graphs are combined for project managers into an earned value chart—Figure 6.18 —where the planned value (PV), actual cost (AC), and earned value completed (actual earned dollars of progress, EV) are plotted. Here we see that the actual cost variance is now substantial, given the poor progress and large schedule variance. Plotted in this manner, one chart will serve to monitor both progress and cost. We can then defi ne three variances: (1) a cost variance equal to the value completed less the actual cost (EV – AC), where a cost overrun is negative; (2) a schedule variance equal to the value completed less the planned value (EV – PV), where “behind” is negative; and (3) a time variance equal to the effective time less the actual time (TE – TA), where a delay is negative.

When these variances are signifi cant, the project manager must identify (or at least attempt to identify) an assignable cause for the variance. That is, he or she must study the project to determine why the variance occurred. This is so that the proper

C O N T R O L L I N G T H E P R O J E C T : E A R N E D V A L U E One of the control systems most widely used in projects is the cost variance report. Cost standards are determined through engineering estimates or through analysis of past performance. They become the target costs for the project. The actual costs are then monitored by the organization ’s cost-accounting system and are compared with the cost standard. Feedback is provided to the project manager, who can exert any necessary control if the difference between standard and actual (called a variance) is considered signifi cant.

As an example, consider the cost–schedule charts in Figure 6.17 . In Figure 6.17 a , actual progress is plotted alongside planned progress, and the “effective” progress time (TE) is noted. Because progress is less than planned, TE is less than the actual time (TA). On the cost chart (Figure 6.17 b ), we see that the apparent variance between the planned value and actual cost at this time (PV – AC) is quite small, despite the lack of progress (earned value, EV). But this is misleading; the variance should be much more given the lack of progress.

Figure 6.17 Cost–schedule reconciliation charts.

TE TA TE TATime

P ro

gr es

s

A m

o u

n t

sp en

t

(a)

Time Time varianceTime variance

(b)

Planned Actual

0

100%

Planned Actual

PV

AC

EV

c06.indd 233c06.indd 233 25/09/12 10:48 AM25/09/12 10:48 AM

234 C h a p t e r 6 : M a n a g i n g P r o c e s s I m p r o v e m e n t P r o j e c t s

1. Frequently, the project ’s tasks are not well defi ned, and there is an urge to “get on with the work,” since time is critical. How serious is it to minimize the planning effort and get on with the project?

2. Contrast the cost–schedule reconciliation charts with the earned value chart. Which one would a project manager prefer?

3. How would a manager calculate the value com- pleted for an earned value chart?

4. Do you think people ’s estimates are more accurate for optimistic or pessimistic activity times?

5. Of the reasons discussed for the growth in project operations, which do you think are contributing most?

6. Why doesn ’t it make sense to think in terms of a critical path when activity times are not known with certainty?

7. In calculating the probability that a project will be fi nished by some specifi ed time, the probabilities of each path are multiplied together, on the assumption

that the paths are independent of one another. How reasonable is this assumption?

8. Is the stretched-S life-cycle project form more com- mon or the exponential form? What other aspects of managing a project are affected by the nature of the project form besides budgeting and early termination?

9. What do you think are the reasons for the topics of confl ict among the project team in each stage of the project?

10. Given the powerful nature of project management software packages today, why should a project man- ager have to know how to construct a PERT chart or work breakdown structure?

11. Given the ease of use of simulation software such as Crystal Ball ® , what other data used in project management should probably be simulated?

12. Describe how to actually calculate earned value.

13. What does the project portfolio illustrate? How might it be useful to management?

remedy can be used to keep the variance from recurring. A corrective action is called for if some ineffi ciency or change in the prescribed process caused the variance.

Variances can be both favorable and unfavorable. A signifi cant favorable variance (e.g., a variance resulting from a large quantity discount on material) will usually not require corrective action, though investigation is still worthwhile so that this better- than-expected performance can be repeated.

E X P A N D Y O U R U N D E R S T A N D I N G

1 2 Month

D o

ll ar

s

Time variance

(10-day delay)

Cost variance

Schedule variance

3

Cost–schedule plan (baseline) Actual cost Value completed

TE TA

PV

AC

EV

Figure 6.18 Earned value chart.

c06.indd 234c06.indd 234 25/09/12 10:48 AM25/09/12 10:48 AM

235A p p l y Y o u r U n d e r s t a n d i n g

A P P L Y Y O U R U N D E R S T A N D I N G

e -Razor, Incorpora ted

After four years as a new start-up, e-Razor has expanded its line of programmable razors to include both linear and rotary models, as well as women ’s and teenagers ’ versions. Program- ming is simply done by the user through icons that adjust the speed of the cutting head, adjust the distance from the skin (to avoid ingrown hairs or give the sexy look of a day ’s growth), activate the trimmer, and so on. All the razors use the same electronics module to keep costs down.

e-Razor is now planning their budget for next year, including their investment in razor re- search, development, and product improvement, as well as investments in their production process to improve quality and lower costs. Eight projects have been proposed for top man- agement ’s consideration, as follows:

A. Marketing would like to see the programmable functions extended to stay ahead of the competition, which might involve a new electronics module, the brain of the razor. Projected cost: $25,000.

B. Manufacturing believes a mixed-model fl ow line would reduce costs and improve quality at the same time. To a large extent, the same production equipment as currently used in their job shop process could be reconfi gured for the line and only a few new items would be needed. Projected cost: $11,000.

C. Marketing would also like to see the programmable feature extended to other health/ beauty products, such as hair dryers, facial saunas, and such. Projected cost to investigate and report: $16,000.

D. Sales has received feedback from customers that an engineering modifi cation to the razor head allowing the razor to be used in the shower would make it much more useful. The production process would need to be modifi ed somewhat. Projected cost: $9000.

E. Engineering has been evaluating a new user interface that would allow many new functions to be easily added. Moreover, it is faster and easier for the user to program. The current production process can be used. Projected cost: $17,000.

F. A consultant in chemistry who has been investigating the properties of tiny strands of cut hair for the fi rm believes there may be a way to keep the shaving head clean without opening the head enclosure and emptying or rinsing the head—a distasteful task for most people. Such a change would substantially change the production process. Projected cost: $22,000.

G. Sales has had inquiries from customers about whether a cheap, limited-use version of the razor is available for business trips and vacations. A new production line would be required. Projected cost: $28,000.

H. Along the same line, customers have asked if a purely battery-driven (nonrechargeable) razor is available. This would primarily involve a change in the type of battery being used; the production process would be only slightly affected. Projected cost: $8000.

Questions

1. Construct an aggregate project plan for this portfolio of projects and place the projects on the diagram with the diameters of the circles representing their projected costs.

2. Analyze the diagram for balance across the categories. Are there any gaps or excesses? What should the distribution of projects look like for a fi rm of this age?

3. The total budget for these projects is limited to $100,000. Which projects would you suggest implementing?

c06.indd 235c06.indd 235 25/09/12 10:48 AM25/09/12 10:48 AM

236 C h a p t e r 6 : M a n a g i n g P r o c e s s I m p r o v e m e n t P r o j e c t s

Nu t r i -Sam

Nutri-Sam produces a line of vitamins and nutritional supplements. It recently introduced its Nutri-Sports Energy Bar, which is based on new scientifi c fi ndings about the proper bal- ance of macronutrients. The energy bar has become extremely popular among elite athletes and others. One distinguishing feature of the Nutri-Sports Energy Bar is that each bar contains 50 milligrams of eicosapentaenoic acid (EPA), a substance strongly linked to reducing the risk of cancer but found in only a few foods, such as salmon. Nutri-Sam was able to include EPA in its sports bars because it had previously developed and patented a process to refi ne EPA for its line of fi sh-oil capsules.

Because of the success of the Nutri-Sports Energy Bar in the United States, Nutri-Sam is considering offering it in Latin America. With its domestic facility currently operating at capac- ity, the president of Nutri-Sam has decided to investigate the option of adding approximately 10,000 square feet of production space to its facility in Latin America at a cost of $5 million.

The project to expand the Latin American facility involves four major phases: (1) concept development, (2) defi nition of the plan, (3) design and construction, and (4) start-up and turnover. During the concept development phase, a program manager is chosen to oversee all four phases of the project and is given a budget to develop a plan. The outcome of the con- cept development phase is a rough plan, feasibility estimates for the project, and a rough schedule. Also, a justifi cation for the project and a budget for the next phase are developed.

In the plan defi nition phase, the program manager selects a project manager to oversee the activities associated with this phase. Plan defi nition consists of four major activities that are completed more or less concurrently: defi ning the project scope, developing a broad sched- ule of activities, developing detailed cost estimates, and developing a plan for staffi ng.

The output of this phase is a detailed plan and proposal for management specifying how much the project will cost, how long it will take, and what the deliverables are.

If the project gets management ’s approval and management provides the appropriations, the project progresses to the third phase, design and construction. This phase consists of four major activities: detailed engineering, mobilization of the construction employees, procurement of production equipment, and construction of the facility. Typically, the detailed engineering and the mobilization of the construction employees are done concurrently. Once these activities are completed, construction of the facility and procurement of the production equipment are done concurrently. The outcome of this phase is the physical construction of the facility.

The fi nal phase, start-up and turnover, consists of four major activities: pre-start-up inspec- tion of the facility, recruiting and training the workforce, solving start-up problems, and deter- mining optimal operating parameters (called centerlining). Once the pre-start-up inspection is completed, the workforce is recruited and trained at the same time that start-up problems are solved. Centerlining is initiated upon the completion of these activities. The desired outcome of this phase is a facility operating at design requirements.

The following table provides optimistic, most likely, and pessimistic time estimates for the major activities.

Activity Optimistic Time

(months) Most Likely Time

(months) Pessimistic

Time (months)

A: Concept development 3 12 24

Plan Defi nition

B: Defi ne project scope 1 2 12

C: Develop broad schedule 0.25 0.5 1

D: Detailed cost estimates 0.2 0.3 0.5

E: Develop staffi ng plan 0.2 0.3 0.6

c06.indd 236c06.indd 236 25/09/12 10:48 AM25/09/12 10:48 AM

237E x e r c i s e s

1. The following AON chart was prepared at the beginning of a small construction project. The duration, in days, follows the letter of each activity. What is the critical path? Which activities should be monitored most closely?

At the end of the fi rst week of construction, it

was noted that activity A was completed in 2.5 days, but activity B required 4.5 days. What impact

does this have on the project? Are the same activi- ties critical?

2. Refer to Exercise 1. Compute the earliest start and fi nish times, the latest start and fi nish times, and the slack times for each activity. Also, calculate the slack for each path.

3. Given the following German autobahn repair project, fi nd the probability of completion by 17 weeks and by 24 weeks.

Activity Optimistic Times (weeks)

Most Likely Pessimistic Required

Precedence

A 5 11 11 —

B 10 10 10 —

C 2 5 8 —

D 1 7 13 A

Activity Optimistic Time

(months) Most Likely Time

(months) Pessimistic

Time (months)

Design and Construction

F: Detailed engineering 2 3 6

G: Facility construction 8 12 24

H: Mobilization of employees 0.5 2 4

I: Procurement of equipment 1 3 12 Start-up and Turnover

J: Pre-start-up inspection 0.25 0.5 1

K: Recruiting and training 0.25 0.5 1

L: Solving start-up problems 0 1 2

M: Centerlining 0 1 4

Questions

1. Draw a network diagram for this project. Identify the paths through the network diagram that could potentially delay the project if its deadline is 40 months.

2. Find the probability that the project can be completed within 30 months. What is the probability that the project will take longer than 40 months? What is the probability that the project will take between 30 and 40 months?

3. 2 Use Crystal Ball® (or any other simulation package) to simulate the completion of this project 1000 times, assuming that activity times follow a triangular distribution. Estimate the mean and standard deviation of the project completion time. Compare your results to your answer to Question 2.

E X E R C I S E S

Start

A, 4

B, 3

End

C, 1

D, 5

E, 4

F, 1

I, 1

G, 3

H, 5

J, 2

2Optional instructor-assigned question.

c06.indd 237c06.indd 237 25/09/12 10:48 AM25/09/12 10:48 AM

238 C h a p t e r 6 : M a n a g i n g P r o c e s s I m p r o v e m e n t P r o j e c t s

Activity Optimistic Times (weeks)

Most Likely Pessimistic Required

Precedence

E 4 4 10 B

F 4 7 10 B

G 2 2 2 B

H 0 6 6 C

I 2 8 14 G, H

J 1 4 7 D, E

If the fi rm can complete the project within 18 weeks, it will receive a bonus of €10,000. But if the project is delayed beyond 22 weeks, it must pay a penalty of €5000. If the fi rm can choose whether or not to bid on this project, what should its decision be if this is normally only a breakeven project?

4. Construct a network for the project below and fi nd its expected completion time.

Activity t e (weeks)

Preceding Activities

a 3 None

b 5 a

c 3 a

d 1 c

e 3 b

f 4 b, d

g 2 c

h 3 g, f

i 1 e, h

5. Estimated activity times and precedences are given below:

Activity Optimistic Times (days) Most Likely Pessimistic

Required Precedence

A 6 7 14 —

B 8 10 12 —

C 2 3 4 —

D 6 7 8 A

E 5 5.5 9 B, C

F 5 7 9 B, C

G 4 6 8 D, E

H 2.7 3 3.5 F

What is the probability that the project will be completed within:

a. 21 days? b. 22 days? c. 25 days?

6. Pusan Iron and Steel, located on the eastern coast of South Korea, is a major supplier of both girder and rolled steel to the emerging construction, appli- ance, and automobile companies of China. Due to growing sales volumes and the need for faster delivery, Pusan is converting its current single weigh station to a larger, multiple drive-through sta- tion. The new drive-through weigh station will con- sist of a heated, air-conditioned building with a large fl oor and a small offi ce. The large room will have the scales, a 15-foot counter, and several dis- play cases for its equipment.

Before erection of the building, the project man- ager evaluated the project using CPM analysis. The following activities with their corresponding times were recorded.

Times

# Activity Optimistic Most Likely Pessimistic

Preceding Tasks

1 Lay foundation 8 10 13 —

2 Dig hole for scale

5 6 8 —

3 Insert scale bases

13 15 21 2

4 Erect frame 10 12 14 1, 3

5 Complete building

11 20 30 4

6 Insert scales 4 5 8 5

7 Insert display cases

2 3 4 5

8 Put in offi ce equipment

4 6 10 7

9 Give fi nishing touches

2 3 4 8, 6

Using CPM analysis, fi nd the expected comple- tion time.

7. As in the situation illustrated in Figure 6.18 , an Irish Web-design project at day 70 exhibits only 35 per- cent progress when 40 percent was planned, for an effective date of 55. Planned value was €17,000 at day 55 and €24,000 at day 70, and actual cost was €20,000 at day 55 and €30,000 at day 70. Find the

c06.indd 238c06.indd 238 25/09/12 10:48 AM25/09/12 10:48 AM

239E x e r c i s e s

time variance, cost variance, and schedule variance at day 70.

8. As in the situation shown in Figure 6.18 , a project at month 2 exhibited an actual cost of $78,000, a planned value of $84,000, and a value completed of $81,000. Find the cost and schedule variances. Estimate the time variance.

9. A project at month 5 had an actual cost of $34,000, a planned value of $42,000, and an earned value of $39,000. Find the cost and schedule variances.

10. Given a network:

Note that four activities can start immediately.

Find the following:

a. Critical path

b. Earliest time to complete the project

c. Slack on activities E, F, and H

11. Given the activities data in the table below:

Activity Times (weeks) Preceding Activities

A 3 —

B 6 —

C 8 —

D 7 A

E 5 B

F 10 C

G 4 C

H 5 D, E, F

I 6 G

a. Draw the network.

b. Find the critical path.

c. Find the slacks on all activities.

12.

Activity Duration Preceding Activities

1 1 —

2 2 —

3 3 —

4 4 3

5 3 2, 4

6 8 3

7 2 2, 4

8 4 1, 5

9 2 17

10 6 2, 4

11 5 6, 10

12 10 7, 8, 11

13 11 7, 8, 11

14 1 6, 10

15 9 12

16 3 6, 10

17 8 12

18 6 13, 14, 15

a. Draw the diagram

b. Find the critical path

c. Find the completion time

13. In the project network shown in the following fi g- ure, the number alongside each activity designates its known duration in weeks. Determine the following:

a. Earliest and latest start and fi nish times for each activity

b. Earliest time that the project can be completed

c. Slack for activities

d. Critical activities

e. Critical path

A, 5

B, 3

C, 6

D, 7

Start End

F, 6

E, 5

G, 10

H, 8

I, 4

Start

A, 2

B, 4

End

C, 3

D, 3

F, 6

E, 5

G, 4

H, 4

KL, 3

J, 2

I, 8

c06.indd 239c06.indd 239 25/09/12 10:48 AM25/09/12 10:48 AM

240 C h a p t e r 6 : M a n a g i n g P r o c e s s I m p r o v e m e n t P r o j e c t s

Abrams , S. “ The Big Dig .” Kennedy School Bulletin (Spring 2003 ): 30 – 35 .

Angus , R. B. , N. A. Gundersen , and T. P. Cullinane . Plan- ning, Performing, and Controlling Projects: Principles and Applications , 3rd ed. Upper Saddle River, NJ : Prentice Hall , 2003 .

Cauchon , D. “The Little Company That Could.” USA Today (October 9, 2005) .

Cleland , D. I. Project Managers ’ Portable Handbook , 2nd ed. New York : McGraw-Hill , 2004 .

Englund , R. L. , and R. J. Graham . “ From Experience: Linking Projects to Strategy .” Journal of Product Innova- tion Management , 16 ( 1 ) ( 1999 ): 52 – 64 .

Gale , S. F. “ The Bottom Line .” PM Network (August 2007 ): 34 – 38 .

Gido , J. , and J. P. Clements . Successful Project Manage- ment (with Microsoft Project CD-ROM). Cincinnati: Thompson/ South-Western, 2008 .

Goldratt , E. M. Critical Chain . Great Barrington, MA : North River Press , 1997 .

Gray , C. F. , and E. W. Larson . Project Management: The Managerial Process , 4th ed. New York : McGraw-Hill/ Irwin , 2008 .

Ibbs , C. W. , and Y.-H. Kwak . “ Assessing Project Manage- ment Maturity .” Project Management Journal (March 2000 ): 32 – 43 .

Kerzner , H. Advanced Project Management: Best Prac- tices on Implementation . New York : Wiley , 2004 .

Kerzner , H. Project Management: A Systems Approach to Planning, Scheduling, and Controlling , 9th ed. New York : Wiley , 2006 .

Kolisch , R. “ Resource Allocation Capabilities of Com- mercial Project Management Software Packages .” Inter- faces , 29 ( July–August 1999 ): 19 – 31 .

Mantel, S. J., Jr., J. R. Meredith, S. M. Shafer, and M. M. Sutton. Project Management in Practice , 4th ed. New York : Wiley , 2011 .

Meredith , J. R. , and S. J. Mantel Jr . Project Management: A Managerial Approach , 8th ed. New York : Wiley , 2012 .

Nicholas , J. M. , and H. Steyn . Project Management for Business, Engineering, and Technology . Englewood Cliffs, NJ : Prentice Hall , 2008 .

Project Management Institute. “A Derailed Vision.” PM Network (April 2004 a): 1 .

Project Management Institute. “Digging Deep.” PM Network (August 2004 b): 1 .

Project Management Institute. “Lack of Support.” PM Network ( January 2005 ): 1 .

Project Management Institute. A Guide to the Project Management Body of Knowledge , 4th ed. Newtown Square, PA: Project Management Institute, 2009 .

Thamhain , H. J. , and D. L. Wilemon . “ Confl ict Manage- ment in Project Life Cycles .” Sloan Management Review (Summer 1975 ): 31 – 50 .

Wheelwright , S. C., and K. B. Clark . “ Creating Project Plans to Focus Product Development .” Harvard Business Review (March–April 1992 ): 2 – 14 .

B I B L I O G R A P H Y

c06.indd 240c06.indd 240 25/09/12 10:48 AM25/09/12 10:48 AM

241

� Supply Chain Management

C H A P T E R 7

ROLE OF OPERATIONS IN ORGANIZATIONS’ COMPETITIVENESS

Operations Strategy

Process Design and Control

Process Improvement

Process Execution

Ch. 1: Operations Strategy and Global

Competitiveness

Ch. 2: Process Planning and

Design

Ch. 3: Controlling Processes

Ch. 7: Supply Chain Management

Ch. 8: Capacity, Scheduling, and

Location Planning

Ch. 4: Process Improvement:

Six Sigma

Ch. 5: Process Improvement: Lean

Ch. 6: Managing Process

Improvement Projects

c07.indd 241c07.indd 241 25/09/12 10:49 AM25/09/12 10:49 AM

242

IN T R O D U C T I O N • While Apple ’s enormous success is most commonly attributed to its ability to design

highly innovative products that are easy to use, the signifi cant contribution its opera- tions makes to its success gets much less publicity. Nevertheless, experts and ana- lysts that closely follow Apple readily acknowledge that Apple ’s operations excellence is as much an asset to Apple as is its product innovation and marketing. Indeed, it is Apple ’s operational capabilities that allow it to pull off its massive, high- volume product launches by managing its inventory effi ciently.

It is not widely known, but Apple ’s focus on improving its supply chain dates back to the return of Steve Jobs in 1997. For example, to ensure there was an adequate sup- ply of Apple ’s new translucent blue iMacs, Apple spent $50 million to acquire all the available holiday air transport capacity. Not only did this ensure that Apple could get its products to the customers, but the move crippled competitors, such as Compaq, that didn ’t recognize the need to ship products by air until it was too late. Based on this example and others like it, Apple has learned that making investments in its supply chain upfront pays for itself in the long run in the form of greater volume. Greater vol- umes also yield additional benefi ts. For example, when the sales volume of iPods increased in 2001, Apple discovered it could air-ship the iPods economically from the Chinese factories directly to its customers’ homes. Not only does this help Apple reduce its investment in inventory, but it provides an added level of service for the customer.

When the organization designs its processes (Chapter 2) to achieve its competitive strategy, this will include the supply chain for its products and/ or services. We now consider the execution, or management, of these processes. In this chapter we address issues related to supply chain manage- ment, often involving relationships with organiza- tions outside the fi rm, and in the following chapter we focus on other important competitive aspects such as capacity and scheduling.

Supply chain management fundamentally involves matching supply with demand and as such is strongly related to a fi rm ’s competitiveness. Important supply chain management topics include designing and restructuring the value chain, outsourcing, and e-commerce. Furthermore, competent management of the supply chain has

major impacts on all the strategic sand cone fac- tors described in Chapter 1: quality, dependability, speed, and cost.

We fi rst defi ne the concept of supply chains and discuss their strategic importance. We then describe the many elements involved in their design, such as logistics, global sourcing, and sup- plier management. From this we move to the role of information technology and provide guidelines for successful supply chain management. We con- clude the chapter with a discussion of closed-loop supply chains. Two supplements to the chapter describe a supply chain classroom exercise used by many MBA classes (Supplement A: The Beer Game) and a quantitative technique that is popu- lar for some MBA classes (Supplement B: The Economic Order Quantity Model).

242

C H A P T E R I N P E R S P E C T I V E

c07.indd 242c07.indd 242 25/09/12 10:49 AM25/09/12 10:49 AM

243I n t r o d u c t i o n

Beyond investing fi nancial resources in its supply chain, Apple invests in its human capital as well. For example, to facilitate the process of translating product prototypes into successful new products, Apple ’s design engineers live in hotels for months to be close to their suppliers in order to help them perfect their production processes. For example, when Apple designed a new MacBook with a case that was made from a single piece of aluminum, Apple ’s design engineers worked with the suppliers to develop the equipment to fabricate the cases.

With a huge cash war chest in excess of $80 billion, Apple planned to almost dou- ble its supply chain capital expenditures in 2011 to $7.1 billion. In part this invest- ment will be used to purchase capacity from its suppliers to ensure the prices and availability of its products. For example, prior to the introduction of the iPhone 4 in June 2010, supplier capacity for screens was being used for iPhones, forcing Apple ’s competitor HTC to scramble for sources of phone screens. Likewise, when Apple launched the iPad 2, it purchased so many of the high-end drills used to produce the tablet ’s internal casing that the lead time for other companies to get these drills extended to as long as six months.

Turning the tables, being selected by Apple to be one of its suppliers can be very profi table. However, this comes at a price. For example, when a potential supplier is asked to provide a price quote for a part or assembly that will go into an Apple product, the supplier is required to submit in great detail how it arrived at the quote, including the specifi c material costs, labor costs, and its estimated profi t. Furthermore, to guard against supply disruptions, Apple requires its suppliers to maintain a two-week supply of inventory within a mile of the Asian assembly plants.

Carefully orchestrated events announcing new products are eagerly anticipated by industry analysts and loyal customers. Here, too, Apple ’s supply chain manage- ment practices play an important role. For example, supplier factories work overtime weeks in advance of new-product launches to build up inventory to meet the often overwhelming demand for new Macs, iPods, iPhones, and iPads. Furthermore, the success of the new-product debuts centers on the secrecy Apple is able to maintain about the features of its new products. To ensure that the secrecy of its new prod- ucts is not breached and to discourage leaks, Apple places electronic monitors in a subset of the boxes of parts that go into its products so that it is able to monitor the parts through the production process. Through this monitoring, Apple is able to track every part hand-off from the loading docks through the distribution centers. And not to leave anything to chance, once the new products are fi nished, they are shipped in plain boxes or even disguised boxes, such as tomato boxes.

A fi nal piece of Apple ’s supply chain that contributes to its operational excellence is its retail stores. Apple tracks the sales at its stores hour by hour and, based on these sales, adjusts its production forecast each day. When a risk of a product short- age is identifi ed, Apple immediately deploys teams, and the added capacity is acquired (Satariano and Burrows 2011).

• Even with the lean inventories that have resulted from the prevalence of just-in-time inventory systems, shifts in economic cycles can still wreak havoc for industry-wide supply chains. The electronics industry during the global recession of 2008–2009 illustrates this well.

c07.indd 243c07.indd 243 25/09/12 10:49 AM25/09/12 10:49 AM

244 C h a p t e r 7 : S u p p l y C h a i n M a n a g e m e n t

At one end of the electronics supply chain are the retailers that sell electronic products to end consumers. With the fi nancial crisis rapidly escalating in the fall of 2008, Minnesota-based retailer Best Buy experienced a signifi cant decline in sales. Best Buy orders electronic products such as DVD players six weeks prior to when they are needed. With the 2008 Thanksgiving shopping season approaching, Best Buy revised its prior forecast and dramatically reduced its orders to its suppliers, such as Japan ’s Toshiba and Korea ’s Samsung Electronics in early October 2008. As the fi nancial crisis was uncharted territory, Best Buy ’s merchandising chief had to make his best guess in deciding how to modify the forecast.

Lacking a direct relationship with the fi nal consumers, Best Buy ’s suppliers were caught off guard by its revised forecast and reduced orders. As expected, these suppliers in turn reduced orders from their suppliers. As an example, Zoran Corp, a designer of specialty chips used in electronic products such as TVs, cameras, cell phones, DVD players, and digital picture frames, saw its revenue decline in the fourth quarter of 2008 by 42 percent. Zoran, which only designs chips, relies on companies like Taiwan Semiconductor Manufacturing Company (TSMC) to produce its chips. Faced with decreased orders for its chips, Zoran slashed its orders to TSMC. In January and February of 2009, TSMC saw its revenue decrease by 58 percent compared to the prior year and was only utilizing 35 percent of its plant capacity.

With decreased demand for its chips, TSMC in turn reduced its orders for chip- making equipment by 20 percent. Applied Materials is one company that makes the equipment used in chip-making factories. With the downturn in demand for chip- making equipment, Applied Materials was forced to lay off 2000 workers and require another 12,000 workers to take an unpaid leave.

With the downturn in its business, Applied Materials reduced orders to its suppli- ers. For example, D&H Manufacturing Company, which makes aluminum parts for chip-making equipment, reduced its employment from 600 to 150 workers in 18 months because of the drop-off in business. It also found itself sitting on a one- year supply of inventory versus its usual three months of inventory.

This example illustrates how the effects and decisions made at one end of the supply chain are often amplifi ed as they cascade to the other end. And because the players at different stages in the supply chain are often caught off-guard, it is not surprising that they frequently overreact to the situation. In this particular case, Best Buy was actually having trouble keeping its shelves stocked in the early part of 2009 despite the decline in demand. In fact, Best Buy estimated in March 2009 that it could have sold more in the preceding three months had its suppliers made less drastic reductions to their production plans (Dvorak 2009).

• To many, the mere mention of inventory management conjures up images of detailed calculations and analysis. However, while inventory management is often considered to be a rather bland and narrow topic, there are a number of areas related to inventory management that are generating signifi cant interest. One such area is radio frequency identifi cation (RFID). With RFID, conventional bar codes are replaced with computer chips or smart tags. These smart tags use wireless technology to track inventory. In addition to labor savings, RFID allows

c07.indd 244c07.indd 244 25/09/12 10:49 AM25/09/12 10:49 AM

245 I n t r o d u c t i o n

organizations to manage their inventory more effectively. In 2005, spending on RFID had already reached $1 billion and ABI Research estimated RFID spending was $4.6 billion in 2010.

One early adopter of RFID was Wal-Mart, a company well known for its invest- ments in supply chain technology. By January 2005, 53 of its top 100 suppliers were sending RFID-tagged goods to its three distribution centers in the Dallas, Texas, area. Wal-Mart ’s goal was to have all top-100 suppliers shipping RFID-tagged goods by the end of February 2005 in addition to 37 other suppliers. In subsequent waves, Wal- Mart planned to have the next 200 suppliers on board by January 2006. In August of 2010, Wal-Mart continued its rollout of RFID and began working with its clothing suppliers to install tags on men ’s jeans and other basic clothing items such as socks and undershirts.

The impetus for Wal-Mart ’s investment in RFID was the lack of visibility it had into its backroom storage areas. Better visibility would translate into better informa- tion on which to base replenishment orders, which in turn would provide a better overall customer experience by helping it get inventory onto store shelves in a more timely manner.

The major drawback to RFID is its cost. In 2011 the cost of passive RFID tags ranged from $0.05 to $5.00, depending on the volume of tags purchased and the environmental factors they are designed to withstand (Lacy 2004; Blanchard 2005).

• Vendor-managed inventory (VMI) is another inventory management topic that is generating a signifi cant amount of interest. With VMI, suppliers are given responsi- bility for managing the inventory carried by their retail or wholesale customers. The customers still own the inventory; however, the suppliers are given the responsibility for managing it. Using point-of-sale data, suppliers determine the timing and quan- tity of inventory replenishment orders.

Rich Products, a $2 billion family-owned food company headquartered in Buffalo, has a partnership with IBM to provide VMI services to the grocery industry for its frozen-food items. With this service, grocery stores provide information daily to IBM electronically about inventory withdrawals and inventory balances. Rich accesses this information and then uses its customers’ own purchasing systems to generate replenishment orders based on service performance agreements. These purchase orders are then sent to Rich electronically. Through the use of VMI, retailers hope to increase their inventory turnover while at the same time reducing the occurrence of stockouts (Richardson 2004).

The concept of supply chain management has taken on the nature of a crusade in U.S. industry, in part because of the tremendous benefi ts that accrue to fi rms partici- pating in a well-managed supply chain. The examples above illustrate this by high- lighting the important role supply chain management plays in an organization ’s competitiveness. In Apple ’s case, its supply chain practices ensure a stable supply of its highly demanded products, which in turn leads to satisfi ed customers and mini- mizes potential lost sales. On the other hand, the Best Buy example illustrates the potential for lost sales and profi ts when the supply chain overreacts.

c07.indd 245c07.indd 245 25/09/12 10:49 AM25/09/12 10:49 AM

246 C h a p t e r 7 : S u p p l y C h a i n M a n a g e m e n t

It is also worth noting that, although the benefi ts of superior supply chain manage- ment are clear for manufacturing and distribution fi rms, even service organizations benefi t from good supply chain management. This is not only because services use supplies and facilitating goods in the delivery of their service (as noted in Chapter 1), but also because they, too, outsource many of their internal functions, such as infor- mation technology, accounting, and human resource management, just like manufac- turers do. Thus, the provision of these services becomes part of another supply chain, a chain of services rather than goods, but nonetheless one requiring the same atten- tion to strategy, purchasing, logistics, and management oversight, just like for goods.

We begin the chapter with some defi nitions of the supply chain and supply chain management. As with any new concept, not everyone envisions supply chain management in the same way. We then discuss some of the important strategic advantages that accrue to wise management of the supply chain. From this overview, we then consider the elements of the supply chain in depth, including purchasing/ procurement, logistics, transportation, global sourcing, and supplier management. An important element of supply chain management is the critical role of information technology as a major catalyst in the supply chain movement. Next, we provide some guidelines for successful supply chain management. We conclude with a dis- cussion of closed-loop supply chains.

D E F I N I N G S U P P L Y C H A I N M A N A G E M E N T The term supply chain generally refers to all the activities involved in supplying an end user with a product or service. The perception of each organization that is involved—the ore refi ners, the transporters, the component producers, the manufac- turer, the wholesaler, the retailer, and the customer—being a link in the process makes the analogy of a chain quite appropriate. In Figure 7.1 , we show the position of a typical company (A) in the chain, with its suppliers to the left of it, all the way “upstream” (as it is often called) to the raw materials, and its customers to the right,

UPSTREAM DOWNSTREAM

Raw materials Consumer

Customer Company B

Supplier

Su pp

lie rs

.. ..

Customer Customer Company A

Supplier Customer

Company C Customer .... .... Supplier Supplier

Suppliers .... ... . C

us to

m er

s .... Custom

ers

Figure 7.1 The supply chain.

c07.indd 246c07.indd 246 25/09/12 10:49 AM25/09/12 10:49 AM

247D e f i n i n g S u p p l y C h a i n M a n a g e m e n t

all the way “downstream” to the ultimate consumer. However, company C in the chain (a downstream “customer” as far as company A is considered) sees the same thing as company A, with its suppliers (including upstream supplier company A) to its left and its customers to its right. And as is seen, company B in the middle is the customer of one fi rm and the supplier to another fi rm, as is the situation of almost all the companies in the chain.

Of course, all these companies typically need multiple materials and services to serve their immediate customer in the chain, so there are really a lot of upstream supplier company links connected on the left side of each link in the chain (only shown with links for company A, arrows for all others). And most fi rms typically sell to more than one customer, so there are also multiple downstream customer links connected on the right side of each link in the chain (again shown only for company A). Clearly, managing all these links—that is, suppliers and customers— even if only those directly connected to your company, is a major task!

Given such a lengthy process, it may behoove companies to store inventories of their outputs (if feasible) for immediate delivery. Moreover, it must be remembered that it is not just goods that are fl owing along the chain but also information, funds, paper, people, and other such items, and they are fl owing in both directions along the chain. In addition, the green revolution encourages recycling, recovery, and reuse of products, so even the used product may be fl owing back up the chain. (We will return to the topic of closed-loop supply chains later in this chapter.) In addi- tion, the supply chain also involves other functional areas and activities such as product/service design, fi nance, accounting, marketing, human resources, and engi- neering. Thus, instead of a chain, we should probably think of the supply process as more of a network, with everyone communicating with, and passing monies and items between, everyone else.

Supply chain management (SCM), then, concerns the process of trying to manage this entire chain from initial receipt of the ultimate consumer ’s order all the way back to the raw materials providers and then ultimate delivery back to the consumer. Note that SCM is not restricted to managing only the links that connect with your company ’s position in the chain, but all the links along the chain, so that savings (or increased value) in any part of the chain can be shared or leveraged by other com- panies along the chain. For example, Toyota is famous for teaching their suppliers how to install and operate their famed Toyota Production System (also known as lean manufacturing ). But the teaching doesn ’t stop there, since Toyota ’s fi rst-tier suppliers can gain additional improvements by teaching their suppliers, the second tier, and so on up the supply chain. The interest in supply chain management has exploded primarily because of the development of new information technologies such as intranets, e-mail, EDI (electronic data interchange), and, of course, the Internet. These technologies, in conjunction with greater global competition, have fostered an interest and ability in improving processes along the entire supply chain, resulting in better performance at reduced cost.

SCM can also be considered to include a number of other managerial thrusts, such as quality management (Chapters 1 and 3), inventory management (discussed later), enterprise resource planning (ERP, also discussed later), and lean production (including just-in-time, Chapter 5). But it is even more comprehensive than that. For example, it includes marketing aspects in terms of communication with the cus- tomer, engineering issues involved in product/service design, fi nancial aspects in terms of payments and fl oat, purchasing elements such as sole-sourcing, and, of

c07.indd 247c07.indd 247 25/09/12 10:49 AM25/09/12 10:49 AM

248 C h a p t e r 7 : S u p p l y C h a i n M a n a g e m e n t

course, technological initiatives such as the omnipresent Internet. To a large extent, this breakthrough in conceptualizing the potential for improvement in customer value by including all elements of the value chain is due to the development of advanced information technologies, such as the Internet.

Other defi nitions of SCM include the following points (Walker and Alber 1999):

• SCM coordinates and integrates all the supply chain activities into a seamless process and links all of the partners in the chain, including departments within an organization as well as the external suppliers, carriers, third-party companies, and information system providers.

• SCM enables manufacturers to actively plan and collaborate across a distrib- uted supply chain, to ensure all parties are aware of commitments, schedules, and expedites. By actively collaborating as a virtual corporation, manufactur- ers and their suppliers can source, produce, and deliver products with mini- mal lead time and expense.

• The goal of SCM is to optimally deliver the right product to the right place at the right time, while yielding the greatest possible profi t.

The SCM objective of attempting to manage activities that lie outside a manager ’s normal realm of internal responsibility (that is, managing second- or third-tier sup- pliers, or downstream customers) is to reduce the costs of delivering a product or service to a user and improve its value. Sometimes a distinction is made between a “value” chain, a “demand” chain, and a narrowly defi ned supply chain that simply manages suppliers to obtain the lowest cost. The conceptualization of the value chain is that it considers other important aspects of customer value besides cost, such as timeliness, quality, and functionality. That is, where the supply chain tends to focus on effi ciency, the value chain focuses on effectiveness. These important issues will be discussed in more detail in the next section.

Also, as many have pointed out (e.g., Lummus and Vokurka 1999), the current con- ceptualization of the supply chain still has many elements of the old “push” system of production based on forecasts of demand. (See the Chapter 8 Supplement: Forecasting for more information on this topic.) The newer “pull” systems, consisting of just-in- time ( JIT) deliveries, lean manufacturing, and so on, dictate a different view of the value chain, called a demand chain . In this conceptualization, a customer order pulls the product through the chain on demand, thereby further improving costs and ben- efi ts. Of course, acting after the fact rather than anticipating demand will put even further stress on the ability of the value chain to respond in a timely manner.

