the long and winding (and frequently bumpy) road to successful client engagement: one team's...

21
The long and winding (and frequently bumpy) road to successful client engagement: one team’s journey Deborah Campbell* Deborah Campbell, formerly with Hewlett-Packard Company as a senior consultant, is currently pursuing a Master of Philosophy degree in system dynamics at the University of Bergen in Norway. Working with senior management teams, she has applied system dynamics to strategic business problems in the areas of human performance, organizational performance, and market positioning. Abstract Successful client engagements require both good model-building process and good facilitation. Using an executive team-building wilderness journey as an analogy, this article describes the process and facilitation of one particularly successful client engagement. At completion, in addition to using the model results to guide their strategic decision process that year, the client took on full ownership of the model, its use as an ongoing decision-making resource, and its maintenance. Yet, between the smooth and relatively easy sales process and the important final results were a myriad of ups and downs. Each obstacle in our road brought learning and insight, and confirmed standard processes or suggested new ones. In the end, we developed a model to guide a specific but not restrictive strategic decision area, facilitated an understanding within the organization of the dynamic complexity of their business without the detail complexity, organized their data in a way that made it useful for them as it had not been before, and built a very strong client–consultant team relationship. Copyright 2001 John Wiley & Sons, Ltd. Syst. Dyn. Rev. 17, 195–215, (2001) Introduction In the U.S.A. during the last ten years, executive team-building programs have become quite popular. The most well known are wilderness journeys, which consist of leaving a group of people in the middle of nowhere with a few critical pieces of equipment and a sketchy map, and letting them find their way through the wilderness back to civilization. By being lost together; fighting over alternate routes to get back; becoming lost again; suffering scraped elbows and knees, or worse; and finally becoming interdependent and successfully making their way out, the team experiences loss, confusion, conflict, hardship, discovery, and success together. The strength of the relationships created on such a journey is renowned. In my opinion, a successful system dynamics project team has many of the same experiences of loss, confusion, conflict, hardship, discovery, and success, on their journey through a system dynamics project. The point of an executive wilderness journey is to provide a learning environment for team problem solving, with the problem being how to return to civilization. The belief is that these strong relationships and problem- solving lessons, when applied back at the office to the company’s strategic business problems, will improve results for the company. Our intent in system dynamics projects is exactly the same: to provide a learning environment for Ł Correspondence to: Deborah Campbell, Department of Information Sciences, University of Bergen, N-5020 Bergen, Norway; E-mail: deborahc@ifi.uib.no System Dynamics Review Vol. 17, No. 3, (Fall 2001): 195 –215 Received August 2000 DOI: 10.1002/sdr.214 Accepted May 2001 Copyright 2001 John Wiley & Sons, Ltd. 195

Upload: deborah-campbell

Post on 06-Jul-2016

216 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: The long and winding (and frequently bumpy) road to successful client engagement: one team's journey

The long and winding (and frequentlybumpy) road to successful clientengagement: one team’s journeyDeborah Campbell*

Deborah Campbell,formerly withHewlett-PackardCompany as a seniorconsultant, iscurrently pursuing aMaster of Philosophydegree in systemdynamics at theUniversity of Bergenin Norway. Workingwith seniormanagement teams,she has applied systemdynamics to strategicbusiness problems inthe areas of humanperformance,organizationalperformance, andmarket positioning.

Abstract

Successful client engagements require both good model-building process and good facilitation.Using an executive team-building wilderness journey as an analogy, this article describes theprocess and facilitation of one particularly successful client engagement. At completion, inaddition to using the model results to guide their strategic decision process that year, the clienttook on full ownership of the model, its use as an ongoing decision-making resource, and itsmaintenance. Yet, between the smooth and relatively easy sales process and the important finalresults were a myriad of ups and downs. Each obstacle in our road brought learning and insight,and confirmed standard processes or suggested new ones. In the end, we developed a model toguide a specific but not restrictive strategic decision area, facilitated an understanding within theorganization of the dynamic complexity of their business without the detail complexity, organizedtheir data in a way that made it useful for them as it had not been before, and built a very strongclient–consultant team relationship. Copyright 2001 John Wiley & Sons, Ltd.

Syst. Dyn. Rev. 17, 195–215, (2001)

Introduction

In the U.S.A. during the last ten years, executive team-building programs havebecome quite popular. The most well known are wilderness journeys, whichconsist of leaving a group of people in the middle of nowhere with a fewcritical pieces of equipment and a sketchy map, and letting them find theirway through the wilderness back to civilization. By being lost together; fightingover alternate routes to get back; becoming lost again; suffering scraped elbowsand knees, or worse; and finally becoming interdependent and successfullymaking their way out, the team experiences loss, confusion, conflict, hardship,discovery, and success together. The strength of the relationships created onsuch a journey is renowned.

In my opinion, a successful system dynamics project team has many of thesame experiences of loss, confusion, conflict, hardship, discovery, and success,on their journey through a system dynamics project.

The point of an executive wilderness journey is to provide a learningenvironment for team problem solving, with the problem being how to returnto civilization. The belief is that these strong relationships and problem-solving lessons, when applied back at the office to the company’s strategicbusiness problems, will improve results for the company. Our intent in systemdynamics projects is exactly the same: to provide a learning environment for

Ł Correspondence to: Deborah Campbell, Department of Information Sciences, University of Bergen, N-5020Bergen, Norway; E-mail: [email protected]

System Dynamics Review Vol. 17, No. 3, (Fall 2001): 195–215 Received August 2000DOI: 10.1002/sdr.214 Accepted May 2001Copyright 2001 John Wiley & Sons, Ltd.

