Competitive Positions and Consumer Choice: A Differentiated Products Model, Study notes of Technology

Download Competitive Positions and Consumer Choice: A Differentiated Products Model and more Study notes Technology in PDF only on Docsity! Competitive Positioning and Generic Strategies: Revisiting the ’Stuck in the Middle’proposition Ron Adner and Peter Zemsky Abstract We explorethe evolution of competitive positions using a formal model of competi- tion with differentiated products in which production technologies improve over time. We show how the interplay between improving technologies and consumers’valuation of the resulting performance improvements affects whether or not the classic generic strategies of Low Cost and Differentiation remain viable in the face of a rival taking an intermediate position between them. We thus clarify when intermediate positions leave a firm’stuck in the middle’and when they allow a firm to ’dominate from the middle.’Our focus on consumer choice and value creation complements the traditional focus in the strategy literature on competition and value capture. 1. Introduction The question of how to choose competitive positions (e.g., Porter, 1980) is core to both strategy thinking and teaching. Porter argues that firms face a choice between positioning as Cost Leaders or Differentiators, where the latter have higher quality offers and higher costs than the former. Those firms that do not choose one of these positions risk being “stuck in the middle” and being out competed. Countering the generic strategy proposi- tion is the observation that some firms that pursue both cost and differentiation advantage simultaneously are, in fact very successful (e.g., Barney, 1997; Besanko et al. 2000; Kim and Mauborne, 2005). Although the notion of generic strategies is a pillar of many strategy lectures and courses, its underpinings have been underexplored in the strategy literature. More generally, a weakness with the received literature on positioning continues to be that “our understanding of the dynamic processes by which firms perceive and ultimately attain superior market positions is far less developed [than our understanding of advantage at a point in time]”(Porter, 1991, p. 95). In this research note we develop an analytic model to test the logic of Porterian generic strategies and identify conditions under which they do and do not hold. Specifically, we explore when an industry segments according to the classic generic strategies of Cost Leader- ship and Differentiation, with firms located at other positions “stuck in the middle”(Porter, 1980), and when such generic strategies are out-competed by a Generalist that dominates from the middle (e.g., Kim and Mauborne, 1997). We focus on three questions: When should firms choose to follow classic generic strategies? When should firms position in the middle? What causes new competitive positions to arise over time?1 We build on Adner and Zemsky (2006), to develop a formal model to consider these questions. At the center of our model is an analysis of competitive advantage that is rooted in consumer choice and firm value creation. Our approach to competitive interactions follows the added-value methodology (e.g., Brandenburger and Stuart, 1996), which assumes that a firm’s ability to capture value is governed by its added value. Thus, we equate competitive advantage with added value. We are concerned with how consumers’relative willingness to pay2 for competing product 1We do not address the generic strategy of Focus in this research note. To do so would require introducing niche segments based on horizontal differentiation, which is beyond the scope of this research note. 2Following Brandenberger and Stuart (1996) the term willingness to pay indicates the maximum price a consumer would be willing to pay for an offer. It is is a distinct construct from market price, which reflects 1 pay for offer i at time t is then wim(t) = amq(xi). We introduce decreasing willingness to pay for quality improvement by assuming that q(xi) = xβi where β ∈ (0, 1).5 Thus, a segment’s willingness to pay for offer i at time t is wim(t) = am[q(xi)] β. Note that, consumers’willingness to pay for an offer increases with the offer’s performance (which itself increases with over time), but at a decreasing rate, (i.e. ∂wi ∂x ,∂wi ∂t > 0 and ∂2wi ∂x2 ,∂ 2wi ∂t2 < 0) 2.3. Value Creation, Competitive Advantage and Consumer Choice Definition 2.1. We define an offer’s value creation for a consumer in a given segment is the difference between the consumer’s willingness to pay and the marginal production cost. We denote this by vmi(t) = wim(t) − ci. We definecompetitive advantage across competing