Jump to ContentJump to Main Navigation
Foundations of Corporate SuccessHow Business Strategies Add Value$

John Kay

Print publication date: 1995

Print ISBN-13: 9780198289883

Published to Oxford Scholarship Online: November 2003

DOI: 10.1093/019828988X.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2019. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use.  Subscriber: null; date: 17 September 2019

Pricing and Positioning, 2

Pricing and Positioning, 2

(p.235) Chapter 15 Pricing and Positioning, 2
Foundations of Corporate Success

John Kay (Contributor Webpage)

Oxford University Press

Abstract and Keywords

Successful companies must strive for stable competitive environments that can be the result of factors such as product homogeneity, stability of cost/demand conditions and limited number of sellers. In this context, product/company positioning ought to be the result of not only the distinctive capabilities of the company but also that of the competitive dynamics.

Keywords:   competition, competitive strategy, corporate success, cost, demand, distinctive capabilities, positioning, pricing, product differentiation

In this chapter I consider further the issues of pricing and positioning discussed in Chapter 14. Successful companies must strive for stable competitive environments, and I describe the conditions that facilitate this, or make it impossible to achieve. Instability of competition does not necessarily mean that prices are bid down to costs unless there are very many firms in the market. Firms still have some discretion in their pricing behaviour but in unstable competitive environments that behaviour must not only reflect appreciation of the positioning of competitors but an analysis of the variety of their possible responses.

Positioning decisions are partly dictated by the firm's distinctive capability. Some firms are naturally suited to up‐ or down‐market positions. This issue was discussed in Chapter 9. But positioning is also a matter of competitive dynamics. When there are few firms in a particular market, clustering behaviour—in which several firms jostle for the same market position—is a likely outcome. As the number of firms in the market increases, they are likely to spread themselves more evenly across the range of available positions.

Increasingly, a product can only be described by reference to a wide range of its characteristics. Many different features of a car are relevant to potential buyers, and a supermarket must decide which selection of many thousands of goods to hold in stock. For commodities like these, successful firms have to make a well‐matched selection of product attributes. For these firms, focus—the consistency of attributes with each other—may be as important as position. Shifts of market position, which require changes in customer perception as well as product specification, need to anticipate competitor responses. They raise special difficulties for multiple‐attribute products.

(p.236) Stable or Unstable Competitive Environments?

If firms are to maximize the added value they earn, they need to be able to price away from cost and towards value. Such strategies can always be undermined by competitors pricing towards cost. Value in use pricing will be sustainable only if at least one of three broad conditions is met:

  1. The product faces limited competition, either because of the firm's strategic assets or because it is strongly differentiated.

  2. Competing producers respect the firm's ‘territory’ (which may be defined in either geographic or product space).

  3. Competing producers recognize the integrity of the pricing structure, and apply it to their own products.

The examples given in Chapter 14 provide a variety of illustrations of each of these possibilities. The price discrimination practised by SNCF is possible because there is only one French railway system. If there were many (a world which is admittedly hard to imagine) then competitors would seek to attract lucrative business travellers away and the margins between different fares would be quickly eroded. If patent protection for pharmaceutical products were less effective, then prices would be generally lower and more uniform across different geographic markets. In both cases, it is the limits to competition which enable firms to charge prices which reflect the value of their goods to their customers.

European automobile markets illustrate both respect for territory and support for the integrity of the price structure. The high level of UK car prices would be reduced if distributors were ready to sell cars outside their designated territories, or if one or more major manufacturers were to sell at levels which reflected their costs rather than the established UK price structure. European insurance retailers and cement companies do not market aggressively across national boundaries. Airlines maintain broadly comparable price differentials between fare types and classes of customer.

Territorial divisions, or tacit acceptance by all firms of price structures which reflect value in use, depend on the existence of formal or informal agreements. Most formal arrangements fall foul of Article 85 of the Rome Treaty, which prohibits agreements between undertakings which restrict or distort competition. Article 85 (3) does provide for an exemption for agreements which enhance technical and economic progress. A number of arrangements have secured such exemption—such as selective distribution agreements in automobile markets. But cartels and collusive price‐fixing, and most geographical or other market‐sharing arrangements, are illegal if they affect trade within the European Community. These practices may also be prohibited by EC states' own national anti‐trust legislation (as in Germany and the United Kingdom) and are generally illegal in the United States and Japan also. But the illegality of a formal agreement does not prevent firms operating within tacitly understood boundaries of the parameters of competition in (p.237) particular markets. Without such understandings, much business would simply be impossible.

