Dynamic Strategic Thinking

Dynamic Strategic Thinking

Peter R. Dickson Knight Ridder Eminent Scholar in Global Marketing Florida International University University Park, 354A, Miami, FL 33199 Tel: 305-348 6120, Fax: 305-348 6331 e-mail: [email protected] Paul W. Farris Landmark Communications Professor of Business Administration Darden Graduate School of Business, University of Virginia Tel: 804-924 0524, Fax: 804-243 7677 (Fax) e-mail: [email protected] Willem J.M.I. Verbeke, Chaired Professor in Sales and Account Management Department of Marketing and Organization, School of Economics University of Rotterdam Burgemeester Oudlaan 50 Erasmus University, The Netherlands. Tel: 31-10-408 1308, Fax: 31-10-453 2669 e-mail: [email protected]

March 2001

1

Abstract

Market analysts and marketing strategists stress understanding the fundamental dynamics of a market but how deeply do they think about the interplay of such fundamentals and what frameworks do they use in such thinking? How do business schools teach managers to think this way? The premise of this paper is that in their strategizing, senior marketing executives, boards of directors, consultants and financial analysts should see the market and the firm’s embeddedness in a market as a moving video rather than a static snapshot. We propose that what makes the video move are fundamental feedback effects that create the evolutionary paths that a market and a firm may travel. A taxonomy of systemic feedback regularities is presented with applications that demonstrate how the taxonomy and proposed soft mapping techniques can be used to construct dynamic mental models that help managers and consultants improve their dynamic strategic thinking and the strategic foresight of firms.

2 Forecasting the way markets will evolve and the way technology will evolve is an excruciatingly difficult job. History is strewn with people who make predictions that proved to be wrong. I would like to see far more firm roots to our judgments. Alan Greenspan, Chairman of the Federal Reserve, 1998.

Strategic management is the investment, redeployment and restructuring of financial, human, organizational and intellectual capital that create flows of revenues and cash beyond the short-term horizon. It requires a very special marketing planning skill, the ability to understand and anticipate the effects of the complex, often chaotic, dynamic interaction between a firm’s deployment of its resources and its evolving business environment (Teece, Pisano and Shuen 1997; Levinthal and Myatt 1994). This requires understanding emergent patterns in the market (sometimes called industry fundamentals by financial analysts) and their “systemic implications” (Stacey 1995; Pascale 1999). It has also been described in more abstract terms as “seeing” the lawfulness in the apparent chaos of market evolution (Hamel 1998; Beinhocker 1999). A lack of such foresight skills is a form of bounded rationality (Simon 1979; 1984; Kahneman and Tversky 1982; Morecroft 1983) called myopia in growth theory (Romer 1986; Anderson et al. 1988; Solow 1994). The consequential behavior of the firm often results in unanticipated and, hence, unintended longer-term consequences. This article is about improving managerial foresight by improving the mental models that managers use in the conjectures they make in their planning and strategizing. The class of mental models we discuss describe market and firm dynamics that drive the evolution of markets, technology, firms and competitive advantage. The two goals of this paper are to: •

Provide a feedback based knowledge framework for understanding the cumulative, often hard to reverse, evolutionary path a firm takes.

•

Demonstrate how feedback dynamics can be visualized and explained.

3

We propose a taxonomy of the general types of positive feedback dynamics that drive fundamental long-term change within markets and firms.

Feedback effects are often associated with the

generation of standards within high-tech markets (Shapiro and Varian 1999).

But they exist in

many additional market domains besides technology standards. To date, the classification of feedback effects and how they produce path dependencies has been somewhat casual (see Arthur 1996). A more precise and comprehensive conceptualization of feedback effects is needed to provide firmer roots for the application of complexity theory to the marketing strategy process. Such an exercise is most timely as the development of a complex system’s school of thought in strategy formation is at a pre-emergent stage. It is neither mentioned by Mintzberg and Lampel (1999) in their review of the field, nor in The Economist’s latest guide to the essentials of business strategy from A to Z, that does not even define basic terms such as systems dynamics and systems feedback (Hindle 1998). These terms are also seldom to be seen in the marketing literature.

This paper also addresses the theory of firm capabilities by responding to the observation that as far as a firm’s dynamic capabilities and strategic management is concerned, “further theoretical work is needed to tighten the framework.” (Teece, Pisano and Shuen 1997, p. 530). For example, we tighten Teece et al.’s explanation of how a firm’s asset positions “shape” competitive advantage. In our theorizing, a firm’s specific asset positions constitute and create an important class of feedback effects that, in turn, create a firm’s evolutionary path.

The inter-firm uniqueness of these

positioning feedback effects (rather than “the” position itself) generates economic rents.

4

We also use the taxonomy to demonstrate that the “dynamic” component of a firm’s capabilities may be best understood by studying the types of learning feedback that determine the “dynamic” in dynamic capabilities. Learning feedback effects specify the likely dynamic changes in the firm’s capabilities that, in turn, create, deploy and re-deploy a firm’s asset positions into ever more competitive and competitively differentiated positions. Solow (1997) describes this more efficient deployment of manufacturing assets as involving two learning processes, learning-by-doing and learning by discrete innovation.

Argote (1999) also describes two learning processes, the learning

of labor deployed in using the asset, and how managers learn to better deploy the capital equipment and technology embedded in the firm's added-value processes. Their theorizing is in accord with our specification that the class of learning feedback effects we describe are a higher-order type of feedback effect within which the other class of feedback effects resulting from the deployment of assets (or more generally capital) operate. The cumulative, often hard to reverse, evolutionary path a firm takes is created by the interaction between the firm’s asset positioning and learning feedback effects and the business environment’s positioning and learning feedback effects. According to Teece et al. (1997), the greatest potential contribution to strategy lies in improving our understanding of such dynamics. Making such a contribution is the first goal of this paper.

The concept of sustainable competitive advantage has also been clearly connected to learning capabilities (de Geus 1988; Dickson 1992; 1996; 1997) and learning clearly involves feedback. While we might argue that learning can also involve negative feedback (don’t do that again), ultimate success will depend on finding at least some positive reinforcement feedback loops (do more of that). The idea that organizations can develop/evolve learning capabilities which either

5

reinforce the competitive advantage or, in of themselves, become competitive advantages can be framed as a positive feedback effect. Learning to learn is an even higher-order feedback loop. It is our claim that this higher-order feedback effect is a fundamental market dynamic. After-all, learning at a slower rate than our competitors, almost by definition, makes us do the wrong thing or do what would have been the right thing at the wrong time. Thus, understanding what our customers want, but being always one step behind the competition must reduce the profitability of the firm and the economic rents that it earns.

The second goal of the paper is to demonstrate how multiple feedback effects can be visualized and explained. The basis for strategists’ insights, instincts and intuition is a superior heuristic or frame (Amit and Schoemaker 1993). Using a feedback system mental model is, "ideal for bringing forth the mental maps that we had sought to influence with scenarios." (de Geus 1997, p. 69). Schoemaker (1995) describes scenario planning as organizing and interpreting complex dynamic systems "into narratives that are easy to grasp." We provide two frameworks for such narratives: feedback maps and market feedback matrices.

The better visualization of feedback effects helps shift the focus of managers' mental models to flows rather than stocks (see also Senge 1990).

Visualization of the dynamics of a system

encourage a paradigmatic shift of focus from states of nature (e.g., market share, quick ratio, brand equity) to rates-of-change of states (e.g., rates of change in market share; Dickson 1992; 1996; 1997).

The scenario planning frameworks we propose can be used as a foundation for the

narratives or stories that describe market and firm scenarios. They also can reduce the frequently

6

reported tendency of managers to become preoccupied with a single feedback effect (Schoemaker 1995; Bower and Christensen 1995): “We suggest that in evaluating the probability of complex events only the simplest and most available scenarios are likely to be considered. In particular, people will tend to produce scenarios in which many factors do not vary at all … …because of the simplified nature of the imagined scenarios, the outcomes of computer simulations of interacting processes are often counter-intuitive (Forrester 1971).” (Tversky and Kahneman 1982, p. 177)

The proposed market feedback matrix can also be used as a diagnostic tool to identify the feedback effects that are driving the market metrics the firm is tracking (or should be tracking) in its market surveillance. Such metrics include sales, average costs, prices and profits of markets and specific firms, consumer and distribution penetration curves, number of articles written on a new technology, organization morale, etc. These feedback effects can then be explained using more detailed feedback maps. Beinhocker (1999), the co-leader of McKinsey & Company’s strategy theory initiative, argues that business strategy is a process of solving complex dynamic problems. Feedback matrices and maps help in this process by surfacing the order in the seemingly chaotic and uncertain evolution of competitiveness in markets (Hamel 1998). In simulating the evolution of a market this intermediate “surfacing” and “validation” step is needed before the system’s dynamic simulators (i.e., complexity consultants) bury the identified dynamics in a string of differential equations that mathematically specify the feedback dynamics of the system. Needless to say, the mathematical modeling is a most challenging and complex task. Explaining the equations to senior management is even more challenging, if not impossible. Consequently, once constructed, it is too late to use senior managers to evaluate the ecological validity of such mathematical models. But even when strategic management does not involve complex systems simulation, we argue that the

7

surfacing of positioning and learning dynamics is of considerable value. This step seems to have close parallels to scenario planning processes that attempt to define the nature, level and scope of “uncertainty” the firm faces in its strategic planning (Courtney, Kirkland and Viguerie 1997).

Feedback Effects and Marketing Strategy Varadarajan and Jayachandran (1999) have provided a comprehensive overview of the most important developments in the evolution of the marketing strategy literature.

There are several

connections in this overview to the themes that we develop here, some of the more important are: competitive reactions, increasing returns, the experience curve, and organizational learning. V&J’s review of the literature is summarized near the beginning of their article in an organizing framework that pictures the environment affecting strategy, resulting in competitive positional advantages, which, in turn, affects market-based performance and financial performance. This key depiction of influences flows from “left to right” (environment to strategy to advantage to performance) but it is clear that performance also is meant to feedback on the environment in the sense that the average and varied performance of the firms in the last period shape the next period’s business environment. While there are no feedback loops connecting the “output to input” in the framework, V&J clearly also believe that the notion of learning as feedback is inherent in understanding how strategy evolves. In concluding sections that highlight organizational learning V&J write, “Analysis of improvisation requires an examination of the temporal sequence in which actions occur in firms. In the continuous feedback system envisaged here, the strategy formulation process, the strategy content, and its implementation occur, in the limit,

8

simultaneously.

Investigation of such theories can benefit from a process view…to

understand better how strategies emerge in organizations.”

We agree that the literature can benefit from a more thorough integration of the idea of positive feedback into theories about the emergence or evolution of competitive advantage. Along with institutional theorists DiMaggio and Powell (1983), and Selznick (1957), V&J propose that: “the actions of firms and the outcomes of these actions are influenced by the knowledge systems, beliefs, and rules that characterize the context of the organization. This dependence of the actions initiated by firms and the outcomes of such actions on the environment or context of the organization is termed embeddedness (Porac and Rosa 1996).” (Varadarajan and Jayachandran 2000, p. 123). (Italics added) By first providing a taxonomy of the class of action-reaction phenomena called feedback effects, then describing a way their interaction can be visualized and thought about (market feedback maps), and third proposing a way the latent market feedback effects (market feedback matrices) can be described and the dynamic capabilities and competitive advantage of rivals compared, we believe we are providing a number of heuristics, frameworks and propositions or rules (that is elements of a knowledge system) for thinking about embeddedness. How practicing strategists actually think is a very different and fascinating empirical issue and question.