firms as the difference in their value creation. It is useful to separate competitive advantage into two parts, relative costs and relative differentiation. We focus our analysis on offer 1 and define the following: the cost (dis)advantage of offer 1 is Ac1 = c2−c1. The differentiation (dis)advantage of offer 1 is Ad1m(t) = w1m(t)− w2m(t). The net competitive advantage is then A1m(t) = v1m(t)− v2m(t) = Ad1m(t) + Ac1. At a point in time, if no firm has positive value creation in a segment (vmi <= 0), consumers in the segment make no purchase. If only one firm’s offer has positive value creation in the segment (e.g., vm1 > 0, vm2 < 0) each consumer in the segment purchases one unit of the value creating firm’s offer. If both firms’offers have positive value creation in the segment, each consumer in the segment purchases one unit of the offer with greatest value creation, which is the offer with net competitive advantage (i.e., the offer for which A1m > 0) offerconsumers purchase the offer (creates value in a segment, consumers in the segment buNote that both an offer’s value creation and its differentiation advantage will vary across segments. 3. Competitive Positioning Following Porter (1980), we assume a trade-offbetween product quality and production costs. Specifically, firms choose di, a level of differentiation, that determines both the quality of 5Note that, mathematically, for xi > 1, DMU is decreasing in β, while for xi < 1 DMU is increasing in β. We restrict attention to xi > 1 so that β has an unambiguous effect on DMU. 4 their offer and their costs.6 As elaborated below, the precise effect of di on costs depends on the nature of the technology that the firm uses. We consider the case of two market segments that vary in their taste for quality. Specifically, there is a low-end segment with a taste for quality of aL and a high-end segment with a taste for quality of aH > aL. We allow for as many as three firms to be active in the market. Our analysis proceeds in three steps. Section 3.1 characterizes the static choice of posi- tioning in a Porterian world where firms that do not follow Cost Leadership or Differentiation strategies are indeed stuck in the middle. Section 3.2 shows how classic generic strategies breakdown in the presence of a suffi ciently scale intensive technology. Following these static analyses, Section 3.3 shows how new positions arise over time as technologies improve. 3.1. Segmentation and Generic Strategies In this subsection we show formally how a market can be segmented by firms using classic generic strategies. We assume that firms only have access to production technology M for which the quality of a firm’s offer is given by xi = bdi and the marginal cost of production is ci = c + di. Because cost and quality are both increasing in di, there exist production possibility frontiers along which production cost and willingness to pay are traded off (Porter 1996; Saloner, et. al., 2001), with a different frontier for each segment.7 Figure ?? illustrates. We define d∗H as the level of differentiation that maximizes value creation, and hence competitive advantage, for the high-end segment. Similarly, we define d∗L as the level of differentiation that maximizes value creation for the low-end segment.8 Because the optimal level of differentiation is increasing in the segment’s taste for quality, we have d∗H > d∗L. Proposition 3.1. Suppose an entrant faces a single Generalist incumbent serving both segments from a middle position dI (i.e., d∗L < dI < d∗H). (i) The optimal position for the entrant is either as a Cost Leader serving only the low-end segment from the position d∗L or as aDifferentiator serving only the high-end segment from the position d∗H . (ii) The relative attractiveness of being a Cost Leader is increasing in the quality level of the incumbent (dI) and the extent of DMU. The attractiveness of being a Differentiator is increasing in the 6Thus, di can be interpretted as an investment in a performance resource. 7The familiar depiction of a single cost-willingness to pay frontier ignores consumer heterogeneity. We show two different frontiers because our segments differ in their taste for quality, and hence in their willingness to pay for a given performance level. 