CASE A “FundCo” – The Chief Executive of amajor fund management firm had reason to worryabout its ability to sustain the exceptional growthrates in earnings and funds under managementthat it had maintained for more than a decade.Like other professional service firms, it dependedcritically on a team of professional staff, which notonly delivered the firm’s services, but alsomaintained strong relationships with clients. Thiscan be a fragile system, dependent upon goodmorale and loyalty amongst the staff. Theprosperity of the firm in question had been builtover many years, and had attracted the very bestrecruits. However, the Board of this firm was onlytoo aware that certain rivals had collapsed veryrapidly, and wished to avoid that fate itself. Figure1 indicates the time path of this firm’s recent historyof staff development, and the alternative futuresits management felt they might face.

Business Strategy Review, 1999, Volume 10 Issue 3, pp 1-16

© Kim Warren

In seeking to build and sustaincompetitive advantage, managers need todevelop strategies which take account oflikely future changes – and which willthemselves change in line withcircumstances. This article starts byoutlining problems with a non-dynamicapproach to formulating strategy andthen lays out the initial frameworks of afact-based method that can help managersunderstand and take control of the time-path of their firm’s performance.

A ubiquitous feature of the strategy challenge facingmanagers is how to tackle dynamic (ie time-related)problems of performance. A typical example is shownon the right (case A – “FundCo”) and two more appearoverleaf. If strategy analysis is to help in such cases, itshould at a minimum provide answers to three basicdynamic questions:

● Why has business performance followed the time-path that it has?

● Where is performance heading into the future undercurrent policies?

● How can we act to alter that future for the better?

Whilst management can do much to adjust short-termfinancial results, there is unavoidable uncertaintyabout medium to long-term outcomes. Nevertheless,managers at all levels are expected to commit to

The Dynamics of StrategyKim Warren

Figure 1Time-chart for staff-losses feared by aprofessional service firm

If this time-path for the staff were to arise, it wouldcoincide with a collapse in clients, funds, earningsand, of course, the share price.

Professionalstaff Preferred

Feared

2000'98'96'94'92

500

400

300

200

100

Business Strategy Review

2 Kim Warren

confident projections. When entrepreneurs seekventure capital or CEOs raise finance for acquisitions,investors expect such time-path forecasts for futureearnings. Vague generalisations about roughly whathas happened, approximately where the business isheading, and possible thoughts about future plans arenot adequate. Senior managers themselves expect noless from subordinates seeking support for budgets,business plans, and new initiatives – executives at alllevels are expected to say with confidence what scaleof performance they will deliver, over what time-scaleinto the future. And this confidence is assumed to besupported by a clear set of intentions as to what willbe done, when, and to what degree across all the majorfunctions of the business in order to bring about thepromised performance.

These questions are so fundamental to theresponsibility of strategic managers that one mightexpect leading strategy books to tackle them head-on. But charts such as Figure 2 are remarkably rare.Why? Is it because these questions simply cannot betackled? This article aims to show that they can. Thereare fundamental structures at work within anybusiness situation that determine how performanceevolves over time. These structures can be understoodand captured by formal analysis, and are amenable tomanagement action. The article defines and illustratesthe first of a set of frameworks in an approach knownas the dynamic resource-system view of strategy(DRSV) that makes this possible.

The Time-path of Strategic PerformanceCase B and Case C, both drawn from recent workwith companies, illustrate the critical importance ofthe questions in Figure 2. The first question – why wehave arrived at today’s level of performance – may

Figure 2The fundamental dynamic questions in strategy CASE B “TelCo” – A dominant telecoms firm

in a deregulating market fears loss of marketshare to new entrants. This firm, a formerlynationalised telecoms operator, faces the openingof its market to free competition. Following theexperience of British Telecom in the UK, the firmknows that its financial performance will sufferfrom losing a proportion of its subscriber-baseover a few years, but wishes to minimise thoselosses, and capture any new subscribers that mayemerge. Figure 3 indicates alternative futuretime-paths for this critical indicator.

Figure 3Time-chart of competitive intrusion facinga dominant telecoms operator

not be relevant in every case: a new venture has nohistory. However, for most firms, the trajectory offuture performance is highly dependent upon theirrecent strategic history. Case C in particular raises deepconcerns for the managers involved – what are ourprospects under current policies, what can we do toimprove those prospects, and what lessons andresources can we bring to bear on the problem frompast experience?

The challenges portrayed in Cases A, B and C arenot merely qualitative questions. In each case:

● the threat or opportunity is substantial in scale;

● the strategic issue will evolve over a certain periodwith speedy response being vital; and

● there is a time-path of progress – the firms’performance will evolve at a varying rate.

Scale of threat/opportunityIn each of our three cases, the difference betweensuccess and failure is considerable. TelCo stands to

Earnings$ million

Today

years

Why?Where?

How?

Subscribersmillions

Preferred

Feared

98 99 2000 01 02

25

20

15

10

5

Autumn 1999

CASE C “GameCo” – A consumer-electronicsmanufacturer wishes to exploit a rapidlydeveloping market opportunity before rivals doso. This firm, facing a challenge similar to thelaunch of the Nintendo 64 against Sega and theSony Playstation, is at an early point in a newphase of the industry’s history, with a consumer-electronics product for which there will be asubstantial market. However, it is critical to buildsales quickly.

Not only is it vital to erode the accumulatingadvantages enjoyed by the rival’s establishedposition, it is also imperative to grow theinstalled base, to drive sales of components andupgrades, to win commitments from suppliersand distributors, and to take the newopportunity before others. Figure 4 indicates twoalternative futures for this launch.

Whilst this is clearly an episode of strategicimportance for the firm, notice the time-scaleover which it has played out – just 12 months.For the Nintendo case, this included a seven-month period in which the price of both its ownand Sony’s product price fell from £250 to £99.Not much use for five-year plans here!

