1 s2.0-s0748575109000402-main

J. of Acc. Ed. 27 (2009) 104–123

Contents lists available at ScienceDirect

J. of Acc. Ed.

journal homepage: www.elsevier .com/locate/ jaccedu

Case

Morgan Systems, Inc.: Application of Six Sigma to thefinance function

Timothy C. Krehbiel a,1, Jan E. Eighme b,*, Phillip G. Cottell b,2

a Department of Management, Miami University, Oxford, OH 45056, United Statesb Department of Accountancy, Miami University, Oxford, OH 45056, United States

a r t i c l e i n f o a b s t r a c t

Keywords:Six SigmaProblem-based learningDMAICFinancial reporting

0748-5751/$ - see front matter � 2009 Elsevier Ltdoi:10.1016/j.jaccedu.2009.11.002

* Corresponding author. Tel.: +1 513 529 6200.E-mail addresses: [email protected] (T.C. K

Cottell).1 Tel.: +1 513 529 4837.2 Tel.: +1 513 529 6200.

This teaching case is based on a Six Sigma project undertaken by asubsidiary of a Fortune 100 company to improve its quarterlyfinancial-reporting process. It is presented as a six-phase Prob-lem-Based Learning (PBL) unfolding problem. The first five phasescorrespond to the stages of the Define–Measure–Analyze–Improve–Control (DMAIC) model, a process-improvement meth-odology used extensively in Six Sigma. The sixth phase focuseson Six Sigma as a way of doing business.

This case can be used in MBA or Master of Accountancy (MAcc)courses. Upon completion, students should be able to explain theDMAIC model stages and identify tools used in each stage; describea project charter; and interpret a suppliers–inputs–process–out-puts–customers (SIPOC) analysis, cause-and-effect (C & E) diagram,and failure modes and effects analysis (FMEA). Students shouldalso be able to calculate and interpret process sigma levels and riskpriority numbers (RPNs).

� 2009 Elsevier Ltd. All rights reserved.

1. Introduction

Six Sigma is a quality-management system that evolved in Motorola during the 1980s. Technicallyspeaking, Six Sigma is the rigorous pursuit of variance reduction leading to the design of business pro-cesses that produce no more than 3.4 defects per million opportunities. As with other quality-manage-ment systems, Six Sigma is a packaging of statistical and managerial methods. Although originally

d. All rights reserved.

rehbiel), [email protected] (J.E. Eighme), [email protected] (P.G.

T.C. Krehbiel et al. / J. of Acc. Ed. 27 (2009) 104–123 105

intended for a manufacturing setting, many companies, including Motorola, GE, and Bank of America,are today reaping bottom-line benefits by using Six Sigma to improve transactional processes, includ-ing those in the accounting and finance functions (Krehbiel, Havelka, & Scharfenort, 2007).

The major objective of this case is for students to learn the applicability of Six Sigma to accountingprocesses. Students will also learn basic Six Sigma terminology and tools, and develop a strongerappreciation for processes and systems thinking. This case is unique because it is a Problem-BasedLearning (PBL) unfolding problem based on an actual Six Sigma project undertaken by a subsidiaryof a Fortune 100 company to improve its quarterly financial-reporting process. Although the namesof the company and the people involved have been changed, the data and all other facets of the caseare represented here as accurately as possible.

PBL is utilized by university faculty interested in enhancing the critical thinking and analyticalskills of their students.3 The essential aspect of PBL is that students confront a problem to be solved be-fore they are given any information as to how to solve it. This turns on its head the typical pedagogywhere students are given material, usually through lecture or demonstration, and then they solve prob-lems. For this reason, the PBL case more closely resembles situations that people in business face whereproblems are encountered before ways to resolve the issues are identified. In the PBL class, students arelooking for information rather than having the material given to them. Information acquired by the stu-dents is perceived as being more useful in the PBL case because it has an immediate impact on solving aproblem that is being faced (Duch, Groh, and Allen, 2001a).

The PBL environment encourages students to learn how to learn. Johnstone and Biggs (1998) pro-pose PBL as a curricular structure that can be implemented to achieve the integration of technicalinformation, practical experience, and life-long learning skills that aid in the development of expertise.Instructors may consult Duch, Groh, and Allen (2001b) for an in-depth perspective about PBL.

Duch (2001) describes the characteristics of good PBL problems. She notes that many PBL problemsare designed with multiple stages. In explaining these multi-stage problems, Bellas, Marshall, Reed,Venable, and Whelan-Berry (2000) call them ‘‘unfolding problems.” Essentially, an unfolding problemdiffers from the traditional classroom problem or case, in that the problem itself does not contain suf-ficient information for the student to solve it. In the initial stage, the problem has a vague and broadrequirement. As students wrestle with the issue, they first determine what information they must ob-tain to fulfill the requirement. The problem ‘‘unfolds” as students are given additional information thatallows them to engage the problem at an increasingly concrete level.

Another critical component of PBL has students working in small groups to solve problems. As thestudents work together on the several stages of a problem, they simultaneously learn both course-re-lated concepts and problem-solving skills. For this purpose, Cooperative Learning (CL) structures, asexplained by Cottell and Millis (1993), are helpful. Instructors wishing an overview of CL may consultJohnson, Johnson, and Smith (1991) or Millis and Cottell (1998).

We believe that the Six Sigma project undertaken by Morgan Systems, Inc. to improve its quarterlyfinancial-reporting process lends itself perfectly to the step-by-step progression of a PBL unfoldingproblem. The natural step-by-step progression of the Define–Measure–Analyze–Improve–Control(DMAIC) model, a five-stage process-improvement model that is used extensively in Six Sigma pro-jects, fits the design of a PBL unfolding problem. Just as Six Sigma teams in industry complete pro-cess-improvement projects one step at a time, students complete this case, when delivered in thePBL format, one step at a time.

This case has been used in an MBA program that focuses on internal and external processes. Insteadof course content covered from the traditional functional perspective, courses are interdisciplinaryand center around different aspects of the internal or external enterprise. The case was used in thecourse ‘‘Process Design and Improvement,” team taught by a management professor, a statistics pro-fessor, and an accountancy professor. The course covered quality-improvement initiatives like Six Sig-ma, change management, and the balanced scorecard. It appeared in the first semester of the programafter the students had completed a boot-camp experience that included a quick survey of statistical

3 Readers interested in seeing other problems in accounting as well as problems in other disciplines may access the PBLClearinghouse at https://chico.nss.udel.edu/Pbl/. Registration is required, but free.

106 T.C. Krehbiel et al. / J. of Acc. Ed. 27 (2009) 104–123

methods and accounting principles. Approximately half of the students held undergraduate degrees ina business discipline, and the remaining students held degrees in education, the humanities, science,or engineering. None of the students had previous training in Six Sigma.

This case has also been used in an elective course for a Master of Accountancy (MAcc) program. Thecourse, entitled ‘‘Statistical Process Analysis and Improvement for Managers,” was taught by a statis-tics professor. The course covered Six Sigma (with a specific focus on accounting/finance functions)and the statistical methods needed to successfully implement Six Sigma. All the students held under-graduate degrees in accountancy and a prerequisite of one statistics course was enforced. As was thecase with the MBA course, none of the MAcc students had previous training in Six Sigma.