195

Page 2: The long and winding (and frequently bumpy) road to successful client engagement: one team's journey

196 System Dynamics Review Volume 17 Number 3 Fall 2001

team problem solving with the goal of improving results for the organization.While we emphasize ‘‘the understanding gained’’ and the ‘‘problem solved’’ inour system dynamics project goals (Richardson and Pugh 1981; Wolstenholme1990), the by-products of our team model-building efforts are often just asimportant: the common language defined, the consensus and commitment to adecision achieved (Winch 1993; Vennix 1996), and the relationships formed.1

Based on both Wolstenholme (1990) and Vennix (1996), I define the goals of ateam-based system dynamics project as:

ž Develop the team’s systemic understanding through team learning (wherelearning implies mental model reformation).

ž Achieve commitment to effective and implementable problem solutions andpolicies through consensus and commitment to the team’s model and itsresults.

A journey is a good analogy for a system dynamics project. In this articleI use the phases of a journey as a way to view and apply the traditionalsystem dynamics model-building process and describe as an example a clientengagement I led when working for the Strategic Planning and Modeling(SPaM) consulting team at Hewlett-Packard.

Richardson and Pugh (1981) identify seven stages in the system dynamicsmodel-building process. Vennix (1996: 49) largely follows Richardson andPugh with his stages:

1. Problem identification and model purpose2. System conceptualization3. Model formulation and parameter estimation4. Analysis of model behavior: testing and sensitivity analysis5. Model evaluation: model validity6. Policy analysis7. Model use or implementation

While these steps are linearly organized, the process itself is necessarilyiterative in nature, to promote and integrate the learning that takes place alongthe way (Randers 1980; Richardson and Pugh 1981; Vennix 1996; Sterman2000).

In using the journey analogy I define four legs of a journey, which map tothese process stages as follows:

1. Planning the journey—stage 1: problem identification and model purpose2. Mapping the terrain—stage 2: system conceptualization3. Traversing the terrain—stages 3–6: all formal model stages2

4. Returning home—stage 7: model use or implementation

Page 3: The long and winding (and frequently bumpy) road to successful client engagement: one team's journey

Deborah Campbell: The Road to Successful Client Engagement 197

In Planning the journey we decide where we want to go and why; we identifya problem (not a system) to model, and define the model purpose.

In Mapping the terrain we develop a map, a diagram of what we knowabout the problem and its causes, or about the terrain. We bring togetherour individual knowledge of this terrain and create and capture on paper ashared map. ‘‘The intention is to broaden the understanding of each person’’(Wolstenholme 1990: 5). This is easier said than done. On an executive team-building trip we may have no experience of the terrain before we go, but notso in a system dynamics project. The members of the team bring their ownmaps (mental models) of the terrain with them, usually detailed in one areaand sketchier in others. Creating a common map from multiple and oftenconflicting points of view is a challenge.

In Traversing the terrain we actually go there and get our shoes dirty. Themap alone can often provide a significant level of understanding about theproblem (Wolstenholme 1990). At the same time, ‘‘Von Clausewitz famouslycautioned that the map is not the territory’’ (Sterman 2000: 89). To further ourunderstanding and explore system behavior, structure, and policy alternativesmore rigorously—to explore our terrain in more detail—we must traverse itourselves. Formal modeling with system dynamics gives us a chance to explorethe terrain in a computer-aided learning environment, just as the wildernessterrain is a learning environment for the executive team. The value that ourmodels and analyses provide over and above a wilderness experience is theability to learn in an environment more closely representative of the real terrain.

In Returning home we bring back the learning that took place and apply it tothe real system, the real terrain. Just as the map is not the terrain, neither is thelearning environment the real environment. We evaluate a system dynamicsproject as we would the executive team-building journey: Did the team learnhow to solve problems together on their journey? Can they apply what theyhave learned back at the office? How can the learning be retained once theteam has left the learning environment? These are all questions that must beconsidered both while we travel and once we return home.

Several system dynamicists believe implementation should not be viewedas a separate stage, but should instead be considered at every stage ofthe model-building process (Roberts 1978, as paraphrased in Vennix 1996:98; Sterman 2000). My experience is that we need both approaches: toemphasize implementation from the start and to recognize the challengesthat face us at project completion. Even if we do our best to bring othersalong with us on our journey, either by participating in the journey or bymeeting up with us at different points along the way, at the end of a projectwe often have results—understandings, diagrams, model results, and policyrecommendations—that must be shared with those who could not make thejourney together with us. We can never recreate the experience of the journeyfor them as if they had been there, yet we must find a way for them to learnsome of what we have learned.

Page 4: The long and winding (and frequently bumpy) road to successful client engagement: one team's journey

198 System Dynamics Review Volume 17 Number 3 Fall 2001

In this introduction I have defined the four legs of a system dynamicsjourney and have mapped them to the traditional system dynamics model-building stages. Now let me tell you about one such journey I took with a clientof mine when I was consulting at Hewlett-Packard and what we learned aboutmodel-building process and facilitation along the way.

Planning the journey—destination: model the system

The General Manager of this Hewlett-Packard (HP) consumer-product businesswas a visionary, a man who could see the potential in new ideas and newapproaches. System dynamics intrigued him. During our initial meeting, thebusiness problem we discussed was market share, ‘‘what specific actions canwe take to increase market share?’’ This General Manager (GM) led a suppliesbusiness. While another business in HP sold a popular consumer product,his business sold the supplies that were required by consumers to use thisproduct. His business sold the ‘‘razor blades’’ for the other businesses’ ‘‘razor’’.Though still young, his business was leading the market, having become quitesuccessful in just a few years. But competition had become strong, retailersless loyal. Several companies had introduced a variety of competing suppliesinto the market. Each one was lower priced than HP’s, yet few were of thesame quality. As retailers were the primary channel for sales, consumers reliedon the retailer to guide them in their purchase decision. Unfortunately forHP, not only were competitors offering lower-priced supplies, they were alsooffering attractive incentives to retailers to sell these competitors’ suppliesto consumers instead of HP’s supplies. As a result, while HP’s market sharewas still high, it was beginning to decline, and the GM was concerned aboutincreased competition in the future taking it even lower. We had a fruitfuldiscussion and agreed to do a project together. Yet later, when the time cameto formalize the problem statement, things began to evolve. Market share wasthe issue, but what was the focus of our project?

