quality management

PLEASE SCROLL DOWN FOR ARTICLE

This article was downloaded by: [Baez, Pablo] On: 14 June 2010 Access details: Access Details: [subscription number 778060948] Publisher Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37- 41 Mortimer Street, London W1T 3JH, UK

Quality Engineering Publication details, including instructions for authors and subscription information: http://www.informaworld.com/smpp/title~content=t713597292

Quality Quandaries: Improving the Invoicing Process of a Consulting Company Tashi P. Erdmanna; Manon de Grootb; Ronald J. M. M. Doesa a Institute for Business and Industrial Statistics, University of Amsterdam, The Netherlands b REAAL Insurances, Utrecht, The Netherlands

Online publication date: 07 June 2010

To cite this Article Erdmann, Tashi P. , Groot, Manon de and Does, Ronald J. M. M.(2010) 'Quality Quandaries: Improving the Invoicing Process of a Consulting Company', Quality Engineering, 22: 3, 214 — 221 To link to this Article: DOI: 10.1080/08982111003771854 URL: http://dx.doi.org/10.1080/08982111003771854

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf

This article may be used for research, teaching and private study purposes. Any substantial or systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material.

Quality Quandaries: Improving the Invoicing Process of a

Consulting Company Tashi P. Erdmann1,

Manon de Groot2,

Ronald J. M. M. Does1

1Institute for Business and

Industrial Statistics, University of

Amsterdam, The Netherlands 2REAAL Insurances, Utrecht,

The Netherlands

INTRODUCTION

Although Lean Six Sigma originates from a manufacturing environment,

over the past decennium the method has been used increasingly for

improvement projects in the service industry (for an introduction see De

Mast et al. 2006). Though projects in the manufacturing industry often

involve experimenting and advanced statistical techniques, in services the

most challenging part is in merely structuring and quantifying the problem.

Once the problem is properly structured and made explicit, the solution is

often quite straightforward and does not require advanced statistics.

Therefore, the key to a successful project in services is to clearly define

and quantify the problem. The first two phases of the define, measure,

analyze, improve, control (DMAIC) roadmap used in Lean Six Sigma, the

phases define and measure, focus on these important activities.

As an illustration, this article discusses an improvement project on the

invoicing process in a large IT and business services company. The goal of

the project was to reduce the throughput and processing time, such that the

client would pay the invoice as soon as possible, and the company would

reduce personnel costs. The yearly benefits of the project were targeted in

advance to be at least 450,000 per year, as a result of increased interest earnings

and reduced operational costs. For reasons of confidentiality and for illustrative

purposes, some of the data and facts have been slightly modified.

DEFINE PHASE

The most important elements of the define phase of a Lean Six Sigma

project are clearly defining the process that is to be improved, stating the

project objectives and analyzing their potential benefits, and setting

up the project organization and time planning.

In the project that is discussed here, the process that is improved is the

process of invoicing consulting fees to clients. The supplier, input, process,

output, and client (SIPOC) model of the process is shown in Figure 1.

Only the invoices regarding consultancy activities by the Financial

Services division are within the scope of the project. At high level, the

invoicing process is organized in the five following steps: First, the consult-

ant fills in his or her time sheet. The consultant’s division checks and

Edited by Ronald J. M. M. Does.

Address correspondence to Ronald J. M. M. Does, IBIS UvA, University of Amsterdam, Plantage Muidergracht 12, 1018 TV Amsterdam, The Netherlands. E-mail: r.j.m.m. [email protected]

Quality Engineering, 22:214–221, 2010 Copyright # Taylor & Francis Group, LLC ISSN: 0898-2112 print=1532-4222 online DOI: 10.1080/08982111003771854

214

D o w n l o a d e d B y : [ B a e z , P a b l o ] A t : 2 2 : 2 7 1 4 J u n e 2 0 1 0

approves the timesheet and then sends it to the

Shared Service Center for processing. The Shared Ser-

vice Center performs several checks and prepares the

invoice. The invoice is then sent to the division for a

check and signature and then finally sent to the client.