The Dynamics of Strategy 3

lose millions of subscribers and hundreds of millionsof dollars in revenue. GameCo expects sales ofhundreds of thousands of units, and desperately needsan installed base to provide the long-term cashflowsfrom sales of upgrades and accessories. Longer-term,pulling off this plan may determine the entire survivalof this multi-billion dollar enterprise. FundCo fearsthat losing just a fraction of its most critical staff could

trigger collapse of a business that is custodian of fundsworth over $75bn.

Evolution of strategic issueIn each case, there is also a time-scale over which thestrategic issue will evolve, and speed of response isvital. GameCo will win or lose its race over a fewmonths, and FundCo could, if it does not act correctly,see staff losses accelerate within a few quarters.Although the competitive threat to TelCo will playout over four or five years, its immediate decisions onpricing, service, network development and marketingwill be powerful determinants of its future prospects.

Time-path of progressFinally, each case exhibits a time-path of progress –the firm’s performance will not just start and end atspecific points, but evolve at a varying rate as its futureunfolds. TelCo may at first lose few subscribers, thensuffer increasingly rapid losses as its rivals build uptheir capacity. The consumer electronics manufacturermay see little absolute growth in early weeks, beforeword-of-mouth accelerates the rate of sales. The fundmanagement firm may initially experience little morethan a stagnation in its staff population, untildisillusioned individuals start to leave, creating ever-faster attrition that could prove catastrophic.

Whilst continuing uncertainties will never permitprecise forecasts (and managers will always need theflexibility to change direction as events unfold),strategy analysis should at least lead to some indicationof such time-paths for future performance. So howmight a management team start to tackle suchchallenges?

Today’s performance depends on today’sstrategic resourcesMost managers understand the importance of buildingand conserving the resources of their business. Thesemay be ‘hard’, tangible resources (cash, plant,customers, products etc) or ‘soft’, intangible factors(product quality, staff morale, or service standards).Furthermore, managers know that resources areinterdependent – consistent product quality can beused to build reputation with customers, and a strongclient base may help attract the best recruits. ‘Ranking’resources by importance misses the point – if any keyresource is in bad shape, the whole business isendangered.

Writers on strategy have long recognised theimportance of strategic resources (eg Grant 1995,

Figure 4Time-chart for rivalry to exploit a new marketfor a consumer-durable product

Unit sales'000/month

Preferred

Feared

1q98

2q 3q 1q99

500

400

300

200

100

4q

Business Strategy Review

Figure 5Strategic resources and firm performance:the simple, immediate connection

Figure 6Strategic resource levels determine performanceat any time – past, present and future

Wernerfelt 1984, Mahoney and Pandian 1992, Peteraf1993). They have also identified the criticalmanagement challenge that arises in trying to buildand maintain the level of each resource (Dierickxand Cool 1989). It is in this process of resource-building that the essence of the strategy dynamicsproblem lies.

There is a puzzle in this ‘resource-based’ view ofstrategy. If we boil it down to its bare essentials, itappears that today’s performance can be preciselycalculated from only a few resources (mostly tangible)and some external conditions: Figure 5, admittedlya caricature, shows how today’s revenue dependson today’s customer-base and price, through the salesvolume that arises, and so on. The implications seemquite profound – we do not need anything else(intangible resources, capabilities, strategic vision orleadership) to explain the performance of the firm.Yet this is clearly nonsense: such items must make adifference, so we need to understand how they impacton the simple observation in Figure 5.

The solution to this puzzle lies in the fact that thetype of analysis represented in Figure 5 is merely asnap-shot of the firm at a moment in time. If thesefew tangible resources explain precisely ourprofitability today, then their scale yesterday explainedour performance then, and their scale tomorrow willexplain precisely our profitability at that moment too(Figure 6). The missing element in a rigorousunderstanding of the dynamics of performance istherefore an explanation of how the level of eachresource changes over time.

An example: BrandCoThe approach to using this insight in practice can beillustrated with a further illustration, again drawnfrom recent case-work (Glucksman et al 1998).

As explained above, the earnings from thisbrand-launch at any point in time will depend onthe resources the firm has then. To simplify, theanalysis here will focus on just three key resourcesfor this business – consumers, stores, and sales force(see Table 1).

Early on, the product will have few consumersand few stores, so the sales revenue will be limitedand more than outweighed by the costs of itssalesforce and other expenditures, notablyadvertising. Later – perhaps, if things go well, bymonth 18 – consumers and stores will be sufficientin scale to provide revenues that exceed the brand’scosts, and the product will be into profit.

4 Kim Warren

Note: ‘Word-and-arrow’ diagrams, common in contemporarymanagement writing, often feature items and connections with awide variety of meanings. In contrast, each element in the figuresin this article has a specific meaning. The boxes simply denotecontainers holding a certain amount of resource. The curved arrowsdo not mean merely that there is some vague relationship betweentwo items: they state that one item can be immediately calculatedfrom another, just like a formula in a spreadsheet cell.

Resourcelevels today

CustomersCapacity

Sales force... etc.

Cost efficiencyResource levels

2 years ago

Resource levelsin 2 years time

Earnings$ million

Today

years

Today's resources...

Average customer demand

PriceSales

Revenue

Cost of goods

Grossprofit

Other costs

Earnings...determine today's earnings...

Customers

Capacity

Cost efficiency

Sales force

Administrationinfrastructure

Autumn 1999

CASE D “BrandCo” – a consumer-productsfirm has developed a new style of spirits andwishes to build a powerful brand. A soundstrategy for this product launch needs a clearview of the time-path that might be achievable,and a clear statement of the strategic resourcesinvolved. From experience with comparableproducts, the firm believes that about five millionconsumers might like the product, and about50,000 stores may feasibly stock it in due course.Typical consumption is about 1 litre/month perperson. Retail prices of about £11/litre ($16) arecommon in this sector, wholesale prices are around£8.50, and direct product costs are about £7.