2. Morgan Systems, Inc.

2.1. Phase 1

You have recently been hired as an accountant in the Corporate Reporting Department of a$750 million manufacturing company, hereafter called Morgan Systems, Inc., (MSI), which is awholly-owned subsidiary of a Fortune 100 company. For consolidated reporting purposes, MSI reportsquarterly financial results to its parent company. One of MSI’s financial managers, Rhonda Edwards,realized that the Corporate Reporting Department was taking too long to complete the quarterlyfinancial-reporting process. She sent you the following note on email:

The CFO has instructed me to convene a Six Sigma project to reduce the cycle time of the quarterlyfinancial-reporting process. The CFO will serve as the project champion. Since you are currently inGreen Belt training, I am placing you in charge and was told that this project should help you toearn your Green Belt designation. A Black Belt from MSI’s Continuous Quality Improvement depart-ment4 will serve as your mentor. In addition, seven corporate reporting team members have beenassigned to the project team. I really don’t know a lot about Six Sigma and have several questions.Isn’t Six Sigma just for manufacturing processes? Why might this quality initiative work when wehad such mixed results with TQM? Also, what is this DMAIC thing the CFO kept talking about?

Ms. Edwards scheduled the first meeting of the team for tomorrow morning and noted that theagenda was simply ‘‘Define Stage.”

2.1.1. Phase 1 directions

1. Write a memo to Ms. Edwards addressing her concerns about applying Six Sigma to the financefunction and briefly describe the DMAIC model to her.

2. What is the objective of the Define Stage in the DMAIC model? In general, what tools (statisticaland/or non-statistical) are useful during the Define Stage of a Six Sigma project? What tools seemmost appropriate for this project?

3. Write a problem statement and a goal statement for the project.

2.2. Phase 2

The project team at MSI began the Define Stage by creating a project charter. The primary compo-nents of a project charter are the problem statement and the goal statement. Both statements concernthe most critical outcome of the quarterly financial-reporting process. This critical outcome, often re-ferred to as a critical-to-quality (CTQ) variable, is the number of hours spent preparing quarter-endfinancial schedules. First, the project team agreed upon the following problem statement:

‘‘Too many hours are being spent preparing quarter-end financial schedules.”

Next, the project team decided to use the following goal statement:

4 Played by your professor.

Exhibit 1. Suppliers–inputs–process–outputs–customers analysis (SIPOC).

T.C. Krehbiel et al. / J. of Acc. Ed. 27 (2009) 104–123 107

‘‘Reduce the hours spent preparing quarter-end financial schedules.”

After preparing the project charter, the project team at MSI solicited input from the quarterly finan-cial-reporting process’s primary customer, Corporate Finance, its parent company’s finance depart-ment. The information received from Corporate Finance was viewed as key voice of the customer(VOC) input. It was determined that Corporate Finance’s most critical requirement was that the quar-ter-end financial schedules be received by the requested date. The project team at MSI concluded thatits project charter was aligned with this key VOC input.

Finally, the team created the suppliers–inputs–process–outputs–customers (SIPOC) analysis pre-sented in Exhibit 1.

2.2.1. Phase 2 directions

1. What information can you glean from the SIPOC analysis? What new questions arise from this newinformation?

2. What improvements, if any, should be made to the problem statement? Why?3. What improvements, if any, should be made to the goal statement? Why?4. What is the objective of the Measure Stage in the DMAIC model? In general, what tools (statistical

and/or non-statistical) are useful during the Measure Stage of a Six Sigma project? What tools seemmost appropriate for this project?

2.3. Phase 3

In the Measure Stage, the project team at MSI quantified the quarterly financial-reporting process’sbaseline cycle-time performance. Then the team established a specific cycle-time goal for the process.Table 1 depicts the baseline performance in terms of labor hours worked to complete the 18 quarter-end financial schedules (grouped into three categories: balance sheet, income statement, and inter/in-tra company).

The process’s cycle-time goal was established by setting preparation-time goals for each of the 18quarter-end financial schedules and adding together the associated times. The individual preparation-

Table 1Pre-improvement quarter-end E-Trans schedules: labor hours worked.

Schedule category Number of schedules Hours worked

Balance sheet 8 64.8Income statement 8 16.5Inter/intra company 2 28.0Total 18 109.3

Exhibit 2. Goal vs. baseline performance*.

108 T.C. Krehbiel et al. / J. of Acc. Ed. 27 (2009) 104–123

time goals were based on intra-company benchmarks. Exhibit 2 shows these goals for all 18 schedulesas well as the actual hours needed to prepare each schedule at the end of the most recent quarter. Herethe schedules are categorized by the day of the week they are due. As illustrated in the exhibit, thebaseline performance for 10 of the 18 schedules failed to meet the established goal.

After completing its process assessment in the Measure Stage, the project team at MSI redefined thegoal statement in the project charter as, ‘‘Reduce direct hours worked for the 18 schedules from over100 h total to 26 h total.”

2.3.1. Phase 3 directions

1. Calculate the process sigma level for the quarterly financial-reporting process.2. What information can you glean from Table 1? What new questions arise from this new

information?

T.C. Krehbiel et al. / J. of Acc. Ed. 27 (2009) 104–123 109

3. What information can you glean from Exhibit 2? What new questions arise from this newinformation?

4. What is the objective of the Analyze Stage of the DMAIC model? In general, what tools (statisticaland/or non-statistical) are useful during the Analyze Stage of a Six Sigma project? What tools seemmost appropriate for this project?

2.4. Phase 4

To begin the Analyze Stage, the project team at MSI discussed the implications of the 1.35 processsigma level calculated in the Measure Stage. Certainly there was much room for improvement. Theteam then brainstormed reasons why the quarterly financial-reporting cycle took so long. The brain-storming session resulted in the cause-and-effect (C & E) diagram presented in Exhibit 3. To identifypossible root causes of excessive cycle time, the team first identified four general causes: (1) too manyhours spent on balance sheet schedules, (2) one-time items were a surprise, (3) E-Trans schedulesstarted late in the day, and (4) a lack of valid references and data-entry errors. For each general causethe team asked ‘‘Why?” until the most basic root causes of the problem were identified. For example(see the upper-right hand portion of the C & E diagram): ‘‘Started E-Trans schedules late in the day.”Why? ‘‘Final Balance Sheet and Income Statement completed after 5:00 p.m.” Why? ‘‘Profit review en-tries completed after 11:00 a.m.” Why? ‘‘Profit review doesn’t start until 8:30 a.m. 4 days after theclosing date (c+4).”

To continue its analysis, the project team began work on a failure modes and effects analysis(FMEA) (see Exhibit 4). The team used brainstorming techniques to determine potential failure modesthat could result in process delays for each high-level process function. It selected one potential failuremode for each of these five functions. The team concentrated on potential failure modes whose cause

High number of hours spent on Balance Sheet schedules

Started E-Trans schedules late in

the day

High number of hours spent in quarterly

E-Trans Submissions

Open position for 3 months

Several openFinance

positions

HR recruitingover-burdened

Balance Sheet assembled only at quarter-endM

ore focuson I/S

Final Balance Sheet and Income Statement complete

after 5:00 p.m.

Profit reviewentries

completed

after 11:00 a.m.

Noreview

of tax prior

toprofit review

Profit reviewat 8:30 a.m.

(C+4)

One-time items were a surprise

Lack of valid references/Data-

entry errors

Transitory items were not resolved

Year-end

issues

not revisiteduntil

quarter-end

Assumed everyone knewwhat was going on

Lack of

communication

G/L entries insufficient or outdated

Insufficient automation

ingenerating

data

Insufficient

automation

ingenerating

data

Balance Sheet schedules do not agree

Lackof

comm

unication

Schedules reviewed

onlyat quarter

-end

Queries updated only at quarter-end

Schedules reviewed

onlyat quarter-end

No ongoing review

No ongoing review

No ongoing review

Exhibit 3. Cause-and-effect diagram.