The project objectives consist of a reduction in the

throughput time, the number of full-time equivalents

(FTEs) involved, and the number of mistakes made in

the invoicing process. The throughput time is especially

important, because each day that the invoices are paid

faster on average will lead to extra interest earnings of

417,000 per year, based on 4183 million of total yearly

revenues. An assumption made here is that extra

throughput time causes the payment to be delayed with

on average the same amount of time, which seems

reasonable. Another important anticipated benefit of

the project is expected to be achieved by reducing

the processing time of time sheets and invoices by the

employees of the Financial Services division and the

Shared Service Center. At the beginning of the project,

the total number of FTEs working on invoices is around

eight: five at the Financial Services division and three at

the Shared Service Center. A third project objective is a

decrease in the number of mistakes. Receiving an incor-

rect invoice is clearly dissatisfying for clients. Addition-

ally, mistakes in invoices lead to delayed payments and

extra work for the consulting company. Currently, 11%

of the invoices sent to clients are incorrect.

The project team consists of three consultants from

the Financial Services division, one of whom is black

belt (i.e., project leader). They report to a champion,

who has given them the assignment to do the project.

The project leader is coached by an external master

black belt, an expert in the Lean Six Sigma method-

ology. The project is to be completed within 6

months, with planned reviews by the champion

and master black belt at the end of each phase.

MEASURE

In the measure phase, the project objectives are

operationalized as requirements on quantifiable

and measurable quality characteristics. Then, a

procedure is established to measure these characteri-

stics, and this measurement procedure is validated.

In Lean Six Sigma, these quality characteristics are

often called critical to quality characteristics (CTQs).

The three CTQs in this project are processing time,

waiting time, and the number of mistakes. Processing

time and waiting time together add up to the total

throughput time of the invoicing process and in this

way influence the time to payment by the client and

the service quality as perceived by the client. Proces-

sing time determines the number of employees

necessary in this process, measured in FTEs. The

number of mistakes influences both processing time

and waiting time and is also a component of service

quality. The way the CTQs relate to the project

objectives and the strategic goals of the company

can be schematically displayed in a so-called CTQ

flowdown (cf. De Koning and De Mast 2007). The

CTQ flowdown for this project is shown in Figure 2.

A common measurement plan used in Lean Six

Sigma projects involving time measurements is to

record processing times and waiting times for

individual files as they make their way through the

process. In this invoicing process, however, whole

batches of files (time sheets and invoices) are

processed together in so-called billing cycles of

approximately 8 working days, which take place

twice per month. Therefore, the black belt decided

FIGURE 2 The CTQ flowdown of the invoicing process.

FIGURE 1 The SIPOC of the invoicing process.

215 Improving the Invoicing Process of a Consulting Company

D o w n l o a d e d B y : [ B a e z , P a b l o ] A t : 2 2 : 2 7 1 4 J u n e 2 0 1 0

to record the total processing times and waiting

times for all files together, rather than those for the

individual files separately.

The CTQs are measured for each process step.

The CTQ processing time is defined as the time

employees are occupied with activities in order to

process all time sheets and invoices. The CTQ wait-

ing time is the time in between steps when no time

sheet or invoice is being processed, for each process

step separately. For steps within the Financial

Services division, the processing time and waiting

time are measured per employee for a single billing

cycle. For steps performed by the Shared Service

Center the times are recorded per billing cycle, based

on computer system reports of the last 3 months,

covering six bimonthly billing cycles. The CTQ

number of mistakes is a count of all mistakes found

by employees or in the computer system reports

during each process step. The requirement for all

three CTQs is that they should be minimized.

There are several issues regarding the validity of

the measurement procedure. A possible problem is

the fact that the processing times and waiting times,

as measured, do not cover part of the rework that is

caused by mistakes. For example, additional proces-

sing and waiting times outside the billing period are

not recorded. Another issue is that mistakes that are

not immediately corrected when they are discovered

are counted more than once, so the number of

mistakes found in each process step does not add

up to the actual total number of mistakes. Further-

more, a disadvantage of this measurement procedure

is that the amount of variation in the total throughput

time of the process cannot be determined, because

times are not recorded for individual files but for

batches of files. Although the measurement pro-

cedure has some drawbacks, it still gives a reason-

ably good insight into the nature of the problem.