There are some academic studies of the problems of achieving co‐operation without communication. In an original experiment by Thomas Schelling, Harvard students were instructed to try to meet someone the next day in New York City. There were no indications as to place or time, and no further information was available. A fair proportion of students given this brief would be found under the clock at Grand Central Station at noon, and therefore the chances of a successful meeting are quite high. The story demonstrates the importance of what Schelling called ‘focal points’ in non‐verbal communication. These are outcomes which, for reasons that may be hard to define precisely, somehow have a natural ‘feel’ to them. Such focal points are often culturally specific. Harvard undergraduates might naturally gravitate to Grand Central Station. Few European visitors would succeed in locating it.

Focal points may be round numbers in prices, or fifty/fifty sharing agreements. Indeed, fifty/fifty sharing provides a striking example, since although there may be little difference in the amount received from an agreement to share something fifty/fifty and a agreement to split it in the proportions 49.9 per cent/50.1 per cent, it is often easy to reach consensus on the first and not at all easy to do so on the second.

If market conditions, and prices and costs, never changed, it would be relatively easy to establish rules of competition that avoided destructive price wars or mutually defeating encroachments into each other's markets. After a few plays of a common strategy, such as tit for tat, there is a good chance that the competitive environment might settle down. But demand conditions, costs, and individual market shares change. Firms then have to move prices, launch new products, invest in new capacity, accept gains and losses of customers. How are they to avoid these moves being misinterpreted?

Focal points are a particularly clear illustration of the importance of simple rule structures in non‐verbal communication. Most markets in which competitors have coexisted for any length of time have well understood conventions of competitive behaviour. Price leadership is a common approach. One firm—usually the largest—initiates a round of price adjustments and, so long as the adjustment is reasonable, others will follow the lead. The car industry follows this pattern almost everywhere, with different price leaders in different markets. Or a price change may follow some external signal. Commercial banks usually move interest rates together in response to a change in the behaviour of the central bank. Indeed, this is such a well‐entrenched convention that governments rely on it in conducting their monetary policies.

The timing of price changes is another standard convention. Many industries have a recognized cycle of price adjustment, in which prices are raised once a year at a regular date. Many fast‐moving consumer goods markets have a tacit rule that price cutting takes the form of extra quantity for the same price, rather than lower prices for the same quantity. Promotions that take this form are understood to be temporary and induce no competitive (p.238) reaction. In some of these markets there is a recognized cycle of whose turn it is to discount next.

If these conventions are to be effective, it is not only necessary that they be simple. It is necessary that breaches of them should be obvious. Some product‐positioning moves that defy conventions are very obvious. Shifts into new geographical markets, or new product launches, are clear statements of aggressive competitive intention. Others may be very hard to identify. What of gradual upgradings of quality, or modest extensions of product lines? Certainly, the range of conventional signals seems much less in the product dimensions of competitive strategy than it is in the pricing dimension. But product complexity hinders the maintenance of competitive discipline, since the number of items to be monitored increases steadily. Yet even this can be overcome. Frustrated in attempts to reach an accommodation in any other way with its rival, Westinghouse, General Electric published a formula book which described exactly how it computed tender prices for electricity generation equipment. The effect was to make it easy for Westinghouse to follow its lead.

Successful Prisoner's Dilemma strategies are, it was suggested in Chapter 3, nice, but responsive, yet forgiving, and that is clearly true in the game of competitive behaviour. You do not assume that your competitor's intentions are all intrinsically hostile. You begin by offering the benefit of the doubt. If your rivals take advantage, your response is identifiable and proportionate to the threat. Yet you do not allow one aggressive action to plunge the industry into a price war for ever.

Despite these opportunities, while some industries establish a stable competitive environment on a sustained basis—retail banking—others are condemned to fight hot wars for ever—as in the personal computer market. There are factors which make stable competition almost impossible in some industries. Competition can rarely be stable if there are many sellers. Not only is it likely that some firm will break the conventions of the game, but even if they do not, the probability that someone will appear to be breaking them is high. And if there are many sellers, instability of individual market shares is almost impossible to avoid.