I had several meetings with the management team, attempting to choose anarea of focus. Finally, the General Manager said, ‘‘Deb, our problem is thatno one, not even me, knows how our business works. We have grown so bigso fast, that no one has the big picture anymore, and we need you to help usdraw it. Without this big picture view, we have no clear context in which toevaluate business tradeoffs between multiple or conflicting strategies.’’ It doesnot take an expert system dynamicist to know that ‘‘model our business’’ is tooambiguous a goal. On the other hand, how could they find a focus without thisunderstanding? What better solution than to use qualitative system dynamicsto pull together knowledge of the business from all parts of the organization andbuild a big picture view for them? Then, using this view, determine the focusfor increasing market share and for a quantitative system dynamics model. Ihad to try.

Page 5: The long and winding (and frequently bumpy) road to successful client engagement: one team's journey

Deborah Campbell: The Road to Successful Client Engagement 199

The management team and I agreed to take a team of 15–20 people fromthe business, including the GM and his staff, through our standard clientengagement phases: (1) develop a high-level descriptive model (a causaldiagram) of the important business and market dynamics, and then, basedon that descriptive model and further research together, (2) develop a formalmodel. The focus of both phases was to understand their business as it was,how it operated in relation to the current consumer, retailer, and competitordynamics, and what variables they could control or influence. The descriptivemodel was to be used as a first step in building the ‘‘big picture’’ view for themand as a communication tool for the organization. The formal model was to beused to deepen the understanding of their current business and its dynamicsand as the foundation for action recommendations. The project team agreedthat a successful conclusion of the project would include alignment of theorganization on the ‘‘big picture’’ and delivery of a recommendation and actionplan for the management team.

Mapping the terrain (a descriptive model)

Mapping experience and reflections

Over the years the system dynamics community has developed a well-definedprocess for creating descriptive models. Our consulting team at HP, SPaM,has our own six-step variation on this process, with the words and techniquestailored to fit our company culture. Yet I discovered during this project, morestrongly than ever before, that my version of this descriptive modeling process,with its neat, clearly delineated steps, was not the process the client team andI were experiencing. We were regularly experiencing confusion. And, while Ihad had client teams in confusion before, it had never been like this. I thoughtfor a while that my inexperience in leading such a large group through theprocess was contributing to the problem. In describing the use of group model-building techniques, Vennix states that the larger the group, the more difficultit is to facilitate (Vennix 1996). But this didn’t explain everything—I decidedthere must be more to our confusion than just group size.

After discussing the formation and reformation of mental models with acouple of learning theory specialists, I confirmed my own suspicion thatconfusion is a normal part of this reformation process. As we discover thatour own mental model does not fit to the behavior of the real world, and wediscover that others have different mental models, which in some dimensionsmore accurately reflect the behavior of the real world, we are forced to resolvethese discrepancies through the formation of a new mental model. This createsconfusion for us, as we let go of our old mental model and reform it from avariety of inputs. This confusion was certainly experienced by the team, andyet there was more.

Page 6: The long and winding (and frequently bumpy) road to successful client engagement: one team's journey

200 System Dynamics Review Volume 17 Number 3 Fall 2001

Not surprisingly, the goal of ‘‘modeling the business’’ created challengestoo. This is a significant business to HP and each team member naturallybelieved his or her department had an important role to play in the business’success. Following the advice I had received from several experienced systemdynamicists, I did everything I could to keep the discussion of dynamics at ahigh level, to ‘‘keep it at a level of detail that was necessary and insightful, andabove a level of detail that did not add value’’. But the client pushed for thedetail every step of the way. Finally, I gave up. I went with the flow. It wastough going for a while, because this detail complexity created confusion too.

The natural confusion experienced during mental model reformation iscompounded by the characteristics of the problems we deal with in systemdynamics projects, problems typically loaded with both detail and dynamiccomplexity. Complexity is messy and takes time to understand. It is bestunderstood not by assimilating it all, but by creating simplified representationsof it. Randers alludes to this when he says, ‘‘the modeler may encounterpowerful organizing concepts that make possible the reformation of the wholestudy in a simpler, more elegant form’’ (Randers 1980: 135). While he saysthis ‘‘may’’ happen, and that if it does it becomes easier for the group todiscuss the problem, I believe it is critical for it to happen. My opinion isthat these ‘‘organizing concepts’’ are the key components of the re-formedmental model. They are the way that simplicity emerges from complexity, thatdynamic complexity is culled from detail complexity. And confusion is thenatural experience during this transition.

So in the end I discovered the reason my client team and I were notexperiencing the six-step descriptive modeling process I described was that Ihad left several key steps out, with each saying something like, ‘‘Step X: Enter astate of confusion here’’. This is what happens every time and when it happensis relatively predictable. For my client team, the time spent in these states wasvery uncomfortable and seemed to take much longer than it actually did. Ithelped to relieve their anxiety when I told them it was normal for confusionto happen at these points in the process and for them to feel uncomfortable.Now I am clearer on the primary reasons why—mental model reformation andsystem complexity.

Finally, in addition to confusion, complexity is intertwined with ownership.My experience, with this client team and others, is that team members are com-mitted to the model only after their own work—the way they impact the busi-ness every day—is included. Now, it may be completely invisible in the end,removed to create a high-level model, but by then, they know where it is ‘‘under-neath’’. They know how their contribution is represented. Then they believe inthe model, and take ownership for it. I have not been able to achieve the sameresult when I keep client teams at a high level, because they do not see how theircontributions fit into the bigger picture. The trade-off is increased complexity.

Vennix, among others, discusses the importance of client ownership of themodel. He emphasizes the need for the group to actively participate during

Page 7: The long and winding (and frequently bumpy) road to successful client engagement: one team's journey

Deborah Campbell: The Road to Successful Client Engagement 201

the model-building process, to have their ideas heard and used. While heappears to favor the approach of group model-building projects starting witha preliminary model rather than from scratch, he explicitly states there can beclient ownership issues with this approach if the group does not feel involved(Vennix 1996: 98, 113, 162). What he and others do not mention, but has beenmy experience, is the connection between ownership and complexity—thatincluding the team’s ideas and arenas of impact to strengthen ownershipnecessarily means increased detail complexity.