ANALYZE

In the analyze phase the current performance of

the CTQs is determined, based on the collected data.

A thorough analysis leads to a diagnosis of the

problem and a list of potential influence factors.

A useful tool for describing the process flow in

detail and visualizing forms of waste in the process

is the so-called value stream map (see Kemper et al.

2010; Womack and Jones 2003). In a value stream

map for each process step the average processing

time and average waiting time is given, and for each

conditional routing the percentage following that

routing is given. Different forms of waste in the

process can be shown in the value stream map.

Two value stream maps for the invoicing process

are shown in Figure 3.

The process is divided into two parts: the first part

of the process takes place in the Financial Services

division, and the second part takes place in the

Shared Service Center. In the value stream map differ-

ent layers are visible, representing the different actors

that play a part in the process: the consultants, who

fill in their time sheets; the secretaries and team

managers, who each perform a check and make sev-

eral corrections to the time sheets and invoices; the

employee of Planning and Control, who creates pro-

ject codes and performs a check; the employees of

the Shared Service Center, who perform six consecu-

tive checks of the time sheets and invoices, make sev-

eral corrections, and create the invoices; the account

managers who perform a check; and, finally, the

clients. Different forms of waste are visualized using

self-explanatory symbols. In Figure 3 there are

symbols for transportation, redundant work and

overprocessing, rework, complexity, and waiting time.

Based on the measurements and with help of the

value stream map, the current performance can be

assessed. The total throughput time of the invoicing

process is currently 32 days on average and can be

divided into three parts: the delivery of time sheets

by the Financial Services division takes 1 day, the

standard process that takes place twice per month

at the Shared Service Center takes 8.5 days, and then

further nonstandard corrections and issues that arise

in the preparation for sending take 22.4 days on

average, including rework after incorrect invoices

are sent back by clients. It is clear that the last part,

consisting of 22.4 days of rework and waiting, is

the largest and should therefore be the focus of the

project. Reducing the number of mistakes will prob-

ably greatly help in reducing this huge amount of

rework. But also the first two parts of the throughput

time, which together form the standard process, are

full of inspections, rework, and waiting time. In the

standard process, altogether nine separate checks

are performed on the time sheets and invoices before

the invoice reaches the client. Combining some of

these inspections and redesigning the process will

T. P. Erdmann et al. 216

D o w n l o a d e d B y : [ B a e z , P a b l o ] A t : 2 2 : 2 7 1 4 J u n e 2 0 1 0

probably reduce the throughput time with several

days.

The total processing time of all time sheets and

invoices in one billing cycle added together is

currently 420 man-hours at the Financial Services

division and 217 man-hours at the Shared Service

Center. This implies that at the Financial Services

division currently 5.3 FTEs are needed and at the

Shared Service Center 2.8, considering that yearly

there are 24 billing cycles and one FTE works 1,886

hours. An overview of the different components

of processing time is given in Table 1.

Much of the processing time is caused by correction

of mistakes, both during the standard process and after

the standard process. After the computer system of the

Shared Service Center creates an error report, the

errors—on average 29% of the 1,200 time sheets have

one error in the error report—have to be corrected

by the Financial Services division, taking around 15

minutes per error. Then after the standard process,

nonstandard corrections and issues that arise in the

preparation for sending cost 30 minutes on average

for each of the 477 invoices, 15 minutes for the

Financial Services division and 15 minutes for the

Shared Service Center. All those corrections are direct

consequences of the large number of mistakes.

The total number of mistakes in all time sheets and

invoices that were found during one billing period

FIGURE 3 Value stream maps of the two processes in the Financial Services division and Shared Service Center, respectively.

217 Improving the Invoicing Process of a Consulting Company

D o w n l o a d e d B y : [ B a e z , P a b l o ] A t : 2 2 : 2 7 1 4 J u n e 2 0 1 0

was 1,172. In this billing period altogether 477

invoices were sent, implying that the average

number of mistakes per invoice was 2.5. Note that

some mistakes may have been counted double,

because they were not corrected the first time. The

numbers of mistakes found in each inspection

separately are given in Table 2.