The first section of the paper defines some basic terminology that helps clarify the meaning of the term increasing returns, not to be confused with the term increasing returns on investment used in traditional economic theory. The second section discusses asset position feedback effects that are

9

associated with supply and demand returns to scale within and along the added-value supply-chain. This is followed by a discussion of types of learning feedback effects that occur within the firm, across firms, within consumers, across consumers and across supply chains. The final section discusses the incorporation of feedback matrices and feedback maps into strategic planning.

The Basic Concept of Feedback Effects

A market feedback effect is a recursive relationship between one changing state of nature in a market and another changing state of nature in a market. It produces an underlying negative or positive trend, pattern, fundamental, systemic dynamic, or serial correlation in relationships between supply and demand constructs and within supply and within demand. The notion that market dynamics are not all stochastic and may have regularities that can be analyzed and partially understood is an essential premise of modern growth theory, industrial dynamics and theories of finance. Thus the concepts of positive and negative feedback effects, as an important class of market dynamics, are core to many contemporary theories of the firm. Negative feedback effects create stability and organization in nature, society, and markets.

Positive feedback effects create

permanent change and growth: the evolution of economic systems and markets (Jervis 1998). The task of the strategist is to understand the ying-yang interplay between these two effects.

Negative Feedback Regularities The best-known negative feedback effect is the equilibrating regularity in neo-classic price theory which is taught in all Economics 101 courses and is specified mathematically by the price elasticity

10

formula. When a supplier raises its prices, some consumers react to the price increase by ceasing to buy the supplier’s goods/services by reducing rate of consumption and frequency of purchase, postponing consumption or switching to a competitor's product. This reaction leads to a fall in the seller's sales which leads the seller to reduce its price back down again. It is called an equilibrating regularity because it tends to bring the market back toward its previous state of production and consumption. In the abstract, a negative feedback effect is when an increase in x (say price) leads to a decrease in y (demand) that then tends to lead to a decrease in x (price). In mechanics a negative feedback effect is called a servo-mechanism because the system has a built-in control mechanism that brings it back to its previous dynamic performance, e.g., an engines revolutions per minute. What is observed in the actions of a servo-mechanism is an equilibrating regularity.

Positive Feedback Regularities A pattern of positive feedback effects that has been called a speculative bubble has been frequently observed in economic history. Investment in gold, tulip bulbs, Florida real estate or Internet stocks can "feed on-itself" and develop a "life of its own." Speculators pay a price for the investment that has nothing to do with market fundamentals – the current and forecast stream of returns from nonspeculative investment in the market. They pay a price they never would otherwise pay because they conjecture that there will be plenty of buyers (that is, greater fools, which is why it is commonly called the “greater fool” theory of stock-market investment) at an even higher price in the near future. This belief creates an endogenous positive feedback effect where rising prices, rather than reducing demand, leads to even greater demand as speculators are attracted by the prospect of making arbitrage profits out of rising prices. The result is more speculative demand and

11

buying that further raises prices, generating further arbitrage profits, which further raises demand. Reality confirms beliefs, and beliefs lead to the predicted reality in a self-fulfilling prophecy. Often analysts at this stage of a speculative bubble feedback effect mask the truth by attempting to rationally define the phenomenon as something else, such as a “new economy”, without ever really explaining their explanation. But when reality finally fails to live up to soaring expectations, as it inevitably must, a rapid reversal in the positive feedback effect will occur. Panic selling is induced by the belief that it is best to get out of what is now perceived as a bursting speculative bubble as fast as possible. This leads to an explosion of selling that rapidly drives down prices that leads to further selling. As we shall see, there are many other types of positive feedback effects that are much more common and economically productive than the positive feedback effect underlying speculative bubbles. In fact speculative bubbles must be categorized as one of the most inefficient positive feedback effects in the capitalist system because of the misallocation of resources they produce.

A positive feedback effect is the opposite of a servo-mechanism because any change in x leads to further change in x in the same direction. The further change sometimes is even greater than the initial first-order effect but it is often less, producing an exponentially shaped "path" for both y and x, over time. This raises the issue of how to define increasing returns. In evolutionary theories of disequilibrium economics such as developed by Arthur (1994), the term increasing returns is a synonym for positive feedback phenomena. A change in state of x or y has an increasing return because second-order feedback effects continue to increase x and y in a particular direction over time.

12

Increasing Returns and Feedback Effects This definition of “increasing returns” should not be confused with managerial theories of equilibrium economics where the term “increasing returns” describes the phenomena where the return on spending on a production factor increases as spending increases. It is very important to keep these two interpretations of “increasing returns” distinct, particularly in any theory that uses both terms. We attempt to do so by using the term positive feedback to describe the dynamic that creates path-dependencies and what Arthur and others who reference his work call “increasing returns.” We prefer to use the term increasing returns in the classic way to describe how, as x and y progress within a feedback relationship, a further increase in x creates an incrementally larger increase in y. In the dynamics of calculus, “increasing returns” is specified by a positive first and second derivative in the relationship between x and y. This may occur for a period of time within a feedback dynamic (during the first turning point of the “S” curve for x and y).

For increasing returns to be the result of positive feedback, the output of the system (network utility, perhaps) has to be connected to some inputs (not just paramorphically in a set of model equations but in the real world that the equations model). Farris and Pfeifer (2000) have noted that when the Interstate Highway System was first constructed, the receipts from gasoline taxes were not devoted to highway construction funds. The feedback loop was not completed. Therefore, even though the U.S. enjoyed the increasing returns from an ever-more-connected Interstate Highway System, which increased utility and usage, an important feedback effect was missing. Output was not connected to input in a sufficiently direct and powerful way. Once the public (or highway construction companies) lobbied congress to devote gasoline taxes to

13

highway construction a market-level positive feedback connecting usage to road spending (that is demand for highway construction) was created. This is a managerial task for social intuitions. Railroad traffic suffered as highway traffic grew. Funds for maintenance, R&D, for the railroads were diminished and the same kind of positive feedback worked against the railway industry and helped change the future. Might we have had a country connected by high-speed railroads with only adequate highways? Of course, the fact that it didn’t happen that way was not necessarily a sign of the inevitable dominance of concrete over rails, but instead could have well been the result of chance, technological momentum, and governmental or managerial intervention at critical points in time. We can imagine, for example, that Eisenhower’s witness of the German Autobahn might have helped create the general vision of a “direction” in which he wanted to move the USA after World War II.

The “experience curve” is a well-known and controversial potential source of a within-firm-level positive feedback effects. If higher share results in higher production volumes that lead to lower costs; firms may then lower their prices and see another round of share and volume increases. Alberts (1989) criticizes, rightfully, we believe, simplistic interpretations of the experience curve. In particular, he argues that share advantages and lower costs are not direct causes of each other, but rather “possible contributing” causes. He argues that firms will generally be better off if they push their costs down through innovation: that is, through Solow’s discrete innovation (Solow 1997) and Argote’s learning to deploy assets better (Argote 1999). While a detailed critique of his analysis is beyond the scope of the this article, we note that Albert’s

14

article is representative of many other that attempt to “debunk” what they regard as overly simple interpretations of cause and effect relationships.

The market share-ROI debate is another example of a literature that is plagued by the same issues (Varadarajan and Jayachandran 1999). While early analyses probably exaggerated the simple, direct effect of increasing share on ROI, there is another side. Much of the literature on market share and profitability failed to analyze the relationship between market share and ROI as part of a longer-term strategic feedback loop. Instead, the focus was too often on small incremental changes in market share and incremental change in profit, concluding that increases in market share itself did not increase profitability. Has the forest been missed for the trees? Albert’s argument seems to boil down to the notion that the learning curve enables innovation, which drives costs lower, and therefore we should concentrate on innovation and not simply achieving volume. Our view is that it is not the inevitability, but the potential, for such positive feedback that should be our focus and that both learning-by-doing and learning by innovation feedback effects can create a firm’s experience curve.

The Basic Taxonomy In the following sections we develop a more comprehensive description and classification of supply and demand feedback effects. It is important to keep in mind that these mechanisms occur within firms and also between firms connecting the actions of firms and their competitors with suppliers, resellers, and consumers. Such feedback loops connect the performance and competitiveness of members of the value-chain through changes in some variables (triggers) that

15

stimulate a response in other variables. These responses ultimately "feedback" to the triggers to cause another change in the trigger variables. Examples of x and y variables that are involved in these feedback loops are investments, costs, prices, demand, sales volume, consumer utility and profits.

According to Arthur (1994) feedback effects are self-reinforcing mechanisms that “are variants of or derive from four generic sources: •

large set-up or fixed costs (which give the advantage of falling unit costs to increased output);

•

learning effects (which act to improve products or lower their cost as their prevalence increases) (Arrow 1962; Rosenberg 1982);

•

coordination effects (which confer advantages to “going along” with other economic agents taking similar action); and

•

self-reinforcing expectations (where increased prevalence on the market enhances beliefs of further prevalence)” (p. 112).

Our conceptualization categorizes feedback effects into two major classes (see Table 1): 1. Types of asset position advantages involving investments in R&D, manufacturing, distribution, and brand equity that create feedback dynamics, and 2. Types of learning dynamics that over time help create asset position advantage effects. We shall demonstrate that Arthur’s coordination effects are actually either positioning advantages that are subject to scale dynamics and/or are a special class of network learning dynamics. We also argue that self-reinforcing expectations are a special type of learning dynamic.

16

Investment and Positioning Feedback Dynamics

If any technology or added-value process has a cost structure that involves heavy investing in expenses or fixed assets, then such an asset “position” can create a market economies-of-scale positive feedback effect (see "Up-front Costs", Arthur 1996). Over a standard time period of say a year (a typical time period used for venture planning and budgeting), the average cost of goods sold is calculated using the following equation: Average cost = Average direct cost + Fixed Costs.(a+i)/Q …………………..(1) where “Q” equals the sales quantity in the year, “a” is a standard yearly amortizing fraction1 and “i” is the rate of return the market expects on the fixed costs deployed in manufacturing and marketing the product. These fixed costs are the financial and human capital that have been invested to date on the project. They reflect the “assets” the company has deployed - its market asset positions (Teece et al. 1997) - and include knowledge assets, manufacturing assets, and complementary assets such as distribution assets, and other marketing assets (e.g. a web-site).