Legend:SEV = Severity of the potential failure modeOCC = Frequency of the potential failure modeDET =Detectability of the potential failure modeRPN = Risk priority number of the potential failure mode (SEV x OCC x DET)

tnerruClaitnetoPtceffElaitnetoPnoitcnuFFailu r e o f S C au s e o f O C o n tr o ls D R

PECeruliaFEeruliaFedoMNTCV

Se n d o u t p r e lim in ar y In co m e State m e n t (c+3)

M an u al in p u t cr e ate s e r r o r s

De lay in p r o ce s s

Fo r m u la e r r o r s , ch an g e s in co m p an ie s , d ata e n tr y

enoNsrorreA s s e m b le f in al In co m e State m e n t

Manual input creates errors

De lay in p r o ce s s

Fo r m u la e r r o r s , ch an g e s in co m p an ie s , d ata e n tr y

enoNsrorre

A s s e m b le f in al Ba lan ce Sh e e t

M an u al in p u t cr e ate s e r r o r s

De lay in p r o ce s s

B/S as s e m b le d o n ly at q u ar te r -e n d , B/S n o t r e vie w e d u n til af te r p r o f it

enoNweiver

Ru n q u e r ie s fo r E-T r an s s ch e d u le s

Qu e r ie s h ave n o t b e e n u p d ate d

De lay in p r o ce s s

Qu e r ie s r e vie w e d o n ly at q u ar te r -e n d , ch an g e s in le d g e r s t r u ctu r e , ch an g e s fr o m co r p o r ate No n e

V alid ate n u m b e r s ag ain s t o th e r s ch e d u le s

Nu m b e r s s u b m it te d ar e in co r r e ct

In co r r e ct d ata to co r p o r ate

No co m m u n icat io n o n w h ich s ch e d u le s /n u m b e r s

enoNeitotdeen

Exhibit 4. Failure modes and effects analysis (FMEA).

Table 2Potential failure modes: severity, occurrence, and detectability.

High-level process function Potential failure mode SEV OCC DET

Send out preliminary I/S 3 days after quarter close (c+3) Manual input creates errors 5 7 6Assemble final I/S Manual input creates errors 4 7 6Assemble final balance sheet Manual input creates errors 8 7 6Run queries for E-Trans schedules Queries have not been updated 8 10 9Validate numbers Numbers submitted are incorrect 7 10 8

110 T.C. Krehbiel et al. / J. of Acc. Ed. 27 (2009) 104–123

of failure was consistent with the root causes identified in the C & E diagram. The team evaluated eachpotential failure mode on three criteria: severity (SEV), frequency of occurrence (OCC), and detectabil-ity (DET). The team assigned scores from 1 to 10 to each criterion. For severity, 1 indicates least severeand 10 most severe. For occurrence, 1 indicates least frequent in occurrence and 10 most frequent inoccurrence. For detectability, 1 indicates easiest to detect and 10 most difficult to detect. Table 2 pro-vides the scores the team assigned to each potential failure mode.

2.4.1. Phase 4 directions

1. What information can you glean from the C & E diagram? What new questions arise from this newinformation?

2. Complete the FMEA in Exhibit 4 by calculating the RPNs. Because the project team at MSI believes itonly has time to work on three of the high-level process functions, which ones do you recommendit work on?

3. How does the C & E diagram relate to the FMEA (Exhibit 4)? How does the C & E diagram relate tothe process-flow diagram in the SIPOC analysis presented in Exhibit 1? How does the FMEA relateto the process-flow diagram in the SIPOC analysis?

4. What is the objective of the Improve Stage of the DMAIC model? In general, what tools (statisticaland/or non-statistical) are useful during the Improve Stage of a Six Sigma project? What tools seemmost appropriate for this project?

T.C. Krehbiel et al. / J. of Acc. Ed. 27 (2009) 104–123 111

2.5. Phase 5

The team hypothesized that three root causes were responsible for much of the excess cycle time:(1) lack of on-going review of balance sheet and inter/intra company schedules, (2) insufficient auto-mation in generating data, and (3) lack of communication in financial reporting. These critical rootcauses helped form the FMEA. More specifically, the first two rows of the completed FMEA (Exhibit5) relate to the root cause ‘‘insufficient automation in generating data.” The next two rows of Exhibit5 relate to the root cause ‘‘lack of on-going review of balance sheet and inter/intra company sched-ules.” The final row relates to the root cause ‘‘lack of communication in financial reporting.”

To begin the Improve Stage, the project team calculated an RPN for each potential failure mode. AsExhibit 5 indicates, the team calculated the RPN by computing the product of the three componentnumbers (the severity of the potential failure mode, SEV; its frequency of occurrence, OCC; and itsdetectability, DET).

Following the calculation of the RPNs, the project team compiled a list of recommended actions (asshown in column 10 of the FMEA in Exhibit 5) to address the root causes of the potential failuremodes. For example, the first two rows of Exhibit 5 indicate that the project team recommended auto-mating I/S reporting to eliminate the problem of ‘‘insufficient automation in generating data.” Rowsthree and four indicate that the project team decided to implement monthly reviews to overcomethe ‘‘lack of on-going review of balance sheet and inter/intra company schedules.” Finally, row five’ssuggestion is to ‘‘Review numbers that are consistent amongst various schedules early in c+4 to assureaccuracy.” By moving this review to an earlier time, MSI was able to minimize communication prob-lems. Results of these actions are documented in Exhibit 6 and Table 3.

2.5.1. Phase 5 directions

1. Calculate the process sigma level from the new data. Is the improvement significant? Why or whynot?

2. What information can you glean from the completed FMEA (Exhibit 5)? What new questions arisefrom this new information?

3. What information can you glean from Exhibit 6? What questions arise from this new information?

Legend:SEV = Severity of the potential failure modeOCC = Frequency of the potential failure modeDET =Detectabilityof the potential failure modeRPN = Risk priority number of the potential failure mode (SEV x OCC x DET)

Function Potential Effect Potential Current Recommended Actions Results of Actions Taken Failure of S Cause of O Controls D R Actions Taken S O D R

PECEPECeruliaFEeruliaFedoMNTCVNTCV

Send out preliminary Income Statement (c+3)

Manual input creates errors

Delay in process 5

Formula errors, changes in companies, data entry

0126enoN7srorre

Automate Income Statement reports through general ledger

In-process

Assemble final Income Statement

Manual input creates errors

Delay in process 4

Formula errors, changes in companies, data entry

8616enoN7srorre

Automate Income Statement reports through general ledger

In-process

Assemble final Balance Sheet

Manual input creates errors

Delay in process 8

B/S assembled only at quarter-end, B/S not reviewed until after profit

6336enoN7weiver

Review Balance Sheet every month, assemble preliminary balance sheet on c+3 Yes 5 7 2 70

Run queries for E-Trans schedules

Queries have not been updated

Delay in process 8

Queries reviewed only at quarter-end, changes in ledger structure, changes from corporate 10 None 9 720

Review quarter end queries every month, make changes as soon as notified by corporate Yes 3 5 6 90

Validate numbers against other schedules

Numbers submitted are incorrect

Incorrect data to corporate 7

No communication on which schedules/numbers need to tie 10 None 8 560

Review numbers that are consistent amongst various schedules early in c+4 to assure accuracy Yes 3 5 7 105

Exhibit 5. Completed failure modes and effects analysis (FMEA).