Another layer of complexity is often added when managing service supply chains, as the customers of the service can also serve as suppliers. For example, you supply the yard to your landscaping service. Likewise, your lifestyle and budget are impor- tant inputs to the architect you hire to design your dream house. Because the custom- ers of a service may also be a supplier, it is likely that these customer-suppliers need to be handled differently than suppliers that are not customers. For example, suppli- ers that are not customers need to be selected, but customer-suppliers need to be attracted.

The dual nature of the customer-supplier role further compounds the complexity of the service supply chain. With a more manufacturing oriented supply chain, the goods tend to fl ow in one direction downstream. In service supply chains and the dual customer-supplier role, services fl ow in both directions, with the customer

c07.indd 248c07.indd 248 25/09/12 10:49 AM25/09/12 10:49 AM

249S u p p l y C h a i n S t r a t e g y

both upstream and downstream from the service provider. Finally, service providers may require additional fl exibility to deal with the added variation that is associated with customer-supplied inputs compared to other situations where the inputs are supplied by a more limited set of suppliers.

Attempts to reduce the costs of supply (previously considered as “purchasing” or “procurement”) have been ongoing for decades, of course. However, management has also realized that there are costs other than strict materials and production costs in the supply chain that can be reduced with better information sharing and tighter management, and these costs are at the forefront of attention in supply chain man- agement. For example, costs of multiple shipments, costs of inappropriate functional- ity, costs of low quality, costs of late delivery—these are all costs that can be eliminated with better information sharing and managerial oversight.

S U P P L Y C H A I N S T R A T E G Y The concept of the value chain was mentioned earlier, and it should be emphasized that an organization ’s supply chain strategy needs to be tailored to meet the needs of its customers, which isn ’t always the lowest cost. In fashion goods, for example, fast response to short fashion seasons is much more important than lowest cost. And in high technology, new functionality (or reliability, or security) may be more impor- tant than cost. Thus, the strategy for building an organization ’s supply chain should focus on maximizing the value to its customers, where value can be considered to be benefi ts received for the price paid, or benefi ts/cost.

In situations where the goods are basic commodities with standard benefi ts (food, home supplies, standard clothing), then cost reduction will be the focus. But in fash- ion goods, timeliness should be the focus of the supply chain, meaning quick deliv- eries, stockpiling of long lead time items, and so on. In new notebook computers, the focus might be on identifying fi rms that offer new functionality; in telecom, the focus might be on reliability; in music, the focus might be on fl exibility to meet quickly changing tastes or talent. Thus, the supply chain needs to be carefully matched to the fi rm ’s market and needs. Where the fi rm operates in multiple mar- kets, or appeals to multiple needs within the same market, it may fi nd it necessary to operate different supply chains for each focus. Although most of the remaining dis- cussion in this chapter is directed toward the traditional supply chain strategy of minimizing costs, which is always an important consideration and probably the major focus of most supply chains today, the other possible strategic purposes should be kept in mind also.

It is also important to point out that many organizations choose to outsource por- tions of the supply chain management function to so-called third-party logistics (3PL) companies. These 3PL companies provide a range of services, including han- dling the distribution of the organization ’s products, receiving incoming materials, managing the organization ’s warehouses, managing the purchasing function, and handling product returns. The balance of activities kept in-house and those out- sourced vary by company and should be driven by the organization ’s strategy and competencies.

There a number of reasons why organizations choose to outsource portions of or the entire supply chain function to a 3PL. First, assuming that supply chain manage-

c07.indd 249c07.indd 249 25/09/12 10:49 AM25/09/12 10:49 AM

250 C h a p t e r 7 : S u p p l y C h a i n M a n a g e m e n t

ment is not the organization ’s core competency, shifting these activities to a 3PL allows the organization to focus more directly on its core competencies. Second, outsourcing these activities reduces the capital investments in the infrastructure needed to support these activities. In effect, the use of a 3PL converts a signifi cant portion of what was a fi xed cost into a variable cost. Finally, by utilizing a 3PL, the organization gains access to the best practices and technologies that it might not be able to afford or develop if the function was kept in-house. 3PLs are able to make the investment to develop these best practices and technologies because these development costs are spread across multiple organizations served by the 3PL.

However, there are also disadvantages in using 3PLs, such as the longer response time and greater risk of disruption in the supply chain when customers are wanting faster, more reliable response. An added danger of all outsourcing is the natural ten- dency for management to measure only the internal response time of the fi rm when the customer is measuring the total time from order to obtaining the good or service. In any outsourcing decision, the added time and risk of delay from outside suppliers need to be considered. This response time is also affected by the production process adopted, since make-to-stock (MTS) will have the fastest response because the order can simply be pulled off the shelf and sent to the customer, assemble-to-order (ATO) is a bit slower, make-to-order (MTO) is slower still, and engineer-to-order (ETO) is slowest of all.

Strategic Need for Supply Chain Management To understand the potential for obtaining strategic advantage from better manage- ment of the supply chain, whether it is kept in-house or outsourced to a 3PL, it is useful to realize that total supply chain costs represent more than half, and in some cases three-quarters, of the total operating expenses for most organizations (Quinn 1997). To understand these values, bear in mind that the broader concept of the sup- ply chain includes the supply, storage, and movement of materials, information, personnel, equipment, and fi nished goods within the organization and between it and its environment. The objective of supply chain management is to integrate the entire process of satisfying the customer ’s needs all along the supply chain. This includes procuring different groups of raw materials from multiple sources (often through purchasing or recycling or recovery), transporting them to various process- ing and assembly facilities, and distributing them through appropriate distributors or retailers to the fi nal consumer. Within this process are a great variety of activities such as packaging, schedule coordination, credit establishment, inventory manage- ment, warehousing, maintenance, purchasing, order processing, and supplier selec- tion and management.

As organizations have continued to adopt more effi cient production techniques such as lean manufacturing, total quality management, inventory reduction techniques to reduce costs and improve the quality, functionality, and speed of delivery of their prod- ucts and services to customers, the costs and delays of procuring the requisite inputs and distributing the resulting goods and services are taking a greater and greater frac- tion of the total cost and time. For example, the cost of just physical distribution itself is now up to 30 percent of sales in the food industry. To achieve quick response with quality goods that accurately satisfy the need at the lowest possible cost requires taking a broad, long-range, integrated perspective of the entire customer fulfi llment process instead of focusing on the little segments and pieces of the chain.

c07.indd 250c07.indd 250 25/09/12 10:49 AM25/09/12 10:49 AM

251S u p p l y C h a i n S t r a t e g y

For instance, if each segment of the supply chain is acting in a way to optimize its own value, there will be discontinuities at the interfaces and unnecessary costs will result. If an integrated view is taken instead, there may be opportunities in the sup- ply chain where additional expense or time in one segment can save tremendous expense or time in another segment. If a broad enough view is then taken, the sav- ings in the one segment could be shared with the losing segment, so everyone would be further ahead. This broad, integrated view of the supply chain is more feasible these days due to the recent capabilities of advanced information technol- ogy and computer processing (e.g., bar codes, computerized manufacturing, the Internet, enterprise resource planning systems, electronic funds transfer).

Other factors are also driving the need to better manage the supply chain:

• Increasing global competition. In addition to increased pressure on cost from global competitors who have lower labor rates, they also frequently offer bet- ter quality, functionality, and customer responsiveness. This is pressuring fi rms to look globally for better or cheaper suppliers, resulting in increased outsourcing and offshoring.

• Outsourcing. Since more organizations are outsourcing and thereby increas- ing the need for transportation, this has pushed up transportation costs.

• E-commerce. The advent of e-commerce and other electronic technologies has made it easier and cheaper to outsource, either domestically or even globally.

• Shorter life cycles. Customers are demanding greater variety, faster response, higher quality, and cheaper prices. One result of these demands is shorter product life cycles, which means constantly changing supply chains and using more chains over the same period of time.

• Greater supply chain complexity. The increased complexity of supply chains requires much more attention and better management of these chains. For example, in early 2001, when the bottom fell out of the telecom market, Solectron Corp., the world ’s biggest electronics contract manufacturer, was holding $4.7 billion of inventory from its 4000 suppliers to fi ll fi rm orders from Cisco, Ericsson, Lucent, and other telecoms. But when the telecoms canceled their orders, no one knew who owned all that inventory (Engardio 2001)!

• Increasing levels of concern for the environment. Addressing environmental concerns impacts virtually all aspects of supply chain management from the sourcing of parts, to the distribution of the product, and even to the disposal of the product once it reaches the end of its useful life. Green sourcing seeks to identify suppliers in such a way that the organization ’s carbon footprint and overall impact on the environment are minimized. Reducing the waste associated with products is another way organizations minimize the negative impact they have on the environment. Along these lines, and as is discussed in Chapter 1, organizations can deploy a strategy referred to as the three R ’s: reduce, reuse, and recycle.

Implementing supply chain management has brought signifi cant documented benefi ts to many companies. Ferguson (2000) reports, for example, that compared to their competitors, such fi rms enjoy a 45 percent supply chain cost advantage, an order-cycle time and inventory days of supply 50 percent lower, and fi nished product

c07.indd 251c07.indd 251 25/09/12 10:49 AM25/09/12 10:49 AM

252 C h a p t e r 7 : S u p p l y C h a i n M a n a g e m e n t

delivery 17 percent faster. Lummus and Vokurka (1998) note that these fi rms operate with 36 percent lower logistics costs, which, by itself, translates into a 4 percent increase in net profi t margins. One fi rm reported a 25 to 50 percent reduction in fi nished product inventories, a 10 percent reduction in cost, and a 10 to 25 percent improvement in production process reliability.

Of course, these are primarily the cost aspects of the SCM process, which are more easily measured than the qualitative benefi ts, such as more loyal customers and a larger market share. There are also signifi cant effects on other important aspects of an organization, such as its ability to learn new procedures and ways of operating, the morale of its employees, and the ability to change direction quickly.

Measures of Supply Chain Performance Better supply chain performance will show up in a number of standard fi nancial measures of a company ’s health. Lower inventories, normally considered an asset, will be refl ected in less need for working capital (WC) and a higher return on asset (ROA) ratio (since assets are reduced). And the lower cost to carry these inventories (as well as other reduced costs in the supply chain) will be seen in a reduced cost of goods sold (CGS) and thus a higher contribution margin , return on sales (ROS), and operating income . Moreover, if the supply chain is also better managed to provide other benefi ts to the consumer, as mentioned earlier, the effect should be seen in higher total reve- nue , since the consumer will be willing to pay more. Lower costs, if used to reduce prices, will also result in higher volumes, which will further increase revenues.

One performance measure that provides managers with a broad view of the sup- ply chain is the cash conversion cycle. This fi nancial performance metric helps a company assess how well it is managing its capital. In effect, the cash conversion cycle is the amount of time the organization ’s cash is tied up in working capital before being returned by customers as they pay for delivered products or services. The key inputs needed to calculate the cash conversion cycle are inventory (I), accounts receivable (AR), and accounts payable (AP). These inputs are readily avail- able from the organization ’s fi nancial statements. Before calculating the cash conver- sion cycle (CCC), the inputs are standardized into days as follows:

I   5   inventory

_______________________ annual cost of goods sold

  3   365

AR   5   accounts receivable _________________ annual net sales

  3   365

AP   5   accounts payable

_______________________ annual cost of goods sold

  3   365

These standardized inputs are used to calculate the cash conversion cycle as follows:

Cash conversion cycle   5   I   1   AR   2   AP

A positive cash conversion cycle represents the number of days the organization ’s capital is tied up waiting for the customer to pay for the products or services. A nega- tive cash conversion cycle represents the number of days the organization is able to

c07.indd 252c07.indd 252 25/09/12 10:49 AM25/09/12 10:49 AM

253S u p p l y C h a i n D e s i g n

receive cash from its sales before it pays its suppliers. Thus, the smaller the cash con- version cycle, including negative numbers, the better the organization is performing.

Dell has reduced their supply time so much that they actually receive payment from the customer before (known as fl oat , another fi nancial term) they have to pay their suppliers for the parts that make up the customer ’s product! In 1998, Dell ’s CCC was 29 days. By 2005, it had improved to 230 days, and by 2009 it was 244 days (Dignan 2002; Magretta 1998).

Beyond these standard fi nancial measures, however, we can also look at some more operations-oriented measures that we typically use to see how well operations is performing, such as defect rates, lead times, inventory turns, productivity ratios, and so on. Since one of the major cost savings in SCM is the cost of inventories, it is worth- while to examine some performance measures related to inventory reduction. One such measure to track is the percent of the fi rm ’s assets represented by inventory. First we calculate the aggregate inventory value (at cost) on average for the year ( AAIV )   :

AAIV   5   raw materials   1   work - in - process   1   fi nished goods

%   Assets in inventories   5   AAIV / total assets

Another inventory measure is the inventory turnover (or “turns,” as it is sometimes called):

Inventory turnover ( “ turns ” )   5   annual cost of goods sold / AAIV

Note that the inventory turnover is based on the same items that make up total annual revenues, but is based on their cost instead of their price. Turnover essen- tially represents how often the average inventory is used up to obtain the total sales for the year. Like ROA, the more the inventory and assets can be reduced and still maintain the same sales, the better! Inverting the equation for turns gives us the same information but through a measure of the proportion of the year ’s sales we are holding in inventory. This is usually expressed in daily (or weekly) periods:

Days of supply   5   AAIV / daily CGS

In some fi rms that have achieved supply chain excellence, they measure their supply in hours instead of days. Dell Computer is one of these fi rms (Dignan 2002; Magretta 1998) due to the outstanding job they have done on fi ne-honing their sup- ply chains.

S U P P L Y C H A I N D E S I G N As shown in Figure 7.2 , the supply chain consists of the network of organizations that supply inputs to the business unit, the business unit itself, and the customer network. Note that the supplier network can include both internal suppliers (i.e., other operating divisions of the same organization) and external suppliers (i.e., operating divisions of separate organizations). Also, note how design activities cut

c07.indd 253c07.indd 253 25/09/12 10:49 AM25/09/12 10:49 AM

254 C h a p t e r 7 : S u p p l y C h a i n M a n a g e m e n t

across the supplier network and the business unit, and how distribution activities cut across the business unit and the customer network. This broader view of the entire process of serving customer needs provides numerous benefi ts. For example, it focuses management attention on the entire process that creates value for the cus- tomer, not the individual activities. When viewed in this way, information is more freely shared up and down the supply chain, keeping all parties informed of one another ’s needs. Furthermore, activities can be performed at the point in the supply chain where they make the most sense. To illustrate, instead of providing Johnson Controls with detailed specifi cations for car seats, car manufacturers provide broad specifi cations and rely on Johnson Controls’ expertise to design and manufacture their car seats.

In this section we will look at each of the major logistical elements of the supply chain to better understand how they operate and interact to deliver value to the fi nal customer: the “bullwhip” effect, transportation, and location. Outsourcing, purchas- ing, supplier management, and the role of information technology are discussed later in the chapter.

Logistics Logistics can be defi ned as planning and controlling the effi cient, effective fl ows of goods, services, and information from one point to another. As such, it consists of inventories, distribution networks, storage and warehousing, transportation, infor- mation processing, and even production—a rather all-enveloping term.

Figure 7.2 Simplifi ed supply chain.

Transformation system

Raw materials Finished

goods

Components

WIP

Customer distribution

center

Customer distribution

center

RetailersFirst TierSecond Tier

Supplier Network

R

RS

S

S

S

S

S

R

R

E n d

C u s t o m e r s

DistributeDesign

Information

R

Business Unit Customer Network

c07.indd 254c07.indd 254 25/09/12 10:49 AM25/09/12 10:49 AM

255S u p p l y C h a i n D e s i g n

In these days of intense worldwide competition, international production in sup- ply chains, and global distribution, logistics is taking on tremendous importance. Labor cost is dropping as a proportion of total output cost, as are manufacturing costs in general, but the costs of acquisition and distribution have remained about the same and now account, as noted above, for up to 30 percent of sales.

Generally speaking, when organizations design their supply chains, they tend to focus on one of two overarching goals: maximizing effi ciency and minimizing the cost of the supply chain versus maximizing the fl exibility and responsiveness of the supply chains. A logical question is: Are there guidelines that can help an organi- zation determine whether its supply chain focus should be on effi ciency or respon- siveness? The good news is that there are indeed guidelines for this and, as you might suspect, the emphasis on effi ciency versus responsiveness depends largely on the demand pattern of the outputs.

More specifi cally, Professor Marshall Fisher of the Wharton School distinguishes two fundamentally different types of outputs that he refers to as functional and innovative products. Functional products tend to be staples that we routinely pur- chase. As such, functional products tend to be more mature products with predict- able demand patterns, long life cycles, and relatively low contribution margins. Examples of functional products include frozen vegetables, batteries, paper towels, soft drinks, and printer paper. At the other extreme, innovative products represent products that are continuously being improved and enhanced with new styles, fea- tures, capabilities, and so on. Because they are continuously updated, innovative products have unpredictable demand, short life cycles, and relatively high contribu- tion margins as well as being offered in many varieties and options. Examples of innovative products include smart phones, tablet computers, tennis rackets, and designer blue jeans.

Using this classifi cation, Professor Fisher suggests that focusing on supply chain effi ciency is appropriate for functional outputs, while focusing on responsiveness is appropriate for innovative products. In fact, when seeking to identify the root cause of supply chain problems, quite often you will fi nd that the problems are the result of the supply chain not being properly aligned with the demand pattern of the product. Most often, this occurs when an organization seeks to offer innovative products but is focusing on the effi ciency of the supply chain. In these cases, the organizations would be well served to either consider marketing their products as functional products or placing greater emphasis on improving the responsiveness of their supply chain as opposed to optimizing its effi ciency.

The Bullwhip Effect

While all products have an underlying demand pattern, the way the supply chain is managed can distort our perception of what the true underlying pattern of demand is. We now have a better understanding of one logistical effect that distorts the demand pattern known as the bullwhip effect , named after the action of a whip where each segment further down the whip goes faster than that above it. Unfortunately, this same effect occurs in a supply chain, but in reverse order, and has been well documented. More specifi cally, in supply chains the bullwhip effect results when the variability of demand increases from the customer stage upstream to the factory stage. This is often the result of different parties in the supply chain being overly reactive in their ordering practices, as in the Best Buy example at the

c07.indd 255c07.indd 255 25/09/12 10:49 AM25/09/12 10:49 AM

256 C h a p t e r 7 : S u p p l y C h a i n M a n a g e m e n t

beginning of the chapter. For example, this happens when a small percentage increase in a retailer ’s orders results in the wholesaler increasing its orders by an amount greater than that of the retailer—a safety stock—just to be covered in case demand is increasing. Then the distribution center sees this greater demand from its wholesalers and increases its orders by some safety percentage, also to be safe. The end result is that the factory sees a huge jump in demand. As it orders more equip- ment, labor, and materials to satisfy this big increase, too much is fed into the pipe- line and the retailer cuts back, with the wholesaler and distribution center likewise cutting back even more. The factory then sees a tremendous drop in demand and reverses the cycle, cutting excessively into production and initiating another round of excessive demand. This boom–bust cycle is particularly prevalent in some indus- tries, such as commercial building. Obviously, both overproduction and underpro- duction are expensive and drive up supply chain costs.

The bullwhip effect can occur whenever any one of three conditions is extreme enough to cause the boom–bust cycle. The fi rst condition is simply long lead times between the stages of the supply chain, so that changes in demand requirements are slow moving up and down the chain, thereby allowing excessive changes to occur in the other stages of the chain. The second condition is large lot sizes with infrequent orders, resulting again in lags in information. And the third condition is the sole trans- mission of information occurring by hand-offs from one link of the chain to the next.

The ways to eliminate the bullwhip effect are to reverse these three conditions. Reducing lead times through just-in-time programs, for example, will result in imme- diate deliveries of the ordered amounts, so safety stocks are unnecessary. Reducing lot sizes means smaller, more frequent deliveries, which again eliminates the need for large safety stocks. And, fi nally, the sharing of information from the retailer throughout the supply chain gives the factory, as well as the other supply chain part- ners, accurate information, so appropriate amounts of items are produced and delivered.

In addition to these three conditions, there are a number of business practices that also contribute to the bullwhip effect. One business practice is the tendency for customers to have a preference for placing all their orders either at the beginning or the end of the week (or month) rather than spacing orders out evenly. This leads to a situation where incoming orders will be bunched up around the beginning and end of the week (or month), thereby increasing the variability of the supplier ’s daily demand beyond the variability of the customers’ daily demand. Furthermore, this problem tends to be amplifi ed as the orders cascade upstream.

Another business practice that contributes to the bullwhip effect is the use of standard batch sizes. For example, if a particular product is packaged in cases of 24 units, then replenishment orders for this product will be done in multiples of 24. This practice further bunches up orders and again results in the supplier ’s daily demand being larger than that of the customers placing the orders.

Trade promotions are yet another practice that contributes to the bullwhip effect. Trade promotions are short-term discounts suppliers offer their customers. These discounts provide customers with an incentive to order more product than they need, called forward buying. Because customers will choose to place their orders when the trade promotion is offered and even delay orders in anticipation of a trade promotion, these trade promotions create another order-bunching problem.

A fi nal practice that contributes to the bullwhip effect is shortage gaming. This practice occurs in situations where a product is in short supply. Anticipating that the

c07.indd 256c07.indd 256 25/09/12 10:49 AM25/09/12 10:49 AM

257S u p p l y C h a i n D e s i g n

supplier will allocate its inventory to its customers, some suppliers infl ate their orders, fearing that they will be shipped less than they ordered. Attempting to game the system in this fashion exacerbates the shortage problem, as some customers end up with less than they can sell because they did not infl ate their orders, while others end up with more than they can sell. In some cases, the suppliers themselves further compound this problem by allowing their customers to return unsold inventory.

There are several actions suppliers can take to mitigate these practices. For exam- ple, suppliers can ask their customers to share information more frequently about actual demand. Likewise, suppliers can coordinate with their customers to eliminate the batching of orders. Alternatively, suppliers can encourage their customers to make greater use of technology such as the Web and electronic data interchange (EDI) to place smaller but more frequent orders. Furthermore, suppliers can elimi- nate the practice of offering trade promotions. Finally, suppliers can enhance the value proposition they provide their customers while at the same time helping smooth out incoming orders by taking over the management of their customers’ inventory, referred to as “vendor-managed” inventory and illustrated by the Rich Products example in the Introduction to this chapter.

Transportation

The four major transportation modes are, historically, water, rail, truck, and air. Water is the least expensive mode and is good for long trips with bulky, nonperishable items. But it is very slow and of limited accessibility. It handles the majority of ton- miles of traffi c. However, railroads handle the most total tons of traffi c and are thus used for shorter hauls than water. They have many advantages: ability to handle small as well as large items, good accessibility, specialized services (e.g., refrigera- tion, liquids, cattle), and still a relatively low cost.

Trucking holds more advantages for short hauls with small volumes to specialized locations. Truck transport has grown at the expense of rail for several reasons, such as growth of the national highway system, better equipment, and liberalized regulations.

Air transport is used for small, high-value, or perishable items such as electronic components, lobsters, optical instruments, and important paperwork. Its main advantage is speed of delivery over long distances. Thus, for the appropriate prod- ucts, it can signifi cantly reduce inventory and warehousing costs, with a correspond- ing improvement in customer service.

Taking all the pros and cons of each mode of transportation into consideration in planning is a complex task. Table 7.1 lists the major considerations that should be factored into the decision. Each particular situation may have additional factors to consider.

Independent of the specifi c mode of transport are additional transportation prob- lems involving such considerations as the number of transporting vehicles, their capacities, and the routes that each vehicle will take. In general, these interrelated problems are frequently included as part of the routing problem . Solving the routing problem involves fi nding the best number of vehicles and their routes to deliver the organization ’s output to a group of geographically dispersed recipients. When only one vehicle is serving all the recipients, the problem is known as the traveling sales- man problem . In this problem, a number of possible routes exist between the organ- ization and all the recipients, but only a few or perhaps just one of these routes will minimize the total cost of delivery. In the routing and traveling salesman problems,

c07.indd 257c07.indd 257 25/09/12 10:49 AM25/09/12 10:49 AM

258 C h a p t e r 7 : S u p p l y C h a i n M a n a g e m e n t

certain procedures are available to minimize either the distance traveled or the cost, but quite often there are other considerations, such as balancing workloads among vehicles or minimizing idle or delay time.

Location

Besides distributing outputs to customers by transporting them, if there is a facilitat- ing good, we can also locate where our customers can easily obtain them. Since service outputs without a facilitating good are generally diffi cult, expensive, or even impossible to transport, service organizations distribute their output primarily by locating in the vicinity of their recipients. Examples of this approach include medical clinics, churches, playgrounds, restaurants, and beauty shops.

Advances in information and telecommunications technology have allowed some pure service organizations (i.e., those without a facilitating good) to reach their recipients through phone, cable, the Internet, or microwave links. Thus, stockbro- kers, banks, and other such service providers may locate in areas removed from their customers or recipients but more economical in other respects, such as proximity to the stock exchange or the downtown business district.

Some pure service organizations, however, do attempt to transport their services, although frequently with a great deal of trouble. These instances occur when the nature of the service (a traveling carnival, a home show) makes it impractical to remain in one fi xed location for an extended duration or, more commonly, when the service (mobile X-ray, blood donor vehicle, bookmobile) is deemed very important to the public but may otherwise be inaccessible.

Product organizations, on the other hand, can generally trade transportation costs for location costs more easily and, therefore, can usually minimize their logistics costs. This allows goods producers to locate in the best global locations for each

T A B L E 7 .1 • Fac tor s to Cons ider in Transpor ta t ion Dec i s ions Cost per unit shipped

Ability to fi ll the transporting vehicle

Total shipment cost

Protection of contents from theft, weather, and the like

Shipping time

Availability of insurance on contents, delivery, and so forth

Diffi culty of arranging shipment (governmental regulations, transportation to shipment site, etc.)

Delivery accommodations (to customer ’s site, transfer to another transportation mode, extra charges)

Seasonal considerations: weather, holidays, and so on

Consolidation possibilities (among multiple products)

Risk (to contents, to delivery promises, to cost, etc.)

Size of product being shipped

Perishability of product during shipment

c07.indd 258c07.indd 258 25/09/12 10:49 AM25/09/12 10:49 AM

259O u t s o u r c i n g a n d G l o b a l S o u r c i n g

stage in their supply chains. In some instances, however, even product organizations are forced into fi xed locations. One of these instances concerns the nature of the fi rm ’s inputs, and the other concerns its outputs.

Processing Natural Resources . Organizations that process natural or basic resources as raw materials or other essential inputs to obtain their outputs will locate near their resource if one of the following conditions holds:

1. There is a large loss in size or weight during processing.

2. High economies of scale exist for the product. That is, the operating cost of one large plant with the same total capacity as two smaller plants is signifi - cantly less than the combined operating costs of the two small plants.

3. The raw material is perishable (as in fi sh processing and canning) and can- not be shipped long distances before being processed.

Examples of these types of industries include mining, canning, beer production, and lumber. In these cases the natural inputs (raw materials) are either voluminous or perishable, and the fi nal product is much reduced in size, thus greatly reducing the cost of transportation to the recipients (either fi nal users or further processors).

Immobile Outputs . The outputs of some organizations may be relatively immo- bile, such as dams, roads, buildings, and bridges. In these cases (referred to as projects), the organization locates itself at the construction site and transports all required inputs to that location. The home offi ce is frequently little more than one room with a phone, secretary, fi les, and billing and record-keeping facilities.

Product organizations may also locate close to their market, not necessarily to minimize transportation costs of distribution, but to improve customer service. Being close to the market makes it easier for the recipient to contact the organization and also allows the organization to respond to changes in demand (involving both quan- tity and variety) from current and new recipients. As in war, the people on the front- lines are closest to the action and are able to respond to changing situations faster than those far away, simply because information about changes is available sooner and is generally more accurate.

O U T S O U R C I N G A N D G L O B A L S O U R C I N G Outsourcing is the process of contracting with external suppliers for goods and serv- ices that were formerly provided internally and offers an important benefi t for SCM. Global sourcing is an important aspect of supply chain outsourcing strategy, and we see it occurring more and more. In the news, we read and hear about the meetings of the WTO (World Trade Organisation), the latest accords of the G-7 major trading nations, the dangers of NAFTA (North American Free Trade Agreement), the job losses due to overseas outsourcing (furniture manufacturers closing U.S. plants and sourcing from Asia, call centers being relocated to India), and so on. When asked on the Lou Dobbs show for the reasons all this outsourcing is occurring now, the economist Paul Craig Roberts responded that two primary factors were responsi- ble: (1) the fall of communism and the economic insulation it had maintained,

c07.indd 259c07.indd 259 25/09/12 10:49 AM25/09/12 10:49 AM

260 C h a p t e r 7 : S u p p l y C h a i n M a n a g e m e n t

and (2) the advent of telecommunications and computer technology, which physi- cally allow work that previously had to be done locally or regionally to now be con- ducted overseas.

The classic example of global outsourcing has been Nike, where the shoes are designed in the United States but all the production is done overseas. The strategic appeal of this lean model of business to other manufacturing and consumer fi rms is multiple. First, overseas production offers the promise of much cheaper labor costs, clearly a strategic benefi t. But equally attractive to many fi rms that are outsourcing, whether globally or domestically, is the ability to dump a large portion of their capital-intensive production assets and staff, thus giving a big boost to their balance sheets, especially their return on assets. In addition, not being burdened with fi xed, unchangeable capital production assets allows fi rms to be more fl exible and respon- sive to their customers’ changing needs.

There is a danger to outsourcing, however, particularly overseas outsourcing, and that is the possibility of being hollowed out , as noted in Chapter 1. To summa- rize, this is the situation where the supplier has been trained to produce, and even sometimes design, the customer ’s product so well that it can simply sell the product under its own brand and compete successfully against its former customer. In many cases, the customer has gone so long without designing or producing its own product—simply slapping its logo on the foreign-produced item—that it has lost the knowledge and skills to even compete in the market. This happened in the 1980s when American manufacturers trained foreign fi rms in how to produce television sets and other electronic goods—and lost those entire industries. Clearly, decisions about outsourcing at this level are strategic ones for the organization, involving great potential benefi ts but also great risks, and should be deliberated thoroughly.

A more recent phenomenon is the trend toward outsourcing the entire produc- tion process to third-party contract manufacturers . In this case, the fi rms often con- clude that their core competency is not in manufacturing per se but rather in system innovation or design. In the electronics industry, this is becoming a major element of SCM strategy for fi rms like Cisco, Apple, IBM, and many others. Cisco, for example, hardly makes any products itself. The big players in this growing industry are Jabil Circuit, Flextronics, and SCI Systems. In fact, in the electronics sector, contract manu- facturing was growing faster than the rate of growth of electronics itself in the late 1990s. In spite of the provision of products, these contract manufacturers consider themselves manufacturing service providers, and, indeed, this is a substantial service they offer their customers. However, in addition to the major impacts outsourcing involves for operations, it also has major impacts on other functional areas of the organization, such as marketing, fi nance, R&D, and human resource management. Moreover, to use this approach successfully requires that the fi rm maintain a strong, perhaps even core, competence in outsourcing. Many failures have resulted when fi rms jumped into outsourcing but didn ’t have the skills to manage it properly.

Outsourcing in general is a strategic element of SCM these days, not just for pro- duction materials but for a wide range of services as well. For example, organiza- tions are coming to realize that many of the activities they perform internally, such as accounting, human resources, R&D, and even product design and information systems, are not part of their core competencies and can be performed more effi - ciently and effectively by third-party providers, often at a fraction of the cost of in- house workers. There is thus a growing movement toward increasing the span of SCM to include the acquisition of these services.

c07.indd 260c07.indd 260 25/09/12 10:49 AM25/09/12 10:49 AM

261O u t s o u r c i n g a n d G l o b a l S o u r c i n g

Purchasing/Procurement Organizations depend heavily on purchasing activities to help them achieve their supply chain strategy by obtaining quality materials and services at the right cost when they are needed. Purchasing is expected to be able to quickly identify and qualify suppliers, negotiate contracts for the best price, arrange for transportation, and then continue to oversee and manage these suppliers. Lately, purchasing has been given the added responsibility in many organizations for also supplying major services to the organization, such as information technology, accounting, human resources, and other previously internal functions.

Another common term for the purchasing function is procurement . Whereas “ purchasing” implies a monetary transaction, “procurement” is the responsibility for acquiring the goods and services the organization needs, by any means. Thus, it may include, for example, scrap and recycled as well as purchased materials. Procurement thus allows the consideration of environmental aspects of obtaining and distributing products. For example, there is often the possibility of recovering certain materials through recycling, reuse, or scrap purchases. And remanufacturing of goods is an inexpensive alternative to virgin production. On the distribution side, the concept of reverse logistics is being practiced in Germany, where packaging must reverse the logistics chain and fl ow back to the producer that originated it, for disposal or reuse.

The purchasing area has a major potential for lowering costs and increasing profi ts—perhaps the most powerful within the organization. Consider the following data concerning a simple manufacturing organization.

Total sales   5   $ 10,000,000

Purchased materials   5   7,000,000

Labor and salaries   5   2,000,000

Overhead   5   500,000

Profi t   5   500,000

To double profi ts to $1 million, one or a combination of the following fi ve actions could be taken.

1. Increase sales by 100 percent

2. Increase selling price by 5 percent (same volume)

3. Decrease labor and salaries by 25 percent

4. Decrease overhead by 100 percent

5. Decrease purchase costs by 7.1 percent

Although action 2 may appear easiest, it may well be impossible, since competi- tors and the market often set prices. Moreover, raising prices almost always reduces the sales volume. In fact, raising prices often decreases the total profi t (through lower volume). Action 5 is thus particularly appealing. Decreasing the cost of pur- chased material provides signifi cant profi t leverage. In the previous example, every 1 percent decrease in the cost of purchases results in a 14 percent increase in profi ts. This potential is often neglected in both business and public organizations.

Furthermore, this logic is also applicable to service organizations. For example, investment fi rms typically spend 15 percent of their revenues on purchases. However,

c07.indd 261c07.indd 261 25/09/12 10:49 AM25/09/12 10:49 AM

262 C h a p t e r 7 : S u p p l y C h a i n M a n a g e m e n t

manufacturing fi rms spend about 55 percent of their revenues for outside materials and services (Tully 1995)! And with factory automation and outsourcing increasing, the percentage of expenditures on purchases is increasing even more. In addition, with lean and JIT programs at so many fi rms (discussed in greater detail in Chapter 5), “just-in-time purchasing” is even further increasing the importance of purchasing and procurement, since delays in the receipt of materials, or receiving the wrong materials, will stop a JIT program dead in its tracks.

SCM programs are putting ever greater emphasis on the purchasing function. Thus, we are seeing multiple new initiatives for cutting purchasing costs, including reverse auctions and joint venture Web sites by organizations that are normally com- petitors. Reverse auctions use a Web site to list the items a company wants to buy and bidders make proposals to supply them, the lowest qualifi ed bidder typically winning the auction. Joint venture Web sites are typically for the same purpose but combine the purchasing power of multiple large players in an industry—automobile manufacturing, aerospace, health care, for example—in order to obtain even bigger cost savings. Such sites are virtual online bazaars, including all the goods and serv- ices the joint partners wish to outsource. But the range and volumes are massive, considering that the old-big-three U.S. auto companies each spent close to $80 bil- lion a year on such purchases.

Value Analysis

A special responsibility of purchasing, or purchasing working jointly with engineering/design and operations (and sometimes even the supplier), is to regu- larly evaluate the function of purchased items or services, especially those that are expensive or used in high volumes. The goal is to either reduce the cost of the item or improve its performance. This is called “value analysis” because the task is to investigate the total value of the item to see if it can be eliminated, redesigned for reduced cost, replaced with a less expensive or more benefi cial item, or even if the specifi cations can be relaxed. Other aspects are investigated, too, such as the pack- aging, the lead time, the transportation mode, the materials the item is made from, whether the part can be combined with another part or parts, and so on.

Recent efforts in this area have extended the reach farther up the supply chain to involve second- and third-tier suppliers, even bringing them in before the product is designed in order to improve its value upfront, called early supplier involvement . Value analysis should be a continuing effort to improve supply chain perfor - mance and increase its value to the ultimate consumer.

Key Elements of Effective Purchasing

Organizations that are highly effective in SCM purchasing seem to follow three practices:

1. They leverage their buying power. The advantages associated with decentralization are typically not achieved when it comes to purchasing. For example, Columbia/HCA combines the purchases of its 200-plus hospi- tals to increase its overall purchasing power. By combining all of its pur- chases for supplies ranging from cotton swabs to IV solutions, for instance,

c07.indd 262c07.indd 262 25/09/12 10:49 AM25/09/12 10:49 AM

263O u t s o u r c i n g a n d G l o b a l S o u r c i n g

it was able to reduce purchasing costs by $200 million and boost profi ts by 15 percent.

2. They commit to a small number of dependable suppliers. Leading sup- pliers are invited to compete for an organization ’s business on the basis of set requirements, such as state-of-the-art products, fi nancial condition, reli- able delivery, and commitment to continuous improvement. The best one to three suppliers are selected from the fi eld of bidders on the basis of the specifi ed requirements. Typically, one- to fi ve-year contracts are awarded to the selected suppliers. These contracts provide the supplier with the opportunity to demonstrate its commitment to the partnership. The customer shares information and technology with the supplier, and the supplier responds in turn. If a supplier is able to consistently improve its perform- ance, the organization reciprocates by increasing the volume of business awarded to that supplier and extending the contract.

3. They work with and help their suppliers reduce total cost. Often, organizations will send their own production people to a supplier ’s plant to help the supplier improve its operating effi ciency, improve its quality, and reduce waste. Additionally, an organization may benchmark key aspects of a supplier ’s operation such as prices, costs, and technologies. If it is discovered that a supplier has slipped relative to the competition, the organization can try to help the supplier regain its lead. If the supplier is unable or unwill- ing to take the steps necessary to regain its leadership position, the organiza- tion may need to fi nd a new partner.

Supplier Management Our discussion of the management of an organization ’s suppliers will focus on three areas: (1) selecting the suppliers, (2) contemporary relationships with suppliers, and (3) certifi cation and auditing of ongoing suppliers.

Supplier Selection and Vendor Analysis

The general characteristics of a good supplier are as follows:

• Deliveries are made on time and are of the quality and in the quantity specifi ed.

