Market-Based Assets Theory of Brand Competition PDF

Journal of Retailing and Consumer Services 76 (2024) 103566 Contents lists available at ScienceDirect Journal of Retailing and Consumer Services...

Journal of Retailing and Consumer Services 76 (2024) 103566 Contents lists available at ScienceDirect Journal of Retailing and Consumer Services journal homepage: www.elsevier.com/locate/jretconser The market-based assets theory of brand competition Byron Sharp, John Dawes, Kirsten Victory * Ehrenberg-Bass Institute, University of South Australia, Level 4 Yungondi Building, North Terrace, Adelaide, SA 5000, Australia A R T I C L E I N F O A B S T R A C T Keywords: Kotler popularised the Segmentation, Targeting, Positioning (STP) theory of brand competition. This theory still Brand competition dominates marketing textbooks. In this article we show how the discovery of scientific laws concerning how Market-based assets brands compete, grow, and decline clash with the STP theory. The contradiction between these empirical reg­ Empirical generalisations ularities and STP theory has led to the recent emergence of a new market-based asset view of brand competition. Segmentation Targeting We show how this theory fits the now well-established empirical laws, and we discuss some promising areas for future research. 1. Introduction activities entered the literature in the 1950s (Smith, 1956). It was in the 1960s that McCarthy’s mnemonic checklist of the ‘4Ps’ appeared. Many disciplines have a central question. In business strategy McCarthy’s (1960) 4Ps then went on to feature in most, if not all uni­ research, the central question is along the lines of ‘why do some firms versity textbooks, particularly after being adopted by Philip Kotler. earn more or less profits than their competitors?’. A candidate for Kotler’s first edition of Marketing Management (1967) included marketing science’s central question is ‘why do some brands sell more or lengthy coverage of the optimisation of the marketing mix, before in­ less than their competitors?’. The fact that rival brands can offer similar dividual chapters on decisions for each of the 4Ps. This optimisation products, at similar prices, but sell vastly different volumes has long material disappeared in later editions, and was replaced by chapters on intrigued scholars, and also marketers and investors. Yet, the traditional segmentation, targeting and positioning, value creation, customer view of brand competition that has dominated marketing education for satisfaction and loyalty. sixty years fails to provide an adequate explanation for why some brands As a theory of brand competitiveness, the marketing mix is an sell far more (or less) than others. Nor does this traditional view fit with extension of economics’ ‘perfect competition’ model but with now four several now well-documented empirical laws. main types of demand driver, not just price. Under the 4P’s view, brand In this article we outline what has been the dominant Segmentation, competition is portrayed as a battle to identify and deliver the marketing Targeting, Positioning (STP) theory of brand competitiveness, pop­ mix that generates highest demand. Marketers are portrayed as opti­ ularised by many, especially Philip Kotler.1 We show how STP theory misers, and use techniques such as marketing mix modelling and choice does not predict, let alone fit with the empirical laws which describe modelling to work out the most attractive and profitable marketing mix how brands compete, grow, and decline. We present extensive evidence, they can offer. that covers many countries and a vast number of product/service cate­ While it would be difficult to overstate the influence of the 4Ps even gories, that supports a market-based asset theory of brand competi­ today, there were immediately rather obvious problems for the “best tiveness. In turn, this theory and evidence provides an explanation of marketing mix wins” model. For example, even well-resourced corpo­ why some brands are much bigger than others. rations regularly launch carefully researched and consumer-tested new products and brands only to see them fail (Victory et al., 2021). Failure is 1.1. The beginnings of segmentation, targeting, positioning theory even common for new product launches that have been judged ‘winners’ by consumers and industry experts (Victory and Tanusondjaja, 2023). The concept of marketers as “mixers” of demand-influencing Even with hindsight it is not always apparent what would have been the * Corresponding author. E-mail address: [email protected] (K. Victory). 1 We use the terms ‘4P’s theory’ and ‘STP theory’. It can be argued that the 4 P’s and Segmentation, Targeting and Positioning are not formalised theories, but they certainly serve as informal theories consistent with the broad idea that a theory is a set of ideas intended to help understand a phenomenon, in this case, how brands compete and succeed. https://doi.org/10.1016/j.jretconser.2023.103566 Received 10 April 2023; Received in revised form 12 September 2023; Accepted 12 September 2023 Available online 5 October 2023 0969-6989/© 2023 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC license (http://creativecommons.org/licenses/by- nc/4.0/). B. Sharp et al. Journal of Retailing and Consumer Services 76 (2024) 103566 better marketing mix in terms of brand competitiveness. Even predicting despite some brands having very similar marketing