D ay G o al P ost-ImpD ue S ch e d u le # S che d u le D e scrip tion H o u rs H o u rs

5.05.0ataDtnemetatSemocnIdetceleS20-SIeuT

Fri B S -01 B alanc e S heet & S tatem ent o f C as h F low s D ata 2.0 3.50.2sgninraEdeniateRfotnemetatS16-SB 2.30.1stnuoccAytiuqE/tnemtsevnI27-SB 1.30.1eriannoitseuQweiveRylretrauQ57-SB 1.30.2emocnIfotnemetatS10-SI 4.0

0.10.1ataDtnemetatSemocnIcsiM60-SI5.15.1esnepxE&emocnIcsiM35-SI0.10.1emocnInosexaTlaredeFrofnoisivorP75-SI

IC -17 Interc om pany R ec eivables and P ayables 2.0 2.0

0.2ataDtroppuSteehSecnalaB20-SBnoM 3.85.15.1sisylanAtnuoccAICO831/331SAF13-SB3.05.0tropeRsnoitubirtnoC39-SI

0.20.2)tiforPssorGurht(gnitropeResihcnarF24-SIeuT

0.20.2noitailicnoceRstnuoccAynapmocretnI52-CIdeW

5.05.0eludehcSnoitamrofnItbeD52-SBuhT5.05.0puorGtcudorPybataDteehSecnalaB031-SB0.30.3)emocnIteNurht(gnitropeResihcnarF04-SI

7.130.62seludehcS81rofsruoHtceriDlatoT 0

Exhibit 6. Goal vs. post-improvement performance.

Table 3Pre-improvement (first quarter) and post-improvement (second and third quarters) quarter-end E-Trans schedules: labor hoursworked.

Schedule category Number of schedules Hours worked

First quarter Second quarter Third quarter

Balance sheet 8 64.8 29.8 14.7Income statement 8 16.5 15.0 13.0Inter/intra company 2 28.0 9.0 4.0Total 18 109.3 53.8 31.7

112 T.C. Krehbiel et al. / J. of Acc. Ed. 27 (2009) 104–123

4. What information can you glean from Table 3? What questions arise from this new information?5. What is the objective of the Control Stage of the DMAIC model? In general, what tools (statistical

and/or non-statistical) are useful during the Control Stage of a Six Sigma project? What tools seemmost appropriate for this project?

2.6. Phase 6

The project team at MSI determined that the process sigma level increased from approximately1.35 to approximately 1.93. During the Control Stage, the team wanted to ensure that the processimprovement was sustained. Staff members were trained to monitor schedule preparation times(and the factors affecting them) using the measurement system defined in the Measurement Stage. Pro-cess procedures were standardized and documented so they could be easily repeated by current andfuture employees.

The team declared that the project was completed and preceded to hand off the project to the pro-cess owner, the CFO, who was also the project’s Champion. Perhaps the most important control step

T.C. Krehbiel et al. / J. of Acc. Ed. 27 (2009) 104–123 113

actually occurred at the beginning of the project (rather than at the end) when the process owner wasselected as the project’s Champion. This insightful decision facilitated a smooth handoff and helped toensure that the process improvements would be maintained. The team disbanded and a gala dinnerparty was held to celebrate their success.

On the following Monday morning you received an email from Rhonda Edwards stating that shewas pleasantly surprised with your results. However, she ended her email by noting that she nowhad even more questions about Six Sigma.

1. What are the challenges of using Six Sigma in accounting and finance processes (as opposed tomanufacturing)?

2. What are the benefits of using Six Sigma in accounting and finance (as opposed to manufacturing)?3. What are some other accounting processes that might benefit from Six Sigma?4. What do we do next?

2.6.1. Phase 6 directionsPlease respond to Ms. Edwards’ questions, by crafting a carefully worded, but succinct, memo.

3. Implementation guidance

3.1. Learning objectives

Having completed the requirements of the case, students will have met the following learningobjectives:

Phase 1

1. Learned the stages of the Define–Measure–Analyze–Improve–Control (DMAIC) model.2. Gained an understanding of a project charter.3. Learned to write a problem statement and a goal statement.4. Learned the objectives of, and the common tools used in, the Define Stage.

Phase 2

5. Interpreted a suppliers–inputs–process–outputs–customers (SIPOC) analysis.6. Learned the objectives of, and some common tools used in the Measure Stage.

Phase 3

7. Calculated a process sigma level.8. Learned the objectives of, and some common tools used in, the Analyze Stage.

Phase 4

9. Interpreted a cause-and-effect (C & E) diagram.10. Calculated and interpreted a risk priority number (RPN).11. Learned the objectives of, and some common tools used in, the Improve Stage.

Phase 5

12. Interpreted a failure modes and effects analysis (FMEA).13. Examined the interrelationships among a process-flow diagram, a C & E diagram, and a

FMEA.14. Learned the objectives and the importance of the Control Stage.

114 T.C. Krehbiel et al. / J. of Acc. Ed. 27 (2009) 104–123

Phase 6

15. Gained an appreciation for the challenges and benefits of applying Six Sigma to accounting andfinance processes.

3.2. Teaching approach

The case requires a pre-reading assignment because typical accounting or statistics textbooks donot contain a thorough (yet concise) presentation of Six Sigma. Students should learn the basicterminology and tools of Six Sigma and also be exposed to transactional (i.e., non-manufacturing)Six Sigma applications. We found that the following reading assignment accomplishes such purposes:Chapter 2 of the text by Evans and Lindsay (2004, pp. 29–50), Brewer and Bagranoff (2004), andRudisill and Clary (2004).

Our students completed this case in structured learning teams of 3–5 students. The professorplayed the role of a Six Sigma Black Belt, but in reality, a professor with just an introduction to Six Sig-ma can facilitate the case. The texts Six Sigma for Green Belts and Champions (Gitlow & Levine, 2005)and An Introduction to Six Sigma and Process Improvement (Evans & Lindsay, 2004) are good referencesfor instructors who have little or no Six Sigma training. There are also several good web sites contain-ing useful information, including one from GE that includes a glossary, www.ge.com/sixsigma/mak-ingcustomers.html. It is also useful to have a business statistics textbook that contains reference toSix Sigma (e.g., Berenson, Levine, & Krehbiel, 2009).

The case takes a minimum of two 50-min periods to complete, but can be expanded up to three75-min classes. The majority of the class time should be spent with the teams working on theiranalyses and the professor roaming to help as needed. A 2-day schedule for the case is outlinedbelow:

Day 1: Phase 1 (team effort).

Day 2: Review Phase 3 (class discussion).Phase 4 (team effort).Phase 5 (team effort).Phase 6 (class discussion).

The structure of the case allows flexibility in scheduling. For example, in a 3-day schedule for anMBA class, we modified the 2-day schedule to begin Day 2 with a review of phases 1–3. Next, teamscompleted and the class discussed phases 4 and 5. We ended Day 2 with the introduction of phase 6and instructed teams to complete phase 6 outside of class prior to Day 3. We devoted Day 3 to a classdiscussion of phase 6 and a wrap-up review of all six phases.

Phases are very short and can be read in a couple of minutes. The start of a phase often gives anoverview of what the MSI Six Sigma team accomplished in the previous phase, which in other words,is the MSI solution to the previous phase. There is often lively class discussion at this point concerningthe differences between students’ solutions and the MSI solution. Sometimes it will be obvious thatthe MSI solution is optimal, sometimes students will provide equally good solutions, and sometimeswe have found that the students’ solutions are superior to the MSI solution. By encountering the caseas a PBL, students are allowed to develop their own solution to each stage, and then see how the teamat MSI did it. Moreover, it allows the student teams to ‘‘re-adjust” if they get off track or start to traveldown a different but equally productive road.