• Prices are fair, and efforts are made to hold or reduce the price. • The supplier is able to react to unforeseen changes such as an increase or

decrease in demand, quality, specifi cations, or delivery schedules—all fre- quent occurrences.

• The supplier continually improves products and services. • The supplier is willing to share information and be an important link in the

supply chain.

However, these are not the only factors to be considered in selecting a supplier. Additional considerations involve the supplier ’s reputation/reliability, its having a nearby location (especially important for JIT delivery), its fi nancial strength, the

c07.indd 263c07.indd 263 25/09/12 10:49 AM25/09/12 10:49 AM

264 C h a p t e r 7 : S u p p l y C h a i n M a n a g e m e n t

Supplier Certifi cation and Audits

As can be seen, these sole-sourcing arrangements are becoming virtual partnerships, with the customer asking the supplier to become more involved even at the design stage and asking for smaller, more frequent JIT deliveries of higher-quality items. This means longer-term relationships, help with each other ’s problems, joint plan- ning, sharing of information, and so on. To do this, suppliers are being certifi ed or qualifi ed so that their shipments do not need to be inspected by the customer—the

strength of its management, and even what other customers and suppliers are involved with it. For example, if we are a relatively small customer, we might be more at risk of not getting a delivery if a larger customer experiences a problem and needs our supplier ’s immediate help. Or if our supplier has weak or unreliable second- or third-tier suppliers, we might encounter a problem getting our supplies through no fault of our direct supplier.

Supplier Relationships

In these days of intense global competition and supply chain management, the rela- tionship between customers and suppliers has changed signifi cantly. In the past, most customers purchased from the lowest bidders who could meet their quality and delivery needs, often maintaining at least two or three suppliers in case one was suddenly unable to meet their needs due to a wildcat strike or delivery problem. As pressure mounted to reduce costs, they often pressured their suppliers to cut costs by promising larger volumes to those that had the lowest costs and provided smaller amounts to other suppliers.

To implement SCM, customers are seeking a closer, more cooperative relationship with their suppliers. They are cutting back the total number of their suppliers by a factor of 10 or 20 and combining their purchases, with those remaining getting the overwhelming volume of all their business. They are also asking suppliers to do a greater portion of assembly, such as with automobile seats and other automotive components, which can then simply be installed as a package rather than assembled fi rst and then installed. Not only does the reduced assembly labor save them cost, but in return for the higher volumes, they are expecting even further reductions in cost from their reduced number of suppliers.

DILBERT: ©Scott Adams/Dist. by United Feature Syndicate, Inc.

c07.indd 264c07.indd 264 25/09/12 10:49 AM25/09/12 10:49 AM

265I n v e n t o r y M a n a g e m e n t

items go directly to the production line. This is often referred to as stockless purchas- ing , because the items do not sit in the stockroom costing capital for holding and securing them. To ensure that the contracted supplies will be available when needed, the customers periodically conduct supplier audits of their vendors, checking for potential production or delivery problems, quality assurance, design competence, process improvement procedures, and the management of corrective actions. Some customers rely on standard industry certifi cations such as ISO 9000 (see Chapter 3) rather than incurring the time and expense of conducting their own certifi cation. Such certifi ed suppliers are sometimes known as world-class suppliers.

Of course, most of the benefi ts of this partnership accrue to the customer rather than the supplier. The main immediate benefi t to the supplier is that it stays in busi- ness, and even grows. If managed properly, it should even become more profi table. However, with the help of its customers, its production processes should improve substantially, both in quality and effi ciency, resulting in cost reductions that are shared between the partners. Toyota is known for helping their suppliers, and even their second- and third-tier suppliers, in this kind of fashion.

In the not-too-distant past, when just-in-time ( JIT) production was still novel, customers were using sole-sourcing as a way to put pressure on their suppliers, forc- ing the supplier to stock inventories of items for immediate delivery rather than holding the stock themselves. Singing the praises of JIT—and insisting that the sup- plier implement JIT so that its deliveries could be made in smaller, more frequent batches—was often just a ploy to accommodate the customers’ own sloppy sched- ules, because they never knew from week to week what they were going to need the following week. Today, fi rms are moving to Lean/JIT (described in detail in Chapter 5) and bringing their suppliers along with them. In many cases, the cus- tomer, like Toyota, is teaching the supplier how to implement effective Lean/JIT programs in their own organizations.

I N V E N T O R Y M A N A G E M E N T A key aspect of supply chain management is the use of inventory. In this section we look at the use of inventory and the factors that help determine the best levels of inventories to hold. We describe the various functions of inventories, the forms of inventories, specifi c inventory-related costs, and the two fundamental inventory decisions all organizations must make. A supplement to the chapter provides addi- tional details on using the economic order quantity model to determine how much inventory should be ordered.

Although inventory is inanimate, the topic of inventory and inventory control can arouse completely different sentiments in the minds of people in various depart- ments within an organization. The salespeople generally prefer large quantities of inventory to be on hand. This allows them to meet customers’ requests without hav- ing to wait. Customer service is their primary concern. The accounting and fi nancial personnel see inventory in a different light. High inventories do not translate into high customer service in the accountant ’s language; rather, they translate into large amounts of tied-up capital that could otherwise be used to reduce debt or for other, more economically advantageous purposes. From the viewpoint of the operations manager, inventories are a tool that can be used to promote effi cient operation of

c07.indd 265c07.indd 265 25/09/12 10:49 AM25/09/12 10:49 AM

266 C h a p t e r 7 : S u p p l y C h a i n M a n a g e m e n t

the production facilities. Neither high inventories nor low inventories per se are desirable; inventories are simply allowed to fl uctuate so that production can be adjusted to its most effi cient level. And top management ’s concern is with the “bot- tom line”—what advantages the inventories are providing versus their costs.

Functions of Inventories There are many purposes for holding inventory, but, in general, inventories have fi ve basic functions. Be aware that inventories will not generally be identifi ed and segre- gated within the organization by these functions and that not all functions will be represented in all organizations.

1. Transit inventories. Transit inventories exist because materials must be moved from one location to another. (These are also known as pipeline inventories .) A truckload of merchandise from a retailer ’s regional warehouse to one of its retail stores is an example of transit inventory. This inventory results because of the transportation time required.

2. Buffer inventories. Another purpose of inventories is to protect against the uncertainties of supply and demand. Buffer inventories—or, as they are sometimes called, safety stocks —serve to cushion the effect of unpredictable events. The amount of inventory over and above the average demand require- ment is considered to be buffer stock held to meet any demand in excess of the average. The higher the level of inventory, the better the customer service—that is, the fewer the stockouts and backorders . A stockout exists when a customer ’s order for an item cannot be fi lled because the inventory of that item has run out. If there is a stockout, the fi rm will usually backorder the item immediately, rather than wait until the next regular ordering period.

3. Anticipation inventories. An anticipated future event such as a price increase, a strike, or a seasonal increase in demand is the reason for holding anticipation inventories. For example, rather than operating with excessive overtime in one period and then allowing the production system to be idle or shut down because of insuffi cient demand in another period, inventories can be allowed to build up before an event to be consumed during or after the event. Manufacturers, wholesalers, and retailers build anticipation inven- tories before occasions such as Christmas and Halloween, when demand for specialized products will be high.

4. Decoupling inventories. It would be a rare production system in which all equipment and personnel operated at exactly the same rate. Yet if you were to take an inspection tour through a production facility, you would notice that most of the equipment and people were producing. Products move smoothly even though one machine can process parts fi ve times as fast as the one before or after it. An inventory of parts between machines, or fl uid in a vat, known as decoupling inventory, acts to disengage the produc- tion system. That is, inventories act as shock absorbers, or cushions, increas- ing and decreasing in size as parts are added to and used up from the stock.

Even if a preceding machine were to break down, the following machines could still produce (at least for a while), since an in-process inventory of parts would be waiting for production. The more inventories management

c07.indd 266c07.indd 266 25/09/12 10:49 AM25/09/12 10:49 AM

267I n v e n t o r y M a n a g e m e n t

carries between stages in the manufacturing and distribution system, the less coordination is needed to keep the system running smoothly. Clearly, there is an optimum balance between inventory level and coordination in the opera- tions system. Without decoupling inventories, each operation in the plant would have to produce at an identical rate (a paced line) to keep the produc- tion fl owing smoothly, and when one operation broke down, the entire plant would come to a standstill.

5. Cycle inventories. Cycle inventories—or, as they are sometimes called, lot- size inventories—exist for a different reason from the others just discussed. Each of the previous types of inventories serves one of the major purposes for holding inventory. Cycle inventories, on the other hand, result from man- agement ’s attempt to minimize the total cost of carrying and ordering inven- tory. If the annual demand for a particular part is 12,000 units, management could decide to place one order for 12,000 units and maintain a rather large inventory throughout the year or place 12 orders of 1000 each and maintain a lower level of inventory. But the costs associated with ordering and receiv- ing would increase. Cycle inventories are the inventories that result from ordering in batches, or “lots,” rather than as needed.

Forms of Inventories

Inventories are usually classifi ed into four forms, some of which correspond directly with the previous inventory functions but some of which do not.

1. Raw materials. Raw materials are objects, commodities, elements, and items that are received (usually purchased) from outside the organization to be used directly in the production of the fi nal output. When we think of raw materials, we think of such things as sheet metal, fl our, paint, structural steel, chemicals, and other basic materials. But nuts and bolts, hydraulic cylinders, pizza crusts, syringes, engines, frames, integrated circuits, and other assem- blies purchased from outside the organization would also be considered part of the raw materials inventory.

2. Maintenance, repair, and operating supplies. Maintenance, repair, and operating (MRO) supplies are items used to support and maintain the opera- tion, including spares, supplies, and stores. Spares are sometimes produced by the organization itself rather than purchased. These are usually machine parts or supplies that are crucial to production. The term supplies is often used synonymously with inventories . The general convention, and the one that we will adopt in this book, is that supplies are stocks of items used (consumed) in the production of goods or services but are not directly a part of the fi nished product. Examples are copier paper, staples, pencils, and packing material. Stores commonly include both supplies and raw materials that are kept in stock or on shelves in a special location.

3. Work-in-process. Work-in-process (WIP) inventory consists of all the mate- rials, parts, and assemblies that are being worked on or are waiting to be processed within the operations system. Decoupling inventories are an example of work-in-process. That is, they are all the items that have left the raw materials inventory but have not yet been converted or assembled into a fi nal product.

c07.indd 267c07.indd 267 25/09/12 10:49 AM25/09/12 10:49 AM

268 C h a p t e r 7 : S u p p l y C h a i n M a n a g e m e n t

4. Finished goods. The fi nished goods inventory is the stock of completed prod- ucts. Goods, once completed, are transferred out of work-in-process inventory and into the fi nished goods inventory. From here they can be sent to distribu- tion centers, sold to wholesalers, or sold directly to retailers or fi nal customers.

As you can see from this discussion, the inventory system and the operations sys- tem within an organization are strongly interrelated. Inventories affect customer service, utilization of facilities and equipment, capacity, and effi ciency of labor. Therefore, the plans concerning the acquisition and storage of materials, or “inven- tories,” are vital to the production system.

The ultimate objective of any inventory system is to make decisions regarding the level of inventory that will result in a good balance between the purposes for holding inventories and the costs associated with them. Typically, we hear inventory manage- ment practitioners and researchers speaking of total cost minimization as the objec- tive of an inventory system. If we were able to place dollar costs on interruptions in the smooth fl ow of goods through the operations system, on not meeting customers’ demands, or on failures to provide the other purposes for which inventories exist, then minimization of total costs would be a reasonable objective. But since we are unable to assign costs to many of these subjective factors, we must be satisfi ed with obtaining a good balance between the costs and the functions of inventories.

Inventory-Related Costs There are essentially fi ve broad categories of costs associated with inventory sys- tems: ordering or setup costs, inventory carrying or holding costs, stockout costs, opportunity costs, and cost of goods. This section looks at these costs in turn.

Ordering or Setup Costs

Ordering costs are costs associated with outside procurement of material, and setup costs are costs associated with internal procurement (i.e., internal manufacture) of parts or material. Ordering costs include writing the order, processing the order through the purchasing system, postage, processing invoices, processing accounts payable, and the work of the receiving department, such as handling, testing, inspec- tion, and transporting. Setup costs also include writing orders and processing for the internal production system, setup labor, machine downtime due to a new setup (e.g., cost of an idle, nonproducing machine), parts damaged during setup (e.g., actual parts are often used for tests during setup), and costs associated with employ- ees’ learning curve (e.g., the cost of early production spoilage and low productivity immediately after a new production run is started).

Inventory Carrying or Holding Costs

Inventory carrying or holding costs have the following major components:

• Capital costs • Storage costs • Risk costs

c07.indd 268c07.indd 268 25/09/12 10:49 AM25/09/12 10:49 AM

269I n v e n t o r y M a n a g e m e n t

Capital costs include interest on money invested in inventory and in the land, buildings, and equipment necessary to hold and maintain the inventory, an item of special interest to both fi nancial and top management. These rates often exceed 20 percent of the cost of the goods. If these investments were not required, the organization could invest the capital in an alternative that would earn some return on investment.

Storage costs include rent, taxes, and insurance on buildings; depreciation of buildings; maintenance and repairs; heat, power, and light; salaries of security per- sonnel; taxes on the inventory; labor costs for handling inventory; clerical costs for keeping records; taxes and insurance on equipment; depreciation of equipment; fuel and energy for equipment; and repairs and maintenance. Some of these costs are variable, some fi xed, and some “semifi xed.”

Risk costs include the costs of obsolete inventory, insurance on inventory, physical deterioration of the inventory, and losses from pilferage.

Even though some of these costs are relatively small, the total costs of carrying items in inventory can be quite large. Studies have found that for a typical manufac- turing fi rm, the cost is frequently as large as 35 percent of the cost of the inventoried items. A large portion of this is the cost of the invested capital.

Stockout Costs

If inventory is unavailable when customers request it, a situation that marketing detests, or when it is needed for production, a stockout occurs. Several costs are associated with each type of stockout. A stockout of an item demanded by a cus- tomer or client can result in lost sales or demand, lost goodwill (which is very dif- fi cult to estimate), and costs associated with processing backorders (such as extra paperwork, expediting, special handling, and higher shipping costs). A stockout of an item needed for production results in costs for rescheduling production, costs of downtime and delays caused by the shortage, the cost of “rush” shipping of needed parts, and possibly the cost of substituting a more expensive part or material.

Opportunity Costs

Often capacity and inventory costs can be traded off for one another. For example, capacity costs can be incurred because a change in productive capacity is neces- sary or because there is a temporary shortage of or excess in capacity. Why would capacity be too great or too small? If, for example, a company tried to meet sea- sonal demand (or any fl uctuations in demand) by changing the level of production rather than by allowing the level of inventory to rise or fall, capacity would have to be increased during high-demand periods and lie idle during low-demand periods. Also, capacity problems are often due to scheduling confl icts. These commonly arise when multiple products have to be produced on the same set of facilities.

Opportunity costs include the overtime required to increase capacity; the human resource management costs of hiring, training, and terminating employees; the cost of using less skilled workers during peak periods; and the cost of idle time if capac- ity is not reduced during periods when demand decreases.

c07.indd 269c07.indd 269 25/09/12 10:49 AM25/09/12 10:49 AM

270 C h a p t e r 7 : S u p p l y C h a i n M a n a g e m e n t

Cost of Goods

Last, the goods themselves must be paid for. Although they must be acquired sooner or later anyway, when they are acquired can infl uence their cost considerably, as through quantity discounts.

Decisions in Inventory Management The objective of an inventory management system is to make decisions regarding the appropriate level of inventory and changes in the level of inventory. To maintain the appropriate level of inventory, decision rules are needed to answer two basic questions:

1. When should an order be placed to replenish the inventory? 2. How much should be ordered?

The decision rules guide the inventory manager or computerized materials man- agement system in evaluating the current state of the inventory and deciding if some action, such as replenishment, is required. Various types of inventory management systems incorporate different rules to decide “when” and “how much.” Some depend on time and others on the level of inventory, but the essential decisions are the same. Even when complexities, such as uncertainty in demand and delivery times, are introduced, deciding “how many” and “when to order” still remains the basis of sound inventory management (refer to Supplement B in the back of this chapter).

R O L E O F I N F O R M A T I O N T E C H N O L O G Y In the not-too-distant past, the primary means of communication between members of a supply chain was paper. Unfortunately, communicating via paper-based transac- tions is slow, often unreliable, and prone to errors. For example, Campbell Soup Company estimated that 60 percent of the fax and phone orders it received con- tained errors (Verity 1996). As a result, salespeople often spent 40 percent of their time correcting these errors rather than making additional sales. To correct this prob- lem, Campbell invested $30 million in the electronic redesign of its order-processing system. The company expected the new system to increase the percentage of paper- less orders it received to 80 percent. Managers estimated the system would reduce costs by $18 million annually while at the same time reducing delivery times.

Some problems with paper-based systems have been the time and money that are wasted rekeying the same information into different computer systems. And, of course, the more times the same information is entered, the more opportunities there are for making mistakes. Some analysts estimate that the use of information technology has reduced the cost of processing a purchase order from $150 to $25.

As Campbell Soup illustrates, electronic information technology is a key element and the primary enabler of effective supply chain management. In today ’s highly com- petitive environment, the effective use of information technology helps organizations

c07.indd 270c07.indd 270 25/09/12 10:49 AM25/09/12 10:49 AM

271R o l e o f I n f o r m a t i o n T e c h n o l o g y

reduce cycle times, adopt more responsive cross-functional organizational structures, capture more timely information, and reduce errors and costs. The ultimate goal of such information systems is to make available to all participants in the supply chain all the information needed at the time it is needed. Such information includes the status of orders, product availabilities, delivery schedules, and other such supply chain data.

Everyone knows that computers are everywhere these days and embedded in all kinds of products that one would not have expected. But why is this, and why now? Professor Richard Chase of the University of Southern California believes that the answer lies in two esoteric laws—one about physical goods and the other about abstract information. The fi rst is the better known of the two: Moore ’s law, which states that computing power doubles every 18 to 24 months. The unstated surprise about Moore ’s law is that this doubling of power comes at the same or lower cost as before the doubling. Clearly, with enough money our big computer companies could double computing power every 18 (or 12 or 6) months, but the size of the computers would grow enormously, as would their costs. Yet this law implies that the cost and size do not increase. As a result, more and more computing power is becoming available for less and less money; hence, it is becoming omnipresent, appearing everywhere we go and in everything we buy.

The second law is less familiar to the public but derives from the fact that infor- mation assets, like knowledge, tend to grow with use rather than dwindle, as with physical assets. This second law is called Metcalfe ’s law, which says that the value of a network is proportional to the square of the number of elements (or users) con- nected to the network. This is why Amazon and Microsoft and eBay have been so successful—with more people in a network, the value of the network to the user is enhanced, so more people join this network. And competing networks with fewer users are of less value and hence fade away.

As a result of these two laws, the growth of computers, which support networks, and networks, which support people ’s needs (business transactions, communica- tion, blogging, etc.), has exploded. This phenomenon has been particularly preva- lent in business, where it has contributed to both increased value (and thus revenues) and reduced costs, thereby having a double impact on increased profi ts. Next we will look at some particular types of information technology that are commonly used in business, especially to support supply chain management.

Electronic business ( e-business ) is the use of electronic information technology to help various groups of businesspeople communicate and conduct business transac- tions. Its three primary advantages are enhanced productivity and reduced costs, speed, and the creation of new value opportunities. The primary enablers of e- business have been electronic data interchange (EDI), e-commerce, intranets and extranets, collaborative software applications, customer relationship management (CRM) systems, and enterprise resource planning (ERP) systems (discussed in more detail at the end of this section).

One early approach to e-business was electronic data interchange (EDI)—the ability of one organization ’s computer to communicate with another ’s computer. With EDI, business documents such as purchase orders and invoices are transferred between the computers of different organizations in a standard format. The benefi ts of EDI include faster access to information, reduced paperwork, less redundancy, improved customer service, and better order tracking. And, of course, the costs of paper transactions, both physical (paper handling, mailing, printing, etc.) and qual- ity (errors), are reduced substantially.

c07.indd 271c07.indd 271 25/09/12 10:49 AM25/09/12 10:49 AM

272 C h a p t e r 7 : S u p p l y C h a i n M a n a g e m e n t

Electronic commerce ( e-commerce ) is the term used to describe the execution of business transactions in a paperless environment, primarily through the Internet. In essence, this consists of two major parts: an Internet portal, such as a Web site (dis- cussed a bit later) for information and communication, and the fulfi llment transac- tions to deliver the product or service. These, in turn, include the use of bar coding and scanning, radio frequency identifi cation (RFID), databases, point-of-sale (POS) terminals, e-mail, electronic funds transfer, the Internet, and Web sites and hubs. Bar coding and scanning technologies permit the rapid collection and dissemination of information throughout the supply chain. For example, Wal-Mart is well known for making available the point-of-sale ( POS ) information it collects at its check-out ter- minals to its supply chain partners. Federal Express uses the same technology to provide its customers with up-to-date tracking information on their packages. Firms also use bar coding and scanning in their production facilities to correctly produce and process items and to keep track of different stock-keeping units (SKUs).

Most people are familiar with bar codes and scanning, but the new RFID tags/ transponders are much less familiar. As mentioned in the example at the beginning of the chapter, Wal-Mart has also become known for initiating the retail use of this new technology through their suppliers in order to better track their products. RFID tags provide much more information than bar codes and come in two versions, pas- sive and active. The passive tags are as small as the head of a pin and as thin as a sheet of paper; readers can interrogate them as they pass by, typically within about 20 feet. Active tags broadcast their information and can be read from much greater distances. However, as noted earlier, the main diffi culty with employing these tech- nologies has been the cost of the tags.

Arguably the most signifi cant information technology development for supply chain management is the Internet , and more specifi cally, its graphical component known as the World Wide Web (Web). Without a doubt, the Web offers enormous opportunities for members of a supply chain to share information. Companies such as IBM, General Electric, Dun & Bradstreet, and Microsoft are rapidly developing products and services that will help make the Web the global infrastructure for elec- tronic commerce (Verity 1996).

For example, as noted earlier in the purchasing discussion, the Web will allow vari- ous forms of purchasing fulfi llment to take place, from placing electronic catalogs on a Web site to holding joint purchasing bazaars, exchanges, and auction marketplaces involving massive amounts of materials. Bazaars and reverse auctions (one buyer, mul- tiple sellers) were discussed earlier, but exchanges are for information transfer (often hosted by third parties, such as mySAP.com), and auction marketplaces (one seller, multiple buyers) are primarily for selling commodities or near-commodities at low prices. Of course, the costs of initiating and executing these forms of purchasing will be almost trivial compared to their paper-based predecessors. For example, updating an electronic catalog can be done instantaneously, rather than waiting until next year ’s printing. In addition, password-protected customized catalogs refl ecting negotiated prices can also be placed on a fi rm ’s Web site for use by individual customers.

Intranets are Web-based networks that allow all employees of a fi rm to intercom- municate. They are usually fi rewall-protected and use existing Internet technologies to create portals for company-specifi c information and communication, such as news- letters, training, human resource information and forms, product information, and so on. Extranets are private networks to allow the organization to securely interact with external parties. They use Internet protocols and public telecommunication systems

c07.indd 272c07.indd 272 25/09/12 10:49 AM25/09/12 10:49 AM

273R o l e o f I n f o r m a t i o n T e c h n o l o g y

to work with external vendors, suppliers, dealers, customers, and so on. Clearly, the extranet would be a major element of a fi rm ’s supply chain information system.

Collaborative software facilitates the work of groups or teams in the organization. Its purpose is communication, collaboration, and coordination (of schedules, work- fl ow, etc.). Most collaborative systems these days are Web-based. Microsoft ’s Netmeeting and Cisco ’s Webex are well-known commercial systems.

Customer relationship management ( CRM ) systems are designed to collect and interpret customer-based data (Ragins and Greco 2003). This could be from internal sources such as marketing, sales, or customer support services or from external sources like market research or the customer. The aim is to develop a proc- ess for improving the fi rm ’s response to its customers’ needs, especially the most profi table customers. CRM systems thus provide comprehensive customer data so the fi rm can provide better customer service and design and offer the most appropri- ate products and services for its customers.

Enterprise Resource Planning (ERP) Enterprise resource planning (ERP) systems greatly facilitate communication throughout the supply chain and over the Internet. The ERP system embodies much more than just the supply chain, however; it also includes all the electronic informa- tion concerning the various parts of the fi rm. These massive systems can not only reduce costs and allow instant access to the entire fi rm ’s database but can also help increase revenues, by up to 25 percent in some cases (Mabert et al. 2001, p. 50).

As the name suggests, the objective of these systems is to provide seamless, real- time information to all employees who need it, throughout the entire organization (or enterprise), and to those outside the organization. Figure 7.3 provides a broad overview of SAP ’s MySAP ERP system. MySAP, announced in 2003, represents the latest evolution of SAP ’s ERP system. SAP introduced its R/2 system in 1979, which was an ERP system that ran on mainframe computers, and its R/3 system for client server computing environments in 1992. MySAP takes the evolution one step further

ERP Modules

SuppliersCustomers

Employees

Central Database

Product Development

and Manufacturing

Sales & Service

Corporate ServicesFinancialsAnalytics

Human Capital

Management

Procurement and

Logistics

Figure 7.3 SAP ’s MySAP ERP system.

c07.indd 273c07.indd 273 25/09/12 10:49 AM25/09/12 10:49 AM

274 C h a p t e r 7 : S u p p l y C h a i n M a n a g e m e n t

and is based on service-oriented architecture (SOA) whereby organizations will be able to access the SAP software via the Internet and thereby have access to the full functionality of the software without having to actually install and deploy the soft- ware throughout the enterprise. With the introduction of MySAP, SAP has announced that they will no longer continue to develop R/3.

As shown in Figure 7.3 , an ERP system consists of a number of modules that provide the functionality to support a variety of organizational processes. These modules all access data from the central database, and changes made via these modules update the central database. Using ERP, each area interacts with a centralized database and servers, so suppliers can check on the latest demands and customers can determine the status of their order or available capacity for new orders. ERP can also handle international com- plications such as differences in taxes, currency, accounting rules, and language. Figure 7.4 provides additional details about the functionality offered by each MySAP module.

With the ERP approach, information is entered once at the source and made avail- able to all stakeholders needing it. Clearly, this approach eliminates the incompati- bility created when different functional departments use different systems, and it also eliminates the need for people in different parts of the organization to reenter the same information over and over again into separate computer systems. Although ERP ties all these areas together, the actual implementation of an ERP system in an organization may include only portions of these modules on an as-needed basis.

Davenport (1998) provides an example that illustrates the opportunity to automate tasks in a business process with an ERP system. In the example, a Paris-based sales

MySAP

Strategic Enterprise

Mgt.

Financials

Financial Supply Chain

Mgt.

Human Capital

Management

Talent Mgt.

Procurement &

Logistics

Procurement

Sales and

Service

Sales Order Mgt.

Prod. Dev. & Manf.

Production Planning

Corporate Services

Real Estate Mgt.

Financial Analytics

Financial Accounting

Workforce Process

Mgt.

Supplier Collaboration

Aftermarket Sales & Service

Manufacturing Execution

Project Portfolio

Mgt.

Operations Analytics

Mgt. Accounting

Workforce Deployment

Inventory Warehouse

Mgt.

Professional Services Delivery

Enterprise Asset Mgt.

Travel Mgt.

Workforce Analytics

Corporate Governance

Inbound & Outbound Logistics

Global Trade

Services

Product Development

Environment, Health, and

Safety

Transporta- tion Mgt.

Incentive & Commission

Mgt.

Life-Cycle Data Mgt.

Quality Mgt.

Analytics

Figure 7.4 Detailed view of MySAP ’s modules.

c07.indd 274c07.indd 274 25/09/12 10:49 AM25/09/12 10:49 AM

275R o l e o f I n f o r m a t i o n T e c h n o l o g y

rep of a U.S. manufacturer prepares a quote for a customer in Paris. After the rep enters the customer information into a notebook computer, the ERP system creates the sales contract in French. Included in the sales contract are important details of the order, such as the product ’s confi guration, quantity ordered, price, delivery date, and payment terms. When the customer agrees to the terms of the quote, the sales rep submits the order electronically with a single keystroke. The system then automati- cally checks the customer ’s credit and accepts the order if it is within the customer ’s credit limit. Upon accepting the order, the ERP system then schedules the shipment of the completed order based on the agreed-upon delivery date and then, based on the delivery date and appropriate lead times, reserves the required raw materials. The system also determines if the required materials will be available and, if not, auto- matically generates the orders for the needed materials from suppliers. Next, the ERP system schedules the actual assembly of the order in one of the organization ’s Asian facilities. In addition, sales and production forecasts are updated, the commission due the rep is calculated and credited to his or her account (in French francs), and the profi tability of the order (in U.S. dollars) is computed. Finally, the business units and corporate fi nancial statements such as balance sheets, accounts-payable, accounts- receivable, cash fl ows, and so on are immediately updated.

As this example illustrates, the integration offered by ERP systems provides organi- zations with the potential to achieve dramatic improvements in the execution of their business processes. Owens Corning achieved this integration by replacing 211 legacy systems with one ERP system. Much of the benefi t associated with this integration stems from having real-time access to operating and fi nancial data. For example, after implementing an ERP system, Autodesk reduced the time it took to deliver an order from an average of two weeks to shipping 98 percent of its orders within 24 hours. Before implementation of an ERP system, it took IBM ’s Storage Systems division fi ve days to reprice all of its products. After implementing an ERP system, it was able to accomplish the same task in fi ve minutes. IBM also reduced the time required to ship replacement parts from 22 days to 3 days and reduced the time to perform credit checks from 20 minutes to 3 seconds! Fujitsu Microelectronics achieved similar ben- efi ts, reducing its order fulfi llment time from 18 days to less than 2 days and reducing the time required to close its fi nancial books from 8 days to 4 days.

Although ERP systems were originally developed for and adopted by manufactur- ing fi rms, employees working in service organizations have the same need for seam- less, real-time information. To meet the needs of service organizations, numerous ERP systems specifi c to the needs of service organizations have been developed. For example, Carroll Hospital Center in Westminster, Maryland, adopted an ERP system to help streamline its operations and reduce costs (Monegain 2009). Carroll Hospital is using the ERP to facilitate a variety of functions from payroll to budgeting and planning. According to the CIO of Carroll Hospital, the ERP system has impacted all aspects of the hospital from how patients receive their care to how employees are paid. Employees at Carroll Hospital appreciate the ERP system ’s ability to provide them with the information they need and eliminate paperwork. Overall, Carroll Hospital has found that the ERP system provides everyone with more timely and accurate information, which in turn has facilitated the work of all employees.

In a similar fashion to the health care industry, a number of specialized ERP systems have been developed for higher education. These ERP systems contain a number of specialized modules that universities can select from for maintaining and developing relationships with alumni, student services such as fi nancial aid and course registration,

c07.indd 275c07.indd 275 25/09/12 10:49 AM25/09/12 10:49 AM

276 C h a p t e r 7 : S u p p l y C h a i n M a n a g e m e n t

fi nance and human capital management, and academic applications for tasks such as monitoring student progress and retention.

S U C C E S S F U L S U P P L Y C H A I N M A N A G E M E N T The basic requirements for successful supply chain management are trustworthy partners, good communication, appropriate performance measures, and competent managers with vision. Innovation to suit the particular situation of the individual organization is particularly desirable. Here are some examples of visionary SCM innovations that have been developed:

• Dell ’s “direct model” (Magretta 1998). • Wal-Mart ’s “ cross-docking ” technique of off-loading goods from incoming

trucks at a warehouse directly into outbound distribution trucks instead of placing them into inventory.

• The relatively common approach used by Dell and many others of “ delayed differentiation ,” where fi nal modules are either inventoried for last-minute assembly to customer order, or differentiating features are added to the fi nal product upon receipt of the customer ’s order.

• Sport Obermeyer ’s and Hewlett-Packard ’s “ postponement ” approach to delayed differentiation, where variety and customization are delayed until as late in the production process as possible, sometimes even arranging with the carrier to perform the fi nal customization (called channel assembly ). In Sport Obermeyer ’s (Fisher et al. 1994) version, those product lines where demand is better known are produced fi rst, while customer demand volume information is being collected on less easily forecast lines whose production has thus been postponed. Similarly, Hewlett-Packard ships generic printers to regional centers around the globe, where local workers add country-specifi c power supplies, power cords, and local language instructions. Another variant of postponement was mentioned in the Dell example cited earlier, where drop shipping arrangements are made with the carrier to deliver third- party- supplied elements of the product (e.g., monitors) to the customer at the same time that the main product is being delivered.

Closed-Loop Supply Chains and Reverse Logistics Guide and Van Wassenhove (2009, p. 10) defi ne closed-loop supply chain management as “the design, control, and operation of a system to maximize value creation over the entire life cycle of a product with dynamic recovery of value from different types and volumes of returns over time.” An important aspect of closed-loop supply chain man- agement is recovering value from returned products. The potential for recovering value from returns is enormous, as it is estimated that commercial returns exceed $100 billion annually (Stock et al. 2002). Large retailers like Home Depot can expect to have 10 percent or more of their sales returned, while Hewlett-Packard estimates that it incurs costs equivalent to 2 percent of its outbound sales in returned merchandise.

Product returns are categorized as commercial returns, end-of-use returns, end- of-life returns, and repair and warranty returns. Commercial returns are typically returns to the reseller and occur within 90 days of purchase. For example, many cell

c07.indd 276c07.indd 276 25/09/12 10:49 AM25/09/12 10:49 AM

277E x p a n d Y o u r U n d e r s t a n d i n g

1. Why is supply chain management such a topic of interest lately, especially multifacility distribution? Why wasn ’t it previously?

2. Will all production eventually reside in China? What exceptions might exist?

3. What appears to be the primary “secret” of success- ful supply chain management?

4. Given that the current conceptualization of the sup- ply chain includes JIT and lean manufacturing, what other elements of SCM need to be changed to move toward the idea of a demand chain?

5. In what way can contract manufacturers consider themselves service providers? Hasn ’t Nike been doing this for years? What ’s the difference?

6. To date it appears that purchasing has been the primary benefi ciary of supply chain management. Why do you think this is so? What do you expect will happen in the future?

7. The bullwhip effect is often blamed for the boom- and-bust cycles in our national economy. Which of the remedies for eliminating this effect in a supply chain might also benefi t the national economy?

8. How does postponement differ from assemble- to-order?

9. Why does the Internet rather than the older EDI now seem to be the information foundation for SCM?

10. Why do you think Wal-Mart has been the pioneer of RFID technology?

11. E-commerce has been supplanted by e-business. What is the basic difference between the two?

12. Contrast SCM systems with ERP systems. Which do you suspect are larger and more costly?

13. What additional information might a retailer such as Wal-Mart be interested in putting on an RFID tag, beyond the basic product identifi cation that a bar code communicates?

phone companies allow customers to return their cell phones for any reason within 30 days of purchase. End-of-use returns occur when a product is returned so that its functionality can be upgraded. For example, in the United States it is estimated that 80 percent of cell phone users upgrade their perfectly usable cell phones annually. End-of-life returns occur when the product still functions but is technologically obsolete. Finally, between commercial returns and end-of-life returns, customers return products to be repaired.

The type of product return has important implications for how the return is han- dled. For example, commercial returns have usually been only lightly used. Therefore, they typically require minor processing, such as cleaning and perhaps some minor repairs. End-of-use returns have been used more heavily, and there is likely to be more variability in the quality of these returns. Given this, these returns will typically require more extensive processing. The focus in end-of-life returns is on parts recov- ery and recycling, since these products are technologically obsolete. In summary, then, commercial returns are repaired, end-of-use returns are remanufactured, and end-of-life returns are recycled.

In addition to providing signifi cant environmental benefi ts, the goal of operating a closed-loop supply chain is to generate more value through the recovery activities than the cost of performing these activities. The steps involved in operating a closed-loop supply chain include acquiring the right quantities of the used product with the right quality and at the right time; using reverse logistics, or moving the product back upstream from the customer to the repair/remanufacturing operations; sorting, testing, and grading the returned products to determine their disposition; repairing/remanufac- turing the returned products; and, fi nally, remarketing the refurbished products. Some products, such as consumer electronics and computers, have short life cycles and therefore lose a signifi cant portion of their value per week. In these cases, a slow reverse supply chain can erode much if not all of the potential value that can be recovered.

E X P A N D Y O U R U N D E R S T A N D I N G

c07.indd 277c07.indd 277 25/09/12 10:49 AM25/09/12 10:49 AM

278 C h a p t e r 7 : S u p p l y C h a i n M a n a g e m e n t

14. Do any of the fi ve functions and four forms of inventories exist in service fi rms? If so, which ones, and why? If not, how are the functions served?

15. Contrast the functions and forms of inventories. Does every form exist for each function and vice versa, or are some more common?

16. In many of today ’s fi rms, the customer ’s computer is tied to the supplier ’s computer so that purchase

orders go directly into the supplier ’s planning system. What are the implications of this close relationship?

17. Discuss the pros and cons of relying on outside expertise in the selection and implementation of an ERP system.

A P P L Y Y O U R U N D E R S T A N D I N G

Peak Nut r i t i on , Inc .

Peak Nutrition, Inc. (PNI) offers a line of premium sports recovery drinks. Its drinks are made from all-natural fruit juices that are supplemented with protein, creatine, vitamins, and miner- als. Each fl avor is offered in both a 12-ounce and 20-ounce bottle. Eighty percent of PNI ’s sales are to two national health food chains, and the remaining 20 percent are to independent health food stores and online retailers.

PNI has a single production and bottling line, which has suffi cient capacity to meet its cur- rent demand. Setting up the production line to produce a particular fl avor requires an entire eight-hour shift. Most of the setup time is related to fl ushing out the equipment in order to not contaminate the new fl avor to be produced with the fl avor that was last produced. Given the long setup times, the production and bottling lines are dedicated to producing a single fl avor for an entire week. The typical production schedule involves setting up the line on Monday, producing 12-ounce bottles on Tuesday and Wednesday, and producing 20-ounce bottles on Thursday and Friday. The plastic bottles and labels are purchased from outside suppliers. There is a one-week lead time for both the bottles and labels. PNI maintains a four-week in- ventory of plastic bottles and orders labels three weeks before they are needed. Changing over the bottling line from 12-ounce to 20-ounce bottles requires about an hour and a half, which also includes changing the labels.

Since each fl avor is produced every six weeks, PNI historically produced an eight-week supply for each fl avor to provide a safety net in the event demand exceeded its forecasts. Despite having an extra two weeks of inventory, PNI often experienced stockouts. Given the problem with stockouts, PNI recently acquired additional warehouse space and now plans to produce ten weeks of demand during each production run. It is expected that producing a ten-week supply may result in the need for a small amount of overtime in some weeks.