If it is better for the stability of competition that there should be few sellers, it is best that there should be many buyers. When buyers are fragmented—as in most consumer goods markets—then it is impossible to advertise price, or position, to customers without also advertising it to competitors. So conventions become easy to enforce. A small number of buyers can play sellers off against each other. They claim to be receiving discounts whether they are or not. Competitive discipline is impossible to maintain.

But while there are some industries where manufacturers sell directly to a diffuse public—petrol, or most services—most consumer goods are sold through retailers. Concentration in retailing implies that manufacturers face few buyers even though there are many final consumers for the goods concerned. It is then difficult to achieve stability in the competitive environment. This is a particular problem for firms producing branded foodstuffs.


Stability in competition depends on similarity of background—as was shown by the Harvard students' choice of meeting‐place. This may be the product of the common education, experience, and social position of senior executives. In many industries this is a very real factor promoting stability of competition—look at the insurance industry in most countries, the brewing industry world‐wide, or the relatively comfortable competitive structures which characterize local industries in many small countries. The globalization of markets and industries is upsetting many of these competitive environments, as new entrants emerge who do not understand, or may not wish to understand, the implicit rules of the game of local competition. This has been particularly true of financial services. Diversity will also be a problem if some firms have markedly different cost structures. So attempts to maintain price discipline in retail petrol markets have repeatedly been thwarted by supermarkets and small importers who face supply conditions very different from those of the oil majors.

Pricing in Unstable Environments

In 1993 or 1994, the thirty‐mile tunnel between Britain and France will open for service. As Britain has become more integrated in the European economy, roll‐on/roll‐off traffic—in which loaded trucks are carried by ferries between Dover and Felixstowe on one side and Calais or Zeebrugge on the other—has become a larger and larger proportion of total freight movements. Today, Dover is Britain's second largest port (the largest is Heathrow Airport).

While some of this traffic will shift to through rail services, most of it will continue to be carried in lorries. But in future lorries will have the option of boarding a shuttle train to be carried through the tunnel instead of a ferry. This is not an ideal service. Time will still be spend loading and unloading lorries at each end, and Dover‐Calais, although short, is not a particularly convenient crossing. The reason the Felixstowe‐Zeebrugge route is so popular is that if a direct road link could be built between the heart of England and the heart of Europe, it would almost certainly cross the sea between those two ports.

At present, it costs an average of around 200 Ecus to take a lorry between (p.240) Britain and continental Europe. Ferries will go on providing this service after 1993. What should the tunnel charge? If the ferries were to continue to set current prices, then the choice Eurotunnel faces is described in Fig. 15.1. If price is too low, market share will be high but margins thin. If price is high then margins will be much greater but the result will be loss of market share. In fact, if the ferries do continue to charge 200 Ecus the tunnel will do best to set a much lower price—perhaps as low as 100 Ecus or so. The reason is that the tunnel's marginal costs of operation are very low. It costs it perhaps 15 Ecus to ship a lorry—until the point, still far distant, at which the tunnel is operating at full capacity.1

                      Pricing and Positioning, 2

Fig. 15.1. Eurotunnel Profits at Alternative Price Levels

But would the ferries continue to stick with a price of 200 Ecus if the tunnel were to reduce its price to half of that? Fig. 15.1 shows how the choices open to the tunnel are changed if the ferry price is lower. The lower the price the ferry charges, the lower should be the price of tunnel services. In fact, it always pays the tunnel to undercut the ferries.

So does this mean that prices will tumble until channel crossings are practically given away? Not necessarily, because that outcome would be unprofitable for everyone. Fig. 15.1 enables us to find, for each ferry price, the best price for the tunnel. And in a precisely analogous way, we can find the best ferry price for each tunnel price. Both these relationships are shown in Fig. 15.2.

(p.241) What this describes is the Nash equilibrium, at B—the pair of strategies in which both the ferry and the tunnel are achieving the best outcome for themselves given the strategies of the other. In this Nash equilibrium, the tunnel price is 120 Ecus and the ferry price 225 Ecus. The ferries continue to do business, because there are still many users who prefer sea crossings which take them closer to their destination, but a large fraction of freight traffic uses the tunnel.

The nature of this Nash equilibrium illustrates that even in unstable environments competitors do not necessarily pursue each other to mutual destruction. The unstable environment is one in which each party takes no account of the interests of the other, but nevertheless considers the responses of the other, and that is the key difference. Not to consider other firms' responses would simply be irrational; but to take account of their interests requires a formal or a tacit agreement between the parties.