Figure 7 shows a time-chart of the hoped-for earnings path, and Table 1 defines just threekey resources: consumers, stores and sales force.

Capturing the likely time-paths in Figure 7therefore requires analysis of the mechanisms that willexplain those resource-levels over time.

Resources build and deplete over timeResources accumulate as new resource ‘flows’ intothe current stock of what we possess – winningcustomers adds to the level of a customer-base,advertising increases the level of market-awareness,training raises the average level of staff skill.Resources also deplete or decay by flowing out ofthe stock – customers defect to rivals, resignationsreduce employee numbers and skil ls , andtechnological progress devalues current staff skills.A frequently used analogy for these processes is tothink of a resource as liquid flowing in or out of atank. This helps explain why it takes time to detectchanges in strategic performance. Even substantialchanges to the in-flows and out-flows have littlevisible impact on the levels of liquid in the tank.Only after some time does it become apparent thatchange has occurred, and is continuing to do so.

Managers usually want more resources, so wishto raise the inflow to the stock and minimise theoutflow. These imperatives are directly captured bythe ‘stock-and-flow’ framework (Figure 8 overleaf)at the heart of the method known as systemdynamics (Forrester 1961). The time-path of theresource level for ‘customers’ is shown on the graphinside the central tank. Customers are being won byan in-flow through the ‘pipe’ entering from the left,and the initial rate at which this resource is growing

Figure 7Expected time-path for profits from a newproduct launch

Resources Units In-flows Units of in- Typical driversand out-flows and out-flows

Consumers people new consumers ‘000 people advertising, productinterested aware per month availability

consumers interest inlosing interest other products

Stores stores new stores stores consumer-demand,stocking the stocking the per month sales force, pricebrand brand

stores de-listing more valuable usethe brand for the shelf space

Salesforce people new hires people per salaries, hiring effor tmonth

resignations pressure of work,sales commission

Table 1Core resources to build a brand

The Dynamics of Strategy 5

Brand profit£million/month

Preferred

Feared

2418126

1.0

0.5

0

-0.5

-1.0

Business Strategy Review

is shown in the oval window onto the pipe. The time-path for this rate of in-flow appears on the time-chartbelow the pipe, and is a constant 15 customers perquarter. Similar diagram elements capture the out-flowof customers from the right of the stock – this loss,though starting at a low rate of five per quarter, isrising steadily.

no way to explain performance at any time exceptby knowing all gains and losses to all resourcesover the entire history of the firm (Forrester 1961;see also equation 2 in the Technical Appendix).

● There is similarly no way to produce a confidentview of future performance without estimatinghow those gains and losses will develop.

● There is no way for management to alter thestrategic performance of the firm except by actionsthat impact on resource-flows (though short-termperformance can be changed by making simpleallocation choices, especially between expenditureand declared profits).

The first observation is not as daunting as it may seem.First, we generally know, or can find out, the levels ofresources at relatively recent points in our history, soit is not necessary in practice to go back to the originsof time! Second, it is often possible to estimate thegains and losses of the key resources over the recentpast. All that is needed then is the effort and patienceto calculate the net gains or losses, and today’sresource-levels and performance are explained.Furthermore, estimates of future rates of gain and lossfor strategic resources will give good forecasts ofresource-levels and earnings – indeed, this is the onlymeans to obtain such forecasts.

Characteristics of strategically valuableresourcesThe characteristics that resources must possess if theyare to provide sustainable advantage are well-established in published research on strategy (seereferences). Resources must be durable, should not bemobile or tradeable, should not be easy for rivals toreplicate or to substitute with alternatives. Finally, theyshould be complementary, ie capable of working welltogether – for example, a great new technologyproduct is not much use if the firm’s distributors lackthe skills to support it and have no access to thecustomer segment that may want it.

These may seem reasonable tests of whether anystrategic resource will offer advantage, but they suffertwo problems. First, none of the criteria is black-and-white – each applies to some degree. Few resourcesare totally durable, absolutely non-tradeable, neverreplicable or impossible to substitute. Second, whethera resource is durable, mobile, replicable andsubstitutable is fundamentally a dynamic question:firms always face the problem of the rate at which

6 Kim Warren

(The units of in-and out-flow are always the units of the resource

‘per time-period’ and the time-slope of the resource at any

moment is the net of in- and out-flows).

Figure 8 starts to explain why the time-path ofperformance is rarely intuitively obvious – it takes onlysimple changes to gains and losses of resources togenerate a quite complex trajectory for any resource-level. Here we start in quarter 1, 1998, enjoying a netgain of ten customers per quarter, by quarter 3, lossesequal gains, and our customer-base is static, and byquarter 1, 1999 we are suffering net losses of tencustomers per quarter – 25 minus 15.

Bearing in mind that accumulation and depletionare happening constantly and simultaneously to allthe firm’s resources, a wide variety of behaviours mayreadily arise – exponential growth or collapse, limitsto growth, boom-and-bust, cyclicality, and so on.

Whilst Figure 8 may seem unfamiliar, the processit describes is very common and well understood. Ifyou start the month with £2,000 in your bank account,receive payments of £5,000 during the month, andpay out £4,000, it should be no surprise that you endthe month with a balance of £3,000.

This simple process applies to anything thataccumulates and depletes, whether cash, customers,staff, capabilities, reputation or morale. It hasprofound implications for explaining firms’performance:

● If performance depends on resource-levels, andthese accumulate and deplete over time, there is

Figure 8Building, and losing, the customer-base resource

25201510

5

Customers

Customers won Customers lost

5/quarterrising!

15 perquarter

q198

Numberper quarter

q2 q3 q4 q199

220210200190

q198

q2 q3 q4 q199

q198

q2 q3 q4 q199

Numberper quarter

25201510

5

q198

q2 q3 q4 q199

q198

q2 q3 q4 q199

5 per quarterand rising!