PNI has limited communication with its customers, primarily consisting of the purchase or- ders it receives from its customers and the invoices and products it sends to them. PNI ’s goal is to meet all orders from its inventory. In this way, it is able to provide its customers with a one- week lead time. When the inventory level is insuffi cient to meet the quantity ordered, lead times increase to an average of two to three weeks, depending on how soon the product is next scheduled for production. Once last year PNI was stocked out of a fl avor for almost four weeks.

At the end of each quarter, PNI offers its customers discounts for orders above certain or- der quantity thresholds. The purpose of the discounts is to provide retailers with an incentive to put the sports drinks on sale and help boost quarterly sales. As a result of these incentives, PNI ’s sales tend to be fi ve to ten times higher in the last two weeks of the quarter compared to other times. In anticipation of the increase in sales, PNI builds up its inventory. However, while on average it has plenty of inventory across all fl avors, it often experiences mismatches in its available supply and demand for specifi c fl avors. In other words, it often fi nds that it has too much inventory of some fl avors and too little of other fl avors.

c07.indd 278c07.indd 278 25/09/12 10:49 AM25/09/12 10:49 AM

279A p p l y Y o u r U n d e r s t a n d i n g

Questions

1. What concerns do you have about PNI ’s supply chain management practices? 2. What would you recommend PNI do to address your concerns? 3. Do you have any concerns about the way PNI determines its level of safety stock? 4. Should PNI focus on enhancing the effi ciency or responsiveness of its supply chain? Why?

Dar t ’s Par t s , Inc .

Z. “Dart” Mitchell leaned forward in his chair to read the e-mail that had just arrived from one of his major customers, Avery Machine Corp. It read as follows:

To all our preferred suppliers—

Due to our commitments to our primary customer, Globus Enterprises, we will in the future be doing all of our supply chain business by way of the Internet, e-mail, and EDI. This includes order preparation, bidding, forecasting, production scheduling, delivery monitoring, cost control, accounts payable and receivable, credit and fi nancing, market and advertising planning, human resource acquisition, engineering specifi cations, and so on. To maintain compatibility with our systems, you will have to invest in a specifi c set of EDI hardware and software, available from GoingBust.com on the Web. Although the hardware and software are expensive, we anticipate that the cost savings and increased business this will provide over the coming years can more than offset the additional cost. Please let us know if we can continue to count on you as one of our preferred suppliers as we move our supply chain into the information age.

J. R. Avery, Chairman Avery Machine Corp.

Dart ’s Parts had been founded in 1974, when the country was coming out of the 1973–1974 recession and the need for machine part fabricators was great. Over the years, Dart had built up the business to where it now had a solid base of major customers and a comfortable back- log of orders. Dart had increased the capacity of the plant substantially over the years, moving from a small rented facility to its own 200,000-square-foot plant, with a separate 50,000-square- foot warehouse located adjacent to the main plant. Although not a “fi rst adopter” when it came to new technology, Dart ’s embraced proven advanced technologies both on the plant fl oor, with innovations such as robots and numerically controlled machine tools, and in the offi ce, with computers, digital copiers, and other such offi ce equipment.

Dart Mitchell had been reading industry magazines about some of these new technologies and had to admit they sounded promising. However, he had read about some horror stories, too, when the much-advertised features turned into a nightmare. In one case, a customer had forced its suppliers to obtain production schedules off its Web site. Initially responding to high growth in a new product line, the fi rm had put its component needs on its Web site, but when a major order was canceled, it was late in changing the Web production schedule. As a result, the suppliers were stuck with hundreds of unneeded components and the company wouldn ’t reimburse them. In another case, a manufacturer had made a bid for electronic parts on a Web auction and won. However, when it received the parts, they were too large to fi t in the standard-sized enclosure it was using and they all had to be scrapped.

Dart believed that this new technology was indeed the future of the industry, but he was concerned about getting in too early and being stuck with the wrong equipment. The new supply chain technology would undoubtedly open avenues to increased business, but it would also result in a number of costs. Of course, it would also save the company ’s reputation with Avery, a major customer. However, obtaining the EDI system would be a major fi nancial investment for the fi rm, particularly if Avery later dropped this approach and went to an all- Internet ERP system like some customers had been talking about doing. At this point, Dart wasn ’t sure what to do.

c07.indd 279c07.indd 279 25/09/12 10:49 AM25/09/12 10:49 AM

280 C h a p t e r 7 : S u p p l y C h a i n M a n a g e m e n t

Ballou , R. H. Business Logistics: Supply Chain Manage- ment , 5th ed. Upper Saddle River, NJ : Prentice-Hall , 2003 .

Bender , P. S. “ Debunking 5 Supply Chain Myths .” Supply Chain Management Review ( March 2000 ): 52 – 58 .

Bjurstrom , T. “ How Outsourcing Supply Chain Services to 3PLs Can Affect Your Company ’s Bottom Line .” Con- tract Management ( February 2008 ): 52 – 55 .

Blackwell , R. D. , and K. Blackwell . “ The Century of the Consumer: Converting Supply Chains into Demand Chains .” Supply Chain Management Review ( Fall 1999 ): 22 – 32 .

Blanchard , D. “ RFID Is Off and Running at Wal-Mart .” Logistics Today , 46 ( February 2005 ): 1 – 2 .

Burt , D. N. , D. W. Dobler , and S. L. Starling . World Class Supply ManagementSM: The Key to Supply Chain Man- agement , 7th ed. New York : McGraw-Hill/Irwin , 2003 .

Bowersox , D. J. , D. J. Closs , and M. B. Cooper . Supply Chain Logistics Management . New York: Irwin/McGraw- Hill, 2002 .

Cachon G. , and C. Terwiesch . Matching Supply with Demand , 2 nd ed. New York : McGraw-Hill/Irwin , 2009 .

Champion , D. “ Mastering the Value Chain .” Harvard Business Review ( June 2001 ): 109 – 115 .

Chopra , S. , and P. Meindl . Supply Chain Management: Strategy, Planning, and Operation . Upper Saddle River, NJ : Prentice-Hall , 2001 .

Coyle , J. J. , E. J. Bardi , and R. A. Novack . Transportation . Cincinnati : South-Western , 2000 .

Cross , G. J. “ How e-Business Is Transforming Supply Chain Management .” Engineering Management Review , 28 (Third Quarter 2000 ): 17 – 19 .

Davenport , T. “ Putting the Enterprise into the Enterprise System .” Harvard Business Review ( July–August 1998 ): 121 – 131 .

Davenport , T. Mission Critical: Realizing the Promise of Enterprise Systems . Boston : Harvard Business School Press , 2000 .

Dignan , L. “Is Dell Hitting the Effi ciency Wall?” MSNBC News.com , July 29, 2002 , pp. 1 – 5 .

Dvorak , P. “ Clarity Is Missing Link in Supply Chain .” Wall Street Journal (May 18, 2009 ) A1 , A14 .

Ellram , L. M. , and B. Liu . “ The Financial Impact of Sup- ply Management .” Supply Chain Management Review (November–December 2002 ): 30 – 37 .

Engardio , P. “ Why the Supply Chain Broke Down .” Business Week (March 19, 2001 ): 41 .

Evans , P. , and T. S. Wurster . “ Getting Real About Virtual Commerce .” Harvard Business Review (November– December 1999 ): 84 – 94 .

Ferguson , B. R. , “ Implementing Supply Chain Manage- ment .” Production and Inventory Management Journal , 40 (Second Quarter 2000 ): 64 – 67 .

Fisher , M. L. “ What Is the Right Supply Chain for Your Product? ” Harvard Business Review (March–April 1997 ): 105 – 116 .

Fisher , M. L. , J. H. Hammond , W. R. Obermeyer , and A. Raman . “ Making Supply Meet Demand in an Uncertain World .” Harvard Business Review (May–June 1994 ): 83 – 93 .

Frohlich , M. T. , and R. Westbrook . “ Arcs of Integration: An International Study of Supply Chain Strategies .” Jour- nal of Operations Management , 19 ( 2001 ): 185 – 200 .

Guide Jr., V. D. R. , and L. N. Van Wassenhove . “ The Evo- lution of Closed-Loop Supply Chain Research .” Opera- tions Research , 57 ( 1 ) ( January–February 2009 ): 10 – 18 .

Hammer , M. “ The Supereffi cient Company .” Harvard Business Review ( September 2001 ): 82 – 91 .

Handfi eld , R. B. , and E. L. Nichols Jr. Introduction to Supply Chain Management . Upper Saddle River, NJ : Prentice-Hall , 1999 .

Hutchison , P. D. , M. T. Farris II , and S. B. Anders . “ Cash- to-Cash Analysis and Management .” The CPA Journal ( August 2007 ): 42 – 47 .

Kanakamedala , K. , G. Ramsdell , and V. Srivatsan . “ Get- ting Supply Chain Software Right .” The McKinsey Quar- terly , 1 ( 2003 ): 78 – 85 .

Questions

1. Identify the tradeoffs facing Dart ’s Parts. 2. What are the pros and cons of each alternative? 3. What additional information would be useful to have? 4. What recommendations would you make to Dart Mitchell?

B I B L I O G R A P H Y

c07.indd 280c07.indd 280 25/09/12 10:49 AM25/09/12 10:49 AM

281B i b l i o g r a p h y

Kanet , J. J. , and A. R. Cannon . “ Implementing Supply Chain Management: Lessons Learned at Becton Dickin- son .” Production and Inventory Management Journal , 41 (Second Quarter 2000 ): 33 – 40 .

Kaplan , S. , and M. Sawhney . “ E-Hubs: The New B2B Marketplaces .” Harvard Business Review (May–June 2000 ): 97 – 103 .

Lacy , S. “Inching Toward the RFID Revolution.” Business Week Online (August 31, 2004 ).

Lee , H. L. , C. Billington . “ Managing Supply Chain Inven- tory: Pitfalls and Opportunities .” Sloan Management Re- view , 33 ( Spring 1992 ): 65 – 73 .

Lummus , R. R. , and R. J. Vokurka . “ Strategic Supply Chain Planning .” Production and Inventory Manage- ment Journal , 39 (Third Quarter 1998 ): 49 – 58 .

Lummus , R. R. , and R. J. Vokurka . “ Managing the Demand Chain Through Managing the Information Flow: Capturing ‘Moments of Information’ .” Production and Inventory Management Journal , 40 (First Quarter 1999 ): 16 – 20 .

Mabert , V. A. , A. Soni , and M. A. Venkataramanan . “ Enterprise Resource Planning: Measuring Value .” Production and Inventory Management Journal , 42 (Third/Fourth Quarter 2001 ): 46 – 51 .

Magretta , J. “ The Power of Virtual Integration: An Inter- view with Dell Computer ’s Michael Dell .” Harvard Business Review (March–April 1998 ): 72 – 84 .

Monegain , B. “Maryland Hospital Implements ERP Tech- nology.” Healthcare IT News , www.healthcareitnews .com/news/maryland-hospital-implements-erp-tech nology, October 26, 2009 .

Quinn , F. J. “ What ’s the Buzz? Supply Chain Management; Part 1 .” Logistics Management , 36 (February 1997 ): 43 .

Ragins , E. J. , and A. J. Greco . “ Customer Relationship Management and e-Business: More Than a Software Solution .” Review of Business ( Winter 2003 ): 25 – 30 .

Richardson , H. L. “ The Ins & Outs of VMI .” Logistics Today , 45 ( March 2004 ): 19 – 21 .

Sampson , S. E. , and C. M. Froehle . “ Foundations and Implications of a Proposed Unifi ed Services Theory .” Production and Operations Management , 15 ( 2 ) ( 2006 ): 329 – 343 .

SAP, “MySAP ERP Solution Overview.” www.sap.com/ solutions/business-suite/erp/pdf/BWP_mySAP_ERP_ Overview.pdf, June 26, 2005 .

Satariano , A. , and P. Burrows . “Apple ’s Supply-Chain Se- cret? Hoard Lasers.” Bloomberg Business Week , www. businessweek.com, November 3, 2011 .

Simchi-Levi , D. , P. Kaminsky , and E. Simchi-Levi . Supply Chain Management , 2nd ed. New York : McGraw-Hill , 2003 .

Shapiro , C. , and H. R. Varian . Information Rules: A Stra- tegic Guide to the Network Economy . Boston : Harvard Business School Press , 1999 .

Stock , J. , T. Speh , and H. Shear . “ Many Happy (Product) Returns .” Harvard Business Review , 80 ( 7 ) ( 2002 ): 16 – 17 .

Tully , S. “ Purchasing ’s New Muscle .” Fortune (February 20, 1995 ): 75 – 83 .

Turban , E. , J. Lee , D. King , and H. M. Chung . Electronic Commerce: A Managerial Perspective . Upper Saddle River, NJ : Prentice Hall , 2000 .

Verity , J. W. “ Invoice? What ’s an Invoice? ” Business Week ( June 10, 1996 ): 110 – 112 .

Walker , W. T. , and K. L. Alber . “ Understanding Supply Chain Management .” APICS—The Performance Advan- tage , 9 ( January 1999 ): 38 – 43 .

c07.indd 281c07.indd 281 25/09/12 10:49 AM25/09/12 10:49 AM

c07.indd 282c07.indd 282 25/09/12 10:49 AM25/09/12 10:49 AM

283

� C H A P T E R 7 S U P P L E M E N T A

The Beer Game 1

1 Adapted from J. Sterman. “Instructions for Running the Beer Distribution Game.” Massachusetts Institute of Technology (October 1984); J. H. Hammond. “The Beer Game: Description of Exercise,” Harvard Business School, 9, 964–104.

c07.indd 283c07.indd 283 25/09/12 10:49 AM25/09/12 10:49 AM

284 C h a p t e r 7 S u p p l e m e n t A : T h e B e e r G a m e

The Beer Game has become a staple of the operations management course in MBA programs across the country. In effect, the game simulates material and information fl ows in a simplifi ed supply chain. As shown in Figure 7SA.1 , the supply chain con- sists of four stages. Moving from the factory downstream, the supply chain consists of a factory, wholesaler, distributor, and retailer. Accordingly, each stage in the sup- ply chain is required to manage its inventory levels given the receipt of orders from its downstream customer through the placement of orders with its upstream sup- plier. The only exceptions to this are that the retailer ’s demand comes from the fi nal consumer and the factory schedules production requests as opposed to placing an order from an upstream supplier.

There is a two-week delay between the retailer, wholesaler, and distributor. Thus, orders from the retailer to the wholesaler in a given week arrive two weeks after the wholesaler ships them. Likewise, orders from the wholesaler to the distributor in a particular week arrive two weeks after the distributor ships it. Production orders at the factory are available to ship three weeks after the production requests.

Your objective in playing the game is to minimize the sum of your total weekly costs. Weekly costs consist of two components: an inventory cost and a backlog cost. More specifi cally, weekly inventory cost is calculated at the rate of $0.50/keg of beer in inventory at the end of the week, while backlog costs are calculated at the rate of $1.00/keg on backlog at the end of the week. Obviously, only one of these costs can be positive in any given week (although it is possible that they both could be zero in a particular week).

Because a supply chain for the beer industry in reality would likely be character- ized by multiple factories, dozens of distributors, hundreds of wholesalers, and tens of thousands of retailers, it is often the case that the only information shared between a supplier and its customer is order information. Therefore, in the game, the only communication you may have with your upstream supplier is the placement of your order.

In terms of the initial conditions, as it turns out, the demand at the retailer stage has been quite stable at four kegs per week for the last several weeks. Therefore, every order placed throughout the entire supply chain has been for four kegs over this period. Furthermore, each stage has maintained an inventory level of 12 kegs, or the equivalent of three weeks of demand. However, as the weather turns warmer in the near future, demand is expected to increase. Also, it is expected that there will be one or more promotions over the coming months.

In playing the game, you will be assigned to one of the four stages in the supply chain. During each week of simulated time, you will be required to perform the fol- lowing fi ve tasks. It is important that these tasks be completed in the order listed below and that each stage in the supply chain complete the task simultaneously with the other stages. Note that only the fi nal task requires you to make a decision.

1. Deliver your beer and advance shipments. Move the beer in the Shipping Delay box (on the right, adjacent to your Current Inventory box) into the Current Inventory box. Next, move beer in the other Shipping Delay box to the right to the now empty Shipping Delay box. (Factories move the inventory from the Production Delay box directly to the right of the Current Inventory box into the Current Inventory box. Then move inventory from the top Production Delay box to the bottom Production Delay box.)

c07.indd 284c07.indd 284 25/09/12 10:49 AM25/09/12 10:49 AM

Incoming order for retailer

Orders sold to customer

Retailer

Current inventory

Current inventory

Current inventory

OC

Order placed by

retailer

OC

* *

* * * * * * * * * * * *

Wholesaler

* * * * * * * * * * * *

Incoming order for

wholesaler

OC

Order placed by wholesaler

OC

Incoming order for

distributor

OC

Order placed by distributor

OC

Incoming order for factory

OC

Production request

OC

Distributor

* * * * * * * * * * * *

Factory

Raw materials

* * * * * * * * * * * *

* *

P.D.

P.D.S.D.S.D.S.D. Current inventory

S.D.S.D.S.D.

* = 1 keg, OC = order card, S.D. = shipping delay, P.D. = production delay

* *

* *

* *

* *

* *

* *

* *

* *

* *

* *

* *

* *

* *

* *

* *

* * * * * * * *

* * * * * ** *

* *

* * * ** *

Figure 7SA.1 The beer game board and initial conditions.

285

c07.indd 285c07.indd 285 25/09/12 10:49 AM25/09/12 10:49 AM

286 C h a p t e r 7 S u p p l e m e n t A : T h e B e e r G a m e

2. Pick up the incoming order from your downstream customer in your Incoming Order box at your top left (retailers read incoming order from the consumer). Fill as much of the order as you can from your current inven- tory by placing the appropriate quantity of kegs in the Shipping Delay box directly to the left of your Current Inventory box. Quantities ordered above your current inventory level become part of your current backlog. More specifi cally, the amount to ship this week is calculated as follows:

Quantity to ship   5   incoming order this week   1   previous week ’ s backlog

3. Calculate and record in Figure 7SA.2 your ending inventory or backlog posi- tion (as a negative number). Count the number of kegs remaining in your current inventory after the shipment for the week has been made. If you get into a backlog situation, the backlog must be accumulated from week to week, since quantities ordered but not shipped must be made up. The week ’s ending backlog position is calculated as follows:

Figure 7SA.2 Data sheet

Week Inventory Order Placed Week Inventory Order Placed

621

722

823

924

035

136

237

338

439

5301

6311

7321

8331

9341

0451

1461

2471

3481

4491

5402

6412

7422

8432

9442

0552

c07.indd 286c07.indd 286 25/09/12 10:49 AM25/09/12 10:49 AM

287C h a p t e r 7 S u p p l e m e n t A : T h e B e e r G a m e

Current week’s backlog   5   previous week ’ s backlog 1 incoming order 2 shipments received this week

4. Advance your order cards. (Factories fi ll their production requests.) Advance the order from the Order Placed box to the Incoming Order box (or, for the factory, read the Production Request and fi ll the Production Delay box from the raw materials inventory). Make sure to keep the order cards face down as you move them.

5. Decide how much to order, write it down on your order card (and in Figure 7SA.2 ), and place the card face down in the Orders Placed box. Factories decide how much to schedule for production, write it down on your order, and place the card face down in the Production Request box.

6. Repeat steps 1–5.

Most likely, your instructor will have the class complete one or more practice runs or go through the fi rst couple of weeks at a slow pace.

c07.indd 287c07.indd 287 25/09/12 10:49 AM25/09/12 10:49 AM

c07.indd 288c07.indd 288 25/09/12 10:49 AM25/09/12 10:49 AM

289

� C H A P T E R 7 S U P P L E M E N T B

The Economic Order

Quantity Model

c07.indd 289c07.indd 289 25/09/12 10:49 AM25/09/12 10:49 AM

290 C h a p t e r 7 S u p p l e m e n t B : T h e E c o n o m i c O r d e r Q u a n t i t y M o d e l

The concept of economic order quantity ( EOQ )   applies to inventory items that are replenished in batches or orders and are not produced and delivered continuously. Although we have identifi ed a number of costs associated with inventory decisions in the chapter, only two categories, carrying cost and ordering cost, are considered in the basic EOQ model. Shortage costs and opportunity costs are not relevant because shortages and changes in capacity should not occur if demand is constant, as we assume in this basic case. The cost of the goods is considered to be fi xed and, hence, does not alter the decisions as to when inventory should be reordered or how much should be ordered.

More specifi cally, we assume the following in the basic EOQ model:

1. Rate of demand is constant (e.g., 50 units per day).

2. Shortages are not allowed.

3. Lead times are known with certainty, so stock replenishment can be sched- uled to arrive exactly when the inventory drops to zero.

4. Purchase price, ordering cost, and per-unit holding cost are independent of quantity ordered.

5. Items are ordered independently of each other.

Let us consider a water distributor that sells 1000 5-gallon bottles per month (30 days) and purchases in quantities of 2000 bottles per order. Lead time for the receipt of an order is six days. The cost accounting department has analyzed inven- tory costs and has determined that the cost of placing an order is $60 and the annual cost of holding one 5-gallon bottle in inventory is $10. 2

Under its current policy of ordering 2000 bottles per order, what is the water dis- tributor ’s total annual inventory cost? Its inventory pattern is represented by the “sawtooth” curve of Figure 7SB.1 . For simplicity, let

Q   5   order quantity

U   5   annual usage

C O   5   cost to place one order

C H   5   annual holding cost per unit

To determine the total annual incremental cost of the distributor ’s current inven- tory policy, we must determine two separate annual costs: total annual holding cost and total annual ordering cost.

The ordering cost is determined by C O , the cost to place one order ($60), and the

number of orders placed per year. Since the distributor sells 12,000 5-gallon bottles per year and orders 2000 per order, it must place six (that is, 12,000/2000) orders per year, for a total ordering cost of $360 ( 6 orders per year   3   $ 60 per order ) . Using our notation, we write the annual ordering cost as

Annual ordering cost   5   U __ Q   3   C O

2Sometimes holding cost is given as a fi xed value per year and other times as a percentage of the value of the inventory, especially when interest charges represent the major holding costs. Then holding cost C

H   5   i C , where C is the cost of the inventory item and i is the interest rate.

c07.indd 290c07.indd 290 25/09/12 10:49 AM25/09/12 10:49 AM

291C h a p t e r 7 S u p p l e m e n t B : T h e E c o n o m i c O r d e r Q u a n t i t y M o d e l

The annual holding cost is determined by C H , the cost of holding one 5-gallon

bottle for one year ($10), and the number of bottles held as “cycle stock.” Notice that the inventory level is constantly changing and that no single bottle ever remains in inventory for an entire year. On average, however, there are 1000 bottles in inven- tory. Consider one cycle of the distributor ’s inventory graph, as shown in Figure 7SB.2 . The inventory level begins at 2000 units and falls to 0 units before the next cycle begins. Since the rate of decline in inventory is constant (i.e., 1000 per month), the average level is 1000 units, or simply the arithmetic average of the two levels: ( 2000   1   0 ) / 2   5   1000 .

If, on the average, there are 1000 bottles in inventory over the entire year, then the annual inventory holding cost is $10,000 ( $ 10 per unit   3   1000 units ) . Or, in our general notation,

Annual holding cost   5   Q

__ 2   3   C H

Adding annual ordering cost and annual holding cost gives the following equa- tion for total annual cost (TAC):

TAC   5   � U __ Q � C O   1   � Q

__ 2 � C H

Figure 7SB.1 Water distributor ’s inventory pattern.

Time

In ve

n to

ry l

ev el

Six-day lead time

Maximum level 2000

Average level 1000

Reorder point

2000

0

2 months

In ve

n to

ry l

ev el

Time

Figure 7SB.2 Water distributor’s inventory graph.

c07.indd 291c07.indd 291 25/09/12 10:49 AM25/09/12 10:49 AM

292 C h a p t e r 7 S u p p l e m e n t B : T h e E c o n o m i c O r d e r Q u a n t i t y M o d e l

For the water distributor, TAC is $ 360   1   $ 10 , 000   5   $ 10 , 360 . Thus, its current inventory policy of ordering quantities of 2000 bottles is costing $10,360 per year. Is this the best policy, or can it be improved?

Finding an Optimal Policy We can graph annual holding cost and annual ordering cost as a function of the order quantity, as shown in Figure 7SB.3 . Since the annual holding cost is ( Q / 2 ) C

H , which

can be written ( C H / 2 ) Q , we see that holding cost is linear and increasing with respect

to Q . Annual order cost is ( U / Q ) C O , which can be rewritten as ( U C

O ) / Q . We can see

that ordering cost is nonlinear with respect to Q and decreases as Q increases. Now, if we add the two graphed quantities for all values of Q , we have the TAC

curve shown in Figure 7SB.3 . Note that TAC fi rst decreases as ordering cost decreases but then starts to increase quickly. The point at which TAC is minimized is the optimal order quantity; that is, it gives the quantity Q that provides the least total annual inven- tory cost. This point is called the economic order quantity ( EOQ ) , and for this inventory problem it happens to occur where the order cost curve intersects the hold- ing cost curve. (The minimum point is not always where two curves intersect; it just happens to be so in the case of EOQ .) From Figure 7SB.3 we can see that EOQ is approximately 400 bottles per order.

We can compute an accurate value algebraically by noting that the value of Q at the point of intersection of the two cost lines is the EOQ . We can fi nd an equation for EOQ by setting the two costs equal to one another and solving for the value of Q :

EOQ   5 � _____

2 U C

O _____ C

H

500

EOQ

1000

TAC Annual holding cost, (CH /2)Q Annual order cost, (UCO)/Q

1500

Q (units)

2000 2500

5000

C o

st (

$ )

10,000

15,000

Figure 7SB.3 Graph of annual inventory costs.

c07.indd 292c07.indd 292 25/09/12 10:49 AM25/09/12 10:49 AM

293C h a p t e r 7 S u p p l e m e n t B : T h e E c o n o m i c O r d e r Q u a n t i t y M o d e l

For the water distributor, we can compute EOQ as

EOQ   5 � ___________

2 ( 12 , 000 ) 60

___________ 10   5   � _______

144 , 000   5   379.5

Obviously, since we cannot order a fraction of a bottle, the order quantity would be rounded to 380 units.

The total annual cost ( TAC )   of this policy would be

TAC   5   � 12 , 000 ______ 380 � 60   1   � 380 ____ 2 � 10   5   1894.74   1   1900   5   $ 3794.74 Note that this is an improvement in total annual cost of $6,565.26 over the present

policy of ordering 2000. Actual inventory situations often exhibit relative “insensitiv- ity” to changes in quantity in the vicinity of EOQ . To the inventory manager, what this means is added fl exibility in order quantities. If, for example, shipping and han- dling was more convenient or economical in quantities of 500 (perhaps the items are shipped on pallets in quantities of 250), the additional 120 units per order would cost the organization only an extra $145.26 per year.

Cautions Regarding EOQ The EOQ is a computed minimum-cost order quantity. As with any model or formula, the GIGO rule (garbage in, garbage out) applies. If the values used in computing EOQ are inaccurate, then EOQ will be inaccurate—though, as mentioned previously, a slight error will not increase costs signifi cantly. EOQ relies heavily on two variables that are subject to considerable misinterpretation. These are the two cost elements: holding cost ( C

H ) and order cost ( C

O ) . In the derivation of EOQ , we assumed that by

ordering fewer units per order, the cost of holding inventory would be reduced. Similarly, it was assumed that by reducing the number of orders placed each year, the cost of ordering could be proportionately reduced. Both assumptions must be thor- oughly questioned in looking at each cost element that is included in both C

H and C

O .

For example, if a single purchasing agent is employed by the fi rm, and orders are reduced from 3000 to 2000 per year, does it stand to reason that the purchasing expense will be reduced by one-third? Unless the person is paid on a piecework basis, the answer is clearly no. Similarly, suppose we rent a warehouse that will hold 100,000 items and that we currently keep it full. If the order sizes are reduced so that the warehouse is only 65 percent occupied, can we persuade the owners to charge us only 65 percent of the rental price? Again, the answer is no. Clearly, then, when costs are determined for computations, only real, out-of-pocket costs should be used. Costs that are committed, or “sunk,” no matter what the inventory level or number of orders is, should be excluded.

Note also that C H and C

O are controllable costs. That is, they can be reduced, if

this is advantageous. This is exactly what the Japanese recognized. They saw the fol- lowing problems with holding inventory:

• Product defects become hidden in the inventory, thereby increasing scrap and rework later in the production system, when defects are harder to repair. Just as important, the problem in the system that led to the defective part cannot be tracked down so easily later.

c07.indd 293c07.indd 293 25/09/12 10:49 AM25/09/12 10:49 AM

294 C h a p t e r 7 S u p p l e m e n t B : T h e E c o n o m i c O r d e r Q u a n t i t y M o d e l

• Storage space takes up precious room and separates all the company ’s func- tions and equipment, thereby increasing problems with communication. Space itself is extremely expensive in Japan (directly increasing the variable C

H ).

• More inventory in the plant means that more control is needed, more plan- ning is required, larger systems are required to move all that stock, and in general more “hassle” is created, which leads to errors, defects, missed deliv- eries, long lead times, and more diffi culty in product changeovers.

Rarely do U.S. fi rms consider these real costs in the EOQ formula. More typically, these costs are considered part of the indirect, overhead, or “burden” costs that are assumed to be uncontrollable. Again, the message is: be very careful about the val- ues used in the EOQ formula.

Also, it should be noted that very small EOQ values (e.g., 2) will not usually be valid, because the cost functions are questionable for such small orders. Last, EOQ reorder sizes should not be followed blindly. There may not be enough cash just now to pay for an EOQ , or storage space may be insuffi cient.

A P P L Y Y O U R U N D E R S T A N D I N G

De l ta Produc t s Inc .

Delta Products, of Omaha, Nebraska, makes a line of door hardware, including doorknob sets and deadbolts. Its product line is particularly well known for excel- lent quality, a high level of security, and ease of installation.

Delta hired Nikki Scott, a fi rst-year MBA concentrating in operations manage- ment, for one of its summer internships. Her task was to spend the summer analyz- ing the operation and usage of the large transfer press, the plant ’s current bottleneck machine. The transfer press stamps out the doorknobs used in Delta ’s door hard- ware and is available 2000 hours per year for this task. At the end of the summer, in a report to the plant manager, Nikki was to submit her recommendations for improv- ing the operation of the transfer press.

Nikki spent the fi rst week familiarizing herself with the operation of the transfer press by observing and questioning the machine operator. By the end of the fi rst week, she determined that Delta uses six unique doorknobs in its door hardware.

She decided that her next task was to get an estimate of the annual demand for the knobs. She studied Delta ’s product catalogs and determined which products each knob was used in. Next, she got from the sales manager a copy of a spread- sheet that contained the complete sales fi gures for each product over the last three years. After casually looking over the data, she observed that the sales were remark- ably stable over the three-year period. Nikki added formulas to the spreadsheet to average the sales data over the three-year period and to calculate the number of knobs of each type that were used.

With her analysis of the demand for knobs completed, Nikki turned her attention to the actual production of the knobs. She spent the next couple of weeks collecting data on the individual processing times of the knobs on the transfer press, the time required to set up the press to produce a new batch of knobs, and the production batch sizes currently being used. She also worked closely with the cost accountant to determine the cost of holding knobs in inventory, and she found that machine

c07.indd 294c07.indd 294 25/09/12 10:49 AM25/09/12 10:49 AM

295C h a p t e r 7 S u p p l e m e n t B : T h e E c o n o m i c O r d e r Q u a n t i t y M o d e l

operators are paid $15 per hour, including fringe benefi ts. Nikki summarized the information for her report in the spreadsheet above.

Her next task was to analyze the information she had compiled and look for ways to improve the operation. Given the information she had available, she began by developing a spreadsheet to calculate the economic order quantity ( EOQ ) . She then developed another spreadsheet to compare the total annual cost ( TAC ) of using the current batch sizes with the economic order quantities she had calculated. Nikki was extremely pleased when she realized she could save Delta almost $7000 per year in just one department if it adopted her recommended economic order quantities. After rechecking her calculations, she was convinced of the validity of her analysis and couldn ’t wait to see the plant manager ’s reaction to her report. She was actually hop- ing that the plant manager would be impressed enough with her work to offer her a full-time position upon graduation. The fi nal spreadsheet she developed to summa- rize the potential savings to Delta is shown next.

Item Current Batch

Size Economic

Order Quantity TAC Current Batch Size TAC EOQ

Knob A 1500 586 2810 1904

Knob B 1500 328 3026 1262

Knob C 1500 741 2567 2039

Knob D 2000 662 4105 2448

Knob E 1500 477 3469 2005

Knob F 2000 923 2729 2077

Total $18,705 $11,737

Annual Savings $6968

On the Monday of her last week, Nikki met with the plant manager, Joe Thomas, and the press department manager, Mike Willis. After complimenting Nikki on a very thor- ough and well-written report, Joe asked Mike what his reaction was. Mike commented:

I also was very impressed with the thoroughness of Nikki ’s analysis. The data that she collected on setup times, production times, and holding costs are the best data we have ever had about our operations. Unfortunately, while I have not had a chance to thor- oughly run the numbers, I think there is a problem with Nikki ’s analysis. Her analysis requires signifi cant reductions in our batch sizes. While I agree that we could save money by cutting the batch sizes, the fact of the matter is that the transfer press she analyzed is one of our major bottlenecks. We are currently using every second of the 2000 hours we

Item Annual Demand

Unit Processing Time (hrs.)

Setup Time (hrs.)

Annual Holding Cost ($/unit)

Current Batch Size

Knob A 6000 0.0500 6.2 3.25 1500

Knob B 3000 0.0420 4.6 3.85 1500

Knob C 7000 0.0400 7.2 2.75 1500

Knob D 10,000 0.0380 5.4 3.70 2000

Knob E 8400 0.0375 3.8 4.20 1500

Knob F 9400 0.0480 6.8 2.25 2000

c07.indd 295c07.indd 295 25/09/12 10:49 AM25/09/12 10:49 AM

296 C h a p t e r 7 S u p p l e m e n t B : T h e E c o n o m i c O r d e r Q u a n t i t y M o d e l

1. Frame-Up, a self-service picture framing shop for tourists in Sevilla, Spain, orders 3000 meters of a certain molding every month. The order cost is €40 and the holding cost is €0.05 per meter per year. What is the current annual inventory cost? What is the maximum inventory level? What is the EOQ ?

2. The Corner Convenient Store (CCS) in northern Chicago receives orders from its distributor in three days from the time an order is placed. Light Cola sells at the rate of 860 cans per day. (It can sell 250 days of the year.) A six-pack of Light Cola costs CCS $1.20. Annual holding cost is 10 percent of the cost of the cola. Order cost is $25.00. What is CCS ’s EOQ ?

3. A Liverpool, England, camera supplies wholesaler purchases £1 million worth of camera equipment each year. It costs £100 to place and receive an order, and the annual holding cost per item is 20 percent of the item ’s value.

a. What is the £ value of the EOQ ?

b. How many months’ supply is the EOQ ?

c. How often should orders be placed?

d. How much will the annual holding cost change if the company orders monthly? How much will the annual ordering cost change?

e. What should the £ value of the EOQ be if the wholesaler doubles its annual purchases? What should it be if (instead) the ordering cost dou- bles? What should it be if the holding cost drops to 10 percent?

4. Wing Computer Corporation of Xiamen, China, uses 15,000 keyboards each year in the production of computer terminals for major computer manu- facturers headquartered in Taiwan. Order cost for the keyboards is 50 renminbi, and the holding cost for one keyboard is 1.5 renminbi per unit per year. What is the EOQ ? If Wing orders 1250 per month, what will be the total annual cost? What is the aver- age inventory level?

have available on the machine. Cutting the lot sizes as Nikki has suggested will require more setups, and we simply don ’t have the time for additional setups.

Nikki was quite distraught by the outcome of the meeting. She still had a week left in her internship, and she desperately wanted to salvage the work she had spent an entire summer on. She was determined to spend her fi nal week fi nding a way to save Delta money while at the same time not exceeding its available capacity.

Questions

1. Verify Nikki ’s calculations of the economic order quantity and total annual cost. 2. Is Mike ’s intuition correct—that using Nikki ’s economic order quantities will exceed the

2000 hours of capacity available on the transfer press? 3. Are there any opportunities for Delta to save money without exceeding its available capacity?

E X E R C I S E S

B I B L I O G R A P H Y

Greene , J. H. Production and Inventory Control Hand- book , 3rd ed. New York : McGraw-Hill , 1997 .

Landvater , D. V. World Class Production and Inventory Management . Newburg, NH : Oliver Wight Publications , 1997 .

Silver , E. A. , D. F. Pyke , and R. Peterson . Inventory Man- agement, Production Planning, and Scheduling , 3rd ed. New York : Wiley , 1998 .

Sipper, D., and R. Bulfi n. Production: Planning, Con- trol, and Integration . New York : McGraw-Hill , 1997 .

Vollmann , T. E. , W. L. Berry , D. C. Whybark , and F. R. Jacobs . Manufacturing Planning and Control Systems for Supply Chain Management , 5th ed. New York : McGraw-Hill, 2004 .

Waters , D. Inventory Control and Management . Hoboken, NJ : Wiley , 2003 .

Zipkin , P. H. Foundations of Inventory Management . New York: Irwin/McGraw-Hill, 2000 .

c07.indd 296c07.indd 296 25/09/12 10:49 AM25/09/12 10:49 AM

297

� Capacity, Scheduling,

and Location Planning

C H A P T E R 8

ROLE OF OPERATIONS IN ORGANIZATIONS’ COMPETITIVENESS

Operations Strategy

Process Design and Control

Process Improvement

Process Execution

Ch. 1: Operations Strategy and Global

Competitiveness

Ch. 2: Process Planning and

Design

Ch. 3: Controlling Processes

Ch. 7: Supply Chain Management

Ch. 8: Capacity, Scheduling, and

Location Planning

Ch. 4: Process Improvement:

Six Sigma

Ch. 5: Process Improvement: Lean

Ch. 6: Managing Process

Improvement Projects

c08.indd 297c08.indd 297 25/09/12 5:20 PM25/09/12 5:20 PM

298

IN T R O D U C T I O N • Bo Burlingham, editor-at-large of Inc . magazine, claims that “small giants,” fi rms that

made a conscious choice to stay small and not expand, go public, or sell out to a larger corporation, are the “heart and soul” of the American economy. They are also “giants” because they are recognized in their industry for their achievements and their employees express high levels of contentment and self-satisfaction with the way they do business. As one owner says, “I ’ve made more money by choosing the right things to say no to than by choosing things to say yes to. I measure [suc- cess] by the money I haven ’t lost and the quality I haven ’t sacrifi ced.”