The team did make it through the descriptive model-building process and,while it took some time, we did not get stuck. Once the complexity hadbeen modeled, the team learned from their own experience what I had beentelling them: it is too difficult to see the important high-level dynamics orto communicate them to anyone with so much detail. So, with a little work,simplicity emerged from our complexity. From the original four-page model,we created a single-page descriptive model that represented the high-leveldynamics and proved to be a valuable communication tool (Figure 1).

While the complexity our team struggled through was very likely morethan necessary, it is my experience that to ‘‘start high level and stay highlevel’’ is an ideal that cannot be met and it is my belief that it is an

probabilitypurchaser buys

HP supplies

our unitsales

ourrevenue

our profit

our investments

retailer perception ofour relative value

proposition

non price attributesto purchaser

our ability tocommunicate

relative value tothe retailer

R end user andpurchaser'ssatisfaction

with HP supplies

end user andpurchaser's

satisfaction withcompetitor supplies

our price

BHP consumer product

installed base

HP consumer product usage rate

industry consumer productusage rate

industry consumer productusage trends

replacementtechnologies

<our price>

non price attributesto retailer

quality of oursupplier

relationships

consumerprice

sensitivity

consumerenvironmental

sensitivity

HP consumer productfunctionality

applicationusage

purchaserawareness of

supplyalternatives

our market share ofsupply sales

B

B

R REVA

current technologiessubstituting for consumer product

B

==

purchaser perceptionof our relative value

propositionour ability tocommunicate

relative value tothe purchaser

==

==

==

//

//

==

O

O

O

O

O

O

O

O

O

Fig. 1. High-level model

Page 8: The long and winding (and frequently bumpy) road to successful client engagement: one team's journey

202 System Dynamics Review Volume 17 Number 3 Fall 2001

ideal that should not be met, because avoiding the transitional confusionthat additional complexity creates threatens the very purpose of our work:systemic understanding through mental model reformation, and consensusand commitment to decisions through model ownership.

Mapping results

This leg of the journey took much too long by anyone’s measure. Yet the teamwas pleased with the high-level descriptive model and with what they hadlearned about the business along the way. For example, early in our modelingeffort it became clear that there was already a solution to the market-shareconcerns under consideration. Many were convinced that increasing usage ofthe other HP business’ product by consumers, thereby increasing the purchaserate of supplies, was the answer. Yet, as the model evolved, we began to seethe consequences of increased usage. As members of the team put it, ‘‘Growingusage is a double-edged sword. When we create change, we drive the industry,not just HP. [In the industry,] the relative importance [to the consumer] ofthe supplies purchase decision goes up as usage goes up, [and puts increasedimportance on] the consumer factors HP may be weakest in.’’ If HP increasedthe usage rate of consumers, we might lose rather than gain with our currentsupplies offering. Additional insights included:

ž ‘‘The model has highlighted the complexities, interdependencies, and gapsof knowledge about our business.’’

ž ‘‘We were reminded of the importance of influencing both supply productattributes and the consumer perception of supply product attributessimultaneously.’’

ž ‘‘The factors enabling consumers to switch [to competitive supplies] are notbeing addressed aggressively enough.’’

ž ‘‘The factors affecting the probability of purchase are complex and impactedby many other factors.’’

As a conclusion to this leg of the journey, client team members heldcommunication sessions with the rest of the organization, using the descriptivemodel to explain the dynamics of their business and receiving feedbackon improvements that could be, and were, made. The model, and theunderstanding that was gained from creating it, formed the basis of twoimportant improvements for the organization. First, as the GM put it, ‘‘theprocess has given us a common language to discuss our business’’. Thislanguage, the words and their definitions, was passed on to the rest ofthe organization during the communication sessions. The management teamperceived that this common language improved communication within theorganization and increased the speed of decision-making.

Page 9: The long and winding (and frequently bumpy) road to successful client engagement: one team's journey

Deborah Campbell: The Road to Successful Client Engagement 203

Second, the descriptive model itself was used as a business map, a commoncontext for discussion and decisions. ‘‘This description phase of applyingsystem dynamics is often sufficient in itself to generate problem understandingand ideas for change’’ (Wolstenholme 1982: 549, as quoted in Vennix 1996:100). Large copies of the model were hung in the hallways and one was hung inthe management team’s meeting room. People making project or expenditureproposals were asked to use it to show how their proposal would influencethe business. As the management team had wanted, the model provideda ‘‘clear context in which to evaluate business tradeoffs between multipleor conflicting strategies’’. Additionally, people in the organization reported aricher understanding of their contribution to the business, with comments suchas: ‘‘It makes you think beyond your job’’ and ‘‘You can see how what you do,and do not do, affects the business’’. As a result, we had met the goals of a team-based system dynamics project at a qualitative level, developing the team’ssystemic understanding through team learning (mental model reformation), andachieving commitment to effective and implementable problem solutions andpolicies through consensus and commitment to the team’s model and its results.

Traversing the terrain (a formal model)

Traversing experience and reflections

When it came time to develop the formal model, we turned the client team intoan advisory team and joined forces with one person from the organization (whoI will call ‘‘our client partner’’) to work closely with us to develop the modeland exchange information with the advisory team. Additionally, the GM turnedleadership of the project over to his Controller. The Controller was a pragmaticman, a man interested in the specifics of our modeling work. He was skepticalof system dynamics, wanting to know how it differed from his finance team’sspreadsheet modeling and whether our model results could be validated. Theswitch from a visionary personality to a pragmatic personality was tough. Itmeant rebuilding the joint vision of the project, and using different methods todo so, because a pragmatic person needs different information from a visionaryperson (Moore 1991).