mixes. Similarly, which advertisements will perform better than others was then, and still unlike the 4Ps theory, STP theory fits with the observed evidence that is now, apparently beyond marketers’ ability (Hartnett et al., 2016). many brands, even small ones, last many years, even decades (e.g., Pears Contrary to the 4P’s theory, and brand competition simulations Soap has been in the market for over 200 years). The STP theory of brand based on it (e.g., BrandMaps, MarkStrat), sales and market share appear competition also fits better with the observation that there are typically surprisingly resistant to changes in the marketing mix. Elasticities, mild reactions (stability in market share) when rival brands change their especially for large brands, for both price (Bijmolt et al., 2005) and marketing mixes. While STP theory recognises brands do compete, the advertising (Sethuraman and Tellis, 1991) turned out to be remarkably ideal way to compete under STP theory is portrayed as rather indirect. In low. In other words, the market showed little and a rather slow response other words, STP theory suggests brands can be thought of as analogous to changes in the marketing mix. Advertising weight tests, for example, to village general stores, with each store serving its village, with little can show a near zero response even when ad spend is doubled (Hu et al., overlap in customer bases. 2007; Hu et al., 2009). Ironically, STP theory can also appear to be somewhat anti-market­ Large-scale studies show that brand market shares tend to be quite ing, or at least anti-sales and anti-marketing departments. This is stable over multi-year periods, including in fast-moving consumer goods because the STP theory suggests brands position themselves to fill their markets (Dekimpe and Hanssens, 1995; Dunn et al., 2021) and also for niches, and then the battle should be over. Kotler (1989) even suggests retailers (KantarWorldPanel, 2021). Market share stability is also seen in that successful targeting will practically negate the need for on-going durables markets such as cars (e.g., Knoema, 2023) and for services such marketing because if done well, then the product should practically as airlines (e.g., Statista, 2023c). This is not to say market share change sell itself. This viewpoint reflects the prominent absolute perspective, in does not happen, but rather when it does, it often occurs as small contrast to a recently introduced concept of brand competitiveness that changes over a number of years. suggests measurement and management of brands relative to other brands (see, Baumann et al., 2017; Winzar et al., 2018). 1.2. Segmentation, targeting, positioning theory 2. Empirical observations on brand competition Brand competitiveness appears to be more mysterious than the “best marketing mix wins”. With this realisation, marketing textbooks quickly Over the same decades that STP theory cemented itself as the embraced the complexity in brand competition, along with rising page dominant way to teach and practice marketing, empirical research has counts, proposing a theory of brand competition that emphasised seg­ documented a number of empirical patterns and scientific laws con­ mentation, targeting and differentiation. Even Philip Kotler declared in a cerning how brands grow, decline, and compete. Like the scientific laws recent interview that no-one should read his original textbook because it discovered in other disciplines, they show that the real world is an un­ did not contain his later theory (Kotler, 2023). The re-developed Kot­ usual place that doesn’t conform to intuitive theory. In this paper we lerian approach to marketing is best illustrated by a quote in Kotler and present a subset of these empirical discoveries that relate to how brands Keller (2021): compete, but that clash with STP theory.2 That is, we examine some of the generalised empirical discoveries that contrast with the key precepts ‘A company discovers different needs and groups of consumers in the of the STP theory of brand competition, namely that segmentation, marketplace, targets those it can satisfy in a superior way, and then de­ targeting and positioning is used to assemble the 4Ps in such a way that velops a value proposition and positions its offerings so the target cus­ brands can. tomers recognize the benefits of its offerings. By clearly articulating its value proposition and positioning, companies can deliver high value and a) serve particular customers better than rival brands, satisfaction, which lead to high repeat purchases and ultimately to greater b) as a result of differentiating the brand, company profitability’ (p. 167). c) which increases customer satisfaction, Lynn describes (Lynn, 2012) this theory and its prevalence d) driving repeat-purchase. ‘Almost any marketing textbook will tell you that the key to successful In our examination of this issue, we specifically highlight examples marketing can be summed up by the STP strategy—that is, segmentation, of how retailing and consumer services brands exhibit these empirical targeting, and positioning. This approach suggests that the mass market patterns. consists of some number of relatively homogeneous groups, each with distinct needs and desires. STP marketers attempt to identify those market segments, direct marketing activities at the segments which the marketers 2.1. Observation #1: competing brands share customers believe that their company can satisfy