                      Pricing and Positioning, 2

Fig. 15.2. Tunnel and Ferry Reactions

Is that Nash equilibrium the likely outcome to this competitive game? Certainly, there is nothing that guarantees it, but if anything other than a Nash equilibrium is achieved then it pays at least one player to do something different. For example, the instinct of both parties is probably that since they are offering very similar services they should charge very similar prices, as at A. But if they do, the tunnel will gain by lowering its price and the ferries by raising theirs, and presumably they will realize this.

At an outcome like C, however, both players might be making higher profits than they achieve in the Nash equilibrium. This might be the outcome of (p.242) disciplined competition, rather than the unstable competition assumed in the model. The players are, after all, engaged in a repeated game. They will set new prices every season, and perhaps more often. Is it possible that they could achieve this outcome? The problem here is that because the cost structures of the players are so different, the temptation to depart from such a disciplined outcome is almost irresistible. The ferries have a tiny market share, at a very high price; surely they can gain a bit of the tunnel's volume by trimming it. And how would the tunnel react if they did? The Nash equilibrium really does seem the most likely outcome here.

A Model of Product Positioning

It is common to talk of product positioning, and product location, and the geographical analogy is a helpful way of describing product strategies. Start by taking the locational model quite literally. Visualize a beach, uniformly lined with tanning bodies. In the course of the day, everyone on it will need an ice‐cream, and will hasten to the nearest vendor. A boy and girl arrive, each with an ice‐cream trolley. Where do they locate?

The answer to this problem is very clear, although most people find it surprising at first sight. They will stand together, in the middle of the beach. One—let it be the boy—will serve all the customers on the left. The other—the girl—will serve all the customers on the right. To see why this is so, consider, as in Fig. 15.3, what would happen if the girl were to move. All the customers to the left of A would now buy their ice‐cream from the boy. The people at the far right of the beach have a shorter walk, but they were buying their ice‐cream from the girl anyway. Indeed, if the girl were to stay where she is, it would pay the boy to move towards her, and capture the customers to the right of A as well. If instead of having mobile trolleys, they have to erect fixed stands, or even to rent their pitches from the local authority in advance, the Nash equilibrium outcome is still for both to choose sites in the middle of the beach. If customers are unwilling to walk all the way from the end of the beach, and would do without an ice‐cream instead, it makes sense for them to move a little way apart. But unless this reluctance is very great, the distance between them will not be large.

This clustering behaviour is a phenomenon which is easy to recognize in the business environment. Look at any market which is dominated by two firms of comparable strength. See Hertz and Avis in car rental; distinguish, if you can, the Ford Taurus from the Opel Vectra. The leading firms in these industries pursue pricing and product strategies which can barely be distinguished from each other. At least one industry seemed designed to mimic the boy and girl on the beach. The Australian government decreed a two‐airline policy for domestic aviation, and split the market initially between two carriers, one privately owned, one public, protecting both from new entry. Ansett and Australian Airlines not only charged identical fares and deployed virtually (p.243) identical fleets, but also came to operate virtually identical schedules. Two planes would depart almost simultaneously for a distant airport, followed by a gap until a further pair took off in the same direction.

                      Pricing and Positioning, 2

Fig. 15.3. The Ice‐Cream Sellers

The explanation of these outcomes is almost precisely that which underpins the behaviour of the ice‐cream vendors. Two firms have broadly equal shares, and one modifies its product range or product formulation, or adopts a new marketing approach. A successful innovation of this kind would give it more than half the market, in which case it must pay to follow it. Or it fails, in which case it should be abandoned. In either case, the two firms will move in step.

But notice also that this story does not hold if there are three or more strong firms. A successful innovator will almost certainly make gains mainly at the expense of one competitor rather than the other. The loser may choose to follow the innovator, in which case it may well pay the third firm to move apart. Identity of strategies is a phenomenon of markets with two dominant competitors. It is much less likely if there are many players.