If the firm is attempting to

grow the market and grow its market share, then as its average cost of goods sold decreases with increased sales, its price will similarly decrease. When it lowers its price in the next period, demand increases and its sales increase. This feedback effect underlies the following belief: "To the leader belong the spoils. Whether you make shoes, semiconductors or hair coloring, it's the same story. He who gets the biggest market share can spread research costs and advertising costs over a broader base and come out with lower unit costs." (Brown 1996, p. 56)

If price is a linear function of the average cost then differentiating this linear function after substituting the right hand term in equation (1) for average cost leads to the following equation:

17

dp/dQ = -f(Assets deployed/Q2) …………………………………………(2) What equation (2) shows is that the larger the assets deployed (on the top line in the bracket), the steeper will be the price reduction path as quantity sold increases and the more powerful can be the asset position feedback effect. The deployment of fixed cost assets in manufacturing, distribution and marketing creates the potential for a firm to drive down its costs that drives down prices that further increase market demand and market share. Therefore, the most obvious way of creating a position feedback effect is to increase investment in the fixed cost component of manufacturing and create a cost structure with a very large fixed cost component. The competitive advantage is then in the competitively differentiated manufacturing cost structure position that produces what Chandler (1990) has called the enduring logic of industrial success. For instance, when Henry Ford built assembly lines and trained his work-force in specialized assembly line tasks, he vastly increased his fixed costs but lowered his variable costs significantly. The more cars he sold, the lower his average cost, which enabled him to lower prices (that increased demand) and/or invest his profits in even more capital intensive production. Position feedback effects are also commonly created around fixed cost investments in distribution, research and development, brand equity and customer service. For example, as firms such as Cisco move their customer service processes to the Internet not only does the cost of servicing each contact drop about five-fold ($15-25 to $3-5), but the nature of the cost shifts from mostly variable (labor) to mostly a fixed cost (investment in the Internet system) thus ensuring that future customer service costs will decline as the firm grows. When firms in an industry learn from an innovator (Hayek 1978) and adopt similar production and distribution processes that involve large fixed cost components (e.g., the various energy industries) then the

18

industry cost structure and the associated industry level feedback effects connecting average industry costs and prices to demand and supply become a structural characteristic of the productmarket.

Positioning Timing Dynamics The investment required in organization design, research and development, manufacturing, distribution, and brand equity assets required to enter a market and challenge the existing market makers constitutes a barrier to entry. But this is too static and defensive a conceptualization of positioning advantage. The more a firm possesses manufacturing capacity, distribution reach, communication reach, and brand equity, the greater is its ability to convert potential demand into actual sales. But it is a firm’s early positioning and deployment of these investments in a market’s evolution that gives it a potentially permanent advantage (see Teece et al. 1997). Its larger and increasing sales base and cost advantage dynamic, if successfully managed, means that rivals deploying similar resources later cannot catch up. Conversely, Tellis and Golder (1996; 2001) found that pioneering innovators do not always achieve a "first-mover" advantage if, at the right time, an imitator commits more resources to building distribution or can take advantage of its established distribution system and brand reputation position. This is a form of installed base, fixedcost feedback effect. The greater the installed base of cooperating distributors and deployable brand equity the greater are the returns from push-and-pull marketing. A follower’s earlier deployment of distribution and brand equity assets can trump a pioneers’ earlier deployment of knowledge and manufacturing assets. What is noteworthy for marketing strategy is that complementary assets such as distribution reach and brand loyalty/equity are more re-deployable and less specific than

19

knowledge and manufacturing assets that are much more closely tied to a specific product technology. Thus, pioneering firms that have built their positioning advantage around technology specific knowledge and manufacturing assets are at a strategic asset position disadvantage compared to firms that possess distribution and brand equity positioning advantages because the latter marketing assets are easier to re-deploy.

Varadarajan (1999) argues that, “although Sony was the first to market with its Betamax format VCR, the ultimate emergence of Matsuhita’s video home system (VHS) as the industry standard was a result of a multiplicity of strategic decisions.” These include: competitive pricing strategy, scale effects, experience effects, strategic alliance effects, and positive network externalities. The VHS marketing alliance that steam-rolled Beta technology out of the market delivered an extensive distribution network and set of brand equities that the licensees already had in place and that could be readily re-deployed and managed to create a powerful marketing assets position feedback dynamic. The emergence of the DVD also shows the evolutionary nature of such technologies. The VHS “Lock-in,” that has so concerned some economists, is a temporary phenomenon.

To this point we have been describing a potential virtuous circle dynamic, where lower prices and growing sales feed on themselves, driven by a positioning feedback effect. But such feedback effects can turn vicious where shrinking sales and rising prices feed on themselves. This is the risk that the dot.coms faced in their large start-up fixed cost investments. If sales decline rather than increase then their average cost-of-sales can increase geometrically and they bleed red ink because they dare not raise their charges and risk a further decline in sales and increase in average cost-of-

20

sales. When prices and markets first collapsed in the early 1930s, firms responded to a reduction in sales by calculating that their average cost of manufacture and selling had risen. This average cost increase was most evident for firms that faced large fixed costs. The application of a cost-plus rule led to firms raising their prices in response to a drop in demand. This action further decreased sales, causing prices to spiral up and demand to spiral down (see Figure 1). This sequence of events could not be explained by contemporary economic theory that predicted the basic negative, equilibrating feedback effect described above: as demand falls, prices fall, which increases demand again. In fact, a special congressional inquiry was held into why prices were being raised as company sales fell. The allegation was that it was being caused by a price conspiracy between the leaders of industry and Congressional hearings were held to investigate the “price-fixing”. In terms of political economy the explanation was less political and more economic in its nature: the activation of a powerful cost structure position – pricing practice feedback effect.2

Distributor Network Investment Dynamics A positioning feedback effect often extends to investments that suppliers or distributors make to support a technology leading to a general supply-chain system feedback effect. The gasoline engine generated demand for gas and the investment by general stores and gas stations in gas pumps made it easier to use gasoline engines. It also encouraged investment in gasoline wholesaling and refining systems with high-fixed cost structures. The advantage of the hybrid electric vehicle (HEV) over the pure electric vehicle is that its capability of making the market for low emission vehicles is potentially so much greater because the support infrastructure (necessary collateral assets) for the

21

HEV is already in place. Short of plugging in at a friend’s house and staying the night, such a support infrastructure does not exist for pure electric vehicles.

A distributor's propensity to stock many brands or only a few dominant brands further moderates accessibility to consumers (Farris et al. 1998) and the ability of an innovation to “make” the market. The need for suppliers and distributors to invest in specific assets such as merchandising equipment, storage plant and employee handling, sales and service training is another type of installed base driver of market evolution. Thus, positioning feedback effects must be studied along the entire supply-chain in strategic decision making.

End-User Investment Dynamics Consumer investment in technology and training creates a similar positioning feedback effect that provides a seller an installed base advantage called "Customer Groove-in" by Arthur (1996) or “lock-in” by Shapiro and Varian (1999).

The firm with the largest number of video game

controllers in homes has a decided position advantage because the consumer faces a high one-time fixed cost of switching to buying a rival's game software. This means that its new, improved games are likely to be more successful and produce higher returns than its rival's similar efforts. Similarly, moving consumers to the new digital camcorders will be a slow process, even though these consumers are far along an existing learning curve (and can be sold on the technological improvements) because they have a sunk cost in their existing camcorder and the playback equipment (their VCR).

The prediction then is that these end-of-the-supply-chain fixed cost

feedback effects will slow down the adoption of digital camcorders, but that once the digital TVs

22

and associated digital VCRs are purchased, these new fixed cost investments will dominate and will accelerate the adoption of digital camcorders. In short, a sound strategic decision is for distribution channels to not invest heavily in marketing digital camcorders until about 20% of households have HDTV sets.

At that time a distributor should go for broke, marketing digital camcorders

manufactured by firms who have driven down the design, manufacturing and supply-chain learning curves for digital camcorders.

Utilities of Scale Network Position Feedback Effects A pure network utility effect occurs when the utility of joining the network increases as the size and scale of a network increases. As businesses and households acquired telephone services, the utility of using the service to communicate around the network increased. This network effect has been replayed with fax machines and Web based e-mail. It is also occurring in less obvious ways such as the growth of the network of English speakers in the global business world and in science. As English grows as the language of communication the value of learning the language increases. As other languages shrink as a language used for global communication, the value of investing in learning the language decreases. The evolution of the VCR standard was subject to such an effect because the purchased or home recorded tapes were shared around friends and relatives. As the network of friends and relatives with VHS players expanded, the perceived utility of such tapes increased. The larger the network, conceptualized as an asset position, the greater the individual utility of the network. This is a utilities-of-scale feedback effect. To the extent that the network’s functioning also involves a high shared fixed cost component (such as with the World Wide Web), such coordinated networks also exhibit economies-of-scale as well as utilities-of-scale feedback

23

effects. Technology standards deliver both types of economies across networks which explains their influence and advantages in high-technology markets (e.g., a common technology standard in the European wireless phone market).

In supply-chains, the greater the B2B vertical integration between supply-chain partners via electronic data interchange, the greater the communication and coordination advantages up and down the added-value chain. The pressures to cooperate and coordinate become huge on laggards in the vertical network because the gains to all increase greatly when the whole network is integrated. While the private ownership of interface software that standardizes transaction and trading communication between businesses computers world-wide via the Web can be questioned (cf. The Microsoft antitrust case), the huge network advantages of developing such a standard cannot. It will, however, reduce the unique asset position advantages of companies such as Campbell Soup that has spent tens of millions of dollars on dedicated communication networks with their trade customers and suppliers. As we shall see below, there are further feedback effects that occur across networks of social and economic groups that are associated with morale, motivation contagion, learning and organization transformation and reconfiguration.

Learning Feedback Dynamics

Understanding the learning dynamics of firms and markets is key to understanding the evolution of firms and markets and long-term competitive and comparative advantage (Dickson 1992, 1996; de Geus 1997). There are five general classes of learning dynamics that can create positive feedback

24

effects within and between firms. We categorize them into motivation dynamics, learning-by-doing dynamics, learning ability dynamics, surveillance dynamics, and routinization dynamics.

Motivation/Drive/Contagion Dynamics The motivation to compete between and within companies often exhibits short-term and long-term feedback effects. One form occurs in price wars and has been observed to occur in the real world and in experimental studies of decision-making. Aggressive competition is likely to lead to aggressive competitive reactions which can lead to an all-out price war (Cassady 1957). Even the “tit-for-tat” decision rule in game theory can lead to either escalating competition or cooperation, depending on initial moves (Axelrod 1984). It can be accompanied by an emotional contagion in an industry that is often observed in social groups and subcultures (Hatfield, Cacioppo and Rapson 1994)3 such as the animosity between senior executives in the airline industry that occurred during the 1980s. The momentum to cut prices also has a rational economic basis. As the number of rivals cut their prices (and particularly when the market leader cuts its prices), the price cutting momentum forces the holdout rivals to cut their prices or lose hard earned customer loyalty (Urbany and Dickson 1991). A price-cutting war, whether rule based, emotionally based or rationally based, also increases motivation to make process efficiencies whose effects accumulate over time, long after a cease fire has been called in the price war. A quality improvement motivation can also feed on itself, as it has in the automobile industry, as rivals strive to out-do each other in fewest reported user defects.

25

For Every Action There is a Competitive Reaction According to Varadarajan and Jayachandran (1999), “Competitive behavior, the actions and reactions of competitors, is central to marketing strategy research and practice.” Competitive reactions have historically been measured in one of two ways. The first is the notion of a “reaction coefficient”. If the focal firm increases its marketing, how will competitors respond? If we can anticipate a response, then the estimated effects of our changes in marketing strategy should incorporate the timing and magnitude of those expected responses. If we add to this reaction coefficient,” the idea that the focal firm will respond to the actions of the competitors, we have a complete feedback loop. Input generates output and the output affects input. Now, the issue of whether this is a stable feedback loop that dampens outside perturbations, or whether it is an amplifying feedback loop is the difference between negative and positive feedback.