Fritz Maytag, brewmaster at Anchor Brewery, one of the “small giants,” says, “Just because [your product] is the best around doesn ’t mean you have to franchise or even expand. You can stay as you are and have a business that ’s profi table and rewarding and a source of great pride.”

Burlingham says these small giants all have a strong commitment to great cus- tomer service and always “go the extra mile.” Yet, in a seeming contradiction, he also points out that what really sets them apart is their belief that the customer comes second—their employees come fi rst! It ’s a contradiction that makes fans wish they worked for one of these small giants instead (Burlingham 2006).

• Milwaukee, Wisconsin, is home to two of the biggest multibillion-dollar manufactur- ers of earth-moving equipment that sells for up to $180 million, Bucyrus International, Inc. and Joy Global, Inc. However, they disagree completely on how to locate their production facilities. Bucyrus makes all their machines in the United States and Europe, where they have developed highly effi cient, low-cost production

Finally, we turn to executing our processes. To do so in an effi cient but also effective manner involves three primary considerations: determining the amount of capacity needed, where it should be located, and how to effectively schedule its use. Having adequate capacity and effectively utilizing it are critical for dependability, speed, and maxi- mizing revenues, while having excess capacity will impair costs—all strategic competitive factors. We begin the chapter with an overview of various measures of capacity and then discuss issues related to long-term capacity planning and location planning strategies.

In the next section, we address the unique aspects of making location decisions for pure services. Following this, we consider issues related to effi ciently using the available capacity through effective schedule management. The chapter con- cludes with a discussion of short-term capacity planning, including process-fl ow analysis, the relationship between capacity and scheduling, and how humans ’ ability to learn affects capacity planning. A supplement to the chapter describes some methods of making short- and long-term forecasts.

298

C H A P T E R I N P E R S P E C T I V E

c08.indd 298c08.indd 298 25/09/12 5:20 PM25/09/12 5:20 PM

processes, and then ships them from there to customers all over the globe. Moreover, they expect the U.S. dollar and the euro to remain relatively weak curren- cies, making manufacturing there affordable. Although they were invited to build a plant in China in the 1970s, they declined to do so because the Chinese government decided to make the production of industrial mining machinery “strategically critical,” meaning that they would be heavily subsidizing their own domestic manufacturers.

In contrast, Joy prefers to design and engineer in the United States but build their plants close to emerging markets and in low-cost developing countries, such as China, where the mining market is growing and customers and domestic suppliers are plentiful. Joy ’s Chinese factory operating costs are fully 20 percent less than in the United States or Europe, though they admit to having had some early quality problems. As Bucyrus ’s CEO says: “It ’s going to be interesting to see how it plays out. One of us is more right than the other” (Matthews 2010).

• The global business environment has become highly volatile, with up and down and up again gasoline prices, on and off regulation, supply chain disruptions, green chal- lenges, terrorism, pandemics, and more. Airtel, India ’s largest telecom carrier, with 100 million subscribers, needed to gain agility between its costs, capacity, and cus- tomers so it could scale up and down rapidly in India ’s volatile business environment. Being agile means being able to break even at as low as 30 percent of capacity!

As a telecom carrier, they decided to outsource their IT management to IBM and lease network capacity from Ericsson and Nokia, allowing them to pay for only the capacity they needed at any given time. They also knew that as customer demands change, the mix of skills needed would change, too, so they began cross-training their employees. And to minimize risk, they also don ’t allow any single customer to represent more than 5 percent of revenues; they also insist on drawing customers from across a variety of business sectors so that if one sector gets into trouble, other sectors could counterbalance it. As a result, they are now probably the lowest-priced carrier in the world and one of the most profi table (Prahalad 2009).

• Waiting in line for service is one of the banes of many service industries, from air- lines to retailing to entertainment to government services. However, better service processes can often make a big difference in customers ’ satisfaction with their wait- ing. Alaska Airlines pioneered the self-service kiosk check-in and baggage-drop process in 2003 and then patented it in 2006. The new process cuts about 80 percent of the time off check-in and gives 50 percent more usable space. And although Alaska Air expected 20 to 30 percent increases in agent productivity, instead it got 100 percent, being able to serve passengers in half the time (Carey 2007).

Recent research on waiting lines has found that up to 3 minutes of waiting, cus- tomers are fairly accurate about how long they ’ve been waiting, but after that the wait estimates multiplied with every passing minute so that 4 minutes seemed like 5 or 6 and 5 minutes seemed like 10. In addition, if the line moved more slowly as a person approached the server, even if it had moved faster initially, the customer was highly dissatisfi ed, but if it speeded up toward the end, the customer felt much more positive. And even though the “cattle-stall” single line with multiple registers has been shown to be three times faster than separate lines for each register, many peo- ple would rather jockey for position between the separate lines.

I n t r o d u c t i o n 299

c08.indd 299c08.indd 299 25/09/12 5:20 PM25/09/12 5:20 PM

300 C h a p t e r 8 : C a p a c i t y , S c h e d u l i n g , a n d L o c a t i o n P l a n n i n g

Both customers and organizations have come up with their own solutions to the problem. Customers may select a line that has customers with less-full carts, no accompanying children, no old people, or a server who is young and doesn ’t gab with the customers. Organizations use some of their own weapons, in addition to the ubiquitous impulse purchases along the waiting line, to distract the customers: adding line expediters to begin scanning upcoming customers ’ purchases, super helpers who can ring up purchases on hand-held devices anywhere in the store, and runners for the cashier. Some insist that cashiers stand out in front of their registers to show that their line is empty, others are installing self-checkout registers, and some put an employee or electronic screen at the line to direct shoppers to the next open register (Smith 2011).

• In industries, such as fashion, that are characterized by highly volatile demand, the combined costs of stockouts and markdowns can be greater than total manufactur- ing costs. One approach to forecasting in these highly volatile industries is to determine what can and cannot be predicted well. Products in the “predictable” category are made farthest in advance, saving manufacturing capacity for the “unpredictable” products so that they can be produced closer to their actual selling season. Using this approach, Sport Obermeyer, a producer of fashionable skiwear, increased its profi ts between 50 percent and 100 percent over a three-year period (Fisher et al. 1994).

• Package Products, in Pittsburgh, Pennsylvania, produces folding carton packing for the bakery and deli industries. A key aspect of Package Products ’ strategy is to be recognized by its customers for quality, reliability, and service. Signifi cant growth during the 1990s greatly complicated the task of managing the company ’s opera- tions. In addition, its customers were becoming more demanding, and the market- place was becoming more competitive. To gain better control over its operations, Package Products implemented a fi nite capacity scheduling (FCS) software package. Before it acquired the FCS system, a Gantt chart was maintained manually to sched- ule jobs. Problems with the manual system included chronic capacity shortages and the fact that key data resided in the heads of people who were scattered throughout the organization. Two criteria used for selecting the FCS software pack- age were that it should work with the company ’s existing business system and that it should not be a “black box,” claiming to provide optimal schedules that no one could really understand. Through the use of the FCS program, overtime has been substantially reduced, on-time delivery has been improved by 32 percent, and back- orders have been reduced by 53 percent. Additionally, customer service can now respond to customers ’ inquiries in an average of 22 seconds—versus taking over- night previously (Trail 1996).

As we will discuss in more detail throughout this chapter, capacity represents the rate at which a transformation system can create outputs. Capacity planning is as important to service organizations as it is to manufacturing organizations. For exam- ple, the transformation process at Burger King is designed so that capacity can be

c08.indd 300c08.indd 300 25/09/12 5:20 PM25/09/12 5:20 PM

301L o n g - T e r m C a p a c i t y P l a n n i n g

quickly adjusted to match a highly variable demand rate throughout the day. In manufacturing, semiconductor fi rms incur enormous costs associated with expand- ing capacity. To further complicate matters for manufacturing businesses, shorter product life cycles mean that they have less time to recoup their investment, espe- cially when the next generation of products make their current products obsolete.

Clearly, capacity and location are important elements of a fi rm ’s competitive strat- egy. In fact, they play a major role in the sand cone competitive dimensions described in Chapter 1: quality, delivery dependability, speed, and cost. For example, if capac- ity is insuffi cient for demand peaks, then confusion and errors may result when attempting to meet excessive demand, lowering the quality of the fi rm ’s outputs. And without capacity and a convenient location, customers cannot depend on the availability of the output and may turn to competitors. In terms of speed, suffi - cient capacity and a convenient location allow the organization to meet demand quickly, whenever and wherever it arises. And fi nally, if the fi rm has insuffi cient capacity, it will cost considerably more to engage the extra resources to meet unex- pected demand, whether the resources are additional labor in the form of overtime or hiring, subcontracting out a portion of the demand, or storing inventory to meet demand peaks.

Capacity planning decisions are driven by projected demand estimates for the organization ’s outputs. (The role of forecasting is discussed in the supplement to this chapter.) Over the long term, capacity and location are interwoven considerations, as will be described later in the chapter. Following the long-term discussion, we move into a description of effi ciently utilizing the available capacity through sched- ule management and then into a description of short-term capacity alternatives, but here we are largely past the point of making a location decision.

L O N G - T E R M C A P A C I T Y P L A N N I N G Capacity and location decisions are highly strategic because they are very expensive investments and, once made, are not easily changed or reversed. Hence, they must be carefully and thoroughly analyzed beforehand, using all available tools at man- agement ’s disposal. Capacity is generally taken to mean the maximum rate at which a transformation system produces outputs or processes inputs, though the rate may be “all at once.” Table 8.1 lists measures of capacity for a number of production sys- tems. Notice that since capacity is defi ned as a rate, measures should be clear about the time dimension . For instance, how meaningful is it to know that a hospital can perform 25 surgeries? Without knowing whether this is simultaneously, per day, per week, or possibly per month, the number is relatively meaningless.

As illustrated in Table 8.1 , airlines often measure their capacity in available seat miles (ASMs) per year. One ASM is one seat available for one passenger for 1 mile. Clearly, the number of planes an airline has, their size, how often they are fl own, and the route structure of the airline all affect its ASMs, or capacity. However, we may also talk about the capacity of a single plane, such as a 550-seat jumbo, and here we clearly mean “all at once.” Nevertheless, this capacity measure is not very useful without knowing to what use the plane may be put, such as constantly being in the air generating ASMs or used as an occasional backup. Similarly, an elementary measure of a hospital ’s capacity is often simply the number of beds it has (for the full

c08.indd 301c08.indd 301 25/09/12 5:20 PM25/09/12 5:20 PM

302 C h a p t e r 8 : C a p a c i t y , S c h e d u l i n g , a n d L o c a t i o n P l a n n i n g

year is implied). Thus, a 50-bed hospital is “small” and a 1000-bed hospital is “large.” And a restaurant may measure its capacity in tables (per hour), a hotel in rooms (per night), and a public service agency in family contacts (per weekday).

Notice that these measures of capacity do not recognize the multiple types of outputs with which an organization may, in reality, be concerned. ASMs say nothing about the freight capacity of an airline, but freight may be a major contributor to profi ts. Similarly, number of beds says nothing about outpatient treatment, ambu- lance rescues, and other services provided by a hospital. Thus, capacity planning must often consider the capacity to produce multiple outputs. Unfortunately, some of the outputs may require the same organizational resources as well as some very specialized resources.

The provision of adequate capacity is clearly a generic problem, common to all types of organizations, but in pure service organizations capacity is a special prob- lem because the output cannot normally be stored for later use. A utility, for exam- ple, must have capacity available to meet peak power demands, yet the average power demand may be much, much lower. Where the provision of the service is by human labor, low productivity is a danger when staffi ng is provided to meet the demand peaks.

Another characteristic of capacity is that, frequently, a variety of restrictions can limit it. For example, the capacity of a fast-food restaurant may be limited not only by the number of order-takers on duty but also by the number of cooks, the number of machines to prepare the food, the amount of food in stock, the space in the res- taurant, and even the number of parking spaces outside. Any one of these factors can become a bottleneck (discussed in a later section of the chapter) that limits the restaurant ’s normal operating capacity to something less than its theoretical or design capacity.

In addition, during the production process there are often natural losses, waste (avoidable), scrap (unavoidable), defects, errors, and so on that again limit the capacity of a system. These losses are considered in a measure known as the yield of the process: the amount of output of acceptable quality emerging from a produc- tion system compared with the amount that entered it. Yield management , also known as revenue management , is a somewhat different topic but of high interest

T A B L E 8 .1 • Example s o f Measure s o f Capac i t y

Production System Measure of Capacity in

Terms of Outputs Produced Measure of Capacity in

Terms of Inputs Processed

Airline Available seat miles per year Reservation calls handled per day

Hospital Babies delivered per month Patients admitted per week

Supermarket Customers checked out per hour

Cartons unloaded per hour

Post Offi ce Packages delivered per day Letters sorted per hour

University Graduates per year Students admitted per year

Automobile assembly plant Autos assembled per year Deliveries of parts per day

c08.indd 302c08.indd 302 25/09/12 5:20 PM25/09/12 5:20 PM

303L o n g - T e r m C a p a c i t y P l a n n i n g

these days, particularly in services. We return to the topic of yield management later in the chapter.

In the process of trying to forecast the long-run capacity needs for the organization, the issue of location of the facility, or facilities, cannot be ignored because the demand may well be a function of where the facility is located. And if there are multiple facili- ties, the capacity needs for any one will certainly depend on how many others are serving the same geographic needs. Moreover, transportation may also be a factor if there is a facilitating good, or product, involved, as well as inventories, warehouses, and other such matters that concern supply chain management . In these days of intense worldwide competition, supply chain management is taking on signifi cantly more importance, as it accounts for a greater and greater proportion of the total cost of all outputs. Although we covered supply chain management in Chapter 7, here we discuss the interplay between capacity and location in more detail.

Capacity Planning Strategies Issues of capacity planning over the long run relate primarily to the strategic issues of initiating, expanding, and contracting the major facilities used in producing the output. Note the interdependence of the capacity decision with the location decision. Every capacity decision requires a corresponding location decision. For example, expanding an existing facility defi nes the location of the new capacity to be an existing facility. This section covers capacity planning strategies in terms of facility size, economies of scale and scope, timing of capacity increments, and capacity for multiple outputs. The following section covers the location aspects and relationships.

Facility Size Planning

Figure 8.1 illustrates the issue of facility size in terms of capacity and unit cost. Product cost curves are shown for fi ve sizes of production facilities. When plants are operated at their lowest-cost production level (A, B, or C), the larger facilities will generally have the lowest costs, a phenomenon known as economies of scale . However, if pro- duction levels must be set at a value other than the lowest-cost level, the advantage of a larger facility may be lost. For example, point D is characterized by congestion

Figure 8.1 Envelope of lowest unit output costs with facility size.

Capacity

U n

it o

u tp

u t

co st

Very large facilities

Large facilities

Medium- size

facility

Small-size facility

A B

C E

D

F

G

Economies of scale Diseconomies of scale

c08.indd 303c08.indd 303 25/09/12 5:20 PM25/09/12 5:20 PM

304 C h a p t e r 8 : C a p a c i t y , S c h e d u l i n g , a n d L o c a t i o n P l a n n i n g

and excessive overtime, and point E by idle labor and low equipment utilization. Points F and G illustrate some of the diseconomies of scale, as described next.

Economies of Scale and Scope

Obtaining lower unit costs through the use of larger facilities is known as economies of scale . Primarily, the economy comes from spreading the required fi xed costs— land, administration, sales force, facilities, and such other factors—over a larger vol- ume of products or services, although there are also economies obtained through stronger purchasing power and learning curve effects (discussed in a later section). However, as illustrated by points F and G in Figure 8.1 , there is a limit to the benefi ts that can be obtained because the inherent ineffi ciencies of large facilities begin to counter their economic benefi ts. This occurs through increased bureaucracy, poor communication, long lines of responsibility and authority, and the like. Many manu- facturers now have a corporate policy that no plant will be larger than 200 to 250 workers, often considered an optimum size.

Managers frequently think in terms of economies of scale when making decisions about where to produce new products or services, or whether or not to extend their line of products and services. However, the focus lost through adding these new production requirements can jeopardize the competitive strength of a fi rm. Managers would be well advised to examine more closely where the economies are expected to come from: sometimes it is from higher volumes, sometimes from the use of com- mon technology, sometimes from availability of off-peak capacity. If the source of the economy results in offsetting diseconomies of scale, as a result of loss of focus or for other reasons, the fi rm should not proceed.

An allied concept related to the use of many advanced, fl exible technologies, such as programmable robots, is called economies of scope . The phrase implies that economies can also be obtained with fl exibility by offering variety instead of vol- ume. However, upon closer examination it is not clear why being fl exible offers any particular economies. The real reason for economies of scope derives from the same economies as those of scale—spreading fi xed costs among more products or serv- ices—but the scale is now obtained over many small batches of a wide variety of outputs, rather than large batches of only a few standard outputs.

Capacity Planning for Multiple Outputs

Realistically, organizations are not always expanding their capacity. We usually focus on this issue because we are studying fi rms in the process of growth, but even suc- cessful organizations often reduce their capacity. Major ways of contracting capacity are to divest the fi rm of operations, lay off workers, outsource functions, and sell or lease equipment and facilities. Most organizations, however, try to contract only capacity that is ineffi cient or inappropriate for their circumstances, owing in part to a felt responsibility to the community. If it appears that organizational resources are going to be excessively idle in the future, organizations often attempt to add new outputs to their current output mix rather than contracting capacity (the latter fre- quently being done at a loss). This entails an analysis of the candidate output ’s life and seasonal demand cycles.

It is traditional in fi re departments to use the slack months for performing build- ing inspections, conducting fi re prevention programs, giving talks on safety, and

c08.indd 304c08.indd 304 25/09/12 5:20 PM25/09/12 5:20 PM

305L o n g - T e r m C a p a c i t y P l a n n i n g

other such activities. The large investment in labor and equipment is thus more effectively utilized throughout the year by adoption of an anticyclic output (an output counter to the fi re cycle)—fi re prevention. For many of the same reasons, many fi re departments have been given the responsibility for the city ’s or county ’s medical rescue service (although rescue alarms are not entirely anticyclic to fi re alarms).

Clearly, many organizations, such as the makers of greeting cards, winter coats, swimming pool equipment, and fi reworks, face this cyclic diffi culty. A classic case of seasonality is that of furnace dealers. For the last 100 years all their business typically was in the late autumn and winter, as illustrated in Figure 8.2 . With the rapid accept- ance of air conditioning in the 1950s and 1960s, many furnace dealers eagerly added this product to their output mix. Not only was it conceptually along the same lines (environmental comfort) and often physically interconnected with the home furnace but, most important, it was almost completely anticyclic to the seasonal heating cycle. As shown in Figure 8.2 , the addition of air conditioning considerably leveled dealers ’ sales throughout the year in comparison with furnace sales alone.

It is important to note that seasonality need not refer solely to the seasons of the year but can equally relate to daily, weekly, or monthly cycles, too. For example, fi re departments are also well aware of the high demand for fi refi ghting during the 3 PM to 9 PM period of the day between when children get out of school and when they go to bed.

In a similar manner, and for much the same reasons, organizations add to their mix outputs that are anticyclic to existing output life cycles . Figure 8.3 illustrates the expected life cycles of an organization ’s current and projected outputs. Total required capacity is found by adding together the separate capacities of each of the required outputs. Note the projected dip in required capacity fi ve years in the future, and, of course, beyond the eight-year R&D planning horizon.

The message of Figure 8.3 should be clear to the organization—an output with a three-year life cycle (appearing similar to the shaded area) is needed between years

Figure 8.2 Anticyclic product sales.

Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec.

100

200

300

400

500

600

Total Air conditioners Furnaces

Sa le

s (u

n it

s)

c08.indd 305c08.indd 305 25/09/12 5:20 PM25/09/12 5:20 PM

306 C h a p t e r 8 : C a p a c i t y , S c h e d u l i n g , a n d L o c a t i o n P l a n n i n g

4 and 7 in order to maintain effi cient utilization of the organization ’s available capac- ity. A priority output development program will have to be instituted immediately. At this point it is probably too late to develop something through R&D; a more effective strategy, especially in light of the relatively low volume and short life cycle, might be an extension of an existing output.

Timing of Capacity Increments

Once the best alternative for obtaining the desired capacity has been determined, the timing and manner must still be chosen. A number of approaches are illustrated in Figure 8.4 . Sometimes there is an opportunity to add capacity in small increments (Figure 8.4 a ) rather than as one large chunk (Figure 8.4 b ), such as an entire plant.

Clearly, small increments are less risky, but they do not offer an opportunity to upgrade the entire production system at one time, as a single chunk does. Other choices are to add capacity before (Figure 8.4 c ) or after (Figure 8.4 d ) the demand has arisen. Adding capacity before demand occurs upstages the competition and enhances customers ’ loyalty but risks the cost of the capacity if the expected demand never materializes. Adding capacity after demand arises encourages the competition to move into the market and take away part of your share. Clearly, the most appropriate strategy must be carefully evaluated for the situation at hand.

Figure 8.3 Forecast of required organizational capacity from multiple life cycles.

Total current output Additional required output between years 4 and 7

Capacity requirements for six different outputs

Now 1 2 3 4 5 6 7 8 9 10

R eq

u ir

ed c

ap ac

it y

Year

c08.indd 306c08.indd 306 25/09/12 5:20 PM25/09/12 5:20 PM

307L o c a t i o n P l a n n i n g S t r a t e g i e s

L O C A T I O N P L A N N I N G S T R A T E G I E S Having determined capacity requirements, we next discuss the most economical way to obtain the inputs needed to produce and deliver the output to the customer. This includes determining the location of the facility relative to suppliers and potential customers. Although we are discussing capacity and location planning sequentially, as noted earlier, these decisions are typically considered simultaneously since every capacity decision requires a location decision (e.g., where to add the new capacity or which plant should be closed).

In general, the decision about location is divided into three stages: regional (including international), community, and site. Sources of information for these stages are chambers of commerce, realtors, utilities, banks, suppliers, transportation companies, savings and loan associations, government agencies, and management consultants who specialize in relocation. For some pure service organizations

Figure 8.4 Methods of adding fi xed capacity.

Time Small capacity increments

U n

it s

Capacity Demand

(a)

Time Large capacity increments

U n

it s

(b)

Time Preceding demand

U n

it s

(c)

Time Following demand

U n

it s

(d)

Si n

gl e

in cr

em en

t

Capacity additions

c08.indd 307c08.indd 307 25/09/12 5:20 PM25/09/12 5:20 PM

308 C h a p t e r 8 : C a p a c i t y , S c h e d u l i n g , a n d L o c a t i o n P l a n n i n g

(e.g., physicians), only the site selection stage may be relevant because they are already focused on a specifi c region and community. Before discussing these stages in detail, however, we fi rst highlight the relationship between the location decision and the development of core capabilities.

Developing Capabilities and the Location Decision In examining the rationale offered by organizations regarding their decisions to relo- cate existing facilities or open new ones, it often appears that these decisions are being driven primarily by short-term considerations such as differentials in wage rates and fl uctuations in exchange rates. In addition to having the appearance of being more band-aid solutions than addressing how to improve long-term competitiveness, these decisions are often dominated by operational factors such as wage rates and transportation costs. The problem with such static and one-dimensional analyses is that conditions change. For example, if one competitor chooses a location based on low wage rates, there is very little to prevent its competitors from locating in the same region. Furthermore, the benefi t of low wages is likely to be short-lived, as the demand for labor will increase when more organizations locate in the region.

An alternative approach to the location decision is to consider the impact these decisions have on the development of key organizational capabilities. In Chapter 1 we defi ned core capabilities as the organizational practices and business processes that distinguish an organization from its competition. Clearly, the way various organ- izational units are located relative to one another can have a signifi cant impact on interactions between these units, which in turn impacts the development of core capabilities.

In order to leverage the location decision to enhance the development of long- term capabilities, Bartmess and Cerny (1993) suggest the following six-step process:

1. Identify the sources of value the company will deliver to its customers. In effect, this translates into identifying the order winners discussed in Chapter 1.

2. Once the order winners have been defi ned, identify the key organizational capabilities needed in order to have a competitive advantage.

3. Based on the capabilities identifi ed, assess the implications for the location of organizational units. For example, if the company determines that a rapid product development capability is needed, then it follows that design needs to be in close contact with manufacturing and leading-edge customers. Alternatively, if operational fl exibility is needed, then it follows that manufac- turing needs to be in close proximity to design, customers, marketing, and management information systems.

4. Identify potential locations.

5. Evaluate the sites in terms of their impact on the development of capabilities, as well as on fi nancial and operational criteria.

6. Develop a strategy for building an appropriate network of locations.

Having highlighted the relationship between the choice of a location and the development of capabilities, we next turn our attention to the actual stages that loca- tion decisions typically progress through.

c08.indd 308c08.indd 308 25/09/12 5:20 PM25/09/12 5:20 PM

309L o c a t i o n P l a n n i n g S t r a t e g i e s

Stage 1: Regional–International

In the regional–international stage, an organization focuses on the part of the world (e.g., North America, Europe, Pacifi c Rim) or perhaps the region of a country (e.g., Southwest, Midwest, Northeast) in which it wants to locate its new facility. For exam- ple, when Mercedes-Benz needed a new facility to produce its new multipurpose vehicle (MPV), it initially decided that its new facility should be located in North America and subsequently further narrowed the region to sites in the southeastern United States. There are four major considerations in selecting a national or overseas region for a facility: proximity, labor supply, availability of inputs , and environment .

To minimize transportation costs and provide acceptable service to customers, the facility should be located in a region in close proximity to customers and suppliers. Although methods of fi nding the location with the minimum transportation costs will be presented later in this chapter, a common rule of thumb in the United States is that the facility should be within 200 miles of major industrial and commercial customers and suppliers. Beyond this range, transportation costs begin to rise quickly.

The region should have the proper supply of labor available and in the correct proportions of required skills. One important reason for the past expansion of American fi rms abroad, particularly to Japan in the 1980s, was the availability of labor there at wage rates much lower than rates at home. Currently, this disparity has been eliminated because of Japan ’s increased wages. However, the real consid- eration should not be wage rates, but rather the productivity of domestic labor rela- tive to productivity abroad. This comparison would thus involve considering level of skills, use of equipment, wage rates, and even work ethics (which differ even between regions within the United States) to determine the most favorable labor supply in terms of output per dollar of wages and capital investment. The organiza- tion of the labor pool should also be given consideration—that is, whether all the skills are unionized or whether there is an open shop. Some states have passed right-to-work laws that forbid any requirement that all employees join a union in order to work in an organization. Often, these laws result in signifi cantly lower wage rates in these states.

The region selected for location of the facility should have the necessary inputs available. For example, supplies that are diffi cult, expensive, or time-consuming to ship and those that are necessary to the organization (i.e., no reasonable substitutes exist) should be readily available. The proper type (rail, water, highway, air) and supply of transportation; suffi cient quantities of basic resources such as water, elec- tricity, gas, coal, and oil; and appropriate communication facilities should also be available. Obviously, many American industries are located abroad in order to use raw materials (oil, copper, etc.) available there.

The regional environment should be conducive to the work of the organization. Not only should the weather be appropriate, but the political, legal, and social cli- mate should also be favorable. The following matters should be considered:

1. Regional taxes

2. Regional regulations on operations (pollution, hiring, etc.)

3. Barriers to imports or exports

4. Political stability (nationalization policies, kidnappings)

5. Cultural and economic peculiarities (e.g., restrictions on working women)

c08.indd 309c08.indd 309 25/09/12 5:20 PM25/09/12 5:20 PM

310 C h a p t e r 8 : C a p a c i t y , S c h e d u l i n g , a n d L o c a t i o n P l a n n i n g

These factors are especially critical in locating in a foreign country, particularly an underdeveloped country. Firms locating in such regions should not be surprised to fi nd large differences in the way things are done. For example, in some countries gov- ernmental decisions tend to move slowly, with extreme centralization of authority. Very little planning seems to occur. Events appear to occur by “God ’s will” or by default. The pace of work is unhurried, and at times discipline, especially among man- agers, seems totally absent. Corruption and payoffs often seem to be normal ways of doing business, and accounting systems are highly suspect. Living conditions for the workers, especially in urbanized areas, are depressing. Transportation and communi- cation systems (roads, ports, phone service) can be incomplete and notoriously unreli- able. Attempting to achieve something under such conditions can, understandably, be very discouraging. When locating in such countries, a fi rm should allow for such dif- fi culties and unexpected problems. In such an environment, Murphy ’s law thrives.

With the escalating use of outsourcing, and especially offshoring, the roles of loca- tion and capacity in the competitive elements of a fi rm ’s strategy take on increased importance. By subcontracting production to another fi rm, an organization can often save substantially on labor costs (especially when offshoring) and at the same time reduce its own asset base tremendously, thereby increasing both its profi t margins and return on assets (ROA). Contract manufacturers such as Flextronics, Selectron, and Jabil Circuit are quick to point out these advantages and others, such as leaving the organiza- tion free to concentrate on its strengths, such as design, brand building, marketing, and strategy. There are, however, also disadvantages in both outsourcing and offshoring. One is the loss of control of the product. Another is a probable reduction in speed of response to customers. A third, which is especially sensitive in communities and is increasingly publicized by the media, is the loss of domestic jobs when the company outsources its work. And outsourcing production is always a dangerous action for two reasons: (1) Engineering and then design typically must follow production overseas, meaning the additional loss of these capabilities within the organization. (2) There is the increased potential that the fi rm is simply training a powerful competitor (especially if engineering and design have also been outsourced), thereby “hollowing itself out.” In the 1980s, many fi rms in the TV and VCR industries outsourced all their production overseas, simply slapping on their own logo to sell their product domestically. Then the foreign producers started introducing their own brands and all the formerly domestic producers went out of business, losing the entire industry to foreign competition.

A model to help make the regional–international location decision is the CVD model we described for helping with the job shop layout in Chapter 2. In this case, we are interested in the total of two sets of costs: the costs of bringing all the materials into the facility and the distribution costs of shipping the fi nal products out to customers. The procedure is to select some initial site for the facility that appears to be good and then sum the products of the transportation rate ( C ), the volume or weights ( V ), and the distance ( D ) over all the locations. Then we can simply consider placing the facility in another site and see if the cost is less, and so on. If all the sites are prespeci- fi ed, then the site with the lowest cost is deemed best (at least on this one measure).

Stage 2: Community

After the region of a new facility has been selected, candidate communities within the region are identifi ed for further analysis. Many of the considerations made at the regional–international stage should also be considered at this next stage. For example,

c08.indd 310c08.indd 310 25/09/12 5:20 PM25/09/12 5:20 PM

311L o c a t i o n P l a n n i n g S t r a t e g i e s

the availability of acceptable sites, attitudes of the local government, regulations, zon- ing, taxes, labor supply, the size and characteristics of the market, and the weather would again be considered. In addition, the availability of local fi nancing, monetary inducements (such as tax incentives) for establishing operations there, and the com- munity ’s attitude toward the organization itself would be additional factors of interest to the organization.

Last, the preferences of the organization ’s staff should play a role in selecting a com- munity. These would probably be infl uenced by the amenities available in the commu- nity, such as homes, religious congregations, shopping centers, schools and universities, medical care, fi re and police protection, and entertainment, as well as local tax rates and other costs. Upper-level educational institutions may also be of interest to the organiza- tion in terms of opportunity for relevant research and development. For example, it was no coincidence that major IBM plants were located in Lexington, Kentucky, in Denver, Colorado, and in Austin, Texas, all of which are also sites of major state universities.

The standard “breakeven” or “cost–volume–profi t” model can be helpful for this stage of the location decision, except that there is no revenue line and there are multi- ple costs lines, each representing a different community ’s costs. We assume that the problem is to choose from among a set of predetermined communities, on the basis of a range of fi xed and variable costs rather than just distribution cost, as calculated by the CVD model just given. That is, distribution cost may be considered, but it is only one factor (perhaps fi xed, perhaps variable with output volume) among many that need to be considered to make a decision. Although the relevant factors for comparison between the communities may be known (e.g., labor costs, taxes, utility charges), their values may be uncertain, particularly if they are a function of the output rate of the facil- ity being located. The various alternatives for location are then compared by graphing total operating costs for each alternative at different levels of demand, as in Figure 8.5 .

Figure 8.5 Breakeven location model.

E

Output demand volume

Fixed costs community 2

Community 2 Community 1

T o

ta l

an n

u al

o p

er at

io n

c o

st

Fixed costs community 1

Variable costs community 1

c08.indd 311c08.indd 311 25/09/12 5:20 PM25/09/12 5:20 PM

312 C h a p t e r 8 : C a p a c i t y , S c h e d u l i n g , a n d L o c a t i o n P l a n n i n g

This is accomplished by dividing the total operating cost into two components— fi xed costs that do not vary with the demand for the output (e.g., land, buildings, equipment, property taxes, insurance) and variable costs such as labor, materials, and transportation—and plotting them on the axes of a graph. At the demand point E (the intersection of the two lines) the costs for the two alternatives are the same; for demand levels in excess of E , community 2 is best, and for levels less than E , community 1 is best. Thus, if the range of uncertainty concerning the output volume is entirely above point E , the manager need not be concerned about which commu- nity to choose—community 2 is best. Similar reasoning holds for any uncertainty existing entirely below point E —community 1 is best.

If the range of uncertainty is closely restricted to point E , then either community may be selected because the costs will be approximately the same in either case. However, if the range of uncertainty is broad and varies considerably from point E in both directions, then the breakeven chart will indicate to the manager the extra costs that will be incurred by choosing the wrong community. Before selecting either community, the manager should probably attempt to gather more information, in order to reduce the range of uncertainty in demand.

Stage 3: Site

After a list of candidate communities is developed, specifi c sites within them are identi- fi ed. The site —the actual location of the facility—should be appropriate to the nature of the operation. Matters to consider include size; adjoining land; zoning; community atti- tudes; drainage; soil; the availability of water, sewers, and utilities; waste disposal; trans- portation; the size of the local market; and the costs of development. The development of industrial parks in some communities has alleviated many of the diffi culties involved in choosing a site, since the developer automatically takes care of most of these matters. Before any fi nal decision is made, a cash-fl ow analysis is conducted for each of the can- didate sites; this includes the cost of labor, land, taxes, utilities, transportation, and so on.

A model that can help with the site selection is the weighted-score model . This approach can combine cost measures, profi t measures, other quantitative measures, and qualitative measures to help analyze multiple locations (as well as any other multicriteria decision). Deciding on a location, whether for products or services, is complicated by the existence of multiple criteria such as executives ’ preferences, maximization of facility use, and customers ’ attitudes. These and other criteria may be very diffi cult to quantify or even to measure qualitatively; if they are important to the decision, however, they must be included in the location analysis.

Locations can be compared in a number of ways. The most common is probably just managerial intuition: which location best satisfi es the important criteria? The weighted-score model is a simple formalization of this intuitive process that is useful as a rough screening tool for locating a single facility. In this model a weight is assigned to each factor ( criterion ), depending on its importance to the manager. The most important factors receive proportionately higher weights. Then a score is assigned to each of the alternative locations on each factor, again with higher scores representing better results. The product of the weights and the scores then gives a set of weighted scores, which are added up for each location. The location with the highest weighted score is considered best. In quantitative terms:

Total weighted score   5   ∑ i W

i S

i

c08.indd 312c08.indd 312 25/09/12 5:20 PM25/09/12 5:20 PM

313L o c a t i o n P l a n n i n g S t r a t e g i e s

where

i 5 index for factors W

i 5 weight of factor i

S i 5 score of the location being evaluated on factor i

The following example illustrates the method.

Communicable Disease Center

A provincial health department is investigating three possible locations for a special- ized control clinic that will monitor acquired immune defi ciency syndrome (AIDS) and other sexually transmitted diseases (STDs). The director of public health for the province is particularly concerned with four factors.

1. The most important consideration in the treatment of STDs is ease of access for those infected. Since they are generally disinclined to recognize their infection and seek treatment, it is foolish to locate a clinic where it is not eas- ily accessible to as many patients as possible. This aspect of location is prob- ably as much as 50 percent more important than the lease cost of the building.

2. Still, the annual cost of the lease is not a minor consideration. Unfortunately, the health department is limited to a very tight budget, and any extra cost for the lease will mean that less equipment and staff are available to the clinic.

3. For some patients it is of the utmost importance that confi dentiality be main- tained. Thus, although the clinic must be easily accessible, it must also be relatively inconspicuous. This factor is probably just as important as the cost of the lease.

4. The director also wants to consider the convenience of the location for the staff, since many of the physicians will be donating their time to the clinic. This consideration is the least important of all, perhaps only half as impor- tant as the cost of the lease.

The three locations being considered are a relatively accessible building on Adams Avenue, an inconspicuous offi ce complex near the downtown bus terminal, and a group of public offi ces in the Civic Center, which would be almost rent-free.

The director has decided to evaluate (score) each of these alternative locations on each of the four factors. He has decided to use a 4-point scale on which 1 represents “poor” and 4 represents “excellent.” His scores and the weights (derived from the relative importance of the four factors) are shown in Table 8.2 . The problem now is somehow to use this information to determine the best location for the clinic.

To determine the weighted score for each location, we multiply each score by the weight for that factor and then the sum over all factors for each location, as illustrated in Table 8.3 . Since higher scores indicate better ratings, the location with the largest score—B, the offi ce near the bus terminal—is best, followed by C, the Civic Center.

Quebec City, Canada, provides a good example of almost exactly this process (Price and Turcotte 1986). The Red Cross Blood Donor Clinic and Transfusion Center of Quebec City was located in a confi ned spot in the downtown area and wanted to

c08.indd 313c08.indd 313 25/09/12 5:20 PM25/09/12 5:20 PM

314 C h a p t e r 8 : C a p a c i t y , S c h e d u l i n g , a n d L o c a t i o n P l a n n i n g

expand in another location. The center ’s main activities affecting the choice of a new location were receiving donors, delivering blood and blood products throughout the community and the province of Quebec, and holding blood donor clinics across the same region.