Although not a requirement to use this case, our classrooms had wireless access. This allowed stu-dents to use their laptops to access the course readings from electronic reserve, as well as surf theWeb. Teams also used statistics textbooks or Six Sigma reference books. During the team time, stu-dents were free to solicit help from the professor playing the role of a black belt.

T.C. Krehbiel et al. / J. of Acc. Ed. 27 (2009) 104–123 115

4. Teaching note

Each phase of the case contains questions for the students to answer. We are providing suggestedanswers and discussion material for these questions.

4.1. Phase 1 solutions

1. Six Sigma has proven its applicability to any type of process, either transactional or manufactur-ing. Because submitting the quarter-end financial schedules to the parent company is a process, the SixSigma approach is applicable. The students should explain that the managerial aspects of Six Sigma areresult-oriented and not as deeply rooted in theoretical management ideals as the systems popularizedin the 1980s (e.g., TQM, which was heavily influenced by the work of W. Edwards Deming). In general,Six Sigma is more prescriptive than TQM and more accountable to bottom-line results than TQM.

Six Sigma is a project-based, process-improvement initiative whose focus is linking together statis-tical and management tools into a logical flow. One of the most predominate aspects of the Six Sigmamethodology is the five-stage DMAIC process-improvement model. The DMAIC model has proven tobe an effective method of data-driven decision making leading to quality improvement and increasedbusiness performance (Montgomery, 2009 pp. 45–59; Brewer & Bagranoff, 2004).

2. The objective of the Define Stage is to clearly state the problem and frame the project’s scope andexpectations. We remind our students to avoid massive ‘‘boil the ocean” or ‘‘world peace” projects, butat the same time, avoid trivial exercises better suited to less sophisticated analyses. Processes need tobe examined and critical outputs defined. Once the critical outputs are defined, the problem statementand goal statement can be articulated. To a large extent, the tools students mention as effective for theDefine Stage will depend on their extent of relevant class work and the assigned pre-readings. Threetools that are appropriate for this stage include benchmarking, a SIPOC analysis, and a Pareto chart.

3. It is not clear whether the problem statement and goal statement should be focused on (1)reducing the number of hours required to complete the quarter-end financial-reporting process, or(2) reducing the number of schedules not completed on time. It is an easy argument to make thatby reducing the number of hours, the number of schedules not completed on time should also be re-duced. If students focus on reducing the number of hours, they should develop problem statementssimilar to the actual statement used by MSI:

‘‘Too many hours are being spent preparing quarter-end financial schedules.”

If students concentrate on reducing the number of schedules not completed on time, their problemstatements will be similar to:

‘‘Too many schedules are not completed by the due date.”

With the limited amount of data gathered at this point, writing the goal statement is tricky.Although we prefer goals to be clearly and numerically defined, here the best students can probablydo is develop a broad goal similar to what the project team at MSI did:

‘‘Reduce the number of hours spent preparing quarter-end financial schedules.”

Or, if they concentrated on the number of late schedules:

‘‘Reduce the number of schedules not completed by the due date.”

Some students may prefer a zero-defects goal:

‘‘Complete all schedules on time.”

Although this last goal might be considered a stretch goal, such lofty ambitions are not uncommonin Six Sigma. If stretch goals are used, positive responses to significant improvements that do not reachthe goal are necessary. At this time, we remind our students that goal statements can be revisited inthe next stage of the DMAIC model. This second stage, Measure, will allow a better understanding ofmetrics and baseline information, which in turn, could lead to a better goal statement.

116 T.C. Krehbiel et al. / J. of Acc. Ed. 27 (2009) 104–123

4.2. Phase 2 solutions

1. Students should spend a substantial amount of time discussing the SIPOC analysis for the quar-terly financial-reporting process displayed in Exhibit 1. The key supplier is MSI’s finance function. Thekey inputs to the process include the general ledger and a software program called E-Trans, which en-ables subsidiaries to transmit financial data to the parent company in a pre-specified format. Otherkey inputs include MSI’s balance sheet and income statement, and a questionnaire that is used byMSI’s parent company to standardize reporting practices across subsidiaries. The process map identi-fies the first step of the process as ‘‘book journal entries for the current month.” The last of the eightkey steps is to ‘‘fax or email non-E-Trans schedules.” The key outputs are the schedules that are sent tothe parent company using E-Trans, email, or fax. Corporate Finance is listed as the key customer.

Because the only customer listed on the SIPOC is Corporate Finance, students could make the argu-ment that the focus is the number of schedules not completed on time, instead of the number of hoursspent to complete the schedules. The direction of this discussion will influence the answers studentsgive to questions 2 and 3 below.

2. We believe that the problem statement agreed upon by the project team at MSI, ‘‘Too manyhours are being spent preparing quarter-end financial schedules,” is acceptable as given. Perhaps partof this reasoning is driven by the fact that this is the problem statement the actual MSI project teamused. However, it is very informative to let students explore a slightly different path. At the core of thisdiscussion lies the difficulty with many accounting processes—lack of appropriate data. Do we mea-sure the hours it takes, and thus collect one continuous data value every 3 months? Or, do we lookat the proportion of the 18 schedules completed on time, and thus collect 18 ‘‘Yes” (completed ontime) or ‘‘No” (not completed on time) values every 3 months? Regardless of the approach that is se-lected, neither provides the large number of data points that are commonly available in the data-richmanufacturing environment from which Six Sigma evolved.

3. Students should recognize that the goal statement developed by the project team at MSI, ‘‘Re-duce the hours spent preparing quarter-end financial statements,” is too vague. The phrase ‘‘Reducehours spent” should incorporate a concrete value such as ‘‘Reduce the number of hours spent to25.” However, until the Measure Stage is complete, it is difficult or impossible to give an exact numer-ical goal.

4. The objective in the Measure Stage is to have the team identify, define, and measure the process’key input, process, and output variables. Key output variables are commonly referred to as critical-to-quality variables (CTQs) in Six Sigma circles. Key input/process variables are those that affect the CTQs.Input/process variables are referred to as X variables. (The X variables are analogous to independentvariables in a regression analysis, and each CTQ is analogous to the dependent variable in a regressionanalysis.) The causal relationships between the Xs and the CTQs are documented by brainstorming,observations, and when possible, correlation and regression analysis.

In the Measure Stage, the Six Sigma team will typically calculate the existing defects per millionopportunities (DPMO). The process sigma level can then be derived from the DPMO using Table 4,or other process sigma-level tables (for example, see Gitlow and Levine (2005), pp. 31–34). Eitherthe DPMO or process sigma level can be used to benchmark the current quality level. Before calculat-ing the DPMO, the team will need to determine a clear definition of defect and defect opportunity. Inmany cases, obtaining consensus definitions is quite difficult.

At this time it is possible to begin an ongoing discussion as to the merit of using the process sigmalevel. What does the process sigma level really indicate? Is a process sigma level of six always prac-tical? An interesting avenue to explore is the impact that the definitions of ‘‘defect” and ‘‘defect oppor-tunity” have on the DPMO and thus ultimately on the process sigma level. Is it acceptable to narrowthe definition of defect opportunity and thus greatly increase the process sigma level?

4.3. Phase 3 solutions

1. A ‘‘defect” is defined as taking more time than the ‘‘goal time” to prepare a schedule, and a ‘‘de-fect opportunity” is defined as a schedule. Since 10 of the 18 schedules failed to meet the establishedgoal, 10 of the 18 defect opportunities resulted in a defect. Therefore the DPMO is (10/

Table 4Conversion of DPMO to process sigma level. This table details Motorola’s conversion from defects per million of opportunities(DPMO) to a process sigma level (see Montgomery (2009) p. 28). Use this table for all processes (one- or two-sided specificationlimits, numerical or attribute data). To use this table: 1. Calculate the DPMO and 2. Look up the corresponding sigma level.