8That is d∗θ = argmaxd(aθ(bd) β − (c+ d)) = [ aθβb β ] 1 1−β for θ = H,L. 5 Figure 3.1: The cost-willingness to pay frontier for both segments when β = 0.5, b = 1, aL = 2, and aH = 2 6 specialist strategies of Cost Leadership and Differentiation using technology M because a firm can optimally trade-off marginal cost and performance (as in Proposition 3.1). With low scalability (0 < f < f1), the economies of scale are insuffi cient to offset the advantages of fine tuning the offer to a single segment and the Generalist is unable to profitably enter the market. With access to a suffi ciently scalable technology (f1 < f), the Generalist is able to out compete the specialists in both segments. Note that a successful Generalist might or might not be located in the middle, depending on the level of scalability of its technology. For intermediate levels of scalability (f1 < f < f2), the Generalist locates in the middle and offering a compromise product. For a suffi ciently high level of scalability (f2 < f), its cost of increasing performance is so low that the performance of the Generalist’s offer exceeds that of a Differentiator. An example of a Generalist leveraging high fixed costs to target the mass of the market is Barnes and Noble book superstores which offered higher performance (e.g., wider selection, knowledgeable staff and in store cafe) than the differentiated independent booksellers that had dominated the high-end of the market prior to its entry. 3.3. The Evolution of Competitive Positions Thus far, we have identified three strategies that can be viable at any point in time: Cost Leader, Differentiator and Generalist. We now consider the emergence of firms using these strategies in a dynamic setting where technologies improve over time. Since xi(t) = bdit the cost-willingness to pay frontiers are shifting outward over time. When t = 0, consumers have a zero willingness to pay for offers, while the marginal cost of production is always at least c. Hence, no firm can profitably enter the market. We are interested in which strategy is used by the firm that pioneers the market and in whether that strategy is sustainable as technology matures. We identify three regimes. Proposition 3.4. Suppose there is pool of potential entrants with access to technologiesM and F and K < K̄. (i) For low levels of scalability (0 < f < f1), a Differentiator pioneers the market and is later joined by a Cost Leader. (ii) For intermediate levels of scalability (f1 < f < f3), a Differentiator pioneers the market and is later displaced by a Generalist; for f1 < f < f2, the Generalist has lower quality than the Differentiator and for f2 < f < f3 the Generalist has higher quality. (iii) For high levels of scalability (f3 < f), a Generalist is the first and only firm to enter the market. 9 Consider first the case where technology F is not very scalable (f < f1). From Proposition 3.3, we know that a Generalist strategy is never used. Because it is easier to create value in the high-end due to its greater taste for quality, a Differentiator is able to enter the market when a Cost Leader’s offer still has negative value creation. As technology improves further, the Cost Leader strategy becomes viable as well and the two strategies coexist in the market. For intermediate levels of scalability (f1 < f < f3), the Generalist strategy dominates in the long-run. Initially, however, the willingness to pay of the low-end segment is too low to justify the broad market deployment that is the hallmark of a Generalist. The Differentiator, unencumbered by fixed costs and focused only on the high-end, is then the first to create value and therefore pioneers the market. Over time, with further technology improvements, the willingness to pay of the low-end segment increases suffi ciently that the Generalist strategy becomes viable and it displaces the Differentiator. For high levels of scalability (f3 < f), the Generalist’s marginal costs are so low that serving both segments is profitable early on, leaving no room for other strategies. What factors determine which regime characterizes a given market? Corollary 3.5. The critical thresholds f1 < f2 < f3 from Propositions 3.3 and 3.4 are increasing in consumer heterogeneity (aH−aL aH ) and the extent of fixed costs (K) and they are decreasing in the size of the segments (s). The extent of DMU decreases f1, increases f3 and does not affect f2. Consumer heterogeneity (aH−aL aH ) reflects the extent to which the segments differ in their taste for quality. As heterogeneity increases the returns to targeting individual segments increase and so the Generalist strategy becomes less attractive. Thus, as Figure ?? illustrates, the thresholds f1 and f3 both increase in consumer heterogeneity. Now consider the effects of DMU. On the one hand, DMU acts to mask heterogeneity between segments by reducing the difference in optimal quality levels (d∗H − d∗L), which shrinks region I, in which the Differentiator is joined by a Cost Leader (i.e., f1 falls in DMU). On the other hand, DMU acts to lower overall willingness to pay, which makes it less attractive to serve both segments