Accordingly, the following criteria for a site were identifi ed:

• Highway access for both clinics and blood deliveries • Ability to attract more donors as a result of improved accessibility and

visibility

• Convenience to both public and private transportation • Ease of travel for employees • Internal fl oor space • Lot size • Acceptability of the site to management and governmental authorities

involved in the decision

The analysis of the problem was very complicated, owing to confl icting require- ments and the unavailability of data. Nevertheless, fi ve sites were fi nally identifi ed and evaluated on the basis of four fi nal criteria. The fi ve sites were then ranked on each of these criteria, and a scoring model was constructed to help management

T A B L E 8 .2 • Po ten t ia l C l in i c S i t e s Potential Locations

F: Factor W: Weight A: Adams Avenue

B: Bus Terminal Complex

C: Civic Center

1. Annual lease cost 2 1 3 4

2. Accessibility for patients 3 3 3 2

3. Inconspicuousness 2 2 4 2

4. Accessibility for personnel 1 4 1 2

Note: Factor scoring scale: 1, poor; 2, acceptable; 3, good; 4, excellent.

T A B L E 8 .3 • Compar i son o f S i t e Fac tor s by the We igh ted Score Me thod

Sites

Factor Weight A B C

1 2 2   3   1   5   2 2   3   3   5   6 2   3   4   5   8

2 3 3   3   3   5   9 3   3   3   5   9 3   3   2   5   6

3 2 2   3   2   5   4 2   3   4   5   8 2   3   2   5   4

4 1 1   3   4   5   4 1   3   1   5   1 1   3   2   5   2

Total 19 24 20

c08.indd 314c08.indd 314 25/09/12 5:20 PM25/09/12 5:20 PM

315L o c a t i n g P u r e S e r v i c e s

L O C A T I N G P U R E S E R V I C E S Although all the material presented so far applies equally to services and product fi rms, some aspects of locating service organizations are worth noting. First, service location decisions are usually based on how the location will help increase the organization ’s service revenues, with particular attention paid to avoiding poor loca- tions, which can be fatal to some services. Since the majority of services are highly dependent on physical interaction with the customer, the most important factor in service location is being close to and easily accessible by customers. And if the serv- ice relocates, it does not want to move too far from its original location. The second major factor is usually access to qualifi ed labor at a reasonable cost. Then come vari- ous other factors such as rent, infrastructure, business climate, competition, and so on. There are some exceptions, such as competitive clustering (auto dealers, motels) and saturation marketing (Walgreens, Starbucks).

There are various approaches to analyzing service locations, depending on some distinctions such as whether the issue is locating a single facility or multiple facilities. Another distinction that we will look at in detail here involves the recipient coming to the facility, as in retailing, as opposed to the facility going to the recipient, as with “alarm” services.

Recipient to Facility In recipient-to-facility situations, the facility draws customers or recipients from an area surrounding it, possibly in competition with other, similar facilities. Research has found that under these circumstances the drawing power of retail facilities is proportional to the size of the facility and inversely proportional to the square (or cube, in some cases) of the average recipient ’s travel time. This assumes that all other factors—such as price and quality—are equivalent or insignifi cant. This type of relationship is known as a gravity method because, like gravity, it operates by drawing nearby objects in.

Next, consider the situation of public services such as health clinics, libraries, and colleges. Apart from the diffi culty of framing a location model is the probably more

determine the best location. The weights were to be determined by management, and they could be modifi ed to determine if changing them would have any effect on the best location. The fi nal scores and rankings, assuming equal weights across the four criteria, are shown in Table 8.4 .

T A B L E 8 .4 • Compar i son o f Quebec C i t y ' s S i t e Fac tor s Site Road Access Bus Access Proximity Availability Rank

1 0.4 0.0 0.4 0.7 1

2 0.2 0.2 0.3 0.7 2

3 0.3 0.3 0.2 0.0 4

4 0.0 0.4 0.1 0.0 5

5 0.1 0.1 0.0 0.7 3

c08.indd 315c08.indd 315 25/09/12 5:20 PM25/09/12 5:20 PM

316 C h a p t e r 8 : C a p a c i t y , S c h e d u l i n g , a n d L o c a t i o n P l a n n i n g

signifi cant problem of choosing a measure, or measures, of service: number of recip- ients served (a “surrogate” measure), change in the recipient ’s condition (a direct measure of benefi t), quantity of services offered (another surrogate), and so on. Some measures recommended in the literature on health clinics, which can be used for trial-and-error procedures, are the following:

1. Facility utilization . Maximize the number of visits to the facilities.

2. Travel distance per citizen . Minimize the average distance per person in the region to the nearest clinic.

3. Travel distance per visit . Minimize the average distance per visit to the nearest clinic. No one measure has been found to work best for all cases of deciding on a location.

Facility to Recipient Facility-to-recipient situations are common among the urban “alarm” services: fi re, police, and ambulance. Again, the problem of measuring a service appropriately involves such factors as number of recipients served, average waiting time for serv- ice, value of property saved, and number of service facilities. Two general cases are encountered in this problem, whether a single- or multiple-facility service is being located:

1. High-density demand for services, where multiple vehicles are located in the same facility and vehicles are often dispatched from one alarm directly to another

2. Widely distributed demand for services, where extreme travel distances require additional facilities

Typical of situation 1 are fi re companies and ambulances. Results in these cases have been basically the same. There is a signifi cant drop-off in the returns to scale as more units are added to the system. Typically, the fi rst three or four will improve all measures by up to 80 percent of the maximum improvement. Each additional unit gains less. A second common fi nding is that optimally located facilities yield only about a 15 percent improvement over existing or evenly dispersed facilities. Last, incremental approaches to selecting additional locations provide slightly poorer service than a total relocation analysis of all the facilities.

E F F E C T I V E L Y U T I L I Z I N G C A P A C I T Y T H R O U G H S C H E D U L E M A N A G E M E N T

An important aspect of capacity worth emphasizing is its close tie to scheduling. That is, poor scheduling may result in what appears to be a capacity problem, and a short- age of capacity may lead to constant scheduling diffi culties. Thus, capacity planning is closely related to the scheduling function. The difference is that capacity is oriented primarily toward the acquisition of productive resources, whereas scheduling concerns the timing of their use. However, it is often diffi cult to separate the two, especially

c08.indd 316c08.indd 316 25/09/12 5:20 PM25/09/12 5:20 PM

317E f f e c t i v e l y U t i l i z i n g C a p a c i t y t h r o u g h S c h e d u l e M a n a g e m e n t

Figure 8.6 shows the resource requirements of the two jobs plotted along a time scale. Such a chart, called a Gantt chart , can be used to show time schedules and capacities of facilities, workers, jobs, activities, machines, and so forth. In Figure 8.6 a , each job was scheduled on the required resource as soon as it fi nished on the previous resource, whether or not the new resource was occupied with the other job. This infeasible schedule is called infi nite loading because work is scheduled on the resource as if it had infi nite capacity to handle any and all jobs. Note that in this way capacity confl icts and possible resolutions can be easily visualized. Shifting the jobs to avoid such confl icts—called fi nite loading —gives the longer but feasible schedule shown in Figure 8.6 b .

The fi rst resource confl ict in Figure 8.6 a occurs at 20 hours, when job 1 fi nishes on resource C and next requires resource A, which is still working on job 2. The second confl ict, again at A, occurs at 35 hours, and the third, on B, at 50 hours. It is quickly found that deferring one job for the other has drastic consequences for con- fl icts of resources later (sometimes adding confl icts and sometimes avoiding them) as well as for job completion times. Another consideration, not specifi ed here, is whether an operation can be stopped for awhile and then restarted (called operation splitting )—for example, to let another job pass through (called preemption )—or, once started, must be worked on until completion. If splitting were allowed for job 2, we could have stopped work at resource A on job 2 at 20 hours to let job 1 begin and then fi nished the work starting at time 50 when job 1 was fi nished on resource

where human resources are involved, such as in the use of overtime or the overlapping of shifts.

As a simple example, suppose that an organization has to complete the two jobs shown in Table 8.5 within two weeks. The table shows the sequential processing operations still to be completed and the times required. (The operations resources may be of any form—a facility, a piece of equipment, or an especially skilled worker.) In total, 60 hours of resource A are needed, 45 hours of B, and 25 hours of C. It would appear that two weeks (80 hours) of capacity on each of these three resources would be suffi cient, and additional capacity would, therefore, be unnecessary.

T A B L E 8 .5 • Sequent ia l Opera t ions Requi red fo r Two Jobs Job Operations Resource Needed Time Required (hours)

1 A 10

C 10

A 30

B 20

C 5

2 B 15

A 10

C 10

A 10

B 10

c08.indd 317c08.indd 317 25/09/12 5:20 PM25/09/12 5:20 PM

318 C h a p t e r 8 : C a p a c i t y , S c h e d u l i n g , a n d L o c a t i o n P l a n n i n g

A. Such operation splitting allows fl exibility for rush work but hurts productivity because machines must be taken down and set up multiple times for the same job.

Schedule Management In most organizations a department (or an individual) is specifi cally responsible for scheduling operations. In product organizations, this function is frequently called pro- duction planning and control or some similar name. The breadth of this department ’s responsibility varies considerably; for example, it may consist only of planning gross output levels or may include all the scheduling activities illustrated in Figure 8.7 .

0 8 16 24 32 40 48 56 64 72 80 88

C

B

A

R es

o u

rc e

Time (hr) (a) Infeasible

Job 1

Job 2

0 8 16 24 32 40 48 56 64 72 80 88

C

B

A

R es

o u

rc e

Time (hr) (b) Feasible

Figure 8.6 Gantt charts for capacity planning and scheduling.

c08.indd 318c08.indd 318 25/09/12 5:20 PM25/09/12 5:20 PM

319E f f e c t i v e l y U t i l i z i n g C a p a c i t y t h r o u g h S c h e d u l e M a n a g e m e n t

This fi gure does not describe a standardized scheduling system, such as might exist in an available computer package; rather, it shows a complex set of activities and terms that are often grouped under scheduling . Many of these have become major activities only since the advent of computerized scheduling. Before that, they were simply a matter of individual judgment (as some of them still are).

The foundation that supports scheduling is, in most cases, the forecast of demand for the upcoming planning horizon. In some industries, however, only minimal fore- casting is needed because customers place orders a year or more ahead of the time when the output will be needed. For example, in the airframe industry, airlines may

Figure 8.7 Relationship of scheduling activities.

Aggregate plan

Demand forecast

Master schedule

Rough-cut capacity planning

Priority planning

Capacity requirements planning (CRP)

Loading

Sequencing

Detailed scheduling

Dispatching

Expediting

Actual orders

Actual orders

Planning

Implementation

Production plan

c08.indd 319c08.indd 319 25/09/12 5:20 PM25/09/12 5:20 PM

320 C h a p t e r 8 : C a p a c i t y , S c h e d u l i n g , a n d L o c a t i o n P l a n n i n g

place orders years ahead of time because of long lead times and backlogs of orders. In these situations, organizational operations are scheduled on the basis of actual orders instead of forecasted demand.

Most organizations do not operate in such a favorable environment, however, and their success often hinges on the accuracy of their forecasts of demand. In these cases the concepts and techniques of forecasting briefl y overviewed in the chapter supplement are especially relevant for scheduling.

It might be noted that forecasts over different periods are used for different pur- poses. For example, long-range forecasts (i.e., two to fi ve years) are used for facility and capacity planning.

The aggregate plan is a preliminary, approximate schedule of an organization ’s overall operations that will satisfy the demand forecast at minimum cost. The plan- ning horizon , the period over which changes and demands are taken into consid- eration, is often one year or more and is broken into monthly or quarterly periods. This is because one of the purposes of aggregate planning is to minimize the effects of shortsighted, day-to-day scheduling, in which small amounts of material may be ordered from a supplier and workers laid off one week, only to be followed by reor- dering more material and rehiring the workers the next week. By taking a longer- term perspective on the use of resources, short-term changes in requirements can be minimized with a considerable saving in costs.

In minimizing short-term variations, the basic approach is to work only with “aggregate” units (i.e., units grouped or bunched together). Aggregate resources are used (such as total number of workers, hours of machine time, and tons of raw mate- rials) as well as aggregate units of output (gallons of product, hours of service deliv- ered, number of patients seen, and so on), totally ignoring the fact that there may be differences between these aggregated items. In other words, neither resources nor outputs are broken down into more specifi c categories; that occurs at a later stage.

The result of managerial iteration and changes to the aggregate plan is the organi- zation ’s formal production plan for the planning horizon used by the organization (e.g., one year). Sometimes this plan is broken down (i.e., disaggregated ) one level into major output groups (still aggregated)—for example, by models but not by colors. In either case, the production plan shows the resources required and changes in out- put over the future: requirements for hiring, limitations on capacity, relative increases and decreases in inventories of materials, and output rate of goods or services.

The driving force behind scheduling is the master schedule, also known in indus- try as the master production schedule (MPS). There are two reasons for this:

1. It is at this point that actual orders are incorporated into the scheduling system.

2. This is also the stage where aggregate planned outputs are broken down into individual scheduled items that customers actually want. These items are then checked for feasibility against lead time (time to produce or ship the items) and operational capacity (whether there is enough equipment, labor, etc.).

The actual scheduling is usually iterative, with a preliminary schedule being drawn up, checked for problems, and then revised. After a schedule has been deter- mined, the following points are checked:

• Does the schedule meet the production plan? • Does the schedule meet the demand forecasts?

c08.indd 320c08.indd 320 25/09/12 5:20 PM25/09/12 5:20 PM

321E f f e c t i v e l y U t i l i z i n g C a p a c i t y t h r o u g h S c h e d u l e M a n a g e m e n t

• Are there confl icts in the schedule involving capacity? • Does the schedule violate any other constraints regarding equipment, lead

times, supplies, facilities, and so forth?

• Does the schedule conform to organizational policy? • Does the schedule violate any legal regulations or organizational or union

rules?

Problems in any one of these areas may force a revision of the schedule and a repeat of the previous steps. The result is that the master schedule then specifi es what end items are to be produced in what periods to minimize costs and gives some meas- ure of assurance that such a plan is feasible . Clearly, such a document is of major importance to any organization—it is, in a sense, a blueprint for future operations.

As a part of checking the feasibility of the master schedule, a simple type of rough-cut capacity planning is conducted. One way of doing this, among many, is as follows. Historical ratios of workloads per unit of each type of product are used to determine the loads placed on the resources by all the products being made in any one period. Then the loads are assumed to fall on the resources in the same period as the demands; that is, the lead times are not used to offset the loads. If a resource ’s capacities are not overloaded (underloads are also checked), it is assumed that suffi cient capacity exists to handle the master schedule, and it is accepted for production.

The term priority planning relates not to giving priorities to jobs (a topic included under sequencing ), but rather to determining what material is needed when . For a master production schedule to be feasible, the proper raw materials, purchased mate- rials, and manufactured or purchased subassemblies must be available when needed, with the top priority going to immediate needs. The key to production planning is the “needed” date. Years ago, scheduling concentrated on launching orders , that is, on when to place the order. Priority planning concentrates on when the order is actually needed and schedules backward from that date. For example, if an item is needed on June 18 and requires a two-week lead time, then the order is released on June 4 and not before. Why store inventory needlessly?

The inventory control system and master schedule drive the capacity require- ments planning (CRP) system. This system projects job orders and demands for materials into requirements for equipment, workforce, and facility and fi nds the total required capacity of each over the planning horizon. That is, during a given week, how many nurses will be required, how many hours of a kidney machine, how many hours in operating rooms?

This may or may not exceed available capacity. If it is within the limits of capac- ity, then the master schedule is fi nalized, work orders are released according to schedule, orders for materials are released by the priority planning system, and load reports are sent to work centers, listing the work facing each area on the basis of the CRP system. Note that external lead times (often longer than internal lead times) from suppliers have already been checked at the stage of priority planning, so the master schedule can indeed now be fi nalized.

If the limits of capacity are exceeded, however, something must be changed. Some jobs must be delayed, or a less demanding schedule must be devised, or extra capacity must be obtained elsewhere (e.g., by hiring more workers or using over- time). It is the task of production planning and control to solve this problem.

c08.indd 321c08.indd 321 25/09/12 5:20 PM25/09/12 5:20 PM

322 C h a p t e r 8 : C a p a c i t y , S c h e d u l i n g , a n d L o c a t i o n P l a n n i n g

Loading means deciding which jobs to assign to which resources. Although the capacity planning system determines that suffi cient gross capacity exists to meet the master schedule, no actual assignment of jobs to resources is made. Some equip- ment will generally be superior for certain jobs, and some equipment will be less heavily loaded than other equipment. Thus, there is often a “best” (fastest or least costly) assignment of jobs to resources.

Even after jobs have been assigned to resources, the order in which to process the jobs—their sequencing —must still be determined. Unfortunately, even this seemingly small fi nal step can have major repercussions on the organization ’s workload capacity and on whether or not jobs are completed on time. A number of priority rules have been researched, and some interesting results are available in the literature.

Once all this has been specifi ed, detailed schedules itemizing specifi c jobs, times, materials, and workers can be drawn up. This is usually done only a few days in advance, however, since changes are always occurring and detailed schedules become outdated quickly. It is the responsibility of production planning and control to ensure that when a job is ready to be worked on, all the items, equipment, facili- ties, and information (blueprints, operations sheets, etc.) are available as scheduled.

All the previous activities constitute schedule planning; no production per se has taken place yet. Dispatching is the physical release of a work order from the produc- tion planning and control department to operations.

Once production planning and control has released a job to operations, the department usually has no more responsibility for it, and it is the production man- ager ’s task to get the job done on time. This task is known as expediting . When jobs fall behind schedule, managers have historically tended to use expediters to help push these “hot” jobs through the operations. Of course, expediting can be done more proactively by monitoring the progress of jobs to ensure that they stay on schedule.

Scheduling Services In this section we consider the scheduling of pure services. Much of what was said previously applies to the scheduling of services as well as products, but here we consider some scheduling issues of particular relevance to services.

Up to now we have dealt primarily with situations where the jobs (or recipients) were the items to be loaded, sequenced, or scheduled. There are, however, many operations for which scheduling of the jobs themselves is either inappropriate or impossible, and it is necessary to concentrate instead on scheduling one or more of the input resources. Therefore, the staff, the materials, or the facilities are scheduled to correspond, as closely as possible, with the expected arrival of the jobs. Such situ- ations are common in service systems such as supermarkets, hospitals, urban alarm services, colleges, restaurants, and airlines.

In the scheduling of jobs, we were primarily interested in minimizing the number of late jobs, minimizing the rejects, maximizing the throughput, and maximizing the utilization of available resources. In the scheduling of resources, however, there may be considerably more criteria of interest, especially when one of the resources being scheduled is staff. The desires of the staff regarding shifts, holidays, and work sched- ules become critically important when work schedules are variable and not all employees are on the same schedule. In these situations there usually exist schedules

c08.indd 322c08.indd 322 25/09/12 5:20 PM25/09/12 5:20 PM

323E f f e c t i v e l y U t i l i z i n g C a p a c i t y t h r o u g h S c h e d u l e M a n a g e m e n t

that will displease everyone and schedules that will satisfy most of the staff ’s more important priorities—and it is crucial that one of the latter be chosen rather than one of the former.

Approaches to Resource Scheduling

The primary approach to the scheduling of resources is to match availability to demand (e.g., 7 P.M. to 12 A.M. is the high period for fi re alarms). By so doing, we are not required to provide a continuing high level of resources that are poorly utilized the great majority of the time. However, this requires that a good forecast of demand be available for the proper scheduling of resources. If demand cannot be accurately predicted, the resulting service with variable resources might be worse than using a constant level of resources.

Methods of increasing resources for peak demand include using overtime and part-time help and leasing equipment and facilities. Also, if multiple areas within an organization tend to experience varying demand, it is often helpful to use fl oating workers or combine departments to minimize variability. On occasion, new tech- nologies, such as 24-hour automated tellers, 24-hour order entry via the Web, and bill payment by telephone, can aid the organization.

As mentioned previously, the use of promotion and advertising to shift demand for resources is highly practical in many situations. Thus, we see off-peak pricing in the utilities and communication industries, summer sales of snowblowers in retail- ing, and cut rates for transportation and tours both in off-peak seasons (fall, winter) and at off-peak times (weekends, nights). Let us now consider how some specifi c service organizations approach their scheduling problems.

Hospitals There are multiple needs for scheduling in hospitals. Although arrivals of patients (the jobs) are in part uncontrollable (e.g., emergencies), they are to some extent controllable through selective admissions for hernia operations, some mater- nity cases, in-hospital observation, and so on. With selective admissions, the hospital administrator can smooth the demand faced by the hospital and thereby improve service and increase the utilization of the hospital ’s limited resources.

Very specialized, expensive equipment such as a dialysis machine is also carefully scheduled to allow other hospitals access to it, thus maximizing its utilization. By shar- ing such expensive equipment among a number of hospitals, more hospitals have access to modern technology for their patients at a reasonable level of investment.

Of all the scheduling in hospitals, the most crucial is probably the scheduling of nurses, as illustrated in the following example describing Harper Hospital (Filley 1983). This is because (1) it is mandatory, given the nature of hospitals, that nurses always be available; (2) nursing resources are a large expense for a hospital; and (3) there are a number of constraints on the scheduling of nurses, such as number of days per week, hours per day, weeks per year, and hours during the day.

Like many other hospitals, Harper Hospital of Detroit was under heavy pressure from Blue Cross, Medicare, and Medicaid to provide more health care at less cost. In addition, it needed to achieve more economies of scale from a merger that had taken place some years before. It also wanted to improve its patient care. One target to help achieve these goals was a better system for scheduling nurses.

Previously, nurses were scheduled on the basis of strict bed counts, problems with inadequate staffi ng during the prior day, and requests for extra help. What was devel- oped was a patient classifi cation system (PCS) that incorporated labor standards

c08.indd 323c08.indd 323 25/09/12 5:20 PM25/09/12 5:20 PM

324 C h a p t e r 8 : C a p a c i t y , S c h e d u l i n g , a n d L o c a t i o n P l a n n i n g

to determine what levels of nursing were needed. At the end of each shift, designated nurses evaluated each area ’s patients by their condition and assigned them to a “care level” ranging from minimal to intensive. An hour before the next shift begain, the patients ’ needs for care were added up—accounting for new admissions, check-outs, and returns from surgery—to determine the total levels of care required. Given the levels in each area, nursing labor standards were used to determine how many nurses were needed on the next shift.

As a result of the new system, both the quality of patient care and the nurses ’ satisfaction went up. Annual labor savings from the new system were estimated as exceeding $600,000. Harper has further fi ne-tuned the PCS system and now recali- brates its standards every two years.

Urban Alarm Services In urban services that respond to alarms—such as police, fi re, and rescue services—the jobs (alarms) appear randomly and must be quickly serviced with suffi cient resources. Otherwise, extreme loss of life or prop- erty may result. In many ways this problem is similar to that of a hospital, since the cost of staffi ng personnel is a major expense, but fl oating fi re companies and police SWAT units may be utilized where needed, and some services (such as fi re inspec- tion) can be scheduled to help smooth demand.

Sometimes a major difference vastly complicates some of these services (particu- larly fi re): duty tours of extended duration, as opposed to regular shifts, run over multiple days. These tours vary from 24 to 48 hours in teams of two to four mem- bers. Common schedules for such services are “two (days) on and three off” and “one on and two off,” with every fi fth tour or so off as well (for a running time off, every three weeks, of perhaps 3   1   2   1   3   5   8 days ). Because living and sleeping-in are considered part of the job requirements, the standard workweek is in excess of 40 hours—common values are 50 and 54 hours. Clearly, the scheduling of such duty tours is a complex problem, not only because of the unusual duration of the tours but also because of the implications concerning overtime, temptations to “moon- light,” and other such issues.

Educational Services Colleges and universities have scheduling requirements for all types of transformations: intermittent (such as counseling), continuous (English 1), batch (fi eld trips), and project (regional conferences). In some of these situations the jobs (students) are scheduled; in some the staff (faculty, administrators) are sched- uled; and in others the facilities (classrooms, convention centers) are scheduled.

The primary problem, however, involves the scheduling of classes, assignment of students, and allocation of facilities and faculty resources to these classes. To obtain a manageable schedule, three diffi cult elements must be coordinated in this process:

1. Accurate forecast of students ’ demand for classes

2. Limitations on available classroom space

3. Multiple needs and desires of the faculty, such as

• Number of “preparations” • Number of classes • Timing of classes • Level of classes • Leave requirements (sabbatical, maternity, etc.) • Release requirements (research, projects, administration)

c08.indd 324c08.indd 324 25/09/12 5:20 PM25/09/12 5:20 PM

325E f f e c t i v e l y U t i l i z i n g C a p a c i t y t h r o u g h S c h e d u l e M a n a g e m e n t

Because of the number of objectives in such scheduling problems, a variety of multicriteria approaches have been used to aid in fi nding acceptable schedules, including simulation, goal programming, and interactive modeling.

In summary, the approach to scheduling services is usually to match resources and forecasted demand. Since demand cannot be controlled, it is impossible to build up inventory ahead of time, and backordering is usually not feasible. Careful sched- uling of staff, facilities, and materials is done instead, with (limited) fl exibility achieved through fl oating part-time and overtime labor and off-peak rates to encour- age leveling of demand. The best schedule is often not the one that optimizes the use of resources or minimizes lateness for the expected demand, but rather the one that gives acceptable results under all likely operating conditions. As described later in this chapter, an important aspect of scheduling services is the queues that tend to build up if capacity is inadequate. Here, queuing theory and psychology concerning waiting can be profi tably applied.

Yield/Revenue Management and Overbooking

Yield management , also called revenue management , is the attempt to allocate the fi xed capacity of a service (although the process is now being used by retailers and manufacturers, too) to match the highest revenue demand in the marketplace. It appears that American Airlines was one of the fi rst to develop this technique, but its use has spread to hotels, cruise lines, and other services that hold a fi xed capacity for revenue-producing customers, jobs, items, and so on. As described by Kimes (1989), yield management is most appropriate under the following circumstances:

1. Fixed capacity. There is only a limited, indivisible number of capacity openings available for the period. There is no fl exibility in either dividing up the capacity or in fi nding additional capacity.

2. Perishable capacity. Once the period passes, the capacity can no longer be used for that period. There is essentially no salvage value for the capacity.

3. Segmentable market. The demand for the capacity must be segmentable into different revenue/profi t classes, such as business versus pleasure, Saturday night stayover or not, deluxe versus budget, and so on.

4. Capacity sold in advance. The capacity is sold by reservation. Using yield management techniques, certain classes of capacity are held back for certain, more profi table classes of reservations or periods of the season. If the profi t- able classes fail to fi ll by a certain time point, some of the capacity is then released for the next-lowest profi t class. This procedure cascades down through both reservation classes and time points as the period in question approaches.

5. Uncertain demand. Although demand for each reservation class may be forecast, the actual demand experienced in each class for each time period is uncertain.

6. Low marginal sales cost, high marginal capacity addition cost. The cost to add a unit of capacity is extremely high, but the cost to sell (rent) a unit of it for the period in question is low.

c08.indd 325c08.indd 325 25/09/12 5:20 PM25/09/12 5:20 PM

326 C h a p t e r 8 : C a p a c i t y , S c h e d u l i n g , a n d L o c a t i o n P l a n n i n g

Suppose that a profi t of $50 is made for each passenger carried, but a cost is incurred if a passenger with a reservation has to be turned away. This cost could be a free ticket, ill will, passage on another airline, or whatever. If the cost is low—say, less than the profi t—then it will be to the airline ’s advantage to overbook quite a bit (although possibly not all the way to 32, since there would then be a 90 percent chance of having an overbooking cost). On the other hand, suppose that the cost is very high—much more than the profi t. Then the airline would be very reluctant to overbook much at all, out of fear of having to pay one or more costs of overbooking. Table 8.7 gives the probabilities of demand for each set of overbookings accepted, as specifi ed in the previous paragraph. Assuming that turning a passenger away costs the airline $20, how many reservations should be accepted? Suppose the cost is $100.

The technique used to determine how to allocate capacity among the different classes is similar to that used for overbooking . Overbooking is an attempt to reduce costs through better schedule management, as illustrated by Scandinavian Airlines (Alstrup et al. 1989). Scandinavian Airlines (SAS) operates a fl eet of DC-9 aircraft with 110 seats each. If SAS accepts reservations for only these 110 seats and “no- shows” (passengers who fail to show up for a fl ight) refuse to pay for their reserva- tions, SAS can lose from 5 to 30 percent of the available seats. If there are 100 fl ights every day, these no-shows can cost the airline as much as $50 million a year. To avoid this loss, all airlines overbook fl ights by accepting a fi xed percentage of reser- vations in excess of what is available.

The management of SAS decided to develop an automated overbooking system to include such factors as class, destination, days before departure, current reserva- tions, and existing cancellations. The objective of the system was to determine an optimal overbooking policy for the different classes on each fl ight, considering the costs of ill will, alternative fl ight arrangements, empty seats, and the upgrading or downgrading of a passenger ’s reserved class.

A number of interesting fi ndings were made in the process of conducting the study. For example, an early fi nding was that the probability that a reservation would be canceled was independent of the time the reservation was made. When the system was completed, it was tested against the heuristics used by experienced employees who had a good “feel” for what the overbooking rate should be. It was found that the automated system would increase SAS ’s net revenue by about $2 million a year.

To better understand the situation and demonstrate the approach to determining a solution, let us assume that the number of seats on a plane is fi xed at 28. How many reservations should the airline accept, given the chances described in Table 8.6 of no-shows? For example, if 32 reservations are accepted, the probability that only 30 passengers show up is 35 percent.

T A B L E 8 .6 • Demand fo r F l i gh t s Number of No-Shows Relative Frequency

4 0.10

3 0.20

2 0.35

1 0.25

0 0.10

1.00

c08.indd 326c08.indd 326 25/09/12 5:20 PM25/09/12 5:20 PM

327E f f e c t i v e l y U t i l i z i n g C a p a c i t y t h r o u g h S c h e d u l e M a n a g e m e n t

The process is repeated for 31 reservations, and the calculations are given in Table 8.9 . Continuing with 30, 29, and 28 reservations (it makes no sense to accept fewer than 28 reservations), we get the values shown in Table 8.10 . Clearly, the maximum profi t is obtained with 31 reservations. The results of the turnaway rate raised to $100 are shown in Table 8.11 . Now the highest profi t is obtained with 29 reservations.

Using the probabilities of no-shows (shown in Table 8.6 ), we can calculate the costs and profi ts according to Table 8.8 . (There is no sense in accepting more than 32 reserva- tions, because this will defi nitely fi ll the plane.) Here we see that the total profi t is $1359.

T A B L E 8 .7 • Demand Proba b i l i t i e s w i th Rese r va t ions ( f rom Ta b le 8 .6) Reservations

Relative Frequency 28 29 30 31 32

0.10 24 25 26 27 28

0.20 25 26 27 28 29

0.35 26 27 28 29 30

0.25 27 28 29 30 31

0.10 28 29 30 31 32

1.00

T A B L E 8 .8 • Expec t ed Pro f i t w i th 32 Rese r va t ions Demand

28 29 30 31 32 Total

Probabilities 0.10 0.20 0.35 0.25 0.10

Seats fi lled ( S ) 28 28 28 28 28

Profi t: $50 S 1400 1400 1400 1400 1400

Turnaways ( T ) 0 1 2 3 4

Cost: $20 T 0 20 40 60 80

Net profi t 1400 1380 1360 1340 1320

Expected net profi t 140 276 476 335 132 $1359

T A B L E 8 .9 • Expec t ed Pro f i t w i th 31 Rese r va t ions Demand

27 28 29 30 31 Total

Probabilities 0.10 0.20 0.35 0.25 0.10

Seats fi lled ( S ) 27 28 28 28 28

Profi t: $50 S 1350 1400 1400 1400 1400

Turnaways ( T ) 0 0 1 2 3

Cost: $20 T 0 0 20 40 60

Net profi t 1350 1400 1380 1360 1340

Expected net profi t 135 280 483 340 134 $1372

c08.indd 327c08.indd 327 25/09/12 5:20 PM25/09/12 5:20 PM

328 C h a p t e r 8 : C a p a c i t y , S c h e d u l i n g , a n d L o c a t i o n P l a n n i n g

S H O R T - T E R M C A P A C I T Y P L A N N I N G In the short term, capacity planning is primarily related to issues of scheduling, labor shifts, balancing of resource capacities, and other such issues instead of location decisions. We will look into a variety of such approaches in this section.

T A B L E 8 .10 • Expec t ed Pro f i t a t $20 Turnaway Cos t

Reservations Expected Profi ts

32 $1359

31 $1372 (best)

30 $1371

29 $1345.5

28 $1302.5

T A B L E 8 .11 • Expec t ed Pro f i t a t $100 Turnaway Cos t

Reservations Expected Profi ts

32 $1195

31 $1280

30 $1335

29 $1337.5 (best)

28 $1302.5

Process-Flow Analysis Earlier, we discussed some factors that might limit the output of a production sys- tem, such as bottlenecks in the system and yield considerations like scrap and defects. Here we will introduce some other terms relating to the use of a production

DILBERT: © Scott Adams/Dist. by United Feature Syndicate, Inc.

c08.indd 328c08.indd 328 25/09/12 5:20 PM25/09/12 5:20 PM

329S h o r t - T e r m C a p a c i t y P l a n n i n g

system. One capacity measure that is commonly used is utilization , which is simply the actual output relative to some expected, designed, or normal output rate. For example, if a machine runs 4 hours a day in an American plant and the maintenance and setup time are usually 2 hours a day, the utilization for that day might be reported as 4 / 6   5   67 % , which is considered to be fairly high utilization for a machine in a job shop. However, if the machine was in a Japanese plant, the utilization would probably be reported as 4 / 24   5   17 % , since the machine could, in theory, have been used for all 24 hours in the day!

Clearly, utilization fi gures do not mean much unless one knows what the “nor- mal” or expected output rate is based on. (When labor utilization measures are used for wage payment plans, this “normal” defi nition is often a heated subject of union negotiations. For example, should mandated “breaks” be included in the base or not? Should sick time be included? Lunch? Inactivity due to lack of materials to work on? And so on.) An advantage of basing the utilization on 24 hours is that this shows how much more could be done with the resource if it were needed. On the other hand, most managers would not like hearing that their expensive machinery was only being 17 percent utilized!

Bottlenecks in a Sequential Process

Another major concept in operations is that of effi ciency versus capacity ( output rate ). Effi ciency was defi ned earlier as output divided by input. Here we measure output as minutes of work embodied in the item being produced, and input as min- utes of resource time spent overall in producing the item. It is important to under- stand what production situations are amenable to simple capacity improvements by adding capital resources and what production situations are not. Normally, we expect that the amount of productive capacity and the capital investment needed to gain this capacity will be about proportional. Suppose blood samples are being ana- lyzed in a spectroscope run by one nurse (and both spectroscope and nurse are constantly busy with this task) at the rate of 10 per hour, and a capacity of 100 per hour is required. Then the resource investment translates directly—10 spectroscopes and 10 nurses will be needed.

However, if the production process is sequential , the resource investment may not translate so directly into the required output. Normally, many workers and machines are required to produce an output. In that case, the direct capacity– investment corre- spondence may not exist because of bottlenecks in the production process. Bottlenecks are places (there may be more than one) in the production process where production slows down because of a slow, or insuffi cient number of, machine(s), or perhaps because of a slow worker, or because the product needs to spend time drying. Fixing such bottlenecks usually only marginally improves the capacity of the system, however, because a new bottleneck arises somewhere else in the production proc- ess (this common problem is called “fl oating bottlenecks”). The following examples illustrate how this happens. For ease of understanding, we use a machine to illus- trate the bottleneck, but any kind of service operation could also be a bottleneck and often is in most services, but this is less obvious.

Assume that King Sports Products produces a variety of tennis rackets sequen- tially on four machines, and the times required on each machine for one typical racket are as shown in Figure 8.8 . Note in this fi gure that the work embodied in each fi nished tennis racket is 4   1   3   1   10   1   2   5   19 minutes , which is also the throughput

c08.indd 329c08.indd 329 25/09/12 5:20 PM25/09/12 5:20 PM

330 C h a p t e r 8 : C a p a c i t y , S c h e d u l i n g , a n d L o c a t i o n P l a n n i n g

time for a racket if there are no other rackets being made—that is, the system is not busy. However, during a busy production day, this throughput time will increase, as we will see next.

To minimize the cost of equipment, King could use one of each machine. The resulting capacity or output rate will then be based on the slowest machine ’s process- ing time of 10 minutes, resulting in 6 units per hour. That is, since every item must go through each of the machines, in order, every racket must wait for machine C, the bot- tleneck, to fi nish before it can proceed. During this wait, the fi rst, second, and fourth machines will be idle 6, 7, and 8 minutes, respectively, out of every 10-minute cycle.

(Also, the throughput time during such a busy period now becomes 10   1   10   1   10   1   2   5   32 minutes . The racket need not wait at the last machine to exit, but the machine must wait 8 minutes for the next racket.) Since the output in this pro- cess embodies 19 minutes of work, whereas the input consists of four machines that spend 10 minutes each during every cycle that produces an item (not all of which is necessarily productive), this gives an overall effi ciency of only 47.5 percent:

Effi ciency   5   output

______ input   5   4   1   3   1   10   1   2 ______________

4 ( 10 )   5   19 ___

40   5   47.5 %  

Note that it does not matter whether the bottleneck is at the end of the sequence, at the beginning, or in the middle. The process is still slowed, on average, to the out- put rate of the slowest machine. The capacity of this process is thus six units per hour, and the cycle time of the process is 10 minutes, or 1/6 of an hour—the output rate and the cycle time are always reciprocals. The process cycle time can be visualized as the time between items coming off the end of the production line, whereas the through- put time can be visualized as the time you would spend in the production process if you attached yourself to the item being produced and rode along through the pro- duction process—there is often no relationship between them! And the fi nal output work time is the productive time the item spends in the process.

If King is willing to invest in another, fi fth machine, it should purchase another machine of type C, since that is the bottleneck. Then it could run machines C1 and C2 concurrently and put out two units every 10 minutes, obtaining an “effective” machine C processing time of 5 minutes for the machines by staggering their pro- duction. Note that machine C is still the bottleneck in the production process, so this effective 5-minute machine processing time is once again the cycle time for the sys- tem. The effect of this single investment would be to double the capacity/output rate to 12 units per hour (5-minute cycle time) and increase the system effi ciency to

4   1   3   1   10   1   2 ______________ 5 ( 5 )

    5   19 ___ 25

  5   76 %  

Raw materials

Machine D Finisher

2 minutes

Machine C Stringer

10 minutes

Machine B Handler

3 minutes

Machine A Framer

4 minutes

Figure 8.8 King Sports product process.

c08.indd 330c08.indd 330 25/09/12 5:20 PM25/09/12 5:20 PM

331S h o r t - T e r m C a p a c i t y P l a n n i n g

Note from the table and fi gure that effi ciency of production does not always increase when machines are added, although the general trend is upward. This is because some systems are fairly well “balanced” to begin with. (For example, the cycles across the machines are quite even with seven machines—2, 3, 3.33, 2—and

Note in this effi ciency calculation that the work output per racket is always 19 minutes, regardless of the number of machines; only the input changes. Now the input is fi ve machines running at a 5-minute cycle time. Continuing in this manner results in the data shown in Table 8.12 and sketched in Figure 8.9 . In developing Table 8.12 , the next machine added was always the machine that currently had the longest machine time. For example, when there were six machines, machine A had the longest machine time. Thus, the seventh machine added was a machine A.