A game theory concept that V&J argue is relevant to understanding competitive dynamics is the notion of a Nash Equilibrium (NE): a point at which none of the competitors would have an incentive to change, because it would make them worse off. The perspective in this paper is that short-term reaction coefficients are unlikely to lead a firm to a specific equilibria that “attracts” all significant competitors. First, our view is that even if there is a NE, it is highly unlikely that that firms would know exactly where it is. Also, we doubt that the NE would remain fixed over time or even evolve in a predictable/knowable way. Even with sophisticated signaling, there will inevitably be noise and error in the systems that will prevent firms from finding or staying at NE points. What seems more likely to us is that firms envision a general “direction” or “path”

26

along which they want to travel and attempt to develop the capabilities to make it profitable: with or without the direct or indirect cooperation of competitors.

The Red Queen Effect. Over and above short-term price or quality wars, there can exist a long-term trend that inexorably increases the motivation to improve performance. The basic dynamics of a market are that individual seller innovations change consumer preferences. Firms learn from the innovator and general supply then re-deploys resources to cater to the emerging new profitable segments of demand created by the changes in consumer preferences. This redeployment creates excess supply capacity targeted at demand that increases the motivation of sellers to experiment and innovate, and benchmark and imitate (Dickson 1992). This leads to further seller innovations feeding back to further changes in consumer preferences and redeployment of supplier resources. With this cycling around of innovation-imitation and remaking of the market, each rival has to increase its motivation and effort to serve the customer through cost reduction and quality improvements. It has been called the Red Queen effect because as the Red Queen explained in Alice and Wonderland, you have to run ever faster just to stay in the same (competitive) position (Barnett and Hansen 1996). It leads to a rivalistic accumulation in motivation to learn and improve performance that is a fundamental trend in free markets.

When innovations in supply change demand preferences, but do not increase overall demand, the motivation to quickly adopt performance improvement innovations can become particularly fierce. For example, consider the introduction of food supplements that increase dairy cow productivity. The sellers that profited most from this innovation were those that adopted them fastest before the

27

increase in total supply drove prices down. Over the last 100 years, the window of opportunity to appropriate economic rents from new production innovations has decreased for farmers because they have collectively learned the fundamentals of being either “quick” or “dead.” This has led the survivors

to become even quicker at learning to improve and innovate-imitate.

A similar

motivation to improve is occurring in distribution amongst wholesalers and retailers where total supply measured in terms of square-footage capacity has outstripped total retail spending. As each new innovation in supply-chain management and efficient consumer response is introduced, the time that the early adopters of the innovation such as Wal-Mart have to appropriate the extra profits from such innovations is less than it used to be. This is because everyone is now more alert to new ideas and motivated to benchmark and imitate the best new practices in B2B distribution logistics.

Within-firm motivation dynamics. A motivation to try-harder feedback effect can occur within firms as well as between firms. Figure 2 presents a service quality - employee morale positive feedback effect within the firm (Gronroos 1984). A decline in employee benefits and wages is created by an external trigger such as an economic downturn or a merger that leads to senior management belt-tightening. This leads to a reduction in employee morale which is contagious and leads to a drop in employee service, that leads to a reduction in demand for the firm’s offerings, that leads to belt-tightening and a further drop in employee benefits and wages, that continues the vicious circle or spiral downwards. The same effect can result in a "virtuous circle" created by a positive pulse in employee benefits and salary that raises morale, employee service, demand, and future employee benefits and wages. Southwest Airlines experienced such an effect throughout the 1990s.

28

Learning-by-doing Dynamics Firms learn by doing (Arrow 1962; Solow 1997; Argote 1999). The more they do, the more they learn. The more they learn, the better they do. The better they do, the more they do. Consumers learn from using (Rosenberg 1982). The more they use, the more they learn. The more they learn the better they use. The better they use, the higher the usage utility. The higher the usage utility, the more they use. This basic experience learning effect applies to all processes within the addedvalue supply channel and adds an important higher-order dynamic to the above positioning dynamics. Experience with assets, such as production machinery or a customer service web site, imparts a learning dynamic to asset utilization that can transform and reconfigure scale economies in ways that accumulate. The functional relationship between quantity produced and sold and average cost (which specifies the economies of scale feedback effect associated with assets employed) is itself subject to a positive feedback effect.

The effect on costs and prices of

positioning feedback effects, nested within learning-by-doing feedback effects, can and has been long tracked by researchers (Wright 1936; Hirsch 1952) and consultants, becoming known as the “experience” curve or progress functions.

Dutton, Thomas and Butler (1984) in a review of the

history of progress functions as a management technology, point out that a great deal of attention has been historically paid to direct-labor learning and only recent studies have focused on the cognitive learning of managers (cf. Alberts 1989).

29

Two questions that have been begged are what is the psychological and sociological nature of learning-by-doing?4, and, what does “experience” mean strategically? Assets are deployed within or across added-value manufacturing and distribution processes. But they are always re-deployed and reconfigured within the context of redesigned added-value processes. It is thus, these processes that are subject to various levels or degrees psychological and sociological “learning-by-doing” process improvement.

The first level is that the design of the process remains the same but its

execution is improved with experience. In Adam Smith's famous example, the pin-maker becomes more productive through muscle memory resulting from the repetition of a specific activity within the pin-making production process. The task is completed faster with fewer defects. Clearly learning is very localized around the specific standard operating procedure/activity and is likely to level off quite quickly. It is also largely motor rather than cognitive learning. The second type of process learning results from an incremental series of small alterations in the design of the standard operating procedure (e.g., use of a better frame for stretching the wire) proposed by the individuals undertaking the activity but that have no impact on other activities within the process. This type of learning implies that the labor is thinking about the process as they work and coming up with creative improvements in the process.

The third type of process improvement occurs when the redesign of an activity affects other activities. It leads to a reconfiguration of the work-flow process, perhaps even involving the elimination of activities, the replacement of manpower by machines, and the replacement of management by computer control systems. This learning is now much broader, localized around a technological system such as occurred when primitive machine tools were introduced into pin-

30

making, along with labor specialization. Finally, a fourth type of process improvement learning occurs when a whole technological system is replaced such as moving to continuous flow steel making where very little of the previous plant, equipment and technology is utilized and even the management and labor skills are only marginally transferable.

Localized Learning Feedback Effects The learning-by-doing process can involve extremely localized experience in executing a routine, experience in executing small localized process improvements (undertaken much less frequently) and experience in executing broad radical transformations of whole systems of processes (undertaken very seldom). Learning localization is actually a continuum, but it has been simplified by some scholars into an exploitation-exploration learning dichotomy: “The essence of exploitation is the refinement and extension of existing competencies, technologies, and paradigms. Its returns are positive, proximate, and predictable. The essence of exploration is experimentation with new alternatives. Its returns are uncertain, distant, and often negative.” (March 1991, p. 120)

In the early stages of the evolution of an organization most of its competencies, technologies and paradigms (standard operating procedures and mental models) were new and learning was exploratory. Today’s assets that are being exploited were yesterday’s exploratory, high-risk asset redeployment and reconfiguration. Two learning trajectory effects come into play here: (1) learning to exploit a specific technology, and, (2) learning that exploitation is a productive strategy, independent of the specific technology. Both are based on the fact that “Humans tend to overexploit ‘good’ actions that pay off well early.” (Arthur 1994, p. 152). This leads to two competency traps that encircle strategic marketing management. The first-order competency trap is associated

31

with exploiting the initial economies of scale of the manufacturing and marketing assets employed. This has been recognized by a number of strategists as localized learning effects made even more inflexible by the specificity of the asset (Teece 1988; Henderson and Clark 1990; March 1991; Cohen and Levinthal 1990; Teece et al. 1997).

What we call the second-order competency trap is independent of the nature of the asset or technology and is associated with the actual learning capability dynamic. If managers have not experienced several exploration-exploitation cycles and their predominant experience is in executing an exploitation strategy then a firm may fixate its learning on exploitation: a strategy of deploying and reconfiguring assets in ways that create small localized improvements that give a positive trajectory to the assets’ initial economies of scale. Managers come to believe in such exploitative learning because that is what most of them spend their time doing and the reinforcement is frequent, proximate and predictable (Brown and Eisenhardt 1998). At the level of the firm, the result is that the accumulated learning around historical asset positions advances through learning-by-doing but within a small range constrained by the higher-order selfreinforcing belief in “exploitation” learning. Exploitation is continuously pursued because the alternative, competency-destroying, asset-shedding “exploration” is seldom ever tried and experienced. This brings us to a more general discussion of learning capability dynamics.

Learning Capability Dynamics The evolution of a preference for more localized exploitative learning over the initial exploratory learning of the firm raises some interesting questions about how the learning ability dynamic

32

capabilities of an organization increases over time. Cohen and Levinthal (1990) define this cumulative learning ability as absorptive capacity: “…the ability of a firm to recognize the value of new, external information, assimilate it, and apply it to commercial ends is critical to its innovative capabilities. We label this capability a firm's absorptive capacity and we suggest that it is largely a function of the firm's level of prior related knowledge.” (p. 128)

One computer company has earned billions from appropriating the economic rents from such a feedback dynamic on the demand side of the PC market. Dell was a late entrant into the PC market but its founder recognized that experienced PC shoppers can be taught and sold new features and technology much more easily and faster by phone or by mail catalog. Such consumers are faster learners because of their accumulated learning. This means that Dell can sell its latest models directly to such experienced users more easily and faster than rivals who market through traditional distribution channels. Dell can also ship its new models faster to the market thus gaining a reputation for having the latest technology because it gains a presence in the market with the latest technology faster than its rivals. In military strategy this is known as getting there the fastest with the mostest. As the new technology generates higher profit margins because it competes on new features, rather than price, Dell wins four ways: •

more effective target marketing to a growing segment of experienced PC buyers,

•

lower marketing and distribution costs,

•

a growing reputation as a leader in deliverying new technology, and

•

higher margins.

33

The general conclusion based on this consumer learning dynamic is that a late entrant in a market can use the Internet to target experienced consumers in markets where consumer cumulative learning curves exist. This may be more productive than targeting new entrants into the market. It is non-intuitive because the conventional wisdom is that experienced consumers are expected to be brand loyal – and harder to sell to than novices.

But a deeper analysis of the cumulative effect of learning on further learning reveals that knowledge acquisition may be subject to at least three underlying sub-dynamics. The first dynamic is that the quality of learning increases with learning because a deep and broad knowledge base provides a greater capacity for absorbing new knowledge: i.e., the absorptive capacity effect. For example, a passionate embracing of the Demming or Juran n-step quality improvement methods can feed on itself leading to a deeper knowledge base.