Our most challenging discussions were about data. He was convinced thatthe organization did not have enough quantitative data about their business todevelop a model that could produce reliable results. While I did not know whatdata they did or did not have, or how accurate it was, it was clear that they werealready making strategy decisions based on it. I knew that bringing whateverthey had together, quantitative and qualitative, was a valuable place to start. Butwe stalled. This was a stumbling block, one of the big bumps on this windingroad. The Controller finally agreed to move forward after it was promised thatthe management team would not use the model results if he was not satisfied

Page 10: The long and winding (and frequently bumpy) road to successful client engagement: one team's journey

204 System Dynamics Review Volume 17 Number 3 Fall 2001

with them. While I took a risk that he would veto the results for any number ofreasons, I decided that it was a risk worth taking. I suspected that once he sawthe results, he would be pleased. I was happy it turned out that way.

Our client partner and my consultant team worked hard to gather both thequantitative and qualitative information that existed in the organization. Wediscovered that they had more quantitative data than they thought. A compre-hensive consumer segmentation study, completed by external consultants theyear before, had not been used. An internal study of consumer habits was justconcluding. And there was more agreement than people had anticipated ontheir qualitative data. As we gathered data, we developed a single, organizeddocument. While this document was important for our modeling effort, it alsobecame one of the most valuable deliverables for the organization.

Traversing results

In the data document and formal model (Figures 2–7, described briefly in theAppendix) we gathered and organized the organization’s data about consumers,retailers, competitors, and industry trends into a single logical structure. Thismade it more useful for them. We had a similar experience to Vennix: ‘‘[Themodel] restructured existing, but scattered, knowledge by putting it in asystemic perspective. It thus revealed relationships between various elementsand in that way created new knowledge for the group’’ (Vennix 1996: 185).As a result of our sensitivity analysis, the organization was able to develop aneffective market research strategy for the following year: of the long list of itemsthe organization wanted researched, the priorities were now clear. Our datadocument also created a new way to look at, evaluate, and use the informationthey had. They began to think and speak about their consumers differently.Seeing information on industry trends next to consumer segmentation in the

Fig. 2. Probability ofpurchase

Probability ofPurchaser Buys

HP Supplies

Purchaser RelativePerception of

Value

PurchaserAwareness ofAlternatives

Reseller RelativePerception of value

<PurchaserPrice

Sensitivity>

<Importance ofReseller Attribute>

Our Market shareof Supply Sales

<Attribute Importance>

ResellerPerception of

Attribute Relativeto Competitors

Perception ofPrice Relative to

Competitors

<Table RelativeValue and

Probability>

Perception ofAttribute

Relative to Competitors

tive to

Page 11: The long and winding (and frequently bumpy) road to successful client engagement: one team's journey

Deborah Campbell: The Road to Successful Client Engagement 205

Fig. 3. Attributes forpurchaser

Dollarper Al

AttributeImportance

Average

PurchaserAwareness ofAlternatives

Desired AttributeAccording to

Strategy

RequiredInvestment in

Attribute

Dollars per Unitof AttributeInvestments

AttributeGap

AttributeImportance

<CustomerEnvironmentSensitivity>

Competit ors

AttributeRelative to

Competitors

Initial Attribute Perception

Time to Adjustthe Gap inAttribute

Perception ofAttribute

Relative toCompetitors

Desired AttributePerception

According toStrategy

AttributePerception

Gap

Decrease ofRelative Attribute

Perception

Required Investmentin AttributePerception

<Cut Back>

Time toAdjust the

Gap inPerception

Investment inAttribute

Perception

Dollarsper API

Dollars perUnit of Future

PerceptionInvestments

PerceptionInvestmentRequired

Time

Lifetime of theAdvantage in

AttributePerception

AttributePerceptionInvestment

Implemented

AttributeInvestment

Required Time

<Quality ofthe Supplier

Relationships>

<Cut Back>

Investmentsin Attribute

Initial Attribute

AttributeInvestmentsImplemented

<Retailer RelativePerception of value>

Decrease ofRelative Attribute

Lifetime of theAdvantage in

Attribute

Fig. 4. Attributes forretailer

Importanceof RetailerAttribute

Lifetime of theAdvantage in

R P A

Desired R P AAccording to

StrategyRequired

Investment inR P A

Investmentsin R A PTime to

Adjust Gapin R P A

R P AInvestmentRequired

Time

R P FInvestmentsImplemented Dollars per

RPAI

Initial R P F

Time toAdjust Gap in

R Attribute

R AttributeInvestmentRequired

Time

R AttributeInvestment

Implemented Dollarsper RAI

Initial Retailer AttributeRetailerAttribute

Relative toCompetitors

RetailerPerception of

Attribute Relativeto Competitors

Investments inR Attribute

Dollars per Unitof R AttributeInvestment

Desired RAttribute

According toStrategy

RequiredInvestment in

R Attribute

R AttributeGap

Lifetime of theAdvantage inR Attribute

Decrease ofRelativeAttribute

<Cut Back>

<Cut Back>

Dollars per Unitof R P A

Investment

R P AGap

Decrease ofR P A

data document reinforced ideas for targeting specific segments. They developednew marketing and product development strategies based on a clearer view oftheir consumers.