To see this most vividly, consider the outcome if a third ice‐cream seller arrives at the beach. The best strategy for the newcomer is to locate immediately to the right or left of an existing supplier. But this leaves one supplier squeezed in the middle. In Fig. 15.3c, the boy in the middle must react, by jumping to one side or the other. And that provokes a further response. What if the new boy, less provocatively, distances himself a little to the left of the established suppliers, as in 15.3d? Then it always pays the two suppliers on the outside to edge towards the centre, until eventually they provoke a (p.244) response. This is a game which has no Nash equilibria, and players of such games are condemned to endless, indecisive jockeying for position.

Certainly it is difficult to think of markets where three players of comparable strength have coexisted for long. The US instant coffee market has been dominated by three products—Nescafé, Maxwell House, Folgers—each with powerful backers—Nestlé, General Foods, Procter and Gamble, and has displayed comparable instability. There were three principal airlines operating between Europe and the United States—Pan American, TWA, and the relevant national carrier, and that was a competitive structure that did not last.

Perhaps the most striking demonstration of the instability of threesomes is to be found in political systems, especially in those countries whose first‐past‐the‐post electoral systems resemble the operation of markets. Two‐party structures endure, and generally show convergence of policies typical of two vendors on the beach. Sometimes, they move apart and if the gap is not narrowed a third seller appears. But if the third seller is to survive, he needs to eliminate one of the two others. Three‐party systems are unstable and do not persist.1

Intriguingly, instability is particularly a problem of threesomes. As the number of players increases, so does the probability of a stable outcome. If there are many ice‐cream vendors, they will spread themselves more or less evenly along the beach. There may still be some clustering and jostling for position at the ends. But convergence and instability are typically characteristic of small number competition.

The beach analogy, although very simple, illustrates some of the critical features of jockeying for position in markets (Illustration 15.2). In applying it, the distinction between horizontal and vertical product differentiation described in Chapter 9 must be borne in mind. Such differentiation is horizontal when the dimension along which producers array themselves is, like the beach, a neutral one. Clothes are manufactured in various sizes as well as in various qualities and some colours and some sizes suit particular individuals, some others. But everyone would agree that good tailoring is preferable to bad, and this is vertical differentiation. To apply the lessons of the beach more widely it is necessary to consider these various dimensions of product quality, and allow for the likelihood that products have multiple attributes. In the next section, this model is used to describe industries as different as national newspapers and food retailers.

Positioning in Newspaper and Food Retailing

What are called quality and popular newspapers are differentiated primarily by the density and difficulty of the material they contain. The measurement of these factors is familiar to educational psychologists and enables us to identify (p.245) in a particularly clear way the positioning of these different products. Table 15.1 illustrates the required reading age for Britain's principal national newspapers. Notice, however, that consumers are not uniformly arrayed along this beach. Interpreting where everyone is positioned on the newspaper beach is slightly tricky. We have information on the reading ages attained by the population of potential newspaper readers. But we do not know what kind of newspaper customers would like to read; we can only infer this from the pattern of what they do read. The ways in which newspapers are financed differ markedly along the quality spectrum. Despite higher fixed costs (they employ journalists), quality newspapers sell much more advertising revenue per copy and break even at considerably lower levels of circulation than popular newspapers whose revenue mostly comes from the cover price.

Table 15.1. The Reading Age of British National Newspapers


Required reading age (April 1988)

The Independent


The Daily Telegraph


The Times


The Guardian


Daily Mail


Daily Express


Today (Launch)


Today (April 1988)


Daily Star


The Sun


Daily Mirror


Four national newspapers have entered this market since 1969. The first (actually a relaunch of an existing title) was the Sun, which was initially positioned with a reading age below that of any existing title. The Sun rapidly (p.246) became Britain's best‐selling newspaper, and adjacent titles moved position towards it (reflecting the re‐emergence of clustering behaviour at the end of the beach). This segment also attracted the second new entrant, the Daily Star, which was positioned to take advantage of excess printing capacity within Express Newspapers. Today was brought to market in 1985 by Eddie Shah, a proprietor who had enjoyed some success with free provincial newspapers and whose disputes with trades unions had won him considerable publicity. It is rather unclear what Today's position was intended to be and the results of the launch were disappointing.

The fourth new entrant was the Independent. This paper was initially intended to fill the gap between the Daily Express and the Daily Mail, on the one hand, and the cluster of quality papers, on the other. As plans progressed, however, News International, owners of The Times, used an opportunity provided by a dispute with printing workers to dismiss the staff concerned and to print the paper at a different site using new technology (p. 188). This created some disaffection among The Times journalists and was associated with an increasingly right‐wing tone in the paper's editorial content. The Independent chose instead to launch a paper directly into the market position which The Times was seen as having vacated, recruiting some of its staff in order to do so.