The second dynamic is that the more often a firm

replicates the assimilation and application of new information, the better it gets at doing such processes.5 An example of this second dynamic, observed by Teece, Pisano and Shuen (1997), is that the ability to reconfigure and transform is itself a learned by doing skill: “The more frequently practiced, the easier accomplished.” (p. 521). “Practice” makes learning and transformation processes both more “perfect” and easier. Extending this insight, the more that reconfiguration and transformation is observed and benchmarked in other firms (that is, learning by watching imitation), the easier and better it is accomplished. The third dynamic comes from the fixed cost component of the knowledge asset. The more that cumulative learning processes share the same fixed costs such as research and development testing equipment and the overheads involved in managing research and development, then the cost per unit of learning and absorbed knowledge declines with

34

increases in the production of knowledge and learning. An example of the third dynamic is that an investment in basic research imparts a dynamic to exploratory research because investment in basic research is an asset that is more involved in exploratory research than more localized, learning-bydoing exploitation research and learning-by-watching, imitation research.

The conclusion of March and other organization learning scholars (March 1991) is that the feedback effects associated with exploitative learning dominate over exploration to such an extent that the functional relation between accumulated knowledge and new knowledge absorbed exhibits decreasing returns to knowledge accumulated (i.e., accumulated knowledge makes the firm blind to new technology). This is at odds with the constant or increasing returns to scale suggested by Cohen and Levinthal’s (1990) theoretical explanation (the greater the capacity to absorb knowledge, the more that knowledge will be absorbed).

Both predictions can be

reconciled and accommodated within the above higher-order learning dynamics framework (and by assuming that feedback effects follow an S curve or path). The implications of such debate about which learning feedback effects dominate a firm’s dynamic capabilities, and at what time, are profoundly strategic. But these implications and proposed solutions have to also be placed within the broader context of within and between firm surveillance dynamics and learning network dynamics. The learning dynamics within and between firms are intertwined and can become quite complex.

35

Surveillance Dynamics Teece, Pisano and Shuen (1997) recommend a panoramic surveillance of the market to identify emerging new technologies and new and different best practices that can be used in the internal and external transformation of the organization. It is a way of breaking out of a specific learning trajectory trap. But this dynamic capability can itself be subject to a feedback dynamic between investment in what we will generally call market intelligence scanning and the gathering and the use of the market intelligence (see Figure 3). Glazer (1991) alludes to such a feedback loop in describing knowledge intensive firm's thirst for even more knowledge.

The latent system in Figure 3 can be triggered by a sudden unexpected change in the market that raises actual and perceived market turbulence and uncertainty. This increases the perceived benefit of market and technology surveillance that may reduce the increased uncertainty. Increasing the perceived benefit of surveillance leads to more market intelligence gathering and use. Increasing the quality and amount of intelligence gathering and use raises the quality of decision making which will, through correct attribution, lead to an increase in the perceived benefit of market surveillance. But increasing the use and quality of market intelligence will also raise the decision-maker's knowledge of the market. Understanding more about the market can lead to an increase in perceived market turbulence (i.e., ignorance was bliss) which further increases the perceived value of even more surveillance. This dynamic, suggested by the elementary principles of cost/benefit analysis applied to the economics of information, might also be triggered by an innovation in decision support software that leads to a greater use of market intelligence and, hence, increases its perceived value.

36

Conversely, this feedback system can be triggered to spiral downward in a vicious circle, reinforcing an “insider view.” Reduction in the use of surveillance intelligence in firm decision making leads to a lowering in the perceived value of market surveillance. Decision makers become less aware of emergent technologies and important changes in market supply (new products and best practices) and demand (consumer preferences and product use). The market is perceived to be more tranquil and in even less need of study.

Network Learning Dynamics Within-firm network learning dynamics. Learning network effects within organizations increase the utility of an innovation as it is shared around the organization because it increases the ability of the organization to seize the opportunity: “Although a subsystem of the organism can obviously perceive what needs to be done without the whole system being able to respond, can we say the whole system – the whole organization – has perceived what needs to be done if it cannot respond effectively to what the subsystem knows?” (Langlois 1997, p. 78).

New information is also likely to be integrated with other information that increases the value of both units of information. The existence of a communal learning network is a knowledge asset that enables the rapid transmission of new knowledge from within and outside the organization: “…technological knowledge is both socially constructed and enacted over long periods of time. With respect to the former, the socially constructed nature of technology means that technological change occurs within communities of practitioners who share standards of evaluation, technology recipes, and skills (Bijker, Hughes & Pinch, 1987)…The rationality of a technical decision is dependent upon and embedded within the community’s collective frame of reference…within a community of actors whose collective beliefs and practices provide the grist for determining what is useful and what is not.” (Porac 1997, p. 130-1)

37

A within-firm intelligence generating system that provides information about customer preferences is a learning capability (Slater and Narver 2000) but it must be accompanied by a knowledge network through which the information is disseminated. But such a network can also direct and limit the nature of the learning by reinforcing asset specific and type-of-learning specific feedback effects. Furthermore, such social networks within firms may encourage and foster an “inside” view of future outcomes which bases the likely success of a strategy on the amount of detailed planning and sequences of actions undertaken by the network in developing the strategy rather than an outside view that benchmarks the strategy against the likely strategies of the best competitors or compares the strategy to past successes and failures (Kahneman and Lovallo 1993). Confirmation of the “inside” social construction of reality across an internal network can be very reassuring and reinforcing, particularly if the inside view emphasizes the value of the firm’s current assets and knowledge capabilities (Brown and Eisenhardt 1998). But in such situations, the network may no longer be an efficient way of communicating new information or innovative ideas that challenge the received representation of the situation, the received wisdom of organization capabilities, and over optimistic forecasts. The collective frame of reference, the collective wisdom and collective organization memory produced by past motivation and learning feedback effects are very much part of the path dependency produced by pursuing and exploiting a particular technology trajectory. To ask the collective network to adopt a “competence-destroying” innovation is to ask the network to destroy the very dynamics and shared learning that made the network successful. Essentially an initial virtuous learning network can become a viciously resistant learning network.

38

Between-firm network learning dynamics. Communities of innovation created with rivals and with members of a supply-chain can accelerate the diffusion of more efficient supply side practices across a market (Alchian 1950; Haunschild and Miner 1997).

Learning networks across

organizations increase the potential of learning by watching effects as well as direct training effects. Depending on the openness, cooperation and coordination of the network (characteristics which are themselves self-reinforcing), the collective breadth and depth of the network can deliver absorptive capacity effects. The collective learning of the network may be subject to less localized learning to the extent that the collective learning is more than the sum of the parts: the parts being the localized learning trajectories of its members. The collective surveillance of the network may also be a great deal broader than any individual firm’s surveillance and, serving as an intelligence generating network, creates a virtuous surveillance feedback effect that increases the dynamic capabilities of the individual firm.

However, there is also a downside to cross-firm learning feedback effects. Kotabe and Swan (1995) found that a supply chain learning network can also exhibit localized learning around existing technology that reduces innovation. This is because each player’s investment and learning in existing technology along the added value supply chain reinforces and amplifies other member’s investment and learning in the existing technology standard. A similar effect is reported by Bower and Christensen (1995) who observed that companies that stay close to their existing customers tacitly create learning networks that exclude learning about new customers who are contemplating using new technologies. Their solution is to create a separate, independent organization that creates a separate, independent learning network with new customers. But this requires that the firm’s

39

market scanning capabilities have not been subject to feedback effects that blind the firm to the very need to create such new learning networks. Stuart et al. (1998) report on a case study of a learning network that somewhat avoided being trapped by its product specific competence by focusing on learning about systems and processes that were not product or customer specific (such as learning about document imaging, Internet access, general benchmarking processes). This enabled the members of the network to leverage each other’s learning without the resulting network learning effects trapping them strategically.

End-user learning-network dynamics At the end of the supply chain of some product-markets exist social networks amongst end users that can create powerful learning dynamics (Rogers 1995; Grannovetter and Soong 1986; Arthur and Lane 1994; Winsor 1995). After their first weekend, new motion picture releases live or die by the power of such recommendation network effects which vary in influence between teenagers, adults and type of movie.6 Note that there may also be a shared utility dynamic present with such entertainment. Part of the utility derived from seeing a movie is talking about it with friends. The more a movie is seen by friends and talked about, the greater the utility of seeing the movie so that you can participate in the discussion.

Almost all products and services are also affected by a training-network positive feedback effect. Consumers do more than spread the word about the superiority of one product over another, they also help their friends to learn to use and maintain the new technology. Thus, one of Microsoft's advantages is that its software is more popular to businesses as the software becomes ever more successful because existing workers, familiar with the software can help train new employees to use

40

the software. A driver of long-standing consumption trends is the intergenerational (cultural) learning that occurs between networks of end consumers. When Singer first marketed its home sewing machines it had to create a company owned distribution network whose primary service was to teach women how to use the new machine. Within a decade or so sister was teaching sister and mother was teaching daughter how to sew. This created an endogenous consumer learning network effect that enabled Singer to dominate the market and to shift from the heavy investment in a specialized training-distribution network to a mass-merchandise distribution strategy, with little or no product training needed or provided. Today, the inexorable decline in the sales of home sewing machines is feeding on itself as ever fewer skilled mothers are teaching their daughters how to sew.

Routine/Rule Dynamics The final class of learning feedback effects that need to be understood in strategic management are those produced when resource allocation decisions are on automatic pilot, impervious to any learning effects either progressive or regressive. They are top-down prescribed codified decision rules or standard operating procedures on the supply side and brand loyalty on the demand side. Examples of such seller routines are inventory replenishment rules, retail shelf-space allocation rules, cost plus pricing (as described above), using commissions to reward salespeople, advertising budgeting rules and at the strategic decision-making level, research and development spending rules and buy/sell investment rules. Such routines can create self-reinforcing feedback effects that lead to the spiraling up and down of the competitiveness and wealth creation of the enterprise (Forrester 1961; Nelson and Winter 1982; Arthur 1994; Farris, Verbeke and Dickson 1998). For example as sales increase, expectations of future sales increases rise. When future advertising budgets and sales

41

commissions are tied by allocation rules or routines to such expectations then marketing effort increases and, ceteris paribus, sales will increase. Buy and sell investment routines and rules based on stock market price increases and decreases create similar self-fulfilling effects. At the micro level of consumer behavior, a common habitual loyalty feedback effect occurs when the likelihood of purchasing brand A increases with each past purchase of brand A. These routines represent a stasis in decision making and learning: a lack of adaptation, innovation or learning. The paradox is that because of their very stasis nature they can create powerful feedback dynamics alone and interacting with positioning and other learning dynamics.

The Use of Feedback Analysis in Marketing Strategy

Strategic management has as its goal the earning of a flow of economic rents (Schoemaker 1990). Combining the above feedback framing of competitive dynamics with such a view of strategy leads us to the following fundamental question: how can a firm appropriate the economic rents generated by market feedback effects? The extant literature recommends two basic strategies: exercising intellectual property rights and simply employing a get-ahead and stay-ahead strategy of deploying technological innovation. Almost always the first step is to attempt to establish intellectual property rights over technology and processes through patents and copyright. A second strategy, that is the focus of our conceptualizing, is for a firm to rely on the difficulty that other firms may have in imitating the entrepreneurial firm’s dynamic capabilities (Teece, Pisano and Shuen 1997). From our perspective, the dynamic capabilities of a firm are its unique asset and learning capabilities that have given it a head start down the learning and experience curves created by harnessing market

42

feedback effects. In discussing the causes of enduring market leadership, Tellis and Golder (1996; 2001) provide strong support for the theory that chance, if it plays any role, favors the firm with the drive to make the mass market through sustained product and process innovation and investment in assets and the leveraging of existing assets such as R&D, manufacturing plant, brand equity (customer loyalty networks), and distribution network plant and goodwill.