Page 12: The long and winding (and frequently bumpy) road to successful client engagement: one team's journey

206 System Dynamics Review Volume 17 Number 3 Fall 2001

Fig. 5. Retailer price

PurchaserRelative

Perception ofValue

Our Market Share ofSupply Sales

Our Net OpProfit

Price Increase

Our Unit Sales

Exchange Rate

<Our Net Margin>

RetailerIncrement

in Price

Probability ofPurchaser Buys

HP Supplies

Table RelativeValue andProbability

SegmentSize

HP Con. ProductInstalled Base

Our Revenue

Installed BaseOver Time

InternetAccess

Exchange Rate Over Time

ApplicationUsage

Usage rate ofHP Consumer

Product

HP ConsumerProduct

FunctionalityIndustryCon. ProductUsage Rate

Price Decrease

Final Pricefor the

Purchaser

RetailerPrice

Fig. 6. Economic valueadded

<Growt

Our Net Op Profit

OurInvestmentCapacity

Supplier PriceCompare to

Retailer Price

Capital Charge

Supplierinvestments

Value andAvailability of

Supplier Products

TableRelationship

GroupInvestments

Percentage ofProfit Assignedto Investments

<Growth>Quality of the

Supplier Relationships

Supplier Profit

Our Net Margin

Our Revenue

Liabilities

Net Assets

SupplierMargin

Our Unit Sales

Supplier Revenue

EVA

Cost of Capital

Total Assets

SupplierPercentage ofInvestments inthis Business

The simulation results also had an impact. Before building the formal model,we insisted on the management team choosing a focus—using the model tosupport an important decision they had to make. They chose a controversialmarketing policy that was under review, a sales incentive program for retailers.The incentive was based on volume. The quantity of HP supplies sold through aparticular retailer determined the margin that retailer was paid on each supply

Page 13: The long and winding (and frequently bumpy) road to successful client engagement: one team's journey

Deborah Campbell: The Road to Successful Client Engagement 207

Fig. 7. Investmentalternatives

Our Investment Capacity

<Our Investments>

Cut Back

<Our Net Op Profit>

Total RequiredInvestments

<RequiredInvestment in R

Attribute><Required

Investment inR P A>

<RequiredInvestment in

AttributePerception><Required

Investmentin Attribute>

<Our Revenue>

Growth

Percentage ofProfit Assigned to

Investments

product sold—the larger the volume, the higher the margin. The business hadbeen making large incentive payments for several years. Unsure of its actualeffect on sales, the management team was deciding whether to continue theprogram or not. The views in the organization on this program were strong,falling into one of two camps, either (1) it would be a critical factor in theirbusiness success or (2) it would have no effect on sales and would become acritical factor in their downfall. The issue divided the organization. The keymeasures we used to evaluate the program were market share and revenue.While we used this issue to guide us in modeling, we also used our knowledgeof their business to keep the model broad and flexible enough to be modifiedand used for future policy evaluations.

The simulations showed a result that the management team had neverconsidered, though it will come as no surprise to those familiar with theconcepts of dynamic systems. It turned out that the incentive program wouldbe successful in the short term, but, with the changing state of the market,would not continue to be successful long term. In fact, if they continued toinvest in this program, they would lose money. My team presented theseresults to the management team immediately after the management team hadparticipated in a particularly heated and unproductive two-hour discussion ofthe same program. The management team was intrigued by the counterintuitiveresults. Before accepting them, they wanted some analysis. ‘‘Why does thathappen?’’ ‘‘How did you define this?’’ What values do you use for that?’’To explain what was happening during model simulation, we used flowdiagrams simplified from the descriptive model to highlight key feedbackloops and explain equations (Figures 8 and 9). We also used the ‘‘causes stripgraph’’ function of Vensim to show the team which variables and relationships

Page 14: The long and winding (and frequently bumpy) road to successful client engagement: one team's journey

208 System Dynamics Review Volume 17 Number 3 Fall 2001

Fig. 8. Modelcalculations

REVENUE

Retailer price Probabilityof purchase

Usage rate Installed base

Allowableinvestments

TABLE OFRELATIVE VALUE

AND PROBABILITY

RETAILERPERCEPTION

OF VALUE

PURCHASERPERCEPTION

OF VALUE

RETAILER

AttributeImportance

AttributePerception

1

0.75

0.5

0.25

00 0.25 0.75 1

-X-

RELATIVEPRICE

RelativePerceptionof Attribute

Importanceof Attribute*

RelativePrice

* Importanceof Price

PURCHASER

ProductAttribute

Importance

ProductAttribute

Perception

CurrentTable Relative Value and Probability

0.5

influenced the final results (Figure 10). In the end, after seeing that every resultled logically back to their data or assumptions, they said, ‘‘Why, this is ourinformation, the same information we use everyday!’’. The ‘‘ah ha’’ hit them atonce, and they were convinced.

The outcome of this meeting was not only the resolution of an importantissue, but the meeting of the goals of our team-based system dynamics projectat a deeper level. We developed the team’s systemic understanding throughteam learning and mental model reformation. The results of the formal modelwere initially counterintuitive to the management team. Our diagrams andgraphs reinforced the relationship between system structure and behavior, andresolved the team’s transitional confusion by helping them develop a mentalmodel that was consistent with the behavior demonstrated by the formalmodel. We also achieved commitment to effective and implementable problemsolutions and policies through consensus and commitment to the team’smodel and its results. The diagrams and graphs allowed the managementteam to explore the internal workings of the formal model. Not only didthey validate the structure of the model by agreeing with the relationshipsand values they saw there, but also seeing their information, ‘‘the sameinformation we use everyday’’, cemented their commitment to the model andits results.

Page 15: The long and winding (and frequently bumpy) road to successful client engagement: one team's journey

Deborah Campbell: The Road to Successful Client Engagement 209

Fig. 9. What reallyhappens in the model

Retailer RelativePerception of Value

RetailerPerception of

Attribute Relativeto Competitors

Perception ofPrice Relative to

Competitors<Margin Incentive>

<Importance ofRetailer Attribute>

[Retailer relative perception of attributei ∗Importance of the retailer attributei]

Importance of the retailer attributei

(1−Margin Incentive (Time)

Standard MarginRelative perception of price (time)

i =attribute 1

i =attribute n∑

i =attribute 1

i =attribute n∑ ∗

)

The formula suggested

No Incentive3% Incentive

Retailer Relative Perception of Value

2

1.5

1

0.5

01995 1998 2001

Time (year)

Margin Program

Returning home (organizational ownership)

A special attribute of the managers in this management team was theircommitment from the beginning to take ownership of the final model. This israre (Richardson and Pugh 1981). Our client partner, in addition to workingclosely with us to develop and validate the formal model, took ownership ofthe ongoing use and minor maintenance of the model. SPaM is a resource tohim for updating the model (e.g. with new market information from marketresearch) and enhancing the model for other policy analyses.