0 2 4 6 8 10 12 14 00

1020

2040

3060

4080

50100

Output Efficiency

H o

u rl

y o

u tp

u t

(u n

it s)

E ff

ic ie

n cy

( %

)

Number of machines

Figure 8.9 Effi ciency and output increase when machines are being added.

Number of Machines

Type of Next Machine A B C D

Cycle Time (min)

Total Hourly Output

Effi ciency (%)

4 — 4 3 10 2 10 6 47.5

5 C 4 3 5 2 5 12 76.0

6 C 4 3 3.33 2 4 15 79.2

7 A 2 3 3.33 2 3.33 18 81.4

8 C 2 3 2.5 2 3 20 79.2

9 B 2 1.5 2.5 2 2.5 24 84.4

10 C 2 1.5 2 2 2 30 95.0

11 D 2 1.5 2 1 2 30 86.0

12 A 1.33 1.5 2 1 2 30 79.2

13 C 1.33 1.5 1.67 1 1.67 36 87.5

14 C 1.33 1.5 1.43 1 1.5 40 90.5

Machine Times (min)

T A B L E 8 .12 • Re turn to K ing fo r Us ing More Mach ines

c08.indd 331c08.indd 331 25/09/12 5:20 PM25/09/12 5:20 PM

332 C h a p t e r 8 : C a p a c i t y , S c h e d u l i n g , a n d L o c a t i o n P l a n n i n g

even more so at 10 machines. Also note that the addition of only one extra machine at such points does not pay for itself.) If points of high effi ciency are reached “early” (as machines are added), these points will tend to be natural operating solutions. For example, a tremendous gain in effi ciency (and in output percentage) is reaped by adding a fi fth machine to the system. Further additions do not gain much. The next- largest gain occurs when the tenth machine is added to the system.

Although this analysis describes the general tradeoffs of the system, no mention has been made of demand. Suppose that demand is 14 units per hour. Then, to minimize risk but still keep an effi cient system, King might use fi ve machines and either work overtime, undersupply the market, or use a number of other strategies, as will be dis- cussed later. Similarly, for a demand of 25 to 35 per hour, the use of 10 machines would be appropriate.

Product and Service Flows

With the role of a bottleneck in a production process in mind, let us now consider the more general procedure of conducting a process-fl ow analysis, also known in service systems as “mapping” or “blueprinting.” The purpose of conducting a process-fl ow analysis is normally to identify bottlenecks, inventory buildup points, and time delays in a production process, which are crucially important in determining the capacity of the process. Standard nomenclature is to use rectangles for tasks/activities, triangles for storage or waiting points, diamonds for decision points, and arrows for fl ows. An activity changes the characteristics of a product or service, whereas a fl ow simply indicates the next step in the process, which may involve a change in position.

A simplifi ed process-fl ow diagram for a manufactured unit composed of two pur- chased parts and one fabricated 25-pound component is shown in Figure 8.10 . Demand is currently 120 units per 8-hour day or 15 units per hour, for an effective process cycle time of one unit every 60 / 15   5   4 minutes. The inputs, on the left, con- sist of 1.5 tons (i.e., 3000 lbs) of raw materials delivered by a 2-ton-capacity truck once a day and 240 parts delivered by a 300-part-capacity truck once a day, both of which immediately go into different storage facilities (with different capacities, as

Fabrication

25 lb/component 25 components/hr

Assembly

20 units/hr

Storage

3500-lb cap.

Storage

500 cap.

375 lb/hr

1.5 tons/ day

240 parts/ day

30 parts/hr

15 units/

hr

80 units/ day

40 units/ day

120 units/ day

Raw material 2-ton capacity

1 load/day

Parts 300 cap.

Packaging A 5 min. process time

Packaging B 10-min. process time

15 components/hr

Figure 8.10 Process fl ow for manufactured unit.

c08.indd 332c08.indd 332 25/09/12 5:20 PM25/09/12 5:20 PM

333S h o r t - T e r m C a p a c i t y P l a n n i n g

shown). The capacities of each stage in the production process are as labeled. Fabrication of the 15 hourly 25-pound components will require 15   3   25   5   375 pounds per hour of raw material from the storage facility. The fabricated compo- nents will then fl ow into assembly, along with 30 parts withdrawn per hour from the parts storage facility. Assembly then produces the 15 units per hour, which fl ow into two separate packaging lines with different processing times due to their age. New line A packages 10 units each hour while old line B packages 5 units an hour, for the required total demand. Although the output demand is currently 120 units a day, management anticipates an increase of perhaps as much as a third in the near future; their concern is whether the system can handle this increase in demand.

As we see from the diagram, there is currently excess capacity throughout the production system, but is there enough at each stage and process to handle the additional 0 . 333   3   120   5   40 units a day? Assembly, at 20 units per hour, could just handle the anticipated demand of 120   1   40   5   160 units a day: 20   3   8   5   160 . However, the raw material storage facility, which can only hold 3500 pounds (enough to produce only 3500 / 25   5   140 units a day), is a bottleneck in the system, since we need 160   3   25 pounds / unit   5   4000 pounds of storage (the limit of the delivery truck ’s capacity). Perhaps we could change our system to deliver a portion of the truckload directly to fabrication, or run out 500 pounds to fabrication as the raw material is unloaded from the truck so there is enough space for the full required 2-ton delivery. Note that any activity, resources or storage, could have been the bot- tleneck in the process. What ’s more, even if we increase the capacity of the storage facility, the bottleneck will shift to the packaging machines, being able to produce only 12 units per hour from A and another six per hour from B, for a total of 18 per hour, or 144 units a day. And if their capacity is increased, the bottleneck will shift to the 300-part truck delivery because we will need 160   3   2   5   320 parts delivered each day. As you can see, the bottleneck shifts around the facility as we solve one prob- lem after another. However, the process-fl ow diagram allows us to anticipate such shifts and head them off before they become real problems.

In a similar manner, Figure 8.11 presents a fl ow diagram for a simple photocopy service. When used for a service process, the process-fl ow diagram also typically illus- trates potential failure points in the process and the line of visibility that divides those activities a customer perceives from those that are conducted out of the customer ’s sight (the “backroom,” as in an auto repair shop, where operations can be conducted with effi ciency). Since products are not usually produced in a service, the diagram is called a service “map” or “blueprint,” as noted earlier, and shows the process times more prominently instead. Note the potential failure points in the photocopy service diagram and the “line of visibility” that divides what the customer sees from the back- room operations. Although Figure 8.11 illustrates a simple service process for illustra- tion purposes, service processes are frequently as complex as those in Figure 8.10 , or even more so, and also involve bottlenecks and combined operations.

Short-Term Capacity Alternatives The problem of short-term capacity is to handle unexpected but imminent actual demand, either less than or more than expected, in an economical manner. It is known, of course, that the forecast will not be perfect; thus, managers of resources must plan what short-term capacity alternatives to use in either case. Such considerations are

c08.indd 333c08.indd 333 25/09/12 5:20 PM25/09/12 5:20 PM

334 C h a p t e r 8 : C a p a c i t y , S c h e d u l i n g , a n d L o c a t i o n P l a n n i n g

usually limited to, at most, the next six months, and usually much less, such as the next few days or hours.

Some alternatives for obtaining short-run capacity are categorized in Table 8.13 . Each of the techniques in the table has advantages and disadvantages. The fi rst set of alternatives concerns simply trying to increase the resource base. The use of over- time is expensive (time and a half), and productivity after eight hours of work often declines. It is a simple and easily used approach, however, that does not require additional investment, so overtime is one of the most common alternatives. The use of extra shifts requires hiring but no extra facilities. However, productivity of second and third shifts is often lower than that of the fi rst shift. Part-time hiring can be expensive and is usually feasible for only low-skilled or unskilled work. Floating

CC 5

20

10 Yes

No

CC 10

F F

Ask specifications:

# of copies type of paper

reduce/enlarge

Inform customer

of cost

Customer decision

Take originals

to back room

Original and copies

to customer

Take payment

Ring up on register

Change, receipt, and

thank customer

Leaves

Line of visibility

Originals and

instructions to operator

45 60

Copier electricity

Set up copier

Run copies

Legend:

Seconds

Failure point?

Activity

CC 5 CC 20

F

CC 30

Greet customer

CC 10

F

CC 25

CC 10

F

F F Customer contact?

Paper, toner

F

F

Figure 8.11 Process-fl ow map for a service.

c08.indd 334c08.indd 334 25/09/12 5:20 PM25/09/12 5:20 PM

335S h o r t - T e r m C a p a c i t y P l a n n i n g

workers are fl exible and very useful but, of course, also cost extra. Leasing facilities and workers is often a good approach, but the extra cost reduces the profi t, and these external resources may not be available during the high-demand periods when they are most seriously needed. Subcontracting may require a long lead time, is con- siderable trouble to implement, and may leave little, if any, profi t.

The second set of techniques involves attempts to fi nd ways to improve the utili- zation of existing resources. For daily demand peaks (seen especially in services, as discussed in the next section), shifts can be overlapped to provide extra capacity at peak times, or staggered to adjust to changes in demand loads. Cross-training the workers to substitute for each other can effectively increase labor fl exibility. And there may be other ways to make labor and other resources adjustable, too. A similar alternative is to simply share resources whenever possible. Especially for services,

T A B L E 8 .13 • Techniques fo r Increas ing Shor t - r un Capac i t y I. Increase resources

1. Use overtime

2. Add shifts

3. Employ part-time workers

4. Use fl oating workers

5. Lease workers and facilities

6. Subcontract

II. Improve resource use

7. Overlap or stagger shifts

8. Cross-train the workers

9. Create adjustable resources

10. Share resources

11. Schedule appointments/reservations

12. Inventory output (if feasible) ahead of demand

13. Backlog or queue demand

III. Modify the output

14. Standardize the output

15. Offer complementary services

16. Have the recipient do part of the work

17. Transform service operations into inventoriable product operations

18. Cut back on quality

IV. Modify the demand

19. Partition the demand

20. Change the price

21. Change the promotion

22. Initiate a yield/revenue management system

V. Do not meet demand

23. Do not supply all the demand

c08.indd 335c08.indd 335 25/09/12 5:20 PM25/09/12 5:20 PM

336 C h a p t e r 8 : C a p a c i t y , S c h e d u l i n g , a n d L o c a t i o n P l a n n i n g

appointment and reservation systems, if feasible, can signifi cantly smooth out daily demand peaks. If the output can be stocked ahead of time, as with a product, this is an excellent and very common approach to meeting capacity needs. If recipients are willing, the backlogging of demand to be met later during slack periods is an excel- lent strategy; a less accurate forecast is needed, and investment in fi nished goods is nil. However, this may be an open invitation to competition.

Modifying the output is a creative approach. Doing less customization, allowing fewer variants, offering complementary services, and encouraging recipients to do some assembly or fi nishing tasks themselves (as at self-service gasoline stations and check-out lines), perhaps with a small price incentive, are frequently employed and are excellent alternatives.

Attempting to alter the demand, partition it, or shift it to a different period is another creative approach. Running promotions or price differentials (“off-peak” pricing), or both, for slack periods is an excellent method for leveling demand, espe- cially in utilities, telephones, and similar services. Prices are not easily increased above normal in high-demand periods, however. One formal method of partitioning both the demand and the resource supply is known as yield or revenue manage- ment, as discussed earlier. Last, the manager may simply decide not to meet the market demand—again, however, at the cost of inviting competition.

In actuality, many of these alternatives are not feasible except in certain types of organizations or in particular circumstances. For example, when demand is high, subcontractors are full, outside facilities and staff are already overbooked, second- shift workers are employed elsewhere, and marketing promotion is already low key. Thus, of the many possible alternatives, most fi rms tend to rely on only a few, such as overtime and, for product fi rms, stocking up ahead of demand.

So far we have primarily discussed increasing capacity in the short run, but fi rms also have a need to decrease short-run capacity. This is more diffi cult, however, and most such capacity simply goes unused. If the output involves a product, some inven- tory buildup may be allowed in order to make use of the available capacity; other- wise, system maintenance may be done (cleaning, fi xing, preprocessing, and so on).

Capacity Planning for Services Capacity planning is often much more diffi cult for pure service operations than for products, and with a service there is a clearer distinction between long- and short- run capacity planning. For services, the more diffi cult aspects of providing capacity occur in the short run, usually because the demand for a service is subject to daily peaks and valleys, and the output cannot be stored ahead of time to buffer this fl uc- tuation. For example, doctors ’ offi ces see demand peaks at 9 A.M. and 1 P.M., and college classes see a peak at 10 A.M. Or there may be weekly peaks, monthly peaks, or yearly peaks, such as Friday ’s demand on banks to deposit (or cash) paychecks and the fi rst-of-the-month demand on restaurants when Social Security checks arrive in the mail. Some services, such as fi re departments, experience multiple peaks, as illustrated in Figure 8.12 a , which shows the regular daily cycle of fi re alarms, with a peak from 3 to 7 P.M., and Figure 8.12 b , which shows the yearly cycle of fi re alarms, with a peak in April.

As noted earlier with regard to products, frequently it is not clear whether a prob- lem is a matter of scheduling or capacity; this is particularly true with services. The

c08.indd 336c08.indd 336 25/09/12 5:20 PM25/09/12 5:20 PM

337S h o r t - T e r m C a p a c i t y P l a n n i n g

primary problem is matching availability of staff to demand in terms of timing and skills, both on a daily basis and over the longer term (such as weekly and monthly). As discussed earlier, service organizations have developed many novel approaches to this problem, as just briefl y described: split shifts, overlapping shifts, duty tours (e.g.,

1 3 5 7 9 11 1 3 5 7 9 11

2

4

6

8

10

12

14 Building fire alarms Total fire alarms

Noon

(a) Hourly

Distribution of fire alarms by month Dade County, Florida

Distribution of fire alarms by 2-hour periods Dade County, Florida

A.M. P.M.

% o

f d

ai ly

t o

ta l

Jan. Feb. Mar. Apr. May June July Aug. Sep. Oct. Nov. Dec.

5

10

15

20 Building fire alarms Outdoor fire alarms

Total fire alarms

(b) Monthly

Dry season Wet season

% o

f an

n u

al t

o ta

l

Figure 8.12 Fire alarm histories (a) hourly and (b) monthly.

c08.indd 337c08.indd 337 25/09/12 5:20 PM25/09/12 5:20 PM

338 C h a p t e r 8 : C a p a c i t y , S c h e d u l i n g , a n d L o c a t i o n P l a n n i n g

48 or 72 hours for fi refi ghters), part-time help, overbooking, appointment systems, and on-call staff. However, for services, a favorite alternative is to share capacity with neighboring units by pooling resources such as generators, police patrols, or hotel rooms. When one organization is temporarily overloaded, the neighbor absorbs the excess demand. Another favorite approach for some services that has even been too successful is that of shifting the demand to off-peak periods. When AT&T offered lower phone rates after 5 P.M., it found that it had to raise the Sunday night 5 to 11 P.M. rate owing to excessive shifted demand.

In many situations, it is almost impossible to measure an organization ’s capacity to produce a service because the service is so abstract. Thus, a more common approach is to measure inputs , rather than outputs, and assume (perhaps with regu- lar checkups) that the production system is successful at transforming the inputs into acceptable services (outputs). For example, organizations that offer plays, art exhib- its, and other such intangible services do not measure their patrons ’ pleasure or relaxation; rather, they measure number of performances, number of actors and actresses, and number of paintings (or painting days, since many exhibits have a rotating travel schedule). Even fi re departments do not attempt to measure their capacity by the number of fi res they can extinguish; instead, they use the number of engines or companies they can offer in response to a call, the service or response time, or the number of fi refi ghters responding.

Clearly, this manner of measuring service capacity can leave a lot to be desired. Do more paintings give greater satisfaction? Do higher-quality paintings give greater satisfaction? Might there be other factors that are equally or more important, such as the crowd, the parking facilities, or the lighting on the paintings? Is a hospital where more deaths occur providing worse service? Is a hospital with more physicians on staff providing better service?

The Learning Curve An extremely important aspect of capacity planning, and an important operations concept in and of itself, is the learning curve effect—the ability of humans to increase their productive capacity through “learning.” This issue is particularly impor- tant in the short-term start-up of new and unfamiliar processes, such as those involv- ing new technologies (e.g., learning to use a new software program), and always occurs in the production ramp-up of new models of automobiles, planes, comput- ers, and so on. Thus, the characteristic of slow, possibly error-prone output initially, followed by better, faster production, should be of major concern to marketing and sales, which are often trying to market the output or have promised a certain volume to a customer by a set date; to accounting, which is checking productivity and yield rates in order to determine a fair cost for the output; and to fi nance, which is con- cerned with the timing of cash fl ows related to purchases, labor, and revenues.

The improvement with experience is not necessarily due to learning alone, how- ever. Better tools, improvements in work methods, upgraded output designs, and other such factors also help increase productivity. Hence, such curves are also known as improvement curves, production progress functions, performance curves , and experience curves . The learning curve effect, from this viewpoint, also affects long- term capacity and should be factored into longer-term planning processes, another issue of interest to marketing and accounting as well as fi nance. The Japanese, in

c08.indd 338c08.indd 338 25/09/12 5:20 PM25/09/12 5:20 PM

339S h o r t - T e r m C a p a c i t y P l a n n i n g

Two forms of the learning curve relationship are used in the literature. In one form M corresponds to the cumulative average labor-hours of all N units, and in the other form M corresponds to the actual labor-hours to produce the N th unit. The second interpretation is more useful for capacity planning and will be used here. For example, then, a learning rate of 90 percent would mean that each time production doubled from, say, N

1 to N

2 , unit N

2 would require 90 percent of the labor hours that

particular, count on increasing the long-term capacity of a facility through the work- ers ’ development of better work methods and improvements in tools.

The derivation of the learning curve began in the airframe manufacturing industry during the 1930s, when it was found that the labor-hours needed to build each suc- cessive airplane decreased relatively smoothly. In particular, the learning curve law was found to be

Each time the output doubles, the labor hours decrease to a fi xed percentage of their previous value .

In the case of plane production, this percentage was found to be 80 percent. Thus, when the fi rst plane of a series required 100,000 labor-hours to produce, the second took 80,000 labor hours, the fourth took 80 , 000   3   0 . 80   5   64 , 000 , the eighth 64 , 000   3   0 . 80   5   51 , 200 , and so on. This type of mathematical relationship is described by the negative exponential function, 1 illustrated for airplanes in Figure 8.13 .

1 4 7 10 13 16 19 22 0.00

20.00

40.00

60.00

80.00

100.00

Labor-hours

Number of units produced, N

La b

o r-

h o

u rs

, M

( 1

0 0

0 )

Figure 8.13 80 percent learning curve for airplane production.

1The function is as follows: M   5   m N r where M 5 labor-hours for the N th unit m 5 labor hours for fi rst unit N 5 number of units produced r 5 exponent of curve corresponding to learning rate 5 log(learning rate)/0.693

c08.indd 339c08.indd 339 25/09/12 5:20 PM25/09/12 5:20 PM

340 C h a p t e r 8 : C a p a c i t y , S c h e d u l i n g , a n d L o c a t i o n P l a n n i n g

N 1 required. The log in the equation for r is the “natural” log (the base e), and 0.693

is the natural log of 2.0. But base 10, or any other base, may be used if divided by the log of 2.0 to the same base. That is, r   5   log ( learning rate ) / log 2 .0 .

A number of factors affect the learning curve rate, but the most important are the complexity of the task and the percentage of human, compared with mechanical, input. The greatest learning—sometimes at a rate as much as 60 percent (lower rates meaning greater learning)—occurs for highly complex tasks consisting primarily of human inputs. A task that is highly machine-automated clearly leaves little opportunity for learning. (Thus, a rate close to 100 percent would apply, because only the human can learn.) In airframe manufacturing the proportion of human effort is about 75 per- cent, and an 80 percent learning rate applies. For similar work of the same complexity and ratio of human-to-machine input, approximately the same rate will apply.

But learning curves are not limited to manufacturing or even to product-oriented organizations. These curves apply just as well to hairdressing, selling, fi nding a park- ing space, and preparing pizza. As indicated, they also apply to groups of individu- als, and systems that include people and machines, as well as to individuals.

The primary question, of course, is what learning rate to apply. If previous experi- ence is available, this may give some indication; if not, a close watch of the time it takes to produce the fi rst few units should give a good indication. Let us illustrate the use of the learning curve, and some learning curve tables, with a simple example.

Learning Curve Tables

It is not usually necessary to solve the learning curve equation every time you run across a learning situation. First, the general law already stated will usually suffi ce for most purposes. Second, the solution to the equation for various learning rates, assum- ing that the fi rst item took 1 time unit, has already been calculated and tabulated in Tables 8.14 and 8.15 . These tables provide the percentage of time the N th unit will require relative to what the fi rst unit required (Table 8.14 ) and the cumulative amount of time that the fi rst N units will take relative to what the fi rst unit took (Table 8.15 ).

To use Tables 8.14 and 8.15 , you multiply the values given in these tables by the labor-hours actually required for the fi rst unit in your situation to get the time for the N th unit, or the cumulative time for units 1 through N , respectively. Returning to our example—the 80 percent learning curve for airplanes—we see in Table 8.14 that unit 2 (left-hand column) under “80%” will require 0.8 of what unit 1 required (100,000 labor-hours), that unit 4 will require 0.64, that unit 8 will take 0.512, and so forth. In addition, we also see that unit 3 will take 0.7021 and unit 6, for example, 0.5617 (i.e., 0 . 5617   3   100 , 000 or 56,170 labor-hours). The total labor-hours to produce two, four, or eight planes can be found by adding the necessary values together or by looking at Table 8.15 , where this has already been done. Again, reading under “80%” for 2, 4, and 8 units, we get 1.8, 3.142, and 5.346 3 100,000, respectively, for 180,000, 314,200, and 534,600 labor-hours, cumulative. We next illustrate the use of the learn- ing curve tables with a simple example, followed by a more complex example.

Following the engineering specifi cations for the assembly of a new motor, a pro- duction team was able to assemble the fi rst (prototype) motor in 3.6 hours. After more practice on the second and third motors, the team was able to assemble the fourth motor in 1.76 hours. What is the team ’s learning rate, and how long will the next motor probably take?

c08.indd 340c08.indd 340 25/09/12 5:20 PM25/09/12 5:20 PM

341S h o r t - T e r m C a p a c i t y P l a n n i n g

T A B L E 8 .14 • Un i t Va lues o f the Learn ing Cur ve Example: Unit 1 took 10 hours, 80% learning rate. What will unit 5 require? Solution: Unit 5 row, 80% column value 5 0.5956. Thus, unit 5 will take 10 (0.5956) 5 5.956 hours.

Improvement Ratios

Units 60% 65% 70% 75% 80% 85% 90% 95%

1 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 2 0.6000 0.6500 0.7000 0.7500 0.8000 0.8500 0.9000 0.9500 3 0.4450 0.5052 0.5682 0.6338 0.7021 0.7729 0.8462 0.9219 4 0.3600 0.4225 0.4900 0.5625 0.6400 0.7225 0.8100 0.9025 5 0.3054 0.3678 0.4368 0.5127 0.5956 0.6857 0.7830 0.8877 6 0.2670 0.3284 0.3977 0.4754 0.5617 0.6570 0.7616 0.8758 7 0.2383 0.2984 0.3674 0.4459 0.5345 0.6337 0.7439 0.8659 8 0.2160 0.2746 0.3430 0.4219 0.5120 0.6141 0.7290 0.8574 9 0.1980 0.2552 0.3228 0.4017 0.4930 0.5974 0.7161 0.8499

10 0.1832 0.2391 0.3058 0.3846 0.4765 0.5828 0.7047 0.8433 12 0.1602 0.2135 0.2784 0.3565 0.4493 0.5584 0.6854 0.8320 14 0.1430 0.1940 0.2572 0.3344 0.4276 0.5386 0.6696 0.8226 16 0.1296 0.1785 0.2401 0.3164 0.4096 0.5220 0.6561 0.8145 18 0.1188 0.1659 0.2260 0.3013 0.3944 0.5078 0.6445 0.8074 20 0.1099 0.1554 0.2141 0.2884 0.3812 0.4954 0.6342 0.8012 22 0.1025 0.1465 0.2038 0.2772 0.3697 0.4844 0.6251 0.7955 24 0.0961 0.1387 0.1949 0.2674 0.3595 0.4747 0.6169 0.7904 25 0.0933 0.1353 0.1908 0.2629 0.3548 0.4701 0.6131 0.7880 30 0.0815 0.1208 0.1737 0.2437 0.3346 0.4505 0.5963 0.7775 35 0.0728 0.1097 0.1605 0.2286 0.3184 0.4345 0.5825 0.7687 40 0.0660 0.1010 0.1498 0.2163 0.3050 0.4211 0.5708 0.7611 45 0.0605 0.0939 0.1410 0.2060 0.2936 0.4096 0.5607 0.7545 50 0.0560 0.0879 0.1336 0.1972 0.2838 0.3996 0.5518 0.7486 60 0.0489 0.0785 0.1216 0.1828 0.2676 0.3829 0.5367 0.7386 70 0.0437 0.0713 0.1123 0.1715 0.2547 0.3693 0.5243 0.7302 80 0.0396 0.0657 0.1049 0.1622 0.2440 0.3579 0.5137 0.7231 90 0.0363 0.0610 0.0987 0.1545 0.2349 0.3482 0.5046 0.7168

100 0.0336 0.0572 0.0935 0.1479 0.2271 0.3397 0.4966 0.7112 120 0.0294 0.0510 0.0851 0.1371 0.2141 0.3255 0.4830 0.7017 140 0.0262 0.0464 0.0786 0.1287 0.2038 0.3139 0.4718 0.6937 160 0.0237 0.0427 0.0734 0.1217 0.1952 0.3042 0.4623 0.6869 180 0.0218 0.0397 0.0691 0.1159 0.1879 0.2959 0.4541 0.6809 200 0.0201 0.0371 0.0655 0.1109 0.1816 0.2887 0.4469 0.6757 250 0.0171 0.0323 0.0584 0.1011 0.1691 0.2740 0.4320 0.6646 300 0.0149 0.0289 0.0531 0.0937 0.1594 0.2625 0.4202 0.6557 350 0.0133 0.0262 0.0491 0.0879 0.1517 0.2532 0.4105 0.6482 400 0.0121 0.0241 0.0458 0.0832 0.1453 0.2454 0.4022 0.6419 450 0.0111 0.0224 0.0431 0.0792 0.1399 0.2387 0.3951 0.6363 500 0.0103 0.0210 0.0408 0.0758 0.1352 0.2329 0.3888 0.6314

Source : Albert N. Schreiber, Richard A. Johnson, Robert C. Meier, William T. Newell, and Henry C. Fischer, Cases in Manufacturing Management (New York: McGraw-Hill, 1965), p. 464. Reprinted by permission of McGraw-Hill, © 1965.

c08.indd 341c08.indd 341 25/09/12 5:20 PM25/09/12 5:20 PM

342 C h a p t e r 8 : C a p a c i t y , S c h e d u l i n g , a n d L o c a t i o n P l a n n i n g

T A B L E 8 .15 • Cumula t i ve Va lues o f the Learn ing Cur ve Example: Unit 1 took 10 hours, 80% learning rate. What will be the total hours required to produce the fi rst fi ve units? Solution: Unit 5 row, 80% column: value 5 3.738. Thus, the fi rst fi ve units will require 10 (3.738) 5 37.38 hours.

Improvement Ratios

Units 60% 65% 70% 75% 80% 85% 90% 95%

1 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 2 1.600 1.650 1.700 1.750 1.800 1.850 1.900 1.950 3 2.045 2.155 2.268 2.384 2.502 2.623 2.746 2.872 4 2.405 2.578 2.758 2.946 3.142 3.345 3.556 3.774 5 2.710 2.946 3.195 3.459 3.738 4.031 4.339 4.662 6 2.977 3.274 3.593 3.934 4.299 4.688 5.101 5.538 7 3.216 3.572 3.960 4.380 4.834 5.322 5.845 6.404 8 3.432 3.847 4.303 4.802 5.346 5.936 6.574 7.261 9 3.630 4.102 4.626 5.204 5.839 6.533 7.290 8.111

10 3.813 4.341 4.931 5.589 6.315 7.116 7.994 8.955 12 4.144 4.780 5.501 6.315 7.227 8.244 9.374 10.62 14 4.438 5.177 6.026 6.994 8.092 9.331 10.72 12.27 16 4.704 5.541 6.514 7.635 8.920 10.38 12.04 13.91 18 4.946 5.879 6.972 8.245 9.716 11.41 13.33 15.52 20 5.171 6.195 7.407 8.828 10.48 12.40 14.61 17.13 22 5.379 6.492 7.819 9.388 11.23 13.38 15.86 18.72 24 5.574 6.773 8.213 9.928 11.95 14.33 17.10 20.31 25 5.668 6.909 8.404 10.19 12.31 14.80 17.71 21.10 30 6.097 7.540 9.305 11.45 14.02 17.09 20.73 25.00 35 6.478 8.109 10.13 12.72 15.64 19.29 23.67 28.86 40 6.821 8.631 10.90 13.72 17.19 21.43 26.54 32.68 45 7.134 9.114 11.62 14.77 18.68 23.50 29.37 36.47 50 7.422 9.565 12.31 15.78 20.12 25.51 32.14 40.22 60 7.941 10.39 13.57 17.67 22.87 29.41 37.57 47.65 70 8.401 11.13 14.74 19.43 25.47 33.17 42.87 54.99 80 8.814 11.82 15.82 21.09 27.96 36.80 48.05 62.25 90 9.191 12.45 16.83 22.67 30.35 40.32 53.14 69.45

100 9.539 13.03 17.79 24.18 32.65 43.75 58.14 76.59 120 10.16 14.11 19.57 27.02 37.05 50.39 67.93 90.71 140 10.72 15.08 21.20 29.67 41.22 56.78 77.46 104.7 160 11.21 15.97 22.72 32.17 45.20 62.95 86.80 118.5 180 11.67 16.79 24.14 34.54 49.03 68.95 95.96 132.1 200 12.09 17.55 25.48 36.80 52.72 74.79 105.0 145.7 250 13.01 19.28 28.56 42.08 61.47 88.83 126.9 179.2 300 13.81 20.81 31.34 46.94 69.66 102.2 148.2 212.2 350 14.51 22.18 33.89 51.48 77.43 115.1 169.0 244.8 400 15.14 23.44 36.26 55.75 84.85 127.6 189.3 277.0 450 15.72 24.60 38.48 59.80 91.97 139.7 209.2 309.0 500 16.26 25.68 40.58 63.68 98.851 151.5 228.8 340.6

Source : Albert N. Schreiber, Richard A. Johnson, Robert C. Meier, William T. Newell, and Henry C. Fischer, Cases in Manufacturing Management (New York: McGraw-Hill, 1965), p. 465. Reprinted by permission of McGraw-Hill, © 1965.

c08.indd 342c08.indd 342 25/09/12 5:20 PM25/09/12 5:20 PM

343S h o r t - T e r m C a p a c i t y P l a n n i n g

Here the actual individual assembly times are given, so we can use Table 8.14 , which tabulates the ratio of what the N th unit took relative to the fi rst unit . First, we need to fi nd the ratio from the given data and then locate that value somewhere in the table. Our ratio for the fourth motor would be: 1.76 / 3.6   5   0.49 . Next, we turn to Table 8.14 and scan across row “4” under the “Units” column. We fi nd the value 0.49 under “70%,” so this is our learning rate for the team (which is pretty good, by the way).

To fi nd out how long the next (fi fth) motor will take, we drop down the “70%” column to the next row that corresponds to the fi fth unit. ( Note : The rows are not always in increments of 1. For example, at 10 they jump by 2, and at 100 by 20.) At the fi fth row the value is 0.4368, which, when multiplied by what the fi rst unit took (3.6 hours), gives: 0.4368   3   3.6   5   1.57 hours. Remember: The tabulated values assume that the fi rst unit took only 1 hour (or minute, or day, or whatever the meas- ure), so if the fi rst unit took something other than “1,” you need to multiply the table value by the actual time it took to produce the fi rst unit.

Next, let us consider a more complex, real-life problem that also requires the use of the cumulative table, Table 8.15 .

Spreadsheet, Inc.

Spreadsheet, Inc., has just entered the growing software training market with a contract from a fi nancial organization to teach spreadsheet modeling techniques to the organiza- tion ’s 10 managers, for purposes of fi nancial and pension planning. The lesson for the last manager has just ended, and the organization, considering the fi rst 10 lessons highly successful, has engaged Spreadsheet to give the same lessons to its staff of 150 agents. The lesson for the fi rst manager was highly experimental, requiring 100 hours in all, but careful analysis and refi nement of the techniques have gradually decreased this time to the point where the average time for all 10 initial lessons was just under half that value, 49 hours each. To properly staff, schedule, plan, and cost out the work for the 150 les- sons, Spreadsheet needs to know how many hours of lessons will be required.

To begin, we can use Table 8.15 to determine the learning rate: the average of 49 hours each, times 10 managers, gives 490 hours, cumulative. This is 4.9 times what the fi rst manager required (490 hours/100 hours). Finding the value 4.9 in Table 8.15 for 10 units will then give the learning curve rate applying to these complex lessons. Reading across the 10-unit row, we fi nd 4.931 (close enough) under the “70%” col- umn. (On occasion, interpolation between columns may be necessary, or alterna- tively the exact quantities can be calculated directly using the formula. Spreadsheets can greatly facilitate the task of manually calculating time estimates based on the learning curve formula.)

Assuming that the lessons are continuous and the teaching techniques are not for- gotten (an important assumption), we can look further down the “70%” column in Table 8.15 to fi nd the value corresponding to the total number of lessons to be given: 10   1   150   5   160 . This value, 22.72, is then multiplied by the amount of time required for the fi rst lesson (100 hours) to give a grand total of 2272 hours for the 160 lessons. Since the initial 10 managers required a total of 490 hours by themselves, the second group, consisting of the agents, will require 2272   2   490   5   1782 hours . The time phasing of this 1782 hours is also available, if desired, from Table 8.14 .

The learning curve is only a theoretical construct, of course, and therefore only approximates actual learning. A more realistic, and typical, learning pattern is illus- trated in Figure 8.14 . Initially, actual labor-hours per unit vary around the theoretical

c08.indd 343c08.indd 343 25/09/12 5:20 PM25/09/12 5:20 PM

344 C h a p t e r 8 : C a p a c i t y , S c h e d u l i n g , a n d L o c a t i o n P l a n n i n g

curve until a “learning plateau” is reached at, perhaps, the tenth unit. At this plateau no signifi cant learning appears to occur, until there is a breakthrough. Learning typi- cally involves a number of such plateaus and breakthroughs. At about 30 units, pro- duction is halted for a period of time and “forgetting” occurs, rapidly at fi rst but then trailing off. When production is resumed, relearning occurs very quickly (as when someone relearns to ride a bicycle after 40 years) until the original effi ciency is reached (at about 33 units). If the conditions are the same at this time as for the ini- tial part of the curve, the original learning curve rate will then hold. After suffi cient time passes, the improvement due to learning becomes trivial in comparison with natural variability in effi ciency, and at that point we say learning has ceased.

Queuing and the Psychology of Waiting An important element in evaluating the capacity of operations to produce either prod- ucts or services concerns the waiting lines, backlogs, or queues that tend to build up in front of the operations. Queuing theory provides a mechanism to determine several key performance measures of an operating system based on the rate of arrivals to the system and the system ’s capacity (specifi ed as the system ’s service rate). With an unpaced production line, for example, buffer inventory between operations builds up at some times and disappears at other times, owing to natural variability in the diffi - culty of the operations. The Wiley Web site for this text (see Preface for URL) includes a discussion of the theory, equations, and some example calculations of queuing.

In the production of services, this variability is even greater because of both the amount of highly variable human input and the variable requirements for services. What is more, the “items” in queue are often people, who tend to complain and make trouble if kept waiting too long. Thus, it behooves the operations manager to provide adequate service to keep long queues from forming. This costs more money for serv- ice facilities and staffs. But long queues cost money also, in the form of in-process

0 1 2 3 4 5

0 10 20 30 30

6

40

7

50

8

60

9

70

10

80

La b

o r-

h o

u rs

/u n

it

No production

Unit

Time (days)

Theoretical curve

Actual curve

Learning plateau

Breakthrough

“Forgetting” curve

Relearning curve

Natural variability exceeds improvement

Continuation of original curve

Figure 8.14 Typical pattern of learning and forgetting.

c08.indd 344c08.indd 344 25/09/12 5:20 PM25/09/12 5:20 PM

345S h o r t - T e r m C a p a c i t y P l a n n i n g

inventory, unfi nished orders, lost sales, and ill will. Figure 8.15 conceptually illustrates, as a function of the capacity of the service facility, the tradeoffs in these two costs:

1. Cost of waiting . In-process inventory, ill will, lost sales. This cost decreases with service capacity.

2. Cost of service facilities . Equipment, supplies, and staff. This cost increases with service capacity.

At some point the total of the two costs in Figure 8.15 is minimized, and it is at this point that managers typically wish to operate. However, before investing resources by adding expensive service facilities, the manager may fi nd it worthwhile trying to reduce the cost of waiting instead. Given that perceptions and expectations may have more to do with customer satisfaction than actual waiting time, David Maister (1984) has formulated eight insightful “principles” of waiting, which, if addressed carefully, may be more effective in reducing the overall cost of waiting to the organization than adding service facilities.

1. Unoccupied time feels longer than occupied time. Give customers something to do while waiting, hopefully something that will facilitate the service that is to come. An example is having customers key in their Social Security number while waiting on the phone so the representatives will have their fi le on screen as they answer the call.

2. Pre-service waiting feels longer than in-service waiting. Using staging areas to complete portions of the service, such as taking a patient ’s tempera- ture and blood pressure, communicates that the service has begun.

3. Anxiety makes waiting seem longer. Offer information to relieve anxiety or distracters (even music, mirrors) to allay anxiety.

4. Uncertain waiting is longer than known, fi nite waiting. Provide cues, or direct announcements, to indicate how soon the service will be coming or fi nishing (especially in the case of a painful procedure).