In effect the firm that develops a sustainable advantage, whether leader or follower, is the firm that is able to exploit crucial supply and demand feedback effects described above (see also Levin et al. 1984; Teece 1988). But to be able to deliberately (c.f. fortuitously) take advantage of such dynamic capabilities, senior management must see and understand such dynamic capabilities in the context of their competitive situation. This requires more than a taxonomic knowledge of market feedback effects.

Market Feedback Matrix Analysis Table 2 presents an example of how market feedback analysis might be used to better understand the strategic position of Amazon.com and Barnes and Noble in the Internet market for books.7 The shading of the cells represents our assessment of positive feedback potential and the comparative dynamic capability of the competing firms to exploit such feedback effects is writ large or small by the superimposed initials. For example the top middle cell is deeply shaded because in this market firms can make a major investment in distribution fixed costs that has the potential of generating a strong asset position feedback effect. This involves warehousing facilities, information systems and automated picking and packing. Granted, there is a large

43

delivery variable cost paid to companies such as USPS, UPS, and FedEx but all players face this cost which neutralizes it as a factor. Barnes and Noble initials are writ large (in the investmentdistribution cell in Table 2) because it is our assessment that it has a substantial advantage. It can build on its existing distribution system serving its network of several hundred superstores (including the potentially significant service advantage of local store, fast pick-up).

Our

assessment, therefore, is that B&N also possesses a greater dynamic capability in terms of experience learning and learning ability but its continuance is somewhat conditional on the formation of effective learning networks between the B&N Internet company, the parent bookstore organization, and the stores and their managers.

If Amazon were to develop a

fulfillment strategic partnership with a major retailer such as Wal-Mart then B&N’s distribution advantage might be greatly reduced.

Amazon possesses some positioning advantages in its e-marketing operations. First, it possesses an advantage in its brand equity (see the brand equity-buyer behavior cell). Amazon.com has become one of the best known brand names in the world, in less than five years, largely because of an avalanche of positive free publicity (that itself has exhibited an increasing returns dynamic by virtue of communication networks in the media). While Amazon has tremendous brand recognition, the depth of its brand equity might be questioned because its observed repeat business brand loyalty is weaker than expected (Hof et al. 2000, p. 39). From the feedback perspective, each shopping experience is not producing increasing affect and loyalty that translates into geometrical increases in sales to same customers over time. The consumer network learning-about-Amazon effect may also have lost some of its energy. The expansion of Amazon.com into other product categories is not a

44

good signal of long-term book retailing growth and viability. The current strength of Amazon’s brand equity is also partially moderated by the strength of the Barnes and Noble brand name amongst heavy buyers of books. The financial markets' belief in Amazon's brand equity advantage has faltered in part because its brand equity is being sustained by heavy advertising on carriers such as AOL. If such advertising continues to be needed to sustain sales then this will be further evidence that Amazon’s brand equity (brand asset position) is not as strong as many have supposed.

Amazon’s asset position advantage in other aspects of e-marketing (e.g., site and support service) is less significant because they can be readily matched and the variable operating costs are quite high. Because of ease of imitation neither firm is likely to possess a strong advantage in e-marketing experience and learning ability feedback effects over the medium and longer term. For example, Amazon’s ability to service customers by suggesting books based on past purchases has been much publicized but this learning ability can be matched by Barnes and Noble purchasing and using the same software. By deploying this earlier, Amazon will have a longer history of customer purchases to “deploy,” but this advantage will depend on how rapidly purchase history information becomes obsolete.

Does Amazon have a strong positive feedback advantage in the motivation of its e-marketing employees driven by generous stock options whose value has soared? This would be a feedback effect similar to that presented in Figure 2. We have two reasons to question this learning motivation advantage.

When the speculative bubble that drove up Amazon stock burst, the

virtuous effect on employee motivation may now have turned into a vicious effect. New employee

45

envy of old employee millionaires must also be factored into the dynamics of this capability. Recent layoffs also are likely to have taken a lot of motivation momentum out of the firm. This feedback advantage is also only an important dynamic capability to the extent that employee motivation makes a significant difference to process output and efficiency in this market. One of the "dynamic capability" advantages Barnes and Noble may possess in e-marketing and distribution operations is in surveillance. Playing the role of market follower and with its extensive network of store managers and contacts in the book trade, it should be more alert to sensing change in the entire book market. Our scenario is that Barnes and Noble is obtaining increasing returns from such surveillance and is ahead and will stay ahead of Amazon on this dynamic capability.

Buyer investment positive feedback effects in this market may be deceptively weak. While a buyer’s investment in a PC is quite high, it is not book buying specific. Book-buyer learning-bydoing and learning ability feedback effects are also not great compared with say the computer game market. The presence of a buyer routine feedback effect is the wild card. If book-buying behavior becomes routinized then Amazon stands to gain a large first-mover advantage. But buyers are not buying the same book time-after-time as they do with say laundry detergent. The future deployment by buyers of intelligent agents that search for the lowest delivered price of a desired book would also destroy such routinized buying behavior and make distribution dynamic capabilities paramount.

Some of the above scenario dynamics are quite speculative. Senior management in both Amazon and Barnes and Noble would likely take issue (perhaps strong issue) with several of our conclusions, and have empirical evidence to support their claims of competitive advantage. Our

46

point is not to defend our view of the competitive dynamics, but rather to illustrate how Table 2 provides a systematic foundation for surfacing important feedback effects and developing several evolutionary scenarios that can be tested against emerging empirical evidence.

To further

demonstrate the versatility and generalizability of the deeper insights revealed by our market feedback matrix framework, we apply it to three alternative extant frameworks for strategizing in conditions of market uncertainty: I. Courtney, Kirkland and Viguerie (1997) have proposed four types of uncertainty scenarios: 1.

Low uncertainty where the dynamics of the market are highly predictable and lead to a fairly clear view of the future.

2.

A future where the market could take several different evolutionary paths but the likelihoods of the events or triggers of a specific path can be estimated.

3.

A future where a whole range of different paths could be taken.

4.

True ambiguity where the evolution of the market is almost impossible to predict (e.g., post 1992 Russia, multimedia markets).

From a feedback theory perspective, the ability to specify a market feedback matrix as presented in Table 2 is greatly reduced as one progresses through Courtney et al.'s uncertainty scenarios. But we would argue that even in the "true ambiguity" scenario, a manager should be able to predict the likely power of different feedback effects presented in Table 1 in an emerging multimedia market. Once some trends start to emerge in the market, a system's thinker's ability to forecast the future may be reasonably good.

II.

In a recent review of scenario planning, Schoemaker (1995) suggested listing each trend in an influence diagram.

We propose that the positive feedback effects that are driving such trends in an industry and in individual firm supply and demand curves be specified in a market feedback matrix. In his example of scenario planning for an advertising agency this would include: 1.

Interpreting globalization in the industry in terms of a fixed cost/overhead positive feedback effect.

47 2.

Greater investment by agencies in integrated marketing management overheads that create new positive feedback effects.

3.

A decline in the power of brand equity feedback effects, including the brand equity of agencies.

4.

Learning ability and learning-by-doing positive feedback effects associated with the new technologies such as the Internet and cable communication channels.

5.

Learning network feedback effects between alliances of specialized advertising service firms that take advantage of economies of scale and learning feedback effects.

6.

Alliances with information companies that result in new surveillance and learning network feedback effects.

7.

A decline in the specific investment of customers in agency relationships that is reducing switching costs and changing fee compensation conventions.

8.

Learning networks within and between buyers of ad agency services about how to change their relationships and operating procedures with their agencies.

A strategic question that arises from such an analysis, not identified by Schoemaker’s scenario analysis, is whether agencies will learn faster than their customers how to manage globally integrated marketing communications using the new technologies. If they do not, the investment positive feedback expected from the global mergers that have occurred in the industry will have a vicious rather than virtuous effect on their profitability and survival.

III. The famous Boston Consulting Group (1972) matrix of stars, cows, dogs, and question marks can also be reevaluated in terms of feedback dynamics. The logic behind the relative share dimension of the matrix was derived from an early form of feedback theory related to learning curves. Companies with the largest shares were deemed to be further down the learning curve because of accumulated experience. Thus, relative share was a surrogate for the cost advantage enjoyed by some companies versus others. The market feedback matrix recognizes that the source of one firm's current and future advantage over another can come from several cost positioning and learning advantages that is not revealed by simply comparing market share. The market growth dimension of the BCG matrix is an indicator of the stage of the product life-cycle occupied by the industry or market. In later stages of the product life cycle the strategic window of opportunity closes and companies who are "behind" have little chance to "catch up." The market feedback

48 matrix is much more diagnostic in that it attempts to identify what is driving market change and what firms are in the best position to convert such market forces into advantage and economic rents.

Visualizing Non-Linear Dynamics Figure 4 is a visual aid that can be used as a general complement to Table 2 to help senior executives visualize the non-linear dynamics that feedback effects can create and identify the feedback effects to be included in any market feedback matrix analysis. It identifies and explains the interacting dynamic coupling of two of the major feedback effects that propelled Microsoft to its current dominance. Metaphorically, Figure 4 describes the economic "engine" that created the Window's path dependency. The skilled manager has to think intuitively about such dynamics. As the cause and effect chain flows through the system, both the outcomes and the relationships between the outcomes change. This process is often called non-linearity. Visually it is akin to circling around Figure 4 several times - this creates the non-linear effect as each cycling accumulates the effect of the initial cycle.

A very strong demand network positive feedback effect existed because the more that Windows 95 was used, the more people were available to train and help new users use it most effectively in the office, school and small business.8 The more Windows 95 was used, the more and faster its capabilities were tested, defects removed and new features improved at users' suggestions. If Windows 98, NT and 2000, the latest Windows versions, also standardize and simplify communication between networks of PCs and between each PC and the Web, then usage utility for each user will increase with each addition of an individual on the network (a network positioning dynamic). Shapiro and Varian (1999) assert that such “network externalities make it virtually

49

impossible for a small network to survive” (p. 184). In fact, they can. Apple’s network positioning and learning network feedback effects in the industrial design and advertising industries (segments of the PC operating system market) enabled it to survive and flourish against Wintel’s general dominance. This demonstrates an oft overlooked fact about market dynamics: market feedback effects can be market-segment specific.

A major supply cost structure positive feedback effect has also greatly benefited Microsoft. The inventing, designing, making and mass marketing of PC software has a very high fixed-cost component. A lot of the cost of each unit of software sold is fixed cost because the per unit variable cost of pre-installed software on PCs is less than a dollar. As Microsoft sold more of its Windows 95 software, its average cost of each unit sold dropped. A further cost of supply effect was generated by third-party suppliers of software who have invested much more in R&D that made their software work on Windows 95 than other systems,9

thus increasing Microsoft's innovation

and cost structure advantage. The general cost structure effect is captured in Figure 4 by the arrow that indicates, as supply increases then the cost per unit of quality decreases, which can also be conceived of as costs/utility, if quality is measured in units or part-worths of utility. The next arrow says that as cost per unit of quality decreases, price per unit of quality decreases. How strong this relationship is depends on the competitiveness of the market, and the general price aggressiveness of Microsoft. To date it has proven to be quite strong, although not strong enough for some critics.