One of the main objectives will be to reduce this

large number of mistakes, because this will result

in higher client satisfaction but also in a reduction

in throughput time and processing time.

Based on this analysis the potential benefits of

the project are recalculated as follows. The average

number of mistakes is targeted to decrease with

about 90% from 2.5 mistakes per invoice to 0.2

per invoice. We hope that with this improvement

the number of incorrect invoices sent to clients

will reduce proportionally from 11 to 1% of the

invoices. This reduction in the number of mistakes

combined with a more efficiently organized process

will (hopefully) cause the total throughput time to

decrease from 32 days to 10 days, leading to

4374,000 extra interest earnings per year. This is

based on a 90% reduction of the 22.4 days of

rework because of the reduction in the number

of mistakes and a 25% reduction of the 8.5 days

of standard process because of a more efficiently

organized process. The total processing time per

billing period is targeted to decrease from 637 to

226 man-hours, such that the total number of FTEs

will decrease from 8.1 to 2.9, saving 4281,000 per

year. This is based on a 90% reduction in proces-

sing time of the process steps that are written in

italics in Table 1, because they are almost entirely

caused by mistakes or considered redundant. Alto-

gether the financial benefits of the projected are

now targeted to be 4655,000: much more than

was expected in the define phase.

The last step of the analyze phase is the creation

of a list of factors that influence the CTQs. In the

invoicing process, the project manager defined five

main influence factors, based on analysis of the

measurement data, analysis of the process flow,

and conversations with many of the employees

working in the process. They are as follows:

1. clarity of operating procedures and responsibilities;

2. adequacy of information management;

3. correctness and timeliness of time sheets;

4. effectiveness of inspections and checks;

5. the adequacy of computer systems.

These five influence factors are thought to be the

main causes of variability in the CTQs processing

time, waiting time, and the number of mistakes. They

will be further explained in the next section. These

influence factors will be adjusted in order to optimize

the process.

TABLE 2 Number of Mistakes per Process Step

Process step

Number of

mistakes found

Check by team secretary 271

Check by team manager 130

Check by Planning and Control 1

Error report by software system 348

Fee rate reports 9

Extra check before validation run 238

Validation run 14

Check invoice proposals 31

Check by account manager 19

Client 116

Total 1,177

TABLE 1 Processing Times of the Process Steps

Process step

Total processing

time in hours

Consultant 160.0

Print and check by secretary 16.8

Correction by secretary 13.1

Check and correction by team manager 20.1

Check by Planning and Control 0.3

Check by account manager 3.6

Corrections after error reports 87.0

Corrections after standard process 119.3

Total at division Financial Services 420.1

Enter in system 29.3

Check fee rate reports and correction 24.0

Extra check before TS03 and correction 2.0

Corrections after TS03 0.4

Corrections of invoice base 29.2

Check invoice proposals and correction,

and create invoice

7.2

Correct errors in invoices 2.4

Prepare invoice and corrections after

standard process

119.3

Send invoice 3.2

Total at Shared Service Center 217.0

Process steps written in italics are almost entirely (a) caused by mis- takes or (b) considered redundant.

T. P. Erdmann et al. 218

D o w n l o a d e d B y : [ B a e z , P a b l o ] A t : 2 2 : 2 7 1 4 J u n e 2 0 1 0

IMPROVE

The purpose of the improve phase is to arrive at a

number of improvement actions that together will

cause the CTQs to satisfy their requirements. The

improvement actions are based on the influence fac-

tors identified in the analyze phase. In the improve

phase these influence factors are examined critically

and are prioritized based on their effect on the CTQs

and their changeability.

In the invoicing process, the effect of the five

influence factors mentioned earlier is quite clear

from the value stream maps in Figure 3 and from

conversations with employees and does not need

elaborate evidence in the form of experiments. The

circumstantial evidence provided by all problems,

disturbances, mistakes, and inefficiencies experi-

enced by the people working in the process is

convincing enough to design and implement

improvement actions directly. In the following, the

five influence factors will be further explained.