Autumn 1999

Consumersinterested

Stores won

Storesstockingthe brand

Potentialstore profit

Salesforce

Consumersinterested

New consumers

Advertisingexpenditure

Storesstockingthe brand

they, or rivals, may be able to change resource-levels,in the manner described in Figure 7.

These established but static criteria for resourceadvantages limit the usefulness of another commonidea in strategy, namely that owning resources creates‘barriers to entry’ against rivals. Characterisingresource-ownership as a barrier to entry is a poordescription of reality – firms frequently participate inan industry to some extent with a little of each strategicresource, compete more strongly with more of eachresource, and build competitive advantage by buildingup these resources. Strategic resources are thereforenot so much barriers to entry as ‘hills’ of varying heightand steepness, which firms must climb and from whichthey can compete to a greater or lesser degree,depending on how far they have climbed.

Complementary ResourcesThe last of the conditions given above for strategicresources to provide advantage – that they work welltogether – is particularly challenging, not least becausethe nature of ‘complementarity’ is not well specified,and analytical methods for capturing interdependencebetween resources are not well developed.

It is possible to shed some light on this question,however, once it is appreciated that managers useresources they already have to develop others theyneed. This is not an expression of choice on the partof managers – it is unavoidable. There is no way tobuild any resource without making use of othersthat already exist. Marketing staff need a credibleproduct to build a customer base, sales peoplecannot sell a product unless cost-effective productioncapacity enables them to offer a competitive price,recruiters need a good reputation in the employmentmarket if they are to hire the necessary staff, and soon. Even for a start-up, the entrepreneur appears tostart with nothing, but nonetheless depends uponsome vital intangible resources, such as credibilitywith investors.

This process of interdependence can be illustratedby returning to the case of BrandCo. It is possible thatthe sales force devoted to this product can be allocatedor reallocated quickly; so, unlike other resources, itslevel can be adjusted immediately. So just tworesources remain to be built – consumers and stores.Consumers are stimulated by advertising expenditures,but also by the brand’s visibility in stores asdistribution widens (Figure 9).

Simultaneously, the rate at which new stores arewon over to stocking the brand depends on the size of

Figure 9The rate at which consumers become interestedin a brand reflects advertising and availability

Figure 10The rate at which stores stock a brand reflectssales effort and consumer interest

The Dynamics of Strategy 7

Figure 11 combines these dependencies in brand-building to create a composite ‘system’ for thebusiness. Although the picture may look daunting atfirst, each connection has the precise and practicalmeanings described with Figures 5 and 8 – the thickflow-arrows (stores won, new consumers) indicate thatthe resource stocks are increased by in-flows, whilstthe thinner connections mean simply that one itemcan be estimated from others. So, for example, thenumber of stores next month is the number this monthplus any gained during the month, whilst sales volumecan be estimated from the number of consumers

the sales force, but also upon the number of consumerswho are interested in the brand - no consumer interestimplies no retail sales, so no profit opportunity forstores (Figure 10).

Business Strategy Review

interested in the brand and the availability of theproduct in stores.

Each month the number of new customers is calculatedfrom the current stock of customers, multiplied bythe word-of-mouth fractional growth rate (0.2 permonth). The values on the right record how the stockaccumulates by simply adding these new customersto those already in place.

Figure 11Interdependence between resources in buildinga brand

Figure 11 is known as the ‘strategic architecture’of the firm, or more formally as a ‘dynamic resourcesystem view’. The performance of this particular brand-building system will be explained in more detail later,but before doing so, two features of interdependencebetween resources need to be clarified.

Complementarity between resources– type R: reinforcing feedbackFigures 9 and 10 offer the means to put some precisionon the notion of ‘complementary’ resources. Eachdescribes separately how the rate of growth for eachresource depends on the current level of otherresources in the system. Such systems have aninteresting and powerful new characteristic to add tothe accumulation and depletion of resource-stocks.Since the growth of each resource is accelerated bythe existence of the others, the system is capable ofreinforcing its own growth.

The power of such feedback can be illustrated witha simpler structure concerning just a single resource –the subscriber-base for an internet-service provider(ISP). Although it has been noted that resource growth-rates depend upon the levels of other resources, it isalso possible for growth to be driven by the currentlevel of the same resource. For an ISP, the mechanismat work is ‘word of mouth’, by which customer-basegrowth depends on the current level of that samecustomer-base.

Figure 12 lays out this reinforcing feedback for theISP, and shows how its dynamics can be quantified.

Figure 12Reinforcing feedback grows the customer baseof an internet service provider

Note: The ‘R’ inside the feedback loop denotes that thestructure ‘reinforces’ its own growth – once it star tsincreasing, growth will accelerate.

Figure 13The time-path of reinforcing growth for an ISP

8 Kim Warren

This system is capable of accelerating quite quickly,exhibiting exponential growth. However, such a firmwould be unlikely to rely solely on word-of-mouth,so it may be interesting to see the effect of marketing

Customerbase

10001200144017282072

New customers per monthper existing customer

0.2(fractional word-of-mouth

growth rate)

Increase incustomer base

per month

New customersfrom word-of-mouth

end ofmonth...01234

+ 200+ 240+ 288+ 346

duringmonth 1

234

Newconsumers

Stores (000)Stores won

Productioncosts

Revenue

Grossmargin Costs Sales

volume

Potentialstore profit

Consumers(million)

Advertisingexpenditure

Brand profit£'000/month

Salesforce

Increase incustomer base

per month

Customerbase

15,000

10,000

5,000New customersfrom marketing

(A) zero or(B) 100/month

New customersfrom word-of-mouth

per monthFractional

word-of-mouthgrowth rate

0.2

A

B

3 6 9 12

B

A

3 6 9 12

Months

Months

2,5002,000

1,5001,000

500

3 6 9 12

2,5002,0001,500

1,000500

B

A

Autumn 1999

Servicequality

Subscribers lost'000/month

Maximumsubscribers forgood service

10,000

New subscribers2,000/month

Capacityincrease rate

zero

Marketingspend

Subscribers'000

Months

0 6 12

321

0 6 12

2015105

Months0 6 12

1.00

0.90

0.80

0.70

efforts too – say sufficient to bring in 100 newsubscribers per month. Figure 13 shows how thiscustomer-base will grow, either with word-of-mouthalone (line A) or with this additional in-flow of newsubscribers from marketing (line B).