better than their competitors, and position their product offering so as to appeal to the targeted segments’ (p. First reported for TV viewing in the science journal Nature (Good­ 353). hardt, 1966), and then later in brand purchasing (e.g., Ehrenberg, 1988), the existence of the Duplication of Purchase Law is an extremely This line of logic is echoed in many other works on marketing inconvenient empirical regularity for the STP theory. This law-like management, strategic marketing, and global marketing, such as: Lamb pattern reveals all competing brands share their customers with all et al., (2015 ch. 5), Pride et al. (2021 ch. 5), Iacobucci, (2021 ch. 4–5), other brands in the category (Ehrenberg, 1988) and the degree of the Proctor (2020 ch. 8) and (Schlegelmilch, 2022 ch. 6). customer sharing between competing brands in a category in any given We refer to this approach as Segmentation, Targeting, Positioning period is simply proportional to the relative market shares or penetra­ (STP) theory. Under this theory, marketers are told to try to serve tion (i.e., the % of buyers buying or using) of each brand. In other words, different customers than competitors, and/or to fulfil different customer needs rather than compete ‘head on’. Marketers are also urged not to compete on price, but instead to instil loyalty (i.e., repeat purchasing) to 2 There are also several other empirical discoveries which clash with the inure buyers to competitor marketing activities. working of how brands compete posed by STP theory, but space constraints The STP theory is not in ‘competition’ with the 4P’s framework but prevent proper coverage of them here. These discoveries include the Negative rather, the latter is more tactical while STP is more strategic in nature Binomial Distribution (NBD)/Ehrenberg’s Law of Buyer Frequencies and the (Kotler, 1989). At face value STP theory fits the real world far better than NBD-Dirichlet model of category and brand purchase propensities (see in, the 4P’s framework when applied on its own. This is because of the Driesener and Rungie, 2022; Ehrenberg et al., 2004; Goodhardt et al., 1984; empirical observation that markets can support many rival brands Sharp et al., 2012). 2 B. Sharp et al. Journal of Retailing and Consumer Services 76 (2024) 103566 the Duplication of Purchase Law demonstrates competing brands do not the small brand Monzo. The Duplication of Purchase Law is plainly ‘own’ their customers. This is a certainly a picture of very direct evident: these brands all share their customers with the other banks in competition, much different to the aim of avoiding competition by the market, in line with the size of those other banks. making the brand’s positioning unique, then targeting the appropriate By reading across Table 2 from left to right, we see that sharing of people. customers declines in-line with brand size. That is, bank brands share The Duplication of Purchase Law can be expressed algebraically as. more of their customers with the other bigger brands, and they share fewer of their customers with smaller brands. The correlation between bY|X = D × bY brand penetration and the average proportion of other brand’s buyers Where bY|X is the percent of buyers of X who have also bought Y; D is the using the brand is near perfect (r = 0.98). The marketing implication is constant known as the duplication coefficient; and bY is the percent of that brands – such as these retailing and consumer services brands - the population who have bought Y. compete largely ‘head on’ against all their competitors, a very different Since the Duplication of Purchase Law was first described in the conclusion than the STP theory of competition. 1960s (Goodhardt, 1966; Goodhardt and Ehrenberg, 1969), thousands The Duplication of Purchase pattern is the norm but there can also be of studies by commercial marketing analysts, as well as academic re­ instances of brands that share more of their customers than expected. An searchers, have shown that even in quite short time periods many con­ example in Table 2 is Tesco Bank, where more Tesco Bank customers use sumers are multi-brand buyers. This empirical regularity has even been any other bank (apart from Monzo). For example, 33% of Tesco Bank observed in categories where extreme loyalty is often expected (e.g., customers also bank with Nationwide, whereas the average rate of Dawes, 2008, 2014). Multi-brand buying was previously interpreted as sharing of all brands with Nationwide is 24%. This deviation is also being due to some consumers changing their loyalties (i.e., preference), shown for Santander, where 28% of Tesco Bank customers also bank defecting from one brand and taking up another. For example, Kotler with Santander, compared to the average rate of sharing of 21%. These (1967, p. 235) presented a hypothetical brand switching table along differences can exist, but the overwhelming pattern is sharing with a discussion about how patterns in brand switching could be used (competing) in line with brand size. to reveal how brands compete (see Table 1). The Duplication of Purchase Law strongly suggests that brand Unfortunately, the table presented in Kotler (1967) was made of growth will come from gaining some more cross-purchasing from all of hypothetical data and did not reflect actual buying behaviour. It showed one’s competitors. Support for this pattern is shown from Table 2 to patterns that would be expected under STP