The clarity of operating procedures and responsibil-

ities is an important influence factor. The different

actors who play a part in the process currently do

not have a clear picture of the process as a whole

nor of their own tasks, authorities, and responsibilities.

There is no central unit managing the process, and

there are no standard operating procedures. Because

of this, the predictability of the process is poor. Fur-

thermore, the responsibilities are not well defined.

Some employees perform activities for which they

are not authorized. The number of different actors

playing a part in the process is large, and this leads

to a lot of communication between different actors,

especially because it is unclear who is responsible

for what. Too many different people are doing the

same or similar tasks. Many agreements are violated

and mistakes are made, because people do not know

what the agreements are. Because of the absence

of clear operating procedures, many tasks are not

organized in an efficient way. For example, there are

no separate procedures for different types of contracts,

although it makes a large difference in average

throughput time whether a contract is a fixed price

or a time and material contract. Both types of contracts

are treated as if they were the same, causing much

rework and large delays after the standard process.

The adequacy of information management is

a second influence factor. At the moment, the

management and general accessibility of information

is very poor. The information necessary is often

unavailable for the persons involved. For example,

when consultants are absent, secretaries sometimes

have to fill in time sheets for the consultants based

on e-mail correspondence, because the consultant’s

time registration is not readily available to them. As

a consequence, many actions and decisions are

taken based on incorrect or incomplete information,

leading to mistakes. Many of the checks and inspec-

tions that are done during the process are imposs-

ible, or can only be done partially, because not all

information that is needed is available to the person

who performs the inspection. For example, infor-

mation about the client contracts, which is one of

the most important inputs for the process, is only

available to account managers and is often not up

to date, and therefore checks on completeness

and correctness of the invoices cannot take place

until the invoices are seen by the account managers

in one of the last process steps. This is one of

the reasons that so many different inspections are

necessary. Another consequence is that a lot of

additional communication has to take place between

the different actors in the process.

A third important influence factor is the correct-

ness and timeliness of time sheets. The input that

the consultants deliver is of poor quality, full of

mistakes, incomplete, and often late. For example,

some consultants do not specify the number of

kilometers they have driven, or they do not follow

the rules for time registration specified in the contract

of a specific client. This has to do with insufficient

access to information, as explained in the previous

paragraph, and with a lack of good instructions for

the consultants. For most consultants it is not clear

which requirements their time sheets have to meet.

Consultants never become aware of their mistakes,

because they are not informed of them. In addition,

the discipline of the consultants is poor, because

they do not face any consequences when their time

sheets do not meet the requirements.

Influence factor number four is the effectiveness of

inspections and checks. Many inspections take place

in the current standard process. Altogether there are

nine checks. Nevertheless, the invoices are still full of

mistakes, despite all the checks. Inspections take

place both manually and automatically. In the man-

ual inspection mistakes are made because of missing

219 Improving the Invoicing Process of a Consulting Company

D o w n l o a d e d B y : [ B a e z , P a b l o ] A t : 2 2 : 2 7 1 4 J u n e 2 0 1 0

information and because of the differences in the

working methods of the actors in the process. In

the automatic inspection mistakes are made, because

of incomplete or incorrect input of parameters on

which the software system bases its checks. Some

of the nine checks seem redundant, because hardly

any mistakes are found, such as the check by

Planning and Control, the check of the fee rate

reports, and the check by the account manager.

The last important influence factor is the adequacy

of computer systems. At the moment there are many

mistakes in the checks performed by the software

system. This is because the data on which the con-

figuration of the system is based are often incorrect

and incomplete. The current software system used

for the checks is more of an accounting system

than an operational management system. It has

insufficient possibilities in processing management

information. The different systems used in the organi-

zation are not compatible with each other, such that

it sometimes occurs that information that is changed

in one system remains unchanged in another system.