The positioning of newspapers reflects the range of tastes among newspaper readers, and so is an example of horizontal differentiation. The positioning of food retailers reflects whether they aim to provide high‐quality products for high‐income households, basic goods for low‐income groups, or to strike a balance between the two. So food retailing is an instance of vertical differentiation. Chapter 2 described the positioning of the major British supermarket chains, and this is illustrated in Fig. 15.4. As was indicated there, these product positions have not always been the same. In the late 1970s Tesco made a very conscious decision to reposition from a down‐market stance to one close to that of Sainsbury's, the market leader. Since then, others have followed, less successfully, leaving Kwik Save dominant in the down‐market segment and eventually attracting new entrants there.

The failure of Woolworth to respond to shifting market conditions was very much in the mind of the management of Tesco. Yet changing the market position of a commodity (the services of a supermarket) which has multiple attributes raises special problems. An evolutionary change runs the risk of loss of focus. This was the fate of another British retailer, British Home Stores, which sought to upgrade its market position by introducing ranges of higher quality merchandise. But the existing customer base of Bhs did not want to buy these goods, and those who did were not customers of Bhs. The repositioning failed.

Tesco instead attempted to change the whole bundle of characteristics, bringing all aspects of its operations in line with its desired new position over a short period of time. It introduced a much wider range of fresh produce, and increased the proportion of private label goods. It raised the quality of merchandise generally and shifted the centre of gravity of its operations from (p.247)

                      Pricing and Positioning, 2

Fig. 15.4. Product Positioning of Supermarkets

city centres to out‐of‐town supermarkets. Its advertising emphasized the ‘new’ Tesco, and its logo and style of store layout changed. The overall objective was to attract a new customer group and accept that the repositioned stores would not be appropriate for many of its traditional shoppers. Taken as a whole, the operation was successful and Tesco now occupies a similar market position to Sainsbury's with similar market share.

It is apparent that such a strategy involves high risks. Coke's attempt to reposition its product with a new formula, preferred by customers in blind tastings, proved a fiasco, and Asda's emulation of Tesco was a failure. Tesco's success reflected the different basis of its competitive advantage. The company had established an internal and external architecture in its retailing systems—a distinctive capability which could be exploited in many different market positions. Its reputation (for cheapness) was of modest value. Such a reputation is easy to acquire. So Tesco sacrificed little of the company's existing distinctive capability, and exploited its new one more fully. Tesco's shift of product position was successful in the sense that it achieved its aims. Would it have been less profitable if it had remained in a more differentiated location? The answer to that question is not obvious.

Pricing and Positioning

Decisions on pricing and positioning often have to be taken together. This was obviously true when Bauer entered the listings market. Even if they are (p.248) separate decisions, as they will be if position is dictated by competitive advantage rather than competitive forces, they are interdependent.

What happens on the beach if boy and girl can charge different prices? Even if they can, they will not. They cannot charge different prices if they are standing together, since one vendor will win all the custom. No one should have been surprised that the Radio Times and TV Times were similar magazines and that each of them cost 50p. There is a general lesson here: that similar products command similar prices in competitive markets.

But what is meant by similar products requires careful interpretation. Consider the Coke/Pepsi game of Table 15.2. Coke and Pepsi are similar products but (in the precise sense defined on p. 135) Coke is of better quality; at equal prices, it takes a 70 per cent market share. Both Coke and Pepsi are heavily advertised.

Table 15.2. The Coke/Pepsi Game




Marginal cost (Ecus)



Share at parity (%)



Price response elasticity


Ad spend response elasticity


Initial values

Price (Ecus)



Ad spend (£000s)



Market share (%)



Profits (000 Ecus)



Nash equilibrium

Price (Ecus)

0.40 (+0.01)

0.35 (−0.04)

Ad spend (£000s)

630 (+0)

430 (−200)

Market share (%)

58 (−12)

42 (+12)

Profits (000 Ecus)

2,350 (−275)

1,050 (+285)

At first sight, Coke and Pepsi might be expected to charge the same price, and advertise at similar levels, and indeed this is what they do in many markets. But this cannot be a Nash equilibrium. Coke should advertise more, and it would not be rational for Pepsi to follow them; Pepsi should reduce prices, and it does not pay Coke to follow. Since the products are close substitutes for each other, price cutting and advertising are equally effective for both players. But the costs of advertising are independent of the volume of sales, while the cost of price cutting is directly proportional to them. That makes advertising a relatively expensive promotional weapon for Pepsi, and price an expensive promotional tool for Coke. In the model described in Table 15.2, the Nash equilibrium outcome has Pepsi more than 10 per cent cheaper than Coke and Coke advertising 50 per cent more.