This experiment offers an interesting observation– using marketing to bring in 100 new subscribers permonth may not seem important in the context of whathas become a firm with 12,500 customers. However,without it, there would have been fewer than 9,000 –an increase of 3,500, although marketing only directlyadded 1,200 customers during the year.

Although this example has demonstrated the powerof reinforcing feedback around a single resource, thesame consequences can arise from positive feedbackwithin a multi-resource system, such as the brandsexample. However, self-reinforcing feedback also hasa dark side to its character – it is just as capable ofdriving exponential decline as it is of causing growth.

What happens, for example, if the number ofBrandCo’s consumers declines for some reason? Thepotential profit available to stores falls, causing someto stop stocking the product. The brand is then lessvisible to consumers, and still more of them forgetabout it. Lower revenues force the firm to cutadvertising and the brand collapses until bothconsumers and stores have forgotten it (Figure 14).

The contrast between these two behaviours ofreinforcing feedback is most starkly demonstrated

where there is close interdependence between twotypes of resource, one of the most common beingthe ties between staff and clients in professionalservice firms. Whilst, say, an advertising agency isdriving forward on a growth path, great creativestaff are joining and great clients are signing up fortheir service. A small reversal, however, such aslosing a major client, can trigger key staff to leave,taking further clients with them, and so on. Thehistories of the advertising agency and publicrelations sectors are replete with dramatic cases ofthis process.

Complementarity between resources– Type B: balancing feedbackWhilst the existence of certain resources can enableothers to grow, complementarity may also arise in theform of one resource constraining the growth ofothers. Consider what happens to our ISP if it under-invests in capacity – servers, bandwidth and so on.Assume we can specify this capacity in terms of the‘maximum number of subscribers who can be providedwith good service’, without worrying about the detailsof hardware requirements.

Figure 15 shows how this firm might performduring the early months, with no word-of-mouth,when marketing is bringing in a steady new stream ofsubscribers to utilise the initial capacity, capable ofserving 10,000 subscribers.

Figure 14Contrasting behaviours of reinforcing feedback

The Dynamics of Strategy 9

Figure 15Resource-balance for an Internet service provider

New consumers

Months12 24 36

Consumersinterested

Storesstockingthe brand

OR

000

Stores won

Potentialstoreprofit

Months

millions

millions OR

12 24 36

Physicalcapacity(static)

Business Strategy Review

Consumersinterested

million Newconsumersper month

New consumersreached byadvertsing'000/month

Advertisingexpenditure

(A=high,B=low)

Fraction ofremainingconsumers

reached(A=25%/month,B=10%/month)

Consumersnot interested

million

54321

12 18 246

54321

12 18 246Month-end Month-end

12 18 246Month-end

1,000800600400200

Any increase in subscribers to a level above thefirm’s capacity causes service quality to drop, sosubscriber-losses rise, thus reducing the subscriber baseback towards a level nearer to the capacity limit. Thisform of interdependence is known as ‘balancingfeedback’ – hence the ‘B’ in the middle of Figure 15 –so-called because, unless prevented, it brings resourcesinto balance.

It may seem puzzling that the resource level hasstabilised above the firm’s level of service capacity.But with no excess resource, there would be nothingto cause an out-flow to occur! Taken to an extreme,this mechanism explains why some firms manage tosustain customer-numbers well beyond their capacityto cope for extended periods of time – the excess issimply ‘churned’ through the system continuously,with angry customers who leave being immediatelyreplaced by others who do not yet know how stretchedthe firm is.

(Those who examine Figure 15 very closely mightconclude that there is actually no feedback at all – theconnections go from subscribers to service quality tosubscribers lost, but no further. The only connectionfrom subscribers lost to subscribers appears to be theflow-arrow, but this is going the wrong way! In fact,the causality implied by the flow-arrow does go theright way, since any outflow causes a decrease in theresource-stock – ‘subscribers today = subscribers lastmonth minus subscribers lost and plus subscribersgained’)

Self-limiting resources – a special case ofbalancing feedbackThe ISP example above illustrated the special case ofreinforcing growth – when it concerns only a singleresource. Balancing feedback too can apply to a singleresource, constraining its growth. Returning to ourbrand-building example, consider what happens if theadvertising efforts are successful for a number ofmonths. At some point, nearly all the consumers whomight like the brand actually are interested in it.Advertising comes up against diminishing returns, andit takes ever-greater efforts to reach the dwindling poolof potential consumers.

This notion of ‘potential’ resources can be used toquantify the growth-limiting effect. Actual consumerscan be developed only from the potential population,so the smaller that pool becomes, the slower is therate at which we can develop them. Figure 16 showsthis effect for two different rates of advertisingexpenditure (assuming both that store-presence is not

helping to build awareness, and that consumers donot lose interest once they are won).

Similar limiting mechanisms constrain manyresources, from tangible items such as customers,qualified staff, or distributors, to the soft, intangibleitems such as staff skills, cost-efficiency, morale, orreputation (eg, the higher one’s reputation, the harderit becomes to drive it any higher).