When the organization took over the model, we agreed that our nowexperienced client partner would run all the simulations and do the analyses.Yet, even if he were to facilitate use of the model, members of the organizationneeded to understand this new tool available to them, and why they wouldwant to work with him. My consultant team and our client partner agreed thatdelivering a hands-on workshop that exposed key members of the organizationto using the model with facilitation, and the power of its results, would createthe necessary pull.

These workshops were designed around a single strategy decision. Threeteams competed to produce the most profitable decision alternative, each onefacilitated by a member of the client/consultant team. The results were positive.‘‘I have had great feedback from those who participated in the workshop. The

Page 16: The long and winding (and frequently bumpy) road to successful client engagement: one team's journey

210 System Dynamics Review Volume 17 Number 3 Fall 2001

Fig. 10. Marginprogram No incentive

3% incentive

Purchaser Relative Perception of Value [Segment 1]40

32.525

17.510

Purchaser Relative Perception of Value [Segment 2]40

32.525

17.510

Purchaser Relative Perception of Value [Segment 3]40

32.525

17.510

Purchaser Relative Perception of Value [Segment 4]40302010

0Purchaser Relative Perception of Value [Segment 5]

4035302520

1995 1996 1998 1999 2001Time (year)

No incentive3% incentive

Probability Purchaser Buys HP [Segment 1]1

0.750.5

0.250

Probability Purchaser Buys HP [Segment 2]0.80.60.40.2

0Probability Purchaser Buys HP [Segment 3]

10.75

0.50.25

0Probability Purchaser Buys HP [Segment 4]

0.80.60.40.2

0Probability Purchaser Buys HP [Segment 5]

10.90.80.70.6

1995 1996 1998 1999 2001Time (year)

No incentive 3% incentive

Market share

1995 1998 2001Time (year)

No incentive 3% incentive

Revenue

1995 1998 2001Time (year)

Margin Program

participants were really using the model to talk about the business, insteadof having the model be the focus.’’ People had fun during the workshop,accomplished something in exploring this particular strategy decision, anddiscovered how they could use the model and facilitator to evaluate futurestrategy decisions.

The end of the road (client reviews and lessons learned)

At the completion of the Mapping the terrain leg of the journey, the feedbackfrom the client organization was positive, with everyone excited about theunderstanding they had gained and the discoveries they had made in theirbusiness. The common context and language they had developed increasedthe effectiveness of their communication and the speed of their decisions. Ilearned an important lesson from them about the results of dialogue:

Lesson 1. Do not underestimate the importance of what we in the communityoften refer to as a by-product of our work, the development of acommon language. Previously I had only viewed this as a step on theway to the final goal. Yet the value it brought this client organizationclarifies its importance as a result in its own right.

Page 17: The long and winding (and frequently bumpy) road to successful client engagement: one team's journey

Deborah Campbell: The Road to Successful Client Engagement 211

The one negative comment I received was how long it took to achieve theseresults. While there are aspects of the work we do that are inherently timeconsuming (e.g. having the dialogue that develops the common language),I learned several lessons that would reduce the time required for a similarproject in the future, including:

Lesson 2. Explicitly define when in the process the client team (and I) will be ina state of confusion, and how we will get ourselves out. Explain whythis happens and describe previous experiences of going through itand making it to the other side. This will prepare client teams forthe inevitable. While it will not make the time of confusion any lessdifficult, it will at least reassure everyone we are on the right track.

Lesson 3. In the beginning, let the complexity come. While many systemdynamicists may not agree with this one, to date this has workedfor me. Let team members get their daily contributions onto thetable and into the model. After it has all been said, work togetherto determine what is most valuable and how to express it in a high-level model. This process creates ownership and understanding thatI have not achieved when keeping models high-level from the start.

At the completion of the entire journey, Traversing the terrain and Returninghome, the feedback was only positive. The ‘‘learning environment’’ providedby the formal model improved the organization’s systemic understanding oftheir business and provided them a context in which to evaluate businesstrade-offs. Additionally, the simulation results resolved a seemingly irresolv-able debate over an important marketing policy. This was an important win,saving the management team hours of future discussion and tens of millionsof dollars per year. As our client partner reported, ‘‘I demonstrated the modelto the General Manager and [his boss] the Group General Manager yesterday.The Group GM was overwhelmed, he really was impressed, and started rightin to measuring the effects of the variables on the business. He wants all thebusiness managers to see the model.’’

Additionally, our work together gave them a basis for strategic marketresearch plans. The market research manager said, ‘‘The model has helped usidentify where we are making assumptions about our business that need tobe verified. [It] is helping us structure new questions about [. . .] the business.As a result, we [have plans] to update our segmentation, expand it to Europe,investigate and quantify supply product attributes that impact loyalty andpurchase probability within segments, and quantify the preference for HPversus competitors among these segments and attributes.’’

The lessons from the Traversing the terrain leg of the journey are focused onthe challenges of the data debate:

Lesson 4. Discover and implement ways to discuss and demonstrate thevalue of system dynamics to pragmatists. I have the techniques

Page 18: The long and winding (and frequently bumpy) road to successful client engagement: one team's journey

212 System Dynamics Review Volume 17 Number 3 Fall 2001

for visionaries, but I need to enhance my toolset for pragmatists.Pragmatists need specific examples of past work similar to the workwe will be doing together. They need to see it, touch it, and knowhow someone else just like them benefited from it.

Lesson 5. Do not underestimate the importance of gathering and organizingthe client organization’s data into a single document. Previously Ihad done this only to support the formal modeling effort. Yet thevalue it brought this client organization clarifies its importance as aresult in its own right.