The interface of the software for hour registration is

felt to miss some aspects, and it would be easy for

the consultants if standard values were given for

the different categories. Another software-related

issue is that authorization by the account manager

and team manager could be given with an Internet-

based system. Signatures on paper by the team

manager and by the account manager appear to be

unnecessary and needlessly take time for printing

and transportation of the physical documents.

Based on the five influence factors described

above, the black belt has designed the following

improvement actions:

1. Standard operating procedures will be written.

These standard operating procedures and respon-

sibilities will ensure that the process is manageable

and will also contain feedback loops that define

actions to be taken when mistakes are made.

Responsibilities will be clearly defined. A central

unit of three FTEs will be established within

the Financial Services division and will serve as

contact point for the Shared Service Center.

2. Information will be made available centrally. This

central database should contain information

regarding time allocation, consultant-specific

agreements, contract administration, and project

code administration and should be accessible to

everyone and kept up to date. The central unit

described above will manage this central database.

3. Instructions to consultants will be improved, and

consultants will receive feedback when they have

made a mistake. To improve discipline, the con-

sultants will receive a reminder by text message

before the start of each billing period, and the

quality and timeliness of the time sheets will

become part of the performance indicators used

in the personal review of the consultants.

4. The central unit, which will have direct access to all

the necessary information, will take over all inspec-

tions and checks formerly done by others. It will

deliver correct invoices to the Shared Service

Center, which will not have to do any inspections

anymore. Several checks will be combined,

and redundant checks will not be done anymore.

5. A software system for operational management

will be implemented, which must be compatible

with the other software systems used. Further-

more, the interface of the hour registration soft-

ware will be designed in a way that makes it

easy for the consultants to fill in their time sheets.

In addition, Web-based authorization will replace

signatures on physical documents.

CONTROL

The last phase of the DMAIC method of Lean Six

Sigma is the control phase. In the control phase the

process control is improved. This is done by docu-

menting the improved process, creating a control

plan, organizing continuous improvement, and

defining roles and responsibilities. Then, finally, the

benefits of the project are determined, and the black

belt is discharged of the project.

In the new situation of the invoicing process, the

central unit is set up with two employees who used

to work at the Shared Service Center and one

employee who used to be a secretary. The number

of mistakes drops dramatically and in the first year

after the implementation of the improvements, on

average around 2% of the invoices sent to clients

contain errors. Because of the reduction in the num-

ber of mistakes and the resulting efficiency improve-

ment, altogether three secretaries are not necessary

anymore and are given jobs at a different division,

leading to personnel cost savings for the Financial

T. P. Erdmann et al. 220

D o w n l o a d e d B y : [ B a e z , P a b l o ] A t : 2 2 : 2 7 1 4 J u n e 2 0 1 0

Services division of around 4153,000. The average

total throughput time reduces from 32 to 10 days,

which increases interest earnings by approximately

4374,000.

CONCLUSION

The project about the invoicing process was a big

success, with huge financial benefits to the consult-

ing company. It shows how the DMAIC roadmap

can help in structuring the problem. In the service

industry, many processes have never been analyzed

before and are full of potential improvements, the

so-called low-hanging fruit. The define and measure

phases of Lean Six Sigma give direction to the

project. The solid structure and clear focus that result

from the step-by-step approach help a great deal in

solving problems. Once the black belt knows where

to look, solutions are easily found.

REFERENCES

De Koning, H., De Mast, J. (2007). The CTQ flowdown as a conceptual model of project objectives. Quality Management Journal, 14(2): 19–28.

De Mast, J., Does, R. J. M. M., De Koning, H. (2006). Lean Six Sigma for Service and Healthcare. Alphen aan den Rijn, The Netherlands: Beaumont.

Kemper, B. P. H., De Mast, J., Mandjes, M. R. H. (2010). Modeling process flow using diagrams. Quality and Reliability Engineering International, 26(4):341–349.

Womack, J. P., Jones, D. T. (2003). Lean Thinking: Banish Waste and Create Wealth in Your Corporation. Florence, MA: Free Press.

221 Improving the Invoicing Process of a Consulting Company

D o w n l o a d e d B y : [ B a e z , P a b l o ] A t : 2 2 : 2 7 1 4 J u n e 2 0 1 0