Sigma Sigma Sigma

Level DPMO Level DPMO Level DPMO

0.1 919,243 2.6 135,666.1 5.1 159.10.2 903,200 2.7 115,069.7 5.2 107.80.3 884,930 2.8 96,800.5 5.3 72.40.4 864,334 2.9 80,756.7 5.4 48.10.5 841,345 3.0 66,807.2 5.5 31.7

0.6 815,940 3.1 54,799.3 5.6 20.70.7 788,145 3.2 44,565.4 5.7 13.40.8 758,036 3.3 35,930.3 5.8 8.50.9 725,747 3.4 28,716.5 5.9 5.41.0 691,463 3.5 22,750.1 6.0 3.4

1.1 655,422 3.6 17,864.4 6.1 2.11.2 617,911 3.7 13,903.4 6.2 1.31.3 579,260 3.8 10,724.1 6.3 0.7941.4 539,828 3.9 8197.5 6.4 0.4801.5 500,000 4.0 6209.7 6.5 0.287

1.6 460,172 4.1 4661.2 6.6 0.1701.7 420,740 4.2 3467.0 6.7 0.1001.8 382,089 4.3 2555.2 6.8 0.0581.9 344,578 4.4 1865.9 6.9 0.0332.0 308,538 4.5 1350.0 7.0 0.019

2.1 274,253 4.6 967.7 7.1 0.0112.2 241,964 4.7 687.2 7.2 0.0062.3 211,855 4.8 483.5 7.3 0.0032.4 184,060 4.9 337.0 7.4 0.0022.5 158,655 5.0 232.7 7.5 0.001

T.C. Krehbiel et al. / J. of Acc. Ed. 27 (2009) 104–123 117

18) � 1,000,000 = 555,555, and the process sigma level is approximately 1.35. Obviously students willrealize that the process needs major improvement to get to a process sigma level of six.

2. Table 1 illustrates that the eight schedules related to the balance sheet took the majority of thetime. However, since only two schedules are in the inter/intra company category, on average, theseschedules were the most time-consuming. Some students might suggest that a Pareto chart mightbe more effective than the simple table given here. In general, students seem to indicate that thereis not a whole lot of information in the table. Seeing the cumulative hours by the three categories,most students would now like to see a breakdown of the 18 schedules.

3. Exhibit 2 gives the breakdown of the time required to prepare each of the 18 schedules at the endof the quarter prior to undertaking the Six Sigma project as well as the goal for each schedule. Theeight schedules completed on time are easily identified. For those missing their goal, some are close(e.g., IS-01 has a goal of 2 h and took 3 h) and others are significantly over target (e.g., BS-61 has a goalof 2 h and took 29!). After studying this table, students start to question what is really happening inschedules like BS-01, BS-61, and IC-17; all have a two-hour goal, but the baseline times are 20–29 h.Are the goals realistic? Is there a one-time explanation for the baseline to be so high?

4. The objective of the Analyze Stage is to determine why defects or excessive variation occurs. Inother words, what is the root cause of the undesirable result? For most processes, brainstorming, de-tailed process maps, FMEA, and C & E diagrams can provide qualitative evidence in search of the why.In situations with lots of data, statistical analyses are typical, especially regression and correlation. Insituations without lots of data, as in many non-manufacturing applications, simulation techniques canhelp identify root causes. Simulation software, such as Crystal Ball by Oracle, can be used to model aprocess. The software can run the model many times creating simulated outcomes and providingmany data points for analysis. Analysis of these data, as well as data generated by changing model

118 T.C. Krehbiel et al. / J. of Acc. Ed. 27 (2009) 104–123

parameters, can identify unexpected causes of defects or variation. For the MSI case, the FMEA and theC & E diagram appear to be the most pertinent tools.

4.4. Phase 4 solutions

1. There are four general causes for why the quarterly E-trans submission process takes so long.Drilling down from these general causes leads to specific root causes. For example, in the lowerleft-hand side of Exhibit 3, we see that the general cause ‘‘one-time items were a surprise” leads to‘‘transitory items were not resolved” leads to ‘‘year-end issues not revisited until quarter-end” whichleads to ‘‘no ongoing review,” a root cause. Also, ‘‘one-time items were a surprise” leads to ‘‘assumedeveryone knew what was going on,” which leads to ‘‘lack of communication,” another root cause.

The root causes ‘‘no ongoing review,” ‘‘lack of communication,” and ‘‘insufficient automation ingenerating data” appear multiple times in the C & E diagram. Based largely on this finding, the MSIteam concluded that these three root causes (the critical root causes) were the primary reasons theCorporate Reporting Department was taking too long to complete the quarterly financial-reportingprocess.

New questions that arise include why was the diagram limited to four general causes? How do weprioritize the four general causes? And, why does the process of asking ‘‘why” stop after fourbranches? Moreover, this C & E looks at discovering the root causes of the high number of hours spentcompleting the E-Trans submissions, but what about the E-Mail and Fax submissions (see Outputs inthe SIPOC illustrated in Exhibit 1)?

2. Students can calculate RPNs by simple multiplication. For example, at the bottom of Exhibit 4,‘‘validate numbers against other schedules,” RPN = 7 � 10 � 8 = 560. Since at the time of writing thecase study, the Six Sigma team at MSI had only instituted three actions, we ask the students whichthree they would focus on. The students will clearly identify the last three functions in Exhibit 4; how-ever, some students may question the subjectivity of the 1 to 10 rating scale. Furthermore, some stu-dents might be familiar with different scales. For example, a five-category 1–2–3–5–10 scale results inRPNs being dominated by the absence or existence of a single rating in the most critical, i.e., 10, cat-egory. A final point of possible discussion is the need for including detectability. If a problem is seriousand persistent, why worry about its detectability?

3. The SIPOC, C & E, and FMEA give us three different ways to analyze the process. The SIPOC is ahigh-level process map that sets a project’s boundaries and helps a project team determine what tomeasure. The C & E is a structured method of generating hypotheses about the root cause(s) of a pro-cess problem. These root-cause hypotheses are usually compared to the SIPOC and a more detailedprocess map to determine whether they are logical. The FMEA explores ways that a process can fail,the causes of failure, and how the failures can be prevented or minimized.

The functions in the MSI project team’s FMEA were selected from the high-level process steps in itsSIPOC analysis. For each function, the MSI team used brainstorming techniques to determine potentialfailure modes (ways that the function could fail) that would result in process delays or incorrect data(incorrect data would indirectly cause a process delay). The team gave preference to potential failuremodes whose cause of failure aligned with one of the three critical root causes it had uncovered in theC & E analysis. For example, the critical root cause ‘‘lack of communication” is listed as a cause of fail-ure for the potential failure mode ‘‘numbers submitted are incorrect.”

4. The objective of the Improve Stage is to develop, implement, and evaluate solutions that are in-tended to eliminate the root causes of problems identified in the Analyze Stage. What changes to the Xvariables (input/process variables) are needed to improve the CTQs? Useful tools in this stage includebrainstorming, consensus building, FMEA, simulation, and design of experiments (DOE). DOE is mostappropriate for manufactured goods that can be produced in a laboratory setting independent of cus-tomer interaction. In an accounting process, however, results cannot be produced without the inter-action of the customer. Furthermore, accounting schedules cannot be run and then discarded orrecycled like goods in an ‘‘R&D” setting. Simulation is possible in manufacturing or transactional sit-uations, and thus would be appropriate here. For this case, the validation of changes through a secondround of FMEA is a logical choice.