The evolution of Microsoft in the PC software as described in Figure 4 is not atypical. Coupled supply-demand feedback effects similar to those described in Figure 4 have also helped the German

50

software company SAP dominate the world-wide niche market for manufacturing software (Cook 1997). Cisco Systems has used supply and demand feedback effects to dominate the market for routers that create networks of private Intranets. But this company is vulnerable to Internet-based networks of Intranets (Hutheesing and Young 1996) and the evolution of Windows NT (Kirkpatrick 1997). This helps explain why Cisco has entered into a strategic alliance with Microsoft to develop NT networking technology - an alliance that creates a powerful combination (Reinhardt 1997) of "rent earning" positive feedback advantages.

CONCLUSIONS

Understanding the various feedback effects that dominate market evolution and development can be useful for managers in at least three ways. First, this understanding may help forecast the different directions in which markets are likely to evolve. Second, understanding the complexity of feedback loops helps managers appreciate and illuminate the nature and degree of uncertainty and risk in a particular market's evolutionary path. The degree of uncertainty, in turn, drives the appropriate strategic stance and options. When uncertainty is greatest, option theory and contingent analyses may be critical. Third, feedback theory and diagrams may reveal "latent" feedback loops that marketing managers can influence and foster.

Such analysis and deep strategic thinking may

alternatively reveal that the direction of the market and the firm is set and the outcome very difficult to change. Companies in this situation are either on the train or on the track.

51

The conjecture of management about what is likely to occur in the future in a particular market and the development and sustainability of a firm's competitive advantage requires thinking about the several, often interacting, feedback effects that are currently driving the complex dynamic systems we call "markets." According to Hamel (1998), what is needed is a deep theory of strategy creation based on using deep rules that explain the order in complex dynamic systems. The rules and principles that underlie the positive feedback effects we have identified and how they can be combined and visualized in practice to explain the order in the evolution of a market and comparative firm success constitute a step (albeit a small step) toward such a “deep” theory. Evans (2000), claims that “While Chaos theory offers a wonderful heuristic that illuminates the nature of complexity, it offers precious little help in shaping it.” We have attempted a modest step in the direction of classifying feedback effects that underlay the emergence of order from chaos for the strategic thinkers who cannot duck this task. Managers should find it a lot easier to understand how to use complex system’s theory in their strategizing, scenario planning and risk analysis when the types of generic feedback effects that are likely to occur in a market are categorized and the reasons why they develop are explained. Their cumulative power also becomes transparently obvious when such feedback effects are identified in figures that explain how they amplify each other. It is easy to see these effects in the history of the VCR and PC operating system software markets because they are pronounced. However, insightful executives who have worked for some time in any industry sector or product-market will be able to readily identify the pertinent feedback effects that have driven the evolution of their markets. Have there been cost structure effects, learning effects, network effects and at what stages of the supply chain have these effects dominated? By answering these questions in reconstructing the historic evolution of markets, company strategists will become

52

more skilled at identifying what feedback dynamics are driving the market now and should be included in market feedback matrix analyses.

Firms led by senior management and marketing managers who can intuitively “see” a market’s latent potential from the perspective of Table 2 and Figure 4, are endowed with a private information advantage over the market. The basis for their insights, instincts and intuition is a superior heuristic or frame (Amit and Schoemaker 1993). Successful financial analysts possess a similar skill. In a sense, they can see the direction of uncertainty.

In a market system where

powerful demand and supply effects are latent or manifestly emerging an investment in supply can generate huge cumulative returns. Such powerful feedback effects have made Bill Gates, by far, the richest man in the world. Strategic managers and investors with such thinking skills are able to acquire firm’s whose current managers and owners do not realize the potential position and learning feedback effects in their market that can be exploited. They also appreciate the interplay between innovation and imitation that places limits on their ability to appropriate the returns from a market’s system of feedback effects (Nelson and Winter 1982; Brown and Eisenhardt 1998; Dickson 1992). They redeploy the resources of the firm to match the requirements of a changing environment by understanding how supply, demand and competitive feedback effects are changing the environment. They change technology paths because they understand how supply and demand feedback effects have been played out and can be played out (see Levinthal and Myatt 1994; Bower and Christensen 1995).

53

Schoemaker (1995) describes scenario planning as organizing and interpreting complex dynamic systems "into narratives that are easy to grasp." He describes how students and managers can hold contradictory beliefs about market trends and that these "incoherent beliefs need to be adjusted." Another way of saying this is they do not understand the ying and yang of crucial positive and negative feedback effects. Simple maps of feedback effects are visual narratives that not only adjust beliefs but explain the circling cause and effect sequence that can then become a theory-in-use (Zaltman et al. 1982) in the mental model of the manager. The development of feedback matrices and circles such as illustrated in this paper can contribute to a consensus understanding of such theories in use. They are what de Geus (1997) calls "soft mapping" techniques to resolve arguments about the feedback structure of the market, feedback effects within the firm and the vision/mission/strategy of the firm.

Future Research Directions Future behavioral decision theory (BDT) research might be productively directed at a fundamental bias in intuitive thinking about feedback effects. It is the "conjunction fallacy" (Tversky and Kahneman 1982; Schoemaker 1995) where explaining a sequence of cause-and-effect relationships can raise the perceived probability of some outcomes and lead to overemphasis on a specific feedback effect in scenario planning. A framework such as offered in Table 2 lists a range of types of supply and demand feedback effects selected from Table 1 that may be active in the market, thus encouraging a more balanced, integrative assessment of their current force and the extent they pose an opportunity/threat to the firm. This reduces management preoccupation with the feedback effects that are easiest to understand and most available (Tversky and Kahneman 1982). The way

54

that managers and students learn to identify, understand and build feedback matrices and maps reveals the depth, breadth and complexity of their systems thinking (Senge 1990). This topic has a great deal of potential as a stream of research that cognitive scientists and behavior decision theorists in business might pursue (Kahneman and Tversky 1982).

Indeed, systems thinking

exercises can be used to identify the managers who most quickly understand the analytics behind the decision processes and the cumulative effects of system dynamics.

All other things being

equal, how rapidly a manager or financial analyst learns to understand and use feedback maps and simulations (including suggesting insightful feedback effects that might be added) would seem to be a useful measure of the manager's understanding of feedback dynamics, conjecturing capability and dynamic strategic thinking (DST) aptitude.

55

Endnotes 1

This is a composite fraction reflecting different asset depreciation routines that have become “standardized.”

2

It actually involved the coupling of a cost structure and a learning routinization effect discussed in a later section.

3

Such an emotional contagion feedback effect was the claimed cause of the disastrous Treaty of Versailles that extracted vengeance and was quite insensitive to its long-term consequences: “The peace conference was not given a deliberate structure. It just happened, acquiring a shape and momentum of its own, and developing an increasingly anti-German pattern in the process, both in substance and, equally important, in form.” (Johnson 1992, p. 25).

4

For an exception see Argote (1993) who distinguishes between individual, group, organizational and system learning curves.

5

This is a generalized restatement of the earlier described outer competency trap.

6

Word-of-mouth positive feedback underlies the effects of chains of negative rumors that damage reputation (see Penchant & Mitroff 1992; Sullivan 1990). This reminds us that the adjective "positive" describes the direction of feedback (positive or negative) rather than whether it is normatively good for the firm or the market.

7

We prefer to provide an example that involves ex-ante analysis, rather than ex-post analysis (e.g., the history of QWERTY or VHSvBeta) because of its greater relevance to future-oriented scenario planning.

8

The copying of software between friends further fed the consumer demand, network effect that hurt Microsoft sales in the short-term but helped it in the long-term.

9

This 95/98 investment advantage works against Microsoft to the extent it is not transferable to new platforms (e.g., NT and Windows 2000). If so Microsoft may increasingly struggle against the effects of its past asset positions.

56

TABLE 1

Types of Feedback Effects Along the Supply-Chain from Raw Material Extraction to Ultimate End User

Asset Position Feedback Dynamics Research and Development Investment Position Economies of Scale Manufacturing Cost Structure Position Economies of Scale Distribution Investment Position Economies of Scale Brand Equity Position Economies of Scale Network Asset Position Utilities of Scale and Economies of Scale

Higher-Order Learning Feedback Dynamics Motivation/Drive/Contagion dynamics Learning-by-doing experience dynamics Learning-to-learn ability dynamics Surveillance capability dynamics Routines/Rules dynamics

____________________________________________________________________

57

References

Alberts, F.W., (1989), “The Experience Curve Doctrine Reconsidered,” Journal of Marketing (July ), pp.36-49.

Alchian, A.E. (1950). “Uncertainty, evolution and economic theory,” Journal of Political Economy, 58 (3), pp. 211-21.

Amit, R. and P.J.H. Schoemaker (1993). “Strategic assets and organizational rent,” Strategic Management Journal, 14, pp. 33-46.

Anderson, P., K. Arrow, and D. Pines (1988). The Economy as an Evolving Complex System, Redwood City, California, Addison-Wesley Publishing Company, Inc.

Argote, Linda (1993), "Group and Organizational Learning Curves: Individual, System and Environmental Components," British Journal of Social Psychology, 32,pp. 31-51.

___________ (1999),

Organizational learning: creating, retaining and transferring

knowledge, Boston, MA.: Kluwer Academic Publishers.

Arrow, K. (1962). “The economic implications of learning-by-doing.” Review of Economic Studies, 29, pp. 166-170.

58

Arthur, W. B. (ed.) (1994). Increasing Returns and Path Dependence in the Economy, Ann Arbor, MI: The University of Michigan Press, pp. 69-97.

Arthur, W. B. and D. A. Lane (1994). “Information contagion,” in Arthur, B. (ed.) Increasing Returns and Path Dependence in the Economy, Ann Arbor, MI: The University of Michigan Press, pp. 69-97.

Axelrod, R. (1984). The Evolution of Cooperation, New York: Basic Books.

Barnett, William P. and Morten T. Hansen (1996), “The red queen in organizational evolution,” Strategic Management Journal, 17, 139-157.

Beinhocker, Eric D. (1999), “Robust adaptive strategies,” Sloan Management Review, Spring, pp. 95-106.

Bijker, W., T. P. Hughes, T.P., and T. Pinch (1987). The Social Construction of Technological Systems, Cambridge, Mass: MIT Press.

Boston Consulting Group (1972), Perspectives on Experience. Boston, MA: Boston Consulting Group

59

Bower, J. L. and C. M. Christensen (1995). “Disruptive technologies: Catching the wave,” Harvard Business Review, January-February, pp. 43-53.

Brown, C. (1996), “What's your favorite color?” Forbes, October 7, pp. 54-58.

Brown, S. L. and K. M. Eisenhardt (1998). Competing on The Edge, Cambridge, MA.: Harvard Business School Press.

Cassady, R., Jr. (1957). “The New York department store price war of 1951: A microeconomic analysis,” Journal of Marketing, 12.1 (July), pp. 3-11.

Chandler, Alfred D. (1990), “The Enduring Logic of Industrial Success,” Harvard Business Review, (March-April), 130-140.