(p.249) So why is this not what these two firms actually do? There are several explanations. One is that one, or both, players are making a mistake. Copycat strategy is a common resort of number two in a market. Number two wants to be number one, and believes that acting like number one will achieve that result. But in any market there is only room for one number one, and hence room only for one number one strategy. An alternative explanation is that equal prices offer a focal point. The observed outcome is not a Nash equilibrium in a one‐off game, but might be an equilibrium for a repeated game. The competitive environment is, despite appearances, stable rather than unstable, and rules which would allow the two players to charge different prices are simply too hard to implement.

This chapter and the last have introduced several models of real business situations and problems. They are heavily over‐simplified models. When firms use models of this kind, there is almost always pressure to make them more realistic. ‘Introduce a smaller competitor.’ ‘Segment the market more.’

Sometimes these changes give greater insight, but more usually it is a mistake to complicate the model in this way. The demand for greater realism often reflects a misunderstanding of the role models play in the analysis of business behaviour. This misunderstanding is shared both by critics and naïve users of models. A model of this kind does not provide a forecasting tool. Nor can it tell a manager what to do. It is designed to focus attention on key issues—issues which may not have been obvious before analysis began. Shouldn't you be pricing to secure volume if you are the lowest cost producer? Can you compete effectively on advertising if you are number two? Models help to direct attention to questions like these in a more precise way. They exist to guide managerial judgement, not to replace it. I discuss this issue further in Chapter 21.

The best empirically orientated survey of the issues covered in these two chapters is Scherer and Ross (1990), but the orientation is strongly towards public policy. Scherer and Ross is particularly strong on the factors which make oligopolistic markets hot, or cold, wars. Porter (1980) covers similar ground in a business‐orientated framework. On pricing generally, see Nagle (1988), Gabor (1977), Devinney (1988).

Tynan and Drayton (1987) is a survey of the marketing literature on market segmentation. The principles of price discrimination were established by Pigou (1920); more recently, see Phlips (1983), and a survey is Scherer and Ross (1990). Implications of geographical market segmentation within the European Community are in Emerson et al. (1988) and Davis et al. (1989). Focal points are due to Schelling (1960), while Axelrod (1984) is the classic empirical study of co‐operative strategies with the Prisoner's Dilemma.

The original economic model of product positioning is Hotelling (1929) but this has been substantially extended and superseded in subsequent work. Eaton and Lipsey (1975); Phlips and Thisse (1982), who introduce the distinction between horizontal and vertical product differentiation; Ireland (1987); Pepall (1990). Positioning as (p.250) a marketing problem is considered by Ries and Trout (1981), Davies and Brookes (1989). Positioning and focus are extensively discussed by Porter (1980) but the perspective here is rather different.

On the principal empirical examples discussed here: the British TV listings market was extensively documented in a case before the copyright tribunal. Today's position was described in McArthur (1988); for price discrimination by British Gas, see Monopolies & Mergers Commission (1988); the Eurotunnel problem is more extensively described in Kay, Manning, and Szymanski (1990). The GE/Westinghouse case is HBS (1980).


(1) I have not ignored the warnings given at pp. 172 and 227 against thinking marginal costs are low because all costs are fixed—this is one instance in which they really are. This is a possible error, however, in thinking about ferry costs. It is true that once a ferry is on the point of departure, the costs of taking an additional vehicle on board are insignificant. But services do not have to be so frequent, and ferries can be sold or deployed on other routes. In the long run, there are almost no fixed costs of ferry operation. Tunnel costs, by contrast, are almost independent of traffic volumes.

(1) This analysis depends on politics being unidimensional, as it is in many—especially Anglo‐Saxon—countries. Where political preferences are arranged on more dimensions, as where a religious one is superimposed on a left‐right spectrum, multi‐party structures prove more enduring.