T.C. Krehbiel et al. / J. of Acc. Ed. 27 (2009) 104–123 119

4.5. Phase 5 solutions

1. The DPMO = (6/18) � 1,000,000 = 333,333, and the process sigma level is now approximately1.93. Students should not be too concerned that the process sigma level is still low. Although signif-icantly short of a Six Sigma quality level, decreasing the number of defects from 10 to six is a mean-ingful improvement. Now is a good time to revisit the implications of the definitions of defect anddefect opportunity, the D and O, respectively, in DPMO. If we define ‘‘opportunity” as one of, say,100 steps in completing a schedule instead of the currently used definition of completing the entireschedule, we could artificially decrease the DPMO and dramatically increase the process sigma level.

Students might argue that the improvement of overall hours from 109.3 to 31.7 is quite significant.Although still short of the 26-h goal, a huge improvement has occurred. Students can also argue thattracking the number of hours is more in alignment with the project’s goal than tabulating the propor-tion of schedules that meet their preparation-time goal. The metric concerning the total number ofhours, however, is hard to quantify in terms of DPMO or a process sigma level. To what extent shouldwe select our metrics to fit well-established Six Sigma norms?

2. Exhibit 5 contains the completed FMEA. Included are the RPNs before and after the actions taken.As shown in the right-hand column of rows 3–5 in Exhibit 6, the implementation of the recommendedactions substantially reduced the RPNs. Students should note that there have been decreases in sever-ity, frequency, and detectability of selected problems, with a few exceptions. Where actions were ta-ken, the RPNs have decreased and now those RPNs are lower than those of the functions not yetaddressed. Luckily, the actions taken have not caused negative outcomes in the other functions.

New questions that arise have to do with the recommended actions. Changing from a manual to anautomated system is a great idea, but probably not a speedy or cheap fix. It seems that we should be-gin earlier than 4 days after closing to look at numbers that are not consistent, but should be, amongthe schedules.

3. The number of schedules that did not meet their preparation-time goal decreased from 10 to 6(see Exhibit 6). Also, for the 6 schedules that did not meet their goals, the total post-improvementhours were 16.2, compared to the baseline (pre-improvement) hours of 64.6 (see Exhibit 2). Eventhough these 6 schedules did not meet their goals, a 75% reduction in preparation time occurred.As a result of these achievements, the quarterly financial-reporting process’s cycle-time decreasedfrom 109.3 h in the first quarter to 31.7 h in the third quarter (see Table 3). Schedules BS-01, BS-61,and IC-17, which previously took 20–29 h, were all completed in 2–3.5 h. This huge improvementindicates that the lofty goals set earlier are indeed obtainable. Students might question why thepost-improvement hours have increased for some schedules. These increases, however, are minorand students should realize that all outcomes are subject to variability, and that slight increases arealways possible in certain sub-processes even when overall process improvement occurs.

4. Table 3 illustrates continuous improvements in all three categories over time. The table gives aquick overall assessment that clearly indicates a reduction in total labor hours worked. One questionthat students have voiced is whether or not past data would show that the second and third quartersare inherently faster. Without this time-oriented data, the improvements might be due, at least inpart, to the differences in quarters and not the actions taken. However, students should conclude thatthe actions taken have been successful because no information has been given that the different quar-ters are substantially different.

The potential for differences between quarters points out that if such improvements could be cap-tured in a controlled experiment using DOE, the improvements could be more directly linked to theactions taken. In other words, if it were possible to complete the schedules using both the new andold approaches, with everything else held constant, we could statistically isolate the effect of thenew approach. The complex realities of applying Six Sigma in a transactional setting, compared to adata-rich manufacturing setting where R&D experimentation is possible, make the Improve Stage lessquantifiable.

5. The objective of the Control Stage is to maintain the improvements gained by completing the firstfour stages of the Six Sigma project. The focus is on ensuring that problems remain fixed. This stagemight include establishing new standards and procedures, training, and implementing controls suchas checklists, control charts, and balanced scorecards (Brewer (2004) discusses the relationship be-

120 T.C. Krehbiel et al. / J. of Acc. Ed. 27 (2009) 104–123

tween balanced scorecards and Six Sigma). The team needs to standardize and document the changesand develop or improve a monitoring system of key input/process variables and output variables. Thelast step is to hand over the project to the process owner, celebrate the successes, and disband theteam.

4.6. Phase 6 solutions

Instructors have several options on how to complete phase 6. A class discussion could be heldimmediately after students are given class time to read the phase. Secondly, class time could be givenfor the teams to discuss among themselves their thoughts, and then reconvene as a whole class and letthe teams present their reflections. Another approach we have used is to hand out phase 6 at the end

Phase 6 Memo

Morgan Systems Inc.

To: Rhonda Edwards

From: Team #1 (Rob, Shaman, and Brent)

Date: December 5, 2005

Re: Application of Six Sigma to the Finance Function

Even though implementing Six Sigma in our accounting and finance processes is challenging, it provides numerous benefits that make it effective. Six Sigma provides a way to identify and define processes within the accounting and finance functional area. Compared to manufacturing operations, these processes are often difficult to define, but Six Sigma defines them and maps relationships between them. It can show how efficient or inefficient the processes are. The pr oject team quantifies and prioritizes processes to form an approach to making them more efficient.

To gain the full benefits from Six Sigma, a company must overcome certain challenges. Six Sigma must be embedded in the culture of the company to ensure proper creation, implementation, and continuous improvement. A solid training program that has the full support of all key leaders is necessary for implementation. Measurement of results within a service and people oriented process can be somewhat difficult, but made possible by the long-term application of Six Sigma principles.

Six Sigma has already helped our company achieve measurable improvements in the preparation of quarterly financial schedules. Other accounting processes that could benefit from Six Sigma include Accounts Payable, Accounts Receivable, and Profit Fo recasting. We should evaluate which process to analyze next.

There is some concern that improvements made to one CTQ can negatively impact another output variable. For instance, errors may increase as a result of striving to increase the speed of processing the quarterly reports. Six Sigma takes this into account and allows a team to visualize positive and negative effects on other processes and to compensate accordingly. The team can decide if any trade-offs are necessary or if changing a certain process is valuable or not.

To ensure that Six Sigma methods become an important part of regular business operations, leaders must cultivate a positive attitude toward them. Six Sigm a must be sponsored from the top down. MSI should provide ongoing training and get everyone involved. Leaders can begin by generating project ideas and serving as project champions.

The next step for MSI is to continue monitoring recent improvements in financial schedule preparation and to look more carefully at the six schedules still not meeting the target goals. Managers can identify new Six Sigma projects that can utilize lessons learned from the first project. Leverage the momentum created by the success of the first project and continue to seek improvement in all operations.

Exhibit 7. Phase 6 memo.

Table 5Survey items and aggregated responses.