Optimal capacity

Cost of waiting Cost of facility Total cost

Minimal cost

C o

st

Service facility capacity

Figure 8.15 The relevant queuing costs.

c08.indd 345c08.indd 345 25/09/12 5:20 PM25/09/12 5:20 PM

346 C h a p t e r 8 : C a p a c i t y , S c h e d u l i n g , a n d L o c a t i o n P l a n n i n g

E X P A N D Y O U R U N D E R S T A N D I N G

5. Unexplained waiting is longer than explained waiting. Keep custom- ers informed about why they are being delayed and how long it will be before they can be serviced.

6. Unfair waiting is longer than fair waiting. Make sure that priority or express customers are handled in a manner transparent to other customers and treated out of sight if possible.

7. Solo waiting is longer than group waiting. In part this refl ects principles 1 (someone else to talk to), 3 (seeing and talking to others can reduce anxi- ety), and 5 (other waiting customers may communicate reasons for the wait- ing), as well as the general principle that there is more security in groups.

8. The more valuable the service, the longer it is worth waiting for. The use of marketing and other means to increase the perception of the value of the service will reduce the impatience with waiting.

1. Why might a decision maker choose a qualitative forecasting method when extensive historical demand data are available?

2. Frequently, simple models such as breakeven are much more appealing to management than more sophisticated models (such as linear programming). Why might this be so?

3. Exactly what decreases in unit cost occur with larger facilities as a result of economies of scale? Might any costs increase with the size of a facility?

4. Why has the concept of economies of scope never arisen before? List where the economies come from.

5. How ethical is it for airlines, hotels, and other serv- ice providers to overbook their limited-capacity facilities intentionally, knowing that at some point they will have to turn away a customer with a “guaranteed” reservation?

6. Describe how the concept of bottlenecks would apply to services as well as products. Give some examples from your experience.

7. What elements would be measured if a product fi rm were to measure its capacity by its inputs, as do some service fi rms?

8. Does the learning curve continue downward forever?

9. Which measures used to locate pure service organi- zations are direct measures of benefi t and which are surrogate measures of benefi t? Can you think of better direct measures? Why aren ’t they used?

10. Would the failure points, line of visibility, and processing times used in service maps be useful in process-fl ow diagrams for products?

11. When might an organization not use all three stages of the location selection process described here?

12. Might the breakeven model be used for the national or site stage of location? Might the weighted-score model be useful in the national or community stage of location? What factors would be used in these models at other stages?

13. Are the principles of waiting captured in the 23 capacity techniques of Table 8.13 ? Which ones?

14. Would a fi rm that simply expanded its current product line gain economies of scope? Might highly fl exible and profi cient labor also offer economies of scope?

15. Many services, such as airlines, conduct their sched- uling in two stages. First, an overall macroschedule is constructed and optimized for costs and service to the customer. This schedule is then considered to be the baseline for detailed scheduling to attempt to achieve. The second, detailed stage is then a real- time schedule to adjust the macroschedule for any necessary changes, emergencies, and so on. Describe how this might work for airlines, hospitals, schools, and urban alarm services. What serious problems might arise with this approach?

16. Referring to the Yield/Revenue Management and Overbooking section, why might an early reserva- tion be canceled? A late reservation?

c08.indd 346c08.indd 346 25/09/12 5:20 PM25/09/12 5:20 PM

347A p p l y Y o u r U n d e r s t a n d i n g

A P P L Y Y O U R U N D E R S T A N D I N G

Banga lore Tra in ing Ser v i ce s (BTS)

BTS was an entrepreneurial start-up developed by Deepa Anand and Monisha Patel, two recent MBA graduates from the United States who had served internships in the summer with a U.S. call center that was considering setting up operations in India but was unsure how to fi nd suitable employees. Their plan was to offer training to Indian men and women in call center activities such as sales, service, and troubleshooting for electronic goods of all sorts, and then to match those employees to the needs of foreign fi rms looking to set up call centers. The training consisted of a dozen sessions covering culture, speaking fl uency, elec- tronic awareness, buying and service behaviors, and other such basic matters that all call centers required.

Questions

Discuss how the following topics from this chapter might be of relevance to Deepa and Monisha in setting up their new fi rm: 1. Capacity planning 2. Learning curve 3. Bottlenecks 4. Psychology of waiting 5. Scheduling 6. Service map/blueprint

Ex i t Manufac tur ing Company

The planning committee of Exit Manufacturing Company (made up of the vice presidents of marketing, fi nance, and production) was discussing the plans for a new factory to be located outside of Atlanta, Georgia. The factory would produce exterior doors consisting of prehung metal over Styrofoam insulation. The doors would be made in a standard format, with 15 dif- ferent insert panels that could be added by retailers after manufacture. The standardization of construction was expected to create numerous production effi ciencies over competitors ’ fac- tories that produced multidimensional doors. Atlanta was felt to be an ideal site because of its location—in the heart of the Sunbelt, with its growing construction industry. By locating close to these growing Sunbelt states, Exit would minimize distribution costs.

The capital cost for the factory was expected to be $14 million. Annual maintenance ex- penses were projected to total 5 percent of capital. Fuel and utility costs were expected to be $500,000 per year. An analysis of the area ’s labor market indicated that a wage rate of $10 per hour could be expected. It was estimated that producing a door in the new facility would re- quire 1.5 labor-hours. Fringe benefi ts paid to the operating labor were expected to equal 15 percent of direct labor costs. Supervisory, clerical, technical, and managerial salaries were forecast to total $350,000 per year. Taxes and insurance would cost $200,000 per year. Other miscellaneous expenses were expected to total $250,000 per year. Depreciation was based on a 30-year life with use of the straight-line method and a $4 million salvage value. Sheet metal, Styrofoam, adhesive for the doors, and frames were projected to cost $12 per door. Paint, hinges, doorknobs, and accessories were estimated to total $7.80 per door. Crating and ship- ping supplies were expected to cost $2.50 per door.

Exit ’s marketing manager prepared the following price–demand chart for the distribution area of the new plant. Through analysis of these data, the committee members felt that they could verify their expectation of an increase from 15 to 25 percent in the current market share, owing to the cost advantage of standardization.

c08.indd 347c08.indd 347 25/09/12 5:20 PM25/09/12 5:20 PM

348 C h a p t e r 8 : C a p a c i t y , S c h e d u l i n g , a n d L o c a t i o n P l a n n i n g

Questions

Develop a breakeven capacity analysis for Exit ’s new door and determine the following:

1. Best price, production rate, and profi t 2. Breakeven production rate with the price determined in Question 1 3. Breakeven price with the production rate determined in Question 1 4. Sensitivity of profi ts to variable cost, price, and production rate

S ta f fo rd Chemica l , Inc .

Stafford Chemical, Inc. is a privately held company that produces a range of specialty chemi- cals. Currently, its most important product line is paint pigments used by the automobile in- dustry. Stafford Chemical was founded more than 60 years ago by Phillip Stafford in a small town north of Cincinnati, Ohio, and is currently run by Phillip ’s grandson, George Stafford. Stafford has more than 150 employees, and approximately three-quarters of them work on the shop fl oor. Stafford Chemical operates out of the same plant Phillip built when he founded the company; however, it has undergone several expansions over the years.

Recently, a Japanese competitor of Stafford Chemical, Ozawa Industries, announced plans to expand its operations to the United States. Ozawa, a subsidiary of a large Japanese indus- trial company, decided to locate a new facility in the United States to better serve some of its customers: Japanese automobile manufacturers who have built assembly plants in the United States.

The governor of Ohio has been particularly aggressive in trying to persuade Ozawa Indus- tries to locate in a new industrial park located about 30 miles from Stafford ’s current plant. She has expressed a willingness to negotiate special tax rates, to subsidize workers ’ training, and to expand the existing highway to meet Ozawa ’s needs. In a recent newspaper article, she was quoted as saying:

Making the concessions I have proposed to get Ozawa to locate within our state is a good business decision and a good investment in our state. The plant will provide high-paying jobs for 400 of our citizens. Furthermore, over the long run, the income taxes that these 400 individuals will pay will more than offset the concessions I have proposed. Since several other states have indicated a willingness to make similar concessions, it is unlikely that Ozawa would choose our state without them.

George Stafford was outraged after being shown the governor ’s comments.

I can ’t believe this. Stafford Chemical has operated in this state for over 60 years. I am the third generation of Staffords to run this business. Many of our employees ’ parents and grandparents worked here. We have taken pride in being an exemplary corporate citizen. And now our governor wants to help one of our major competitors drive us out of business. How are we supposed to compete with such a large industrial giant? We should be the ones who are getting the tax break and help with workers ’ training. Doesn ’t 60 years of paying taxes and employing workers count for something? Where is the governor ’s loyalty? It seems to me that the state should be loyal to its long-term citizens, the ones who care about the state and community they operate in—not some large industrial giant looking to save a buck.

Average Sales Price ($/door) Area Sales (in units)

$90 40,000

$103 38,000

$115 31,000

$135 22,000

c08.indd 348c08.indd 348 25/09/12 5:20 PM25/09/12 5:20 PM

349E x e r c i s e s

Questions

1. How valid is George Stafford ’s argument? How valid is the governor ’s argument? Is Stafford Chemical being punished because it was already located within the state?

2. How ethical is it for states and local governments to offer incentives to attract new businesses to their localities? Are federal laws needed to keep states from competing with one another?

3. Does the fact that Ozawa is a foreign company alter the ethical nature of the governor ’s actions? What about Ozawa ’s size?

4. What are George ’s options?

1. Three professors are grading a combined fi nal exam. Each is grading different questions on the test. One professor requires 3 minutes to fi nish her portion, a second takes 6 minutes, and the third takes 2 minutes. Assume there is no learning curve effect.

a. What will be their hourly output?

b. If there are 45 tests to grade, how long will the grading take?

c. If each professor were to grade the exams sepa- rately in 18 minutes, how long would it take to grade the 45 tests? How long if another profes- sor (who also required 18 minutes) joined them?

d. If another professor pitches in just to help the second professor in the original arrangement, how long will it take the four of them to grade the tests?

e. If a fi fth professor offers to help, what might happen?

2. A toy fi rm produces drums sequentially on three machines—A, B, and C—with cycle times of 3, 4, and 6 minutes, respectively.

a. Determine the optimum effi ciency and output rates for adding one, two, . . . , six more machines.

b. Assume now that two identical lines are operat- ing, each with machines A, B, and C. If new machines can be shared between the lines, how should one, two, and then three new machines be added? What are the resulting effi ciencies and outputs of the two lines? Is it always best to equally share extra machines between the two lines?

3. If the production system for a product has a utiliza- tion of 80 percent and a yield of 75 percent, what capacity is needed to produce 1000 units a year?

4. If unit 1 requires 6 labor-hours and unit 5 requires 1.8324, what is the learning rate? What will unit 6

require? What have the fi rst fi ve units required in total?

5. A production lot of 25 units required 103.6 hours of effort. Accounting records show that the fi rst unit took 7 hours. What was the learning rate?

6. If unit 1 required 200 hours to produce and the labor records for an Air Force contract of 50 units indicate an average labor content of 63.1 hours per unit, what was the learning rate? What total addi- tional number of labor-hours would be required for another Air Force contract of 50 units? What would be the average labor content of this second con- tract? Of both contracts combined? If labor costs the vendor $10 per hour on this Air Force contract and the price to the Air Force is fi xed at $550 each, what can you say about the profi tability of the fi rst and second contracts, and hence the bidding process in general?

7. All the reports you wrote for one class had three sections: introduction, analysis, conclusion. The times required to complete these sections (includ- ing typing, etc.) are shown below in hours.

Report Introduction Analysis Conclusion

1 1.5 6 2

2 — (lost data) —

3 1 3 0.8

The class requires fi ve reports in all. You are now starting report 4 and, although you are working faster, you can afford to spend only 1 hour a day on these reports. Report 5 is due in one week (7 days). Will you be done in time?

8. Use the CVD model to evaluate the following three locations in terms of access to fi ve destinations. Site I is located 313, 245, 188, 36, and 89 feet, respec- tively, from the fi ve destinations; site II, 221, 376,

E X E R C I S E S

c08.indd 349c08.indd 349 25/09/12 5:20 PM25/09/12 5:20 PM

350 C h a p t e r 8 : C a p a c i t y , S c h e d u l i n g , a n d L o c a t i o n P l a n n i n g

92, 124, and 22 feet; and site III, 78, 102, 445, 123, and 208 feet.

9. Reevaluate Exercise 8 if the number of trips to each of the destinations is, respectively, 15, 6, 12, 33, and 21.

10. The location subcommittee ’s fi nal report to the board has focused on three acceptable communi- ties. Table 15b in the appendix to the report indi- cates that the cost of locating in communities 1, 2, and 3 is approximately €400,000, €500,000, and €600,000 per year (respectively), mortgaged over 30 years. Paragraph 2 on page 39 of the report indicates that the variable cost per unit of product will increase 15 percent in community 1 but decrease 15 percent in community 3, owing to dif- ferences in labor rates. As plant manager, you know that variable costs to date have averaged about €3.05 per unit and sales for the next decade are expected to average 20 percent more than the last 10 years, during which annual sales varied between 40,000 and 80,000 units. Which location would you recommend?

11. Nina is trying to decide in which of four shopping centers to locate her new boutique. Some cater to a higher class of clientele than others, some are in an indoor mall, some have a much greater volume than others, and, of course, rent varies considera- bly. Because of the nature of her store, she has decided that the class of clientele is the most impor- tant consideration. Following this, however, she must pay attention to her expenses; and rent is a major item—probably 90 percent as important as clientele. An indoor, temperature-controlled mall is a big help, however, for stores such as hers, where 70 percent of sales are from passersby slowly stroll- ing and window-shopping. Thus, she rates this as about 95 percent as important as rent. Last, a higher volume of shoppers means more potential sales; she thus rates this factor as 80 percent as important as rent. As an aid in visualizing her location alterna- tives, she has constructed the following table. “Good” is scored as 3, “fair” as 2, and “poor” as 1. Use a weighted score model to help Nina come to a decision.

Location

1 2 3 4

Class of clientele Fair Good Poor Good Rent Good Fair Poor Good Indoor mall Good Poor Good Poor Volume Good Fair Good Poor

12. What is the design capacity of a production system that produces 753 good units a year with a utiliza- tion of 90 percent and yield of 85 percent?

13. A defense contractor is bidding on a military con- tract for 100 radar units. The contractor employs 30 machine operators who work 165 hours a month each. The fi rst radar unit required 1145 operator- hours, and the learning curve for this type of work is known to be 75 percent. It takes a month to order and receive raw material components, which cost $500 per radar unit. The material is then paid for in the month it is received. Fixed costs include a month to tool up, which costs $10,000, and then $5000 per month for every month of production. Direct labor and variable overhead are $8 per hour. The contrac- tor can deliver only completed units and is paid the following month. Profi t is set at 10 percent of the bid price. Find the bid price, derive the production schedule, and calculate the cash fl ow schedule.

14. Is Clarton or Uppingham the best location for a production volume of 600 services? The fi xed costs of Clarton total £6000 (pounds, United Kingdom) per year, while those of Uppingham total only £4500. However, the variable costs of Clarton are £8, while those of Uppingham are £10.

15. The head of the Campus Computing Center is faced with locating a new centralized computer center at one of three possible locations on the campus. The decision is to be based on the number of users in each department and the distance of the various departments from each possible location. Which location should be chosen?

Distance by Location

Department Number of Users 1 2 3

1 25 0 3 5 2 30 5 4 3 3 10 2 0 1 4 5 3 2 0 5 14 6 2 3

16. A new product involves the following costs associ- ated with three possible locations. If demand is fore- cast to be 3900 units a year, which location should be selected?

Location

A B C

Annual cost $10,000 40,00 25,000 Unit variable cost $10.00 2.50 6.30

c08.indd 350c08.indd 350 25/09/12 5:20 PM25/09/12 5:20 PM

351B i b l i o g r a p h y

B I B L I O G R A P H Y

17. A fancy Swiss restaurant has 30 tables. If it accepts N reservations, the probability that N will arrive is 0.1; N –   1 is 0.2; N –   2 is 0.3; and N –   3 is 0.4. If each unfi lled table costs F20 (Swiss francs) but a cus- tomer turned away costs F10, fi nd how many reser- vations to accept. Solve again, assuming that a customer turned away costs F25.

18. The Arms Hotel in South Africa has only 56 rooms. An unfi lled room represents R500 (rands) a night in lost profi t, whereas every turnaway due to a fi lled room costs R300 in ill will. If N reservations are accepted, the probability of N , N –   1 , and N –   2 guests actually showing up is 0.2, 0.5, 0.3, respec- tively. How many reservations should be accepted?

Alstrup , J. , S. E. Andersson , S. Boas , O. B. G. Madsen , and R. V. V. Vidal . “ Booking Control Increases Profi t at Scandinavian Airlines .” Interfaces ( July–August 1989 ): 10 – 19 .

Armstrong , J. S. Long Range Forecasting: From Crystal Ball to Computer . New York : Wiley , 1995 .

Ballou , R. H. Business Logistics Management , 4th ed. Upper Saddle River, NJ : Prentice Hall , 1998 .

Bartmess , A. , and K. Cerny . “ Building Competitive Ad- vantage Through a Global Network of Capabilities .” California Management Review , 35 ( 1993 ): 78 – 103 .

Belkaoui , A. The Learning Curve . Westport, CT : Quorum Books , 1986 .

Blackstone , J. H. Capacity Management . Cincinnati, OH : South-Western , 1989 .

Burlingham , B. Small Giants: Companies That Choose to Be Great Instead of Big . New York : Portfolio , 2006 .

Carey , S. “Case of the Vanishing Airport Lines.” Wall Street Journal , August 9, 2007 .

Crandall , R. E. , and R. E. Markland . “ Demand Management—Today ’s Challenge for Service Industries .” Production and Operations Management , 5 (Summer 1996 ): 106 – 120 .

Drezner , Z. , and H. Hamacher . Facility Location: Appli- cations and Theory . Berlin : Springer Verlag , 2002 .

Durrande-Moreau , A. , and J.-C. Usunier . “ Time Styles and the Waiting Experience: An Exploratory Study .” Journal of Service Research , 2 (November 1999 ): 173 – 186 .

Ferdows , K. “ Making the Most of Foreign Factories .” For- tune (March–April 1997 ): 73 – 88 .

Filley , R. D. “ Putting the ‘Fast’ in Fast Foods: Burger King .” Industrial Engineering ( January 1983 ): 44 – 47 .

Fisher , M. L. , J. H. Hammond , W. R. Obermeyer , and A. Raman . “ Making Supply Meet Demand in an Uncertain World .” Harvard Business Review , 72 (May–June 1994 ): 83 – 93 .

Fitzsimmons , J. A. , and M. J. Fitzsimmons . Service Man- agement: Operations , Strategy, and Information Tech- nology , 5th ed. New York: Irwin/McGraw-Hill, 2006 .

Francis , R. L. , L. F. McGinnis Jr. , and J. A. White . Layout and Location: An Analytical Approach , 2nd ed. Upper Saddle River, NJ : Prentice-Hall , 1998 .

Goldratt , E. Y. , and J. Cox . The Goal : A Process of On- Going Improvement . Croton-on-Hudson, NY: North River, 1984 .

Hartvigsen , D. SimQuick: Process Simulation with Excel , 2nd ed. Upper Saddle River, NJ : Prentice-Hall , 2004 .

Heskett , J. “ Note on Service Mapping .” Harvard Reprint No . 9-693-065 , November 1992.

Hunt , V. D. Process Mapping . New York : Wiley , 1996 .

Kimes , S. E. “ Yield Management: A Tool for Capacity- Constrained Service Firms .” Journal of Operations Man- agement , 8 ( 4 )( 1989 ): 348 – 363 .

Kimes , S. E. , and R. B. Chase . “ The Strategic Levers of Yield Management .” Journal of Service Research , 1 (November 1998 ): 156 – 166 .

Kirchmier , B. , and G. I. Plenert . Finite Capacity Schedul- ing . New York : Wiley , 2002 .

Klassen , K. J. , and T. R. Rohleder . “ Combining Operations and Marketing to Manage Capacity and Demand in Serv- ices .” The Service Industries Journal , 21 ( 2 )( 2001 ): 1 – 30 .

Krajewski , L. J. , and, L. P. Ritzman . “ Shift Scheduling in Banking Operations: A Case Application .” Interfaces (April 1980 ): 1 – 7 .

Maister , D. H. “ The Psychology of Waiting Lines .” Harvard Reprint No . 9-684-064 , May 1984.

Marshall , P. W. “ A Note on Process Analysis (Abridged) .” Harvard Reprint No . 9-689-032 , September 1994.

Matthews , R. G. “Hometown Rivals Diverge on Best Way.” Wall Street Journal , March 3, 2010 .

Metters , R. , and V. Vargus . “ Yield Management for the Nonprofi t Sector .” Journal of Service Research , 1 ( February 1999 ): 215 – 226 .

c08.indd 351c08.indd 351 25/09/12 5:20 PM25/09/12 5:20 PM

352 C h a p t e r 8 : C a p a c i t y , S c h e d u l i n g , a n d L o c a t i o n P l a n n i n g

Niebel , B. , and A. Freivalds . Methods, Standards, and Work Design , 11th ed. New York : McGraw-Hill , 2002 .

Prahalad , C. K. “In Volatile Times, Agility Rules.” Busi- nessWeek , (September 21, 2009 ): 80.

Price , W. L. , and M. Turcotte . “ Locating a Blood Bank .” Interfaces (September–October 1986 ): 17 – 26 .

Pullman , M. E. , J. C. Goodale , and R. Verma . “Service Capacity Design with an Integrated Market Utility-Based Method.” In J. A. Fitzsimmons and M. J. Fitzsimmons (eds.), New Service Development . Thousand Oaks, CA : Sage , 2000 : 111 – 137 .

Radas , S. , and S. M. Shugan . “ Managing Service Demand: Shifting and Bundling .” Journal of Service Research , 1 (August 1998 ): 47 – 64 .

Ramani , K. V. “ Scheduling Doctors ’ Activities at a Large Teaching Hospital .” Production and Inventory Manage- ment Journal (First and Second Quarter, 2002 ): 56 – 62 .

Sanders , N. R. , and K. B. Manrodt . “ Forecasting Practices in U.S. Corporations: Survey Results .” Interfaces , 24 (March–April 1994 ): 91 – 100 .

Sanders , N. R. , and L. P. Ritzman . “ Bringing Judgment into Combination Forecasts .” Journal of Operations Management. 13 ( 1995 ): 311 – 321 .

Schmenner , R. W. “The Location Decisions of New Serv- ices.” In J. A. Fitzsimmons and M. J. Fitzsimmons (eds.), New Service Development . Thousand Oaks, CA : Sage , 2000 : 216 – 238 .

Smith , B. C. , J. F. Leimkuhler , and R. M. Darrow . “ Yield Management at American Airlines .” Interfaces , 22 ( January–February 1992 ): 8 – 31 .

Smith , R. A. “Find the Best Checkout Line.” Wall Street Journal , December 8, 2011 .

Teplitz , C. J. The Learning Curve Deskbook . Westport, CT : Greenwood Publishing Group , 1991 .

Trail , D. T. “ Package Products Capitalizes on Data .” APICS—The Performance Advantage (August 1996 ): 38 – 41 .

c08.indd 352c08.indd 352 25/09/12 5:20 PM25/09/12 5:20 PM

353

� Forecasting

C H A P T E R 8 S U P P L E M E N T

c08.indd 353c08.indd 353 25/09/12 5:20 PM25/09/12 5:20 PM

354 C h a p t e r 8 S u p p l e m e n t : F o r e c a s t i n g

F O R E C A S T I N G P U R P O S E S A N D M E T H O D S There is usually a close relationship between competing successfully and being able to predict key aspects of the future accurately. Clearly, it is not practical to try to plan without some prediction of the future. Even planning a simple party requires predict- ing how many people will show up, how much they will eat and drink, what kind of snacks and beverages they will enjoy, and how long they will stay. A business intro- ducing a new service needs to predict the demand for the service, how prices and advertising will affect this demand, how competitors will respond, and so on.

Thus, we see that an accurate estimate of demand for the output is crucial to the effi cient operation of the production system and, hence, to managing the organiza- tion ’s resources. For example, a supermarket chain that is contemplating the addi- tion of a new store must have a reasonable estimate of demand in order to determine how big the store and the parking lot should be, what ancillary departments (such as a bakery, pharmacy, deli, and bank) should be included, and how many shopping carts and check-out lanes should be specifi ed in the plans. Once the facility is con- structed, a more specifi c, perhaps weekly, forecast of demand will be needed so that the manager will be able to schedule workers and order merchandise. The same is true for decisions about capacity, scheduling, and staffi ng in a product organization. Capacity (obtaining the proper level of resources) and scheduling (the timing of resource usage) both require forecasting, whether or not it is a formal procedure.

As an aside, it is worth noting that it is not only demand for the output that can be forecast. The tools of forecasting can also be used to predict the development of new technology, national and international economic conditions, and even many factors internal to the organization, such as changes in lead time, scrap rates, cost trends, personnel growth, and departmental productivity. Here, however, we will restrict our discussion to the uses of forecasting for long- and short-term capacity planning.

Forecasts are used in organizations for four primary purposes, the fi rst two of which are strategic and long range, and the last two of which are more tactical and short range:

1. To decide whether demand is suffi cient to justify entering the market. If demand exists but at too low a price to cover the costs that an organization will incur in producing an output, then the organization should reject the opportunity.

2. To determine long-term (2- to 5-year) capacity needed, in order to design facili- ties. An overall projection of demand for a number of years into the future serves as the basis for decisions related to expanding, or contracting, capacity to meet the demand. Since there is competition, even in the not-for-profi t sector, an organization is courting disaster if it produces ineffi ciently, because of excess idle capacity, or insuffi ciently to meet demand, because of too little capacity.

3. To determine midterm (3-month to 18-month) fl uctuations in demand, in order to avoid shortsighted decisions that will hurt the company in the long run. To illustrate, if a company planned its staffi ng solely on the basis of its weekly forecast, each week it might adjust the level on the basis of a forecast for the coming week. Thus, in some weeks it might lay off workers only to rehire them in the following week. Such weekly adjustments would most likely lower morale and productivity. A better approach is to base staffi ng on a longer-term perspective.

c08.indd 354c08.indd 354 25/09/12 5:20 PM25/09/12 5:20 PM

355F o r e c a s t i n g P u r p o s e s A n d M e t h o d s

4. To ascertain short-term (1-week to 3-month) fl uctuations in demand for the purposes of production planning, workforce scheduling, materials planning, and other such needs. These forecasts support a number of operational activities and can have a signifi cant effect on organizational productivity, bot- tlenecks, master schedules, meeting promised delivery dates, and other such issues of concern to top management and to the organization as a whole.

Forecasting Methods Forecasting methods can be grouped in several ways. One classifi cation, illustrated in Figure 8S.1 , distinguishes between formal forecasting techniques and informal approaches such as intuition, spur-of-the-moment guesses, and seat-of-the-pants predictions. Our attention here will obviously be directed to the formal methods.

Life-cycle analysis

Surveys

Delphi method

Qualitative

Expert opinion

Consumer panels

Test marketing

Historical analogy

Causal

Quantitative

Informal (intuitive)

Formal

Forecasting methods

Time series analysis

Multiple regression

Econometric

Box-Jenkins

Exponential smoothing

Moving average

Simple regression

Figure 8S.1 A classifi cation of forecasting methods.

c08.indd 355c08.indd 355 25/09/12 5:20 PM25/09/12 5:20 PM

356 C h a p t e r 8 S u p p l e m e n t : F o r e c a s t i n g

In general, qualitative forecasting methods are often used for long-range fore- casts, especially when external factors (e.g., an especially cold winter) may play a signifi cant role. They are also of use when historical data are very limited or nonex- istent, as in the introduction of a new product or service.

Some of the most signifi cant decisions made by organizations, frequently strategic decisions, are made on the basis of qualitative forecasts. These often concern either a new product or service or long-range changes in the nature of the organiza- tion ’s outputs. In both cases, relevant historical data on demand are typically not available.

Qualitative forecasts are made using information such as telephone or mail sur- veys of consumers ’ attitudes and intentions, consumer panels, test marketing in lim- ited areas, expert opinion and panels, and analyses of historical demand for similar products or services—a method known as historical analogy . One example of his- torical analogy would be the use of demand data for CD-ROMs to predict the demand curve for DVDs, or Broadway shows to predict the demand for movies.

A special type of expert panel uses what is called the Delphi method. The RAND Corporation developed the Delphi method as a group technique for forecasting the demand for new or contemplated products or services. The intent was to eliminate the undesirable effects of interaction between members of the group (such as loud and dominating individuals) while retaining the benefi ts of their broad experience and knowledge. The method begins by having each member provide individual written forecasts, along with any supporting arguments and assumptions. These forecasts are submitted to a Delphi researcher, who edits, clarifi es, and summarizes the data. These data are then provided as feedback to the members, along with a second round of questions. This procedure continues, usually for about four rounds, when a con- sensus among panel members can often be reached on some of the issues.

Another qualitative device often used in forecasting is called life-cycle analysis . Experienced managers who have introduced several new products are often able to estimate how long a product will remain in each stage of its life cycle. This forecast, coupled with other market information, can produce reasonably accurate estimates of demand in the medium to long range.

Quantitative forecasting methods are generally divided between methods that simply project the past history or behavior of the variable into the future ( time series analysis ) and those that also include external data ( causal ). Time series analysis is the simpler of the two and ranges from just using an average of the past data to using regression analysis corrected for seasonality in the data. Simple projection techniques are obviously limited to, and primarily used for, very short-term forecast- ing. Such approaches often work well in a stable environment but cannot react to changing industry factors or changes in the national economy.

Causal methods, which are usually quite complex, include histories of external factors and employ sophisticated statistical techniques. In addition to using spread- sheets, many “canned” software packages are available for the quantitative tech- niques, both time series analysis and causal.

Factors Infl uencing the Choice of Forecasting Method Which method is chosen to prepare a demand forecast depends on a number of fac- tors. First, long-range (2- to 5-year) forecasts typically require the least accuracy and

c08.indd 356c08.indd 356 25/09/12 5:20 PM25/09/12 5:20 PM

357T i m e S e r i e s A n a l y s i s

are only for general (or aggregate) planning, whereas short-range forecasts require greater accuracy and are for detailed operations. Thus, the most accurate methods are usually used for short-term needs and, fortunately, the data in the near term is usually the most accurate.

Second, if the data are available, one of the quantitative forecasting methods can be used. Otherwise, nonquantitative techniques are required. Attempting to forecast without a demand history is almost as hard as using a crystal ball. The demand his- tory need not be long or complete, but some historical data should be used if at all possible.

Third, the greater the limitation on time or money available for forecasting, the more likely it is that an unsophisticated method will have to be used. In general, management wants to use a forecasting method that minimizes not only the cost of making the forecast but also the cost of an inaccurate forecast; that is, manage- ment ’s goal is to minimize the total forecasting costs. Costs of inaccurate forecasting include the cost of over- or understocking an item, the costs of under- or overstaff- ing, and the intangible and opportunity costs associated with loss of goodwill because a demanded item is not available.

Fourth, with the advent of computers, the cost of statistical forecasts based on historical data and the time required to make such forecasts have been reduced sig- nifi cantly. It has therefore become more cost-effective for organizations to develop more sophisticated forecasts.

In the remainder of this supplement, we briefl y overview several of the quantita- tive forecasting methods. In the next section, time series analysis is addressed. Then, in the following section, causal methods are discussed.

T I M E S E R I E S A N A L Y S I S A time series is simply a set of values of some variable measured either at regular points in time or over sequential intervals of time. We measure stock closing prices at specifi c points in time and quarterly sales over specifi c intervals of time. If, for example, we recorded the number of books sold each month of the previous year at Amazon.com and kept those data points in the order in which they were recorded, the 12 numbers would constitute a 12-period time series. Time series data can be collected over very short intervals (such as hourly sales at a fast-food restaurant) or very long intervals (such as the census data collected every 10 years).

Components of a Time Series We analyze a time series because we believe that knowledge of a variable ’s past behavior might help us understand (and therefore help us predict) its behavior in the future, normally just the next period. In some instances, such as the stock mar- ket, this assumption may be unjustifi ed, but in planning many operational activities, history does (to some extent, at least) repeat itself and past tendencies continue. Our goal is to fi nd a forecasting model that is easy to compute and use, responsive to changes in the data, and accurate in its predictions. To begin our discussion of time series analysis, let us consider the component parts of any time series.

c08.indd 357c08.indd 357 25/09/12 5:20 PM25/09/12 5:20 PM

358 C h a p t e r 8 S u p p l e m e n t : F o r e c a s t i n g

To analyze time series data, it is often helpful to think of it as being comprised of four components:

1. Trend, T

2. Seasonal variation, S

3. Cyclical variation, C

4. Random variation, R

The trend is the long-run direction of the time series, including any constant amount of demand in the data. Figure 8S.2 illustrates three fairly common trend lines showing changes in demand; a horizontal trend line would indicate a constant, unchanging level of demand.

A straight-line or linear trend (showing a constant amount of change, as in Panel A of Figure 8S.2 ) could be an accurate fi t to the historical data over some limited range of time, even though it might provide a rather poor fi t over an entire time series. Panel B in the fi gure illustrates the situation of a constant percentage change. Here, changes in a variable depend on the current size of the variable (rather than being constant each period, as in Figure 8S.2 A). The trend line shown in Panel C of Figure 8S.2 resembles the life cycle or “stretched-S” growth curve that describes the demand many products and services experience over time.

Seasonal fl uctuations are fairly regular fl uctuations that repeat within one year ’s time, or whatever period encompasses the full set of seasonals. Seasonal fl uctuations result primarily from nature, but they are also brought about by human behavior. Sales of heart-shaped boxes of candy and pumpkins are brought about by events that are controlled by humans. Snow tires and antifreeze enjoy a brisk demand during the winter months, whereas sales of golf balls and bikinis peak in the spring and summer months. Of course, seasonal demand often leads or lags behind the actual season. For example, the production season for meeting retailers ’ demand for Christmas goods is August through September. Also, seasonal variation in events need not be related to the seasons of the year. For example, fi re alarms in New York City reach a “seasonal” peak at 7 P.M. and a seasonal low at 7 A.M. every day. And restaurants reach three sea- sonal demand peaks every day at 7:00 A.M., 12:30 P.M., and 7 P.M.

The cycle or cyclical component is obvious only in time series that span several sets of seasonals or more. A cycle can be defi ned as a long-term oscillation, or a swing of the data points about the trend line over a period of at least three com- plete sets of seasonals. National economic cycles of booms as well as depressions and periods of war and peace are examples of such cycles.

Cycles, particularly business cycles, are often diffi cult to explain, and economists have devoted considerable research and speculation to their causes. Identifi cation of a cyclic pattern in a time series requires the analysis of a long period of data. For most decision-making situations, forecasting the cyclic component is not considered, since long-term data are typically unavailable to determine the cycle. In addition, cycles are not likely to repeat in similar amplitude and duration; hence, the assump- tion of repeating history does not hold.

Random variations are, as the name implies, without a specifi c assignable cause and without a pattern. Random variations are the fl uctuations left in the time series after the trend, seasonality, and cyclical behaviors have been accounted for. Random fl uctuations can sometimes be explained after the fact, such as an increase in the

c08.indd 358c08.indd 358 25/09/12 5:20 PM25/09/12 5:20 PM

359T i m e S e r i e s A n a l y s i s

consumption of energy owing to abnormally harsh weather, but cannot be system- atically predicted and, hence, are not included in time series models.

The objective of time series analysis is to determine the magnitude of one or more of these components and to use that knowledge for the purpose of forecasting

1

E F G H I J K

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

(A) Constant Change

(B) Constant Percent Change

(C) Standard “S” Curve

40

20

0

0

800

600

400

200

0

5

10

15

20

25

30

V ar

ia b

le o

f In

te re

st V

ar ia

b le

o f

In te

re st

1 2 3 4 5

Time

V ar

ia b

le o

f In

te re

st

6 7 8 9 10 11 12 13 14 15

1 2 3 4 5

Time

6 7 8 9 10 11 12 13 14 15

1 2 3 4 5

Time

6 7 8 9 10 11 12 13 14 15

Figure 8S.2 Three common trend patterns.

c08.indd 359c08.indd 359 25/09/12 5:20 PM25/09/12 5:20 PM

360 C h a p t e r 8 S u p p l e m e n t : F o r e c a s t i n g

the next period. In the remainder of this section we will consider three models of time series analysis:

1. Moving averages (trend component of the time series)

2. Exponential smoothing (trend component of the time series)

3. Linear trend, multiplicative model (trend and seasonal components)

Moving Averages

The moving average technique is one of the simplest ways to predict a trend. It gen- erates the next period ’s forecast by averaging the actual demand for only the last n time periods ( n is often in the range of 4 to 7). That is:

Forecast   5   average of actual demand in past n periods

Any data older than n are thus ignored. Note also that the moving average weights old data just the same as more recent data. The value of n is usually based on the expected seasonality in the data, such as 4 quarters or 12 months in a year; that is, n should encompass one full cycle of data. If n must be chosen arbitrarily, then it should be based on experimentation; that is, the value selected for n should be the one that works best for the available historical data.

Mathematically, a forecast using the moving average method is computed as

F t � 1

  5   1 ___ n t

∑ i 5 ( t 2 n 1 1 )

A i

where

t 5 period number for the current period F

t   �   1 5 forecast for the next period

A i 5 actual observed value in period i

n 5 number of periods of demand to be included in the moving average (known as the “order” of the moving average)

To illustrate the use of the moving average, data were collected on Intel ’s monthly stock closing price, as shown in Figure 8S.3 . In the spreadsheet shown, a four-period moving average was computed by entering the formula 5 AVERAGE(B2:B5) in cell C6 and copying it to cells C7:C26. The graph with the actual time series and moving aver- age illustrates how a moving average smoothes out the fl uctuations in the time series.

The plot also demonstrates one of the weaknesses associated with using moving averages. Specifi cally, whenever there is an upward or downward trend in the data, a forecast based on the moving average approach will always lag the time series. Therefore, the moving average approach is most appropriate for situations where the decision maker would like to simply smooth out fl uctuations around an assumed horizontal trend.

A refi nement of the moving average approach is to vary the weights assigned to the values included in the average. Such an approach is called a weighted moving average , with the newer data typically weighted more heavily, rather than using equal weights. The reason for weighting the newer data more heavily is that since it is more current, it is often considered to be more representative of the future.

c08.indd 360c08.indd 360 25/09/12 5:20 PM25/09/12 5:20 PM