This observation that there is nothing more to begained from further efforts, once all potential has beenachieved, begs the question as to why BrandCo cannotsimply stop its advertising, training, or other resource-building efforts. Usually, such efforts must besustained, because there is continuous decay of theseresources. BrandCo cannot simply stop advertising,because all the time that it is making these efforts topush new consumers into the ‘interested’ category,others are losing interest again and flowing back alongthe pipe in Figure 16.

Figure 16Balancing feedback limits the growth of a singleresource

Adding such attrition to resource-building time-paths has two implications. First, the higher thestock of the resource, the greater in absolute termsis the back-flow – eg if 10% of consumers loseinterest each month, we lose 500,000 from aconsumer-base of 5 million but only 100,000 froma consumer-base of 1 million. Secondly, the fastersuch ‘forgetting’ takes place, the more effort mustgo into replenishment. This is why, for example,staff training consumes considerable, continuingeffort and costs in sectors such as fast-food, where

10 Kim Warren

AB

A B

A

B

Autumn 1999

staff attrition is high, since departing staff take theirskills with them.

Performance of the resource-systemFigure 11 is more than just a picture of links in thefirm’s strategic architecture – it is an active tool towhich numerical estimation can be applied. Toillustrate, Table 2 and Figure 17 quantify the story ofa specific brand launch strategy in the contextdescribed earlier. (The scenario includes further effects,not covered in detail here, notably: diminishing returnsto advertising; diminishing returns to sales efforts asthe largest stores are exploited by the sales forceleaving only smaller stores to be won; consumers losinginterest in the brand and some delay for advertisingreach to build up.)

At first, the firm invests in advertising, but allocatesonly a small salesforce, believing that consumer pullalone should make the brand take off. However, afternine months (A), there is little sign of up-take by stores,so management doubles the salesforce allocated to thebrand. This increases costs somewhat, keeping brandlosses at approximately £0.5m per month.

By month 15 (B), stores are starting to take thebrand on, but the salesforce complains that lowconsumer demand provides too little retail profit forstores to find the brand attractive to stock. Since thebrand is becoming profitable by this point,management decide it is worth the risk to doubleadvertising to £0.8m/month. Growth in consumerinterest is stimulated once more, enabling thesalesforce to continue gaining stores.

By month 24, (C) management decides thatconsumer interest is getting about as high as theymight hope to sustain, and looks to improveprofitability by cutting advertising spend to £0.5m/month. In spite of this reduction, consumer interestdoes not fall, being supported by the brand’sincreasing presence in stores. (The mutualreinforcement between growth of consumers andstores is now working strongly enough to counteractthe tendency of either consumers or stores to loseinterest in the product.) The brand stabilises, withsubstantial awareness and store-penetrationdelivering profits of £1.7m per month.

Two key issues arise from Figure 17.

● First, this strategic architecture reflects the earlierobservations about financial performance - theprofits of this product ‘hang off the side’ of thebrand’s strategic architecture, rather than being

a part of it. (For this firm, the cash-flows fromthis one product are not a life-and-death matter.Where cash-flow is genuinely critical – not justimportant – as for an entrepreneur’s new ventureor a desperate turn-round case, cash can andmust be included. Cash is treated with the samestock-and-flow framework used for all otherresources.)

● Second, the time-paths for consumers and storesdon’t seem to show any interesting or importantdynamics – they just grow over time. However,this is an inevitable feature of resource-stocks, giventhat any change is incremental to what has beenaccumulated previously. It becomes clear thatseveral important things are indeed happening tothese resources when one looks at the flow-rates(Figure 18).

Experience in applying the dynamic resource-system view (DRSV) method to solve real strategychallenges has exposed the fact that very few firmshave good information on these rates of gain and loss,not just over history, but even currently. For many

Time-period Adver tising Sales force£’000/month people

Months 1-9 400 2510-15 400 5016-24 800 5025-36 500 50

Table 2

Advertising and sales effort scenario

Figure 17Quantifying the rate at which a brand builds,depending on advertising and sales policies

The Dynamics of Strategy 11

Stores (000)

Productioncosts

Revenue

Grossmargin Costs

Brand profit£'000/month

Stores won

Potentialstore profit

Salesvolume

New consumers

Advertisingexpenditure

Consumers(million)

Months

1,5001,000

5000

-500-1,000

Months

Months

Salesforce

201510

5

4321

12 24 36A

CBA

CB

AC

B

12 24 36

12 24 36

Business Strategy Review

resources there may be two or more important flows(one or more inflow and one or more outflow). Lackof good information on flows thus becomes still moreserious: if we know only the history of our totalcustomer-base, but have no idea what happenedseparately to gains and losses, we are in no positionto make well-reasoned policy choices to start toimprove this resource over time.

These two issues together explain why managerscan typically do l itt le to alter underlyingperformance in the short term. It is difficult to makesubstantial changes to the levels of strategicresources, and without substantial change,underlying performance simply reflects these ‘sticky’resource levels. The only immediate discretion is toallocate revenues between expenditures anddeclared earnings (the financials to the left of Figure17). Any such allocations will, of course, have long-term implications for future resource-levels, sincethey affect rates of accumulation and depletion.(Incidentally, this is a worrying aspect of theincreasing pressure on managers to keep declaringimproved earnings – short-term financial reallocationsare always possible, with little immediate effect onresource levels, and hence underlying performance.However, if such inattention to resource-depletioncontinues for an extended time, the underlyingresource-base is damaged, and performancebecomes unsustainable.)

The time-paths of consumers won and stores wonin Figure 18 show with great clarity why the twodominant resources have followed their specifictrajectory over the product’s 36-month history.