Mean median range

Total ACC MBA(n = 23) (n = 8) (n = 15)

1. The MSI PBL case provided a realistic context forstudying Six Sigma

4.09 3.88 4.20

4 4 43–5 3–4 3–5

2. The MSI PBL case provided a realistic context forstudying the DMAIC model

4.30 4.13 4.40

4 4 43–5 3–5 4–5

3. The MSI PBL case provided a realistic context forexploring the challenges and opportunities ofapplying Six Sigma to transactional processes

4.09 3.75 4.27

4 4 53–5 3–4 3–5

4. The MSI PBL case enhanced my understanding ofa project charter

3.65 3.75 3.60

4 4 42–5 3–5 2–5

5. The MSI PBL case enhanced my understanding ofa process sigma level

4.09 4.00 4.13

4 4 42–5 4–4 2–5

6. The MSI PBL case enhanced my understanding ofa risk priority number (RPN)

4.04 3.88 4.13

4 4 43–5 3–5 3–5

7. The MSI PBL case enhanced my understanding ofa SIPOC analysis

4.13 4.0 4.20

4 4 43–5 3–5 3–5

8. The MSI PBL case enhanced my understanding ofa FMEA

4.00 3.88 4.07

4 4 43–5 3–5 3–5

9. The MSI PBL case enhanced my understanding ofa cause-and-effect diagram

4.00 3.88 4.07

4 4 43–5 3–5 3–5

10. Structuring the case as a six-step PBL case was amore effective learning experience than if it waspresented in a traditional case-study format

4.30 4.38 4.27

4 4.50 42–5 3–5 2–5

11. Overall, the MSI PBL case study was a beneficiallearning experience

4.17 4.13 4.20

4 4 42–5 4–5 2–5

Scale: 1: strongly disagree; 2: disagree; 3: neutral; 4: agree and 5: strongly agree.

T.C. Krehbiel et al. / J. of Acc. Ed. 27 (2009) 104–123 121

of phase 5 discussions and give the teams until the next class time to complete a memo to Ms. Ed-wards. Memos were submitted to the instructor prior to class, and the instructor was able to lead adiscussion based on what was learned from reading the students’ memos. The nature and depth ofthe responses to this phase are greatly dependent upon the amount of time given to the studentsand the type of course where the case is used. MAcc and MBA students will provide their own uniqueperspective to Ms. Edwards’ request. Exhibit 7 contains an un-edited memo from one of our MBAteams.

122 T.C. Krehbiel et al. / J. of Acc. Ed. 27 (2009) 104–123

5. Validation

Several steps were taken to assess the effectiveness of the case. The authors administered a surveyto the students in a MAcc class and an MBA class. The survey, designed to gather student perceptions,consisted of 11 items using a 5-point Likert scale, and two open-ended questions. All 15 students inthe MBA class and all eight students in the graduate accountancy class completed the survey. Table 5lists the survey items and descriptive statistics for our response data.

Overall, the students perceived the case to be a beneficial learning experience (mean score of 4.17on item 11) and more effective than a traditional case-study (mean score of 4.30 on item 10). The PBLcase was perceived as a realistic context for studying Six Sigma and the DMAIC model, and for explor-ing the challenges and opportunities for applying Six Sigma to a transactional process (mean scores of4.09, 4.30, and 4.09 on items 1, 2, and 3, respectively). Items 4–9 suggest that the case was successfulin reaching our stated learning objectives (see Section 3.1).

When asked to ‘‘Briefly describe the best aspects of the PBL case,” most students focused on thebenefits of the unfolding nature of the exercise. Typical comments included, ‘‘The six stages were ex-tremely helpful: If you weren’t sure about something early on, it didn’t prevent you from finishing la-ter stages because we had discussion after each stage.” Other comments included ‘‘It was also easier tounderstand DMAIC when it was used in a real accounting-based example.”

Evidence of the realistic nature of the PBL case was also obtained by sharing the PBL case with twoindividuals currently providing Six Sigma training in industry. They both indicated that the PBL casedelivered extremely important material in a real-world context.

Finally, the un-edited solution in memo form in Exhibit 7 provides some evidence that our studentsdid grasp the important issues when applying Six Sigma to an accounting process.

Acknowledgements

Some of the material in this case is discussed in the following article co-authored by the corre-sponding author for this manuscript: Brewer and Eighme (2005). Using Six Sigma to improve the fi-nance function. Strategic Finance (May), 27–33. Specifically, Exhibits 1 and 3–5 are adapted fromthat article. We would like to thank Kathy Williams, the editor of Strategic Finance, and the Instituteof Management Accountants for granting us permission to use these exhibits here.

We would like to thank our colleague Peter C. Brewer whose interest in Six Sigma and Problem-Based Learning inspired us to write this case. We would also like to thank all of our students for theirinspiration and critical reflection. We greatly appreciate Rob Chrysler, Shaman D’Souza, and Brent Nel-son for letting us use their memo in the teaching note. We are also grateful for the insight provided bySarah Baxter. Finally, we would like to thank the Editor-in-Chief, James E. Rebele, and an anonymousreviewer for their comments and encouragement.

References

Bellas, C. J., Marshall, J., Reed, M. M., Venable, J. M., & Whelan-Berry, K. S. (2000, October). PBL in business—Understandingproblem-based learning and trying PBL in your business discipline: An interactive approach. Paper presented at PBL 2000,Birmingham, AL.

Berenson, M. L., Levine, D. M., & Krehbiel, T. C. (2009). Basic business statistics: Concepts and applications (11th ed.). Upper SaddleRiver, NJ: Prentice Hall.

Brewer, P. C. (2004). Six Sigma helps a company create a culture of accountability. Journal of Organizational Excellence (Summer),45–59.

Brewer, P. C., & Bagranoff, N. A. (2004). Near zero defect accounting with six sigma. The Journal of Corporate Accounting & Finance(January/February), 67–72.

Brewer, P. C., & Eighme, J. E. (2005). Using Six Sigma to improve the finance function. Strategic Finance (May), 27–33.Cottell, P. G., & Millis, B. J. (1993). Cooperative learning structures in the instruction of accounting. Issues in Accounting Education,

8(1), 40–59 [Spring].Duch, B. J. (2001). Writing problems for deeper understanding. In B. J. Duch, S. E. Groh, & D. E. Allen (Eds.), The power of problem-

based learning. Sterling, VA: Stylus Publishing.Duch, B. J., Groh, S. E., & Allen, D. E. (2001a). Why problem-based learning? A case study of institutional change in

undergraduate education. In B. J. Duch, S. E. Groh, & D. E. Allen (Eds.), The power of problem-based learning. Sterling, VA:Stylus Publishing.

Duch, B. J., Groh, S. E., & Allen, D. E. (Eds.). (2001b). The power of problem-based learning. Sterling, VA: Stylus Publishing.

T.C. Krehbiel et al. / J. of Acc. Ed. 27 (2009) 104–123 123

Evans, J. R., & Lindsay, W. M. (2004). An introduction to Six Sigma and process improvement. Mason, OH: Thomson Learning.Gitlow, H. S., & Levine, D. M. (2005). Six sigma for greenbelts and champions: Foundations, tools, cases, and certification. Upper

Saddle River, NJ: Prentice Hall.Johnson, D. R., Johnson, E. J., & Smith, K. A. (1991). Cooperative learning: Increasing college faculty instructional productivity. ASHE-

ERIC higher education report no. 4. Washington, DC: The George Washington University School of Education and HumanDevelopment.

Johnstone, K. M., & Biggs, S. F. (1998). Problem-based learning: Introduction, analysis, and accounting curricula implications.Journal of Accounting Education, 16(3/4), 407–427.

Krehbiel, T. C., Havelka, D., & Scharfenort, M. (2007). Process monitoring in accounting: Implementing pre-control charts. TheJournal of Applied Business Research (4th quarter), 93–103.

Millis, B. J., & Cottell, P. G. (1998). Cooperative learning for higher education faculty. Phoenix, AZ: Oryx Press.Montgomery, D. C. (2009). Introduction to statistical quality control (6th ed.). New York, NY: John Wiley & Sons.Rudisill, F., & Clary, D. (2004). The management accountant’s role in Six Sigma. Strategic Finance (November), 35–39.