Cohen, W. M. and D. A. Levinthal (1990). “Absorptive capacity: A new perspective on learning and innovation,” Administrative Science Quarterly, 35 (March), pp. 128-152.

Cook, W. J. (1997). “What makes IBM run?” U.S. News & World Report, June 9, pp. 48-49.

Courtney, Hugh, Jane Kirkland, and Patrick Viguerie (1997), "Strategy under uncertainty,” Harvard Business Review, (November-December), pp. 67-79.

60

de Geus, A.P. (1988), “Planning as Learning,” Harvard Business Review, (March-April),pp. 70-74.

de Geus, A.P. (1997), The Living Company: Habits for Survival in a Turbulent Business Environment, Boston, MA.: Harvard Business School Press.

Dickson, P.R. (1992). “Toward a general theory of competitive rationality,” Journal of Marketing, 56, (January), pp. 69-83.

Dickson, P.R. (1996). ‘”The static and dynamic mechanics of competition: A comment on Hunt and Morgan's comparative advantage theory,” Journal of Marketing, 60 (October), 102-106.

Dickson, P.R. (1997). Marketing Management, Fort Worth, Texas: The Dryden Press.

Dutton, J.H., A. Thomas and J.E. Butler (1984), “The History of Progress Functions as a Managerial Technol.ogy,” Business History Review, 58.2(Summer), pp. 204-233.

Evans, Philip (2000), “The End of the End Game,” Business Strategy, (21.6).

Farris, Paul W.., Willem Verbeke and Peter R. Dickson, (1998). “Path dependencies and the long-term effects of marketing decisions,” Marketing Letters, 9.3, pp. 247-268.

61

Farris, Paul W. and Pfeifer, Phillip, (2000) “All You Needed to Know About the New Economy You Learned from Playing Monopoly,” Darden School Working Paper.

Forrester, Jay W. (1961). Industrial Dynamics, New York: John Wiley & Sons, Inc.

Glazer, R. (1991),

“Marketing in an information-intensive environment:

Strategic

implications of knowledge as an asset,” Journal of Marketing, 55 (October 1991), pp. 1-19.

Grannovetter, M. and R. Soong, (1986), ‘”Threshold models of interpersonal effects in consumer cemand,” Journal of Economic Behavior and Organization, 7, pp. 83-99.

Gronroos, C. (1984).

Strategic Management and Marketing in the Service Sector,

Cambridge, MA: Marketing Science Institute.

Hamel, G. (1998). “Opinion: Strategy innovation and the quest for value,” Sloan Management Review, Winter, pp. 7-14.

Hatfield, E., J.T. Cacioppo and R.L. Rapson (1994). Emotional Contagion, New York: Cambridge University Press.

Haunschild, P. R. and A. S. Miner (1999). “Modes of interorganizational imitation: The effects of outcome salience and uncertainty,” Administrative Science Quarterly, forthcoming.

62

Hayek, Frederick A. (1978), New Studies in Philosophy, Politics, Economics and the History of Ideas, London: Routledge and Kegan Paul. Chapter 12.

Henderson, R. and K.B. Clark (1990). “Architectural innovation: The reconfiguration of existing product technologies and the failure of established firms,” Administrative Science Quarterly, 35, pp. 9-30.

Hindle, Tim (1998), Pocket Strategy: the Essentials of Business Strategy from A to Z, London, UK: The Economist Books.

Hirsch, W.Z. (1952). ‘”Manufacturing progress functions,” Review of Economics and Statistics, 34 (May), pp. 143-155.

Hof, Robert, Debra Sparks, Elllen Neuborne and Wendy Zellner (2000), "Can Amazon Make It?" Business Week, July 10, pp. 38-43.

Hutheesing, N. and J. Young (1996). “The curse of the market leader,” Forbes, July 29, pp. 78-81.

Jervis, Robert (1998). System Effects, Princeton, NJ: Princeton University Press. Johnson, P. (1992). Modern Times, New York: Harper Perennial.

63

Kahneman, D. and A. Tversky (1982). “The simulation heuristic,” in Daniel Kahneman, Paul Slovic and Amos Tversky (eds.) Judgement Under Uncertainty: Heuristics and Biases, Cambridge, UK: Cambridge University Press, pp. 201-208.

Kahneman, D. and Lovallo, D. (1993). “Timid choice and bold forecast: A cognitive perspective on risk taking,” Management Science, 39, pp. 1.

Kirkpatrick, D. (1997). “He wants ALL your business - and he's starting to get it,” Fortune, May 26, pp. 58-68.

Kotabe, M. and K.S. Swan (1995), ‘The role of strategic alliances in high-technology new product development,’ Strategic Management Journal, 16 (November), pp. 621-636.

Langlois, R. N. (1997), “Cognition and capabilities: Opportunities seized and missed in the history of the computer industry,” in Technological Innovation: Oversights and Foresights, R. Garud, P. R. Nayyar and Z. B. Shapira (eds.), New York, NY: Cambridge University Press, pp. 7194.

Levin, R., A. Klevorick, N. Nelson and S. Winter, (1984), “Survey research on R&D appropriability and technological Opportunity,” unpublished manuscript, Yale University.

64

Levinthal, D. and J. Myatt (1994). “Co-evolution of capabilities and industry: The evolution of mutual fund processing,” Strategic Management Journal, 15, pp. 45-62.

March, J. G. (1991). “Exploration and exploitation in organizational learning,” Organization Science, 2.1 (February).

Mintzberg, Henry and Joseph Lampel (1999), “Reflecting on the strategy process,” Sloan Management Review, Spring, pp. 21-30.

Morecroft, J. D.W. (1983), “System dynamics: Portraying bounded rationality,” Omega, 11, pp. 131-42.

Nelson, R. R. and S. G. Winter (1982). An Evolutionary Theory of Economic Change, Cambridge, MA: Harvard University Press.

Pascale, Richard T. (1999), “Surfing the edge of chaos,” Sloan Management Review, Spring, pp. 83-94.

Penchant, T.C. and I.L. Mitroff (1992). Transforming the Crisis-Prone Organization, San Francisco, CA: Jossey-Bass Publishers.

65

Porac, J. F. (1997). “Local rationality, global blunders, and the boundaries of technological choice: Lessons from IBM and DOS,” in Technological Innovation: Oversights and Foresights, R. Garud,

P. R. Nayyar and Z. B. Shapira (eds.), New York, NY: Cambridge University Press.

Reinhardt, A. (1997). “Bulking up for battle,” Business Week, July 7, pp. 47.

Rogers, E. M. (1995). Diffusion of Innovations, 4th ed., New York: The Free Press.

Romer, P. M. (1986). “Increasing returns and long-run growth,” Journal of Political Economy, 94 (October), pp. 1002-37.

Rosenberg, N. (1982). Inside the Black Box: Technology and Economics. Cambridge, UK: Cambridge University Press.

Schoemaker, P. J. H. (1990). “Strategy, complexity and economic rent,” Management Science, 36.10 (October), pp. 1178-1192.

Schoemaker, P. J. H. (1995). “Scenario planning: A tool for strategic thinking,” Sloan Management Review, Winter, pp. 25-39.

66

Shapiro, C. and H. R. Varian (1999). Information Rules, Cambridge, MA.: Harvard Business School Press.

Simon, H. A. (1984), “On the behavioral and rational foundations of economic dynamics,” Journal of Economic Behavior and Organization, 5, pp. 35-55.

Slater, Stanley F. and John C. Narver, (2000), “Intelligence Generation and Superior customer Value,” Journal of the Academy of Marketing Science (Winter), pp. 120-127.

Solow, R.M. (1994). “Perspectives on growth theory,” Journal of Economic Perspectives, 8 (Winter), pp. 45-54.

___________ (1997), Learning from "Learning-by-doing”, Stanford: CA: Stanford University Press.

Stacey, R. D. (1995), “The science of complexity: An alternative perspective for strategic change processes,” Strategic Management Journal, 16, pp. 477-495.

Stuart, I., P. Deckert, D. McCutcheon and R. Kunst (1998), “Case study: a leveraged learning network,” Sloan Management Review, Summer, pp. 81-93.

67

Sullivan, M, (1990), “Measuring image spillovers in umbrella-branded products,” Journal of Business, 63.3, pp. 309-329.

Teece, D. J. (1988), “Capturing value from technological innovation: Integration, strategic partnering, and licensing decisions,” Interfaces, 18.3 (May-June), pp. 46-61.

Teece, D. J., G. Pisano and A. Shuen (1997), “Dynamic capabilities and strategic management,” Strategic Management Journal, 18.7, pp. 509-533.

Tellis, G. and P. Golder, (1996), “First to market, first to fail? Real causes of enduring market leadership,” Sloan Management Review, 37.2 (Winter), pp. 65-75.

Tellis, G. and P. Golder, (2001), Will and Vision: How Latecomers Came to Dominate American Business, McGraw-Hill.

Tversky, A. and D. Kahneman, (1982). “Availability: A heuristic for judging frequency and probability,” in D. Kahneman, P. Slovic and A. Tversky (eds.). Judgement Under Uncertainty: Heuristics and Biases, Cambridge, UK: Cambridge University Press, pp. 177.

Urbany, J. E. and P. R. Dickson (1991), “The effects of price-cutting momentum and consumer search on price setting in the grocery market,” Marketing Letters, 2.4, 393-402.

68

Varadarajan, P. Rajan (1999), Strategy Content and Process Perspectives Revisited, Journal of the Academy of Marketing Science, (Winter) 27.1, pp. 88-100.

Varandarajan, P. Rajan and Satish Joyachandran, (1999), “Marketing Strategy: An Assessment of the State of the Field and Outlook,” Journal of the Academy of Marketing Science (Spring), pp. 120-143.

Winsor, R. D. (1995), “Marketing under conditions of chaos: Percolation metaphors and models,” Journal of Business Research, 34.3 (November), pp. 181-189.

Wright, T.P. (1936), “Factors affecting the cost of airplanes,” Journal of Aeronautical Science, 3 (February), pp. 122-128.

Zaltman, G., K. Lematers, and M. Heffring (1982). Theory Construction in Marketing: Some Thoughts on Thinking, John Wiley & Sons.

73

FIGURE 1 How Price Can Increase When Demand Falls Exogenous Pulse

-

Market Sales

Positive Negative

Price

Fixed Costs

P

Average Cost P

-

Use of a cost plus pricing rule as recommended by convention. The overall dynamic is driven by the coupling of this routinized rule application and a fixed cost investment feedback effect.

70

FIGURE 2 A Within Firm Motivation Feedback Effect Employee Service

Employee Morale

Demand

Employee Benefits/wages

Exogenous Pulse

71

FIGURE 3 A Within Firm Surveillance Feedback Effect Market Intelligence Gathering

Perceived Value of Market Intelligence

Decision support software innovation exogenous pulse

Market Intelligence Use Decision-Making Quality Perceived Market Turbulence

Market change exogenous pulse

72

FIGURE 4 Generic Market Feedback Effects

Usage Utility: an endogenous measure of system efficiency

Supply

-

Demand-side positive feedback effect

Costs/Quality: an endogenous measure of system efficiency

Demand Buyer innovation exogenous pulses

Supplier innovation exogenous pulses

-

Supply-side positive feedback effect Price/Quality: an endogenous measure of system efficiency