Finally, note that every number reported here is areflection of the specific, quantified relationships thatapply only to the case of BrandCo. There are nogeneral conclusions that can be transferred to othersituations from this or any other model (eg ‘it is alwaysbest to invest first in heavy advertising, then commitsales effort later’). This observation may seemdispiriting, since check-lists of best practice, drawnfrom high-profile case-stories, are so often offered todemonstrate standard solutions to widespreadproblems. DRSV suggests instead that every case isunique, and has its own high performance solutions(Glucksman et al 1998). Therein lies a liberatingmessage – if every firm’s situation is unique, and theperformance differences between good and not-so-good strategies are considerable, opportunities forradically-improved performance may be found frommastering the strategic architecture and using it to seekgood strategies for the future.

How to apply this approachThe principles illustrated above can be applied to anytype of profit or not-for-profit enterprise. Whilst thestrategic architectures for the ISP provider and thebrands business were chosen for their compactness, itis possible in most cases for relatively simple, high-level architectures to capture the essence of firms’performance over time. Indeed, the dynamics sucharchitectures highlight can provide more insight thaneven the most detailed and sophisticated spreadsheetplanning models.

The major steps in applying DRSV to practicalcases are as follows:

● Specify clearly the time-path of the strategicchallenge confronting the firm, whether anopportunity to be taken or a problem to beconfronted (eg see figures 1-4 above).

● Identify and define the strategic resources that mustbe developed, defended and connected if thechallenge is to be met.

● Select from this list the three or four tangible resourcesat the core of the business model that must be builtand sustained (see Table 1). Making this selectioncan be tricky. One tip is to avoid abstract or obscureitems – ‘customers’, ‘staff’ and ‘products’ are muchmore concrete, useful items than ‘brand’,

Figure 18Understanding performance dynamics ishighlighted by tracking resource flows through time

12 Kim Warren

Stores wonper month

Stores (000)

12 24 36

Months

12 24 36

1,000

800

600

400

200

20

15

10

5

New consumers'000 per month

Consumers(mil lion)

200

150

100

50

4

3

2

1

12 24 36

12 24 36Months

A

C

B

A CB

A

C

B

AC

B

Autumn 1999

‘commitment’ or ‘focus’ (which are hard to define asresources in any case). A second tip is that the selectionis likely to include one or two resources associatedwith supply of the product or service (production-capacity, staff etc) and one or two resourcesdetermining demand (customers, clients, dealers, etc)Whilst it is not possible to provide here a precise listfor every eventuality, Table 3 offers some resource-selections that may apply in typical cases.

● For each item, specify and measure the inflows andoutflows, if possible separating these two items andcollecting recent history on each.

● For each resource, identify which of the otherresources drive or constrain its gains or losses (seeFigures 9 and 10), and identify two or three otherkey forces driving these flows, whether policychoices, such as spending on training or advertising,or exogenous items such as disposable income orfinal-product demand.

● Combine these pictures into a composite resource-map (as in Figure 11).

● Add the time-charts for as many items on theresource-map as possible, seeking to identify in

particular the dependency of each item on thosethat feed it (see Figures 17/18, but with time-chartsfor additional items in the picture). This stage ofagreeing with the management team the reasonswhy items change as they do can be challenging –a typical reaction is ‘how can you possibly knowwhy customer-gains or staff losses behave as theydo?’ The response is simple – every time a managermakes a decision about pricing, marketing, hiring,product development and so on, she is makingimplicit assumptions about exactly suchrelationships: all we are doing here is getting thoseassumptions out in the open. Not only is this oftena novel experience for the team, it is often novel alsofor the individual, so no-one should feel embarrassedat having no instant answers to such questions.

● Identify the key decision-levers in the system. ForBrandCo, these include advertising spend,salesforce and pricing. For the ISP, levers includemarketing and increases to capacity. Evaluatealternative ‘stories’ of co-ordinated sequences ofdecisions through time, as in the example under‘Performance of the system’ above, and Figure 17.In the process, look out for unintendedconsequences, such as stimulating customer-

Resources Drivers of gains and losses

Publishing/media Readers/viewers Quality of content from editorial staffAdver tisers Readership number and qualityEditorial/production staff Editorial policy

Installed base Reputation for equipment performanceProduction capacity New capital expenditureInstallation capacity Pressure on installation staff

Clients Quality of current work, reputationProfessional staff Pressure of work, advancementServices Knowledge acquired from clients

Banking Account-holders Interest rates, quality of serviceBranches Financial viabilityService staff Pressure of work, training supportProducts New product development effor ts

Insurance Policy-holders Sales staff, quality of policyadministration

Agents Potential customer-basePolicy administration staff Hiring and training

Telecoms Subscribers Call tariffs, switching costsNetwork capacity Obsolescence, new investment

Table 3

Typical core resources for a range of sectors

The Dynamics of Strategy 13

Capitalequipment

manufacturing

Professionalservices

Business Strategy Review

demand that cannot be fulfilled, or building upwork that staff cannot cope with.

It is vital throughout this stage that the team continuesto focus on the scale and timing of this emerging story,both for the decisions that will be taken and theconsequences of the plan. By this we mean: who willdo what, when, and how much, watching out forwhich indicators that their part of the plan is on track,and with what resulting time-path of businessperformance (Figures 8 and 17).

further ‘soft’ category concerns the firm’scapabilities in key resource-building tasks, such asmarketing, product-development and training.These and other soft variables are increasinglybeing captured by firms. They are essentialelements of the strategic architecture, and theirimpact on performance dynamics can be estimated.

● A robust structure for the firm’s strategic architecturepoints strongly to both the high-leverage decisionpoints and key indicators of future performance(often resource-flows). The business resource-system can exhibit complex and often counter-intuitive behaviours. Consequently, choosingappropriate performance measures, both overalland for individual components, becomes tricky. Itis similarly difficult to arrive at simple goals andpolicies for growth of the firm and its parts. DRSVoffers an integrated picture of the whole enterprise,enabling the key performance indicators to beidentified, and pointing to goals and policies thatare likely to realise the potential of the business.