This was the most challenging, and in many ways the most exciting andsatisfying engagement I worked on in my four years of using system dynamicsat HP. The goals were ambitious, the process sometimes disconnected fromreality, and the team interactions challenging. Yet, in the long run, thejourney we took together led us to exciting new terrain, for the clientorganization and for me. The client organization credits several improvementsin their effectiveness to this engagement. My consulting practice has alsoimproved.

Throughout this article I have spoken about a system dynamics project asa journey, similar to a team-building wilderness journey in which a team ofpeople experience hardship and discovery together as they make their wayback to civilization. I mentioned that the strength of the relationships createdon such a wilderness journey is renowned. A system dynamics project is noexception.

There was one final lesson I learned. After my consulting team completedthe formal model analysis and presented the results to the management team,the Controller was heard to say, ‘‘We would never have made it this far if Debhad not been so persistent.’’ Later, when an important potential client askedfor a reference before agreeing to a contract, I referred them to this Controllerand as a result we won the contract quickly.

Lesson 6. Never underestimate the value of a relationship created by experi-encing hardship and discovery together.

Appendix: Formal model—sector highlights

Figure 2. Probability of purchase

ž Probability Purchaser Buys HP Supplies is the hub of the model dy-namics.

ž Purchaser Relative Perception of Value is derived from consumer perceptionof attributes, the importance of those attributes to the consumer, consumerperception of price, and consumer price sensitivity.

Page 19: The long and winding (and frequently bumpy) road to successful client engagement: one team's journey

Deborah Campbell: The Road to Successful Client Engagement 213

Figure 3. Attributes for purchaser

ž Purchaser (consumer) attributes are supply product features and services(e.g. product quality).

ž Consumer Perception of Attribute Relative to Competitors is influenced bythe actual Attribute Relative to Competitors, but is not the same.

ž Consumer Perception of Attribute Relative to Competitors is also influencedby the Retailer Perception of Value (retailers guiding consumer choice) andby HP’s investments in perception (e.g. advertising).

ž Consumer Perception of Attribute Relative to Competitors is key todetermining Probability Purchaser Buys HP Supplies.

Figure 4. Attributes for retailer

ž Retailer Attributes are service features HP provides the retailer, of interestonly to the retailer (e.g. inventory monitoring).

ž Retailer Perception of Attribute Relative to Competitors is influenced by theactual Retailer Attribute Relative to Competitors, but is not the same.

ž Retailer Perception of Attribute Relative to Competitors is also influenced byHP’s investments in perception (e.g. trade shows).

ž Retailer Perception of Attribute Relative to Competitors is key to determiningPurchaser Awareness of Alternatives, which influences Probability PurchaserBuys HP Supplies.

Figure 5. Retailer price

ž Retailer Price is the price HP sells the supply to the retailer. The retailers settheir own price to the consumers.

ž HP’s market share, the current exchange rate, and the margin incentiveprogram determine Retailer Price.

Figure 6. Economic value added

ž EVA was calculated using the client’s formula.ž The impact of an increase or decrease in the supplier’s investments in this

business was evaluated.

Figure 7. Investment alternatives

ž Both constrained (by management) and unconstrained investments wereevaluated as actions to improve market share.

ž Delays and gains related to investments in attributes were difficult to trackdown and quantify over time.

Page 20: The long and winding (and frequently bumpy) road to successful client engagement: one team's journey

214 System Dynamics Review Volume 17 Number 3 Fall 2001

Acknowledgements

I would like to acknowledge the valuable contributions of my consulting team, AdolfoCrespo Marquez of the Escuela Superior de Ingenieros de Sevilla, Spain, Carol Blancherof SPaM at Hewlett-Packard Company, and our client partner, who shall remainanonymous, to the development and delivery of the formal modeling and ownershipstages of this project. In addition, because of a family tragedy, I could not participatein the final presentation and workshops given to the client organization. Adolfo, Carol,and our client partner took full responsibility for these final presentations and theydeserve the credit for their successful outcomes. I also thank the reviewers for theircomments, our client partner for his reviews, and the Executive Editor for his commentsand suggestions.

Notes

1. For further benefits and processes involved in group model building withsystem dynamics, see the special issue of System Dynamics Review editedby Vennix et al. (1997).

2. Combining the formal modeling efforts into a single stage has a precedentin Wolstenholme (1990), who articulates two model-building stages:Qualitative System Dynamics and Quantitative System Dynamics. WhileI agree that the activities in the traditional stages 3–6 must be carried outwhen implementing a formal model, they share a common purpose for thejourney and this is seen more clearly when they are grouped together.

References

Moore GA. 1991. Crossing the Chasm: Marketing and Selling High-Tech Products toMainstream Customers—Revised Edition. HarperBusiness: New York.

Randers J. 1980. Elements of the System Dynamics Method. Productivity Press:Cambridge, MA.

Richardson GP, Pugh AL. 1981. Introduction to System Dynamics Modeling. Productiv-ity Press: Portland, OR.

Roberts EB. 1978. Strategies for effective implementation of complex corporate models.In Managerial Applications of System Dynamics, Roberts EB (ed). Productivity Press:Portland, OR.

Sterman JD. 2000. Business Dynamics: Systems Thinking and Modeling for a ComplexWorld. Irwin/McGraw-Hill: Boston, MA.

Vennix JAM. 1996. Group Model Building: Facilitating Team Learning Using SystemDynamics. Wiley: Chichester.

Vennix JAM, Anderson DF, Richardson GP (eds). 1997. Special issue on Group ModelBuilding. System Dynamics Review 13(2).

Winch GW. 1993. Consensus building in the planning process: benefits from a ‘‘hard’’modeling approach. System Dynamics Review 9(3): 287–300.

Page 21: The long and winding (and frequently bumpy) road to successful client engagement: one team's journey

Deborah Campbell: The Road to Successful Client Engagement 215

Wolstenholme EF. 1982. System dynamics in perspective. Journal of the OperationalResearch Society 33: 547–556.

Wolstenholme EF. 1990. System Enquiry: a System Dynamics Approach. Wiley:Chichester.