early on, which shrinks region III, in which the market is pioneered by a Generalist (i.e., f3 increasing in DMU). The dashed lines in Figure ?? show the effects of an increase in DMU (moving from β = .5 to β = .4). Finally, consider the effects of market size (s) and fixed costs (K). The larger the size of the market, the more attractive is the Generalist strategy due to its scale economies. Hence, 10 Figure 3.2: The evolution of competitive positions for different levels of scalability and consumer heterogeneity for β = .5 (solid lines) and β = .4 (dashed lines) and K = 1.25s. growing market size can trigger a shift to a Generalist strategy. Conversely, the larger the fixed costs, the less attractive is the Generalist strategy. The spectacular failure of many internet companies at the bursting of the technology bubble highlights the importance of these factors. For example, the large fixed costs in advertising and warehouses required for firms such as Etoys and WebVan were too large relative to the size of their markets. 4. Discussion Porter’s generic strategies are simultaneously a cornerstone of many core strategy courses and a source of intellectual friction in the field. On the one hand, they have served as concise yet powerful characterization of firms’strategic postures. On the other hand, they sit uncomfortably as simplified generalizations in a research field whose trajectory has been to clarify contingent relationships rather than make blanket claims. Explaining the long list of exceptions to the ’stuck in the middle’claim is emblematic of the problem. In this paper we try to resolve this tension by clarifying the contingent nature of the viability of generic strategies. Instead of positing the emergence of a ’new frontier’along which enlightened 11 than those of the incumbent for any level of d: s(c + df) + fdK > s(c + d) for K > s. Moreover, given the scale economies in technology F , either the potential entrant serves all customers (in both segments), or it stays out of the market altogether. We proceed by assuming that the entrant is serving both segments using technology F and check whether or not this is profitable. Incumbents reduce their prices to marginal cost in an effort to fight off entry. Following the logic used to derive Proposition ??, the profits of the entrant for any given d are πE(d) = s(vFH(d) + vFL (d)−max{v∗H , 0} −max{v∗L, 0})̇, where v∗H = vH(d∗H) is the value creation of the Differentiator in the high-end and v∗L is the value created by the Cost Leader in the low-end and where vFθ (d) = aθ(bd)β − [c + df + dfK/(2s)] for θ = H,L is the value created by the entrant when it serves a customer in segment θ. The level of differentiation which maximizes the entrant’s profits is then d∗E(f) = [ aH + aL 2− f(2−K/s)βb β ] 1 1−β . Let v∗E(f) = vFH(d∗E(f)) + vFL (d∗E(f)) be the entrants maximum possible value creation for one customer from each segment. We have v∗E(f) = γ ( aH + aL (2− 2f + fk)β ) 1 1−β − 2c, v∗H = γ(aH) 1 1−β − c, v∗L = γ(aL) 1 1−β − c, where γ = (1− β)(bβ) β 1−β . Note that d∗E/∂f > 0 and ∂v∗E/∂f > 0. Given that there are incumbents at d∗L and d∗H we assume that v∗L, v ∗ H ≥ 0. Then πE(d∗E) = s(v∗E(f)− v∗L − v∗H). Let f1 be such that v∗E(f1) = v∗L + v∗H , which yields f1 = ( 1− 1 2 ( aL + aH ((aL)1/(1−β) + (aH)1/(1−β))1−β )1/β)/( 1− K 2s ) . 14 Let f2 be such that d∗E(f2) = d∗H , which yields f2 = 1 2 ( aH − aL aH )/( 1− K 2s ) . It follows that f1 < f2 and that there exists a K̄ ∈ (s, 2s) such that f1 < 1 iff K < K̄. For K > K̄, f1 > 1 and a Generalist is never viable. Proof of Proposition 3.4 This proof builds closely on the arguments and definitions in the proof of Proposition 3.3. The strategies that exist in the market at any point in time are those that have positive and superior value creation. Recall that v∗θ = γ(aθ) 1 1−β − c for θ = H,L where γ = (1− β)(βb) β 1−β . Hence, v∗H > v∗L and both are increasing over time with b(t) from an initial value of v∗L = v∗H = −c. Let tH be the critical time at which v∗H = 0 and a Differentiator becomes willing to enter the market. At this time, γ = c/(aθ) 1 1−β and hence v∗E(f) > 0 is equivalent to f > f3 = ( 1− ( aL + aH 2aH )1/β)/( 1− K 2s ) where f3 > f2. Thus, for f > f3 the market is pioneered by a Generalist, otherwise by a Differentiator. We have that v∗E(f) > v∗L + v∗H is equivalent to f > f1 where f1 is independent of t. For f < f1, the Generalist never enters and the Differentiator is joined by a Cost Leader. For f > f1, the Differentiator is displaced by the Generalist before the Cost Leader would have entered. Proof of Proposition 3.5 The comparative statics follow from the expressions for f1, f2 and f3 in Propositions 3.3 and 3.4. 15