● Finally, DRSV can be readily extended to dealwith issues that cross the multiple activities oflarger corporations. For clarity, this article hasfocused entirely on capturing the mechanismsdriving performance dynamics for a single-businessfirm. However, diversification, vertical integration,mergers and acquisitions, alliances and geographicexpansion, can all be tackled, along with thecontrol and co-ordination mechanisms that directthe strategies of such complex corporate entities.

ConclusionThis article illustrates the core concepts of the dynamicresource-system view of strategy using some simplecases. It is nevertheless hoped that some of thepotential power of this rigorous, fact-based approachto developing strategy is apparent. Even this core ofthe strategic architecture is capable of capturing twocritical features of business reality for manyorganisations:

● that performance depends upon strategic resources,whose behaviour over time depends on rates ofgain and loss, and

● that performance of the entire system reflects whatcan be a complex web of interdependencies betweenthese resources in a manner specific to each case.

Strategic plans and reports often fail to capture eitherof these fundamentals. That many companies do,

14 Kim Warren

Warning!It is imperative to beware of a serious risk –capturing the current strategic architectureinevitably focuses attention on the status quo,so the team may fail to explore possibilities toadapt or redesign that architecture into a newform capable of radically improved performance.In every case, the team should challenge whetherthe strategic architecture that emerges from thisprocess is indeed the best architecture to dealwith the issues and opportunities they face. If itseems that it may not be the best, the stages listedabove can be repeated for novel architecturesand tested ‘on paper’ before committing toradical, possibly risky innovations.

Wider implications and further developmentsFrom the core frameworks described above, a widearray of further developments become possible:

● The dynamics of rivalry can be captured andquantified in order to improve the firm’s influenceover explicitly competitive challenges. Theseinclude the race to develop and capture newcustomers, to encourage rivals’ customers toswitch, and to win the battles for other contestedresources, such as staff and distribution channels.For fragmented industries, firms may be groupedinto clusters with similar resource-attributes andpolicies in order to capture evolving competitiveconditions and industry structures.

● The critical influence of certain intangible itemson firm performance can also be captured andassessed. Examples include ‘quality’ features oftangible resources, such as customer-value, staff-experience and product-functionality. Otherintangibles, though, are more independent, suchas morale, reputation and investor-support. A

Autumn 1999

Technical appendixThe principles outlined in this article can be formalisedmathematically, as follows:

1. ‘Profitability at time T depends on the levels ofstrategic resources R

1 to R

n, to which the firm has

access at that time.’

Diagrammatically:

2. ‘The current level of any resource R at time Treflects its historic rates of accumulation r sincetime t=0’

Diagrammatically:

The value of resource R at time T is equal to the totalarea under the curve for its net inflow since time 0.

Mathematically:Eq.2 ∫ +=

T

iii RdttrTR0

)0()()(

Kim Warren is a Teaching Fellow in Strategicand International Management at LondonBusiness School. He is indebted to his colleagueProfessor John Morecroft for introducing himto the concepts on which DRSV is built, andfor long-sustained support, encouragement andcontribution to the ideas in this article.

ReferencesBarney, J.B. (1991) Firm Resources and SustainedCompetitive Advantage, Journal of Management, 17:99-120.Dierickx, I. and Cool, K (1989) Asset StockAccumulation and Sustainability of CompetitiveAdvantage, Management Science, 35, pp1504-1511.Forrester, J.W. (1961) Industrial Dynamics,Cambridge MA: Productivity Press.1 Glucksman M., Driek, D., Finskud, L., Marshall,N.H., Reyner, M.J. and Warren, K. (1998) The Endof Voodoo Brand Management, McKinsey Quarterly,August 1998.

Mathematically:Eq.1 [ ])(),..,()(

1TRTRfT

n=Π

The Dynamics of Strategy 15

Grant, R.M. (1995) Contemporary Strategy Analysis,.2nd Edn (Chapter 5), Cambridge MA: Blackwell.Mahoney, J. and Pandian, J.R. (1992) The Resource-Based View within the Conversation of StrategicManagement, Strategic Management Journal 13:363-80.Peteraf, M.A. (1993) The Cornerstones of CompetitiveAdvantage: a Resource-Based View, StrategicManagement Journal 14: 179-192.Wernerfelt, B. (1984) A Resource-Based View of theFirm, Strategic Management Journal, 5, 171-180

somehow, manage to perform reasonably well is atribute more to the skill and intuition of experiencedmanagers than to the value of many strategy tools. Itis no longer sufficient to rely on the intuition of airlinepilots to take us safely between the continents.Similarly, managers now need to adopt the dynamicapproach to strategy more formally than in the past ifthey are to guide the enterprises on which people’slivelihoods and careers, even their health and familystability, depend.

Rate of increasein resource i

Time

Resource i

0 T

Time0 T

Earnings

Resource 3

Resource n

Resource 2

Resource 1

3. ‘The rate of accumulation ri of resource R

i at time

T is a function of all resources to which the firmhas access at that time, including R

i itself.’

Diagrammatically:

Mathemetically:Eq.3 [ ])(),..,()(

1TRTRfTr

nii=

16 Kim Warren

These three equations taken together specify thesimplest representation of the firm as a dynamicresource system. A more complete representationrequires additional formulations to capture rivalry andcapabilities. The values of the variables at the pointswhere curved connecting arrows meet is given by thesingle composite function for that variable (Eqs 1 and3). For ease of estimation, these functions may bebroken down into further sub-functions – eg

earnings = revenue – costswhere:

revenue = a function of certain resources …and:

costs = some function of other resources

This build-up of the functions determining resource-flows is captured diagrammatically by intermediatevariables in the structure, such as ‘potential storeprofit’ in Figure 11.

Rate of increasein resource i

Time

Resource i

0 T

Time0 T

Time0 T

Resource 3

Resource n

Resource 2

Resource 1