theory, but such a pattern Table 3 for Monzo, a small brand with only 4% initial penetration. rarely, if ever, appears in the real world. This example hypothesised that Monzo grew to a 9% penetration brand by 2022, and we can see in zero customers switch from Brand C to Brand A, but 10% of Brand A’s Table 3 it did so by getting more cross-purchasing from every other bank. customers switch to Brand C; and more customers switch from Brand B For instance, we see Monzo went from having 5% of Nationwide cus­ to Brand C than stay with Brand B. These switching patterns are highly tomers also banking with it in 2019 to 8% in 2023, and 4% of Santander infeasible, and violate the empirical patterns observed using the customers also banking with it in 2019, to 10% in 2023, and so on. In Duplication of Purchase Law. other words, Monzo grew by attracting customers ‘across the board’ Even the concept of switching turned out to be largely incorrect, from competitors. This growth pattern goes completely against the idea although it is a more apt model in subscription markets like insurance that brands must necessarily identify and target a segment. One could where consumers tend to have smaller repertories and greater sole brand say there might have been a segment ‘out there’, waiting for a brand like loyalty (Sharp et al., 2002). In reality, consumers are multi-brand buyers Monzo to tap into, but that would mean that previously all the other and sole-brand loyalty is rare (e.g., Cannon et al., 1970), instead banks had attracted some of the same segment, which again does not fit (repertoire) polygamous multi-brand loyalty is overwhelmingly the with STP theory. norm (e.g., Ehrenberg and Scriven, 1997; Zhang et al., 2017). This The pattern seen in Tables 2 and 3 also goes against the idea that means consumers do have repertoires of brands which they show brands primarily grow by developing a unique positioning or value considerable loyalty to, but the loyalty they demonstrate is a very long proposition. If these banks had grown to their current level based on way from being exclusive. unique value propositions or positioning, that implies they each would We now illustrate this empirical fact in Table 2 using data from retail have customer bases that desire or respond to a specific value proposi­ banking in the United Kingdom, a product category not intuitively tion. It is difficult, therefore, to see how a growing competitor would thought to be a repertoire (polygamous loyalty) market. The data was acquire customers from all of them. So, the evidence here, and in sourced from YouGov, who run a large, demographically representative extensive past work using Duplication of Purchase analysis (e.g., Anes­ survey panel in the UK (YouGov, 2023) comprising over one million bury et al., 2021; Ehrenberg, 1988), suggests brands compete rather consumers. We use two annual time periods, 2019 and 2022, to directly against all competitors rather than by targeting and creating demonstrate the broad pattern of multi-brand loyalty, then examine the semi-protected niches or sub-markets (see also Lynn, 2013). growth pattern of a new entrant, Monzo. Table 2 presents the 2019 data The outcome of this law-like empirical pattern suggests that a from the twelve leading retail banks, in penetration (size) order. The brand’s targeting and positioning is not affecting which other brands it table shows the cross-bank usage. For example, looking at the row for competes more (or less) closely with. Direct tests comparing brands’ Nationwide, 21% of consumers bank with it, and of those, 23% also bank image positioning using consumer perceptual maps to customer overlap with Santander, 22% bank with Halifax, while only 5% also bank with shows the Duplication of Purchase Law followed by geographical loca­ tion (where the brands are sold) explain inter brand competition, not each brand’s image positions (e.g., Sharp et al., 2003; Sharp and Sharp, Table 1 1997). Together, these points begin to raise questions about how well Kotler’s (1967) Hypothetical brand switching example. STP theory explains brand competition. For marketers, including those To: involved with retailing and consumer services brands such as banks, this empirical evidence indicates that competition is very direct and ‘head Brand A Brand B Brand C From: Brand A 0.70 0.20 0.10 on’. Managers should recognise this fact when constructing their mar­ Brand B 0.17 0.33 0.50 keting strategy and plan to grow by targeting category buyers rather Brand C 0.00 0.50 0.50 than by fulfilling a narrowly defined niche. The hypothetical patterns of brand switching are at odds with the cumulative evidence of how customers are shared among competing brands. Table adapted from Kotler (1967, p. 235). 3 B. Sharp et al. Journal of Retailing and Consumer Services 76 (2024) 103566 Table 2 Duplication of retail banking customers, UK 2019. % banking with: % also banking with … Nationwide Santander Halifax Barclays Lloyds Natwest HSBC Tesco RBS First Dir Co-op Monzo Nationwide 21 23 22 18 14 15 12 15 5 7 6 5 Santander 19 25 23 18 13 13 11 14 5 6 5 4 Halifax 19 25 23 21 15 15 14 17 6 7 5 4 Barclays 18 22 20 22 15 14 12 16 6 6 5 6 Lloyds 14 23 18 21 19 11 10 15 4 5 4 4 Natwest 14 22 18 20 18 12 10 13 4 5 4 6 HSBC 12 22 18 21 18 11 12 15 4 7 5 8 Tesco Bank 10 33 28 30 27 20 18 17 8 12 6 3 RBS 6 21 19 20 20 10 11 9 14 6 6 5 First Direct 5 27 23 25 19 12 13 15 22 6 5 6 Co-op 5 26 19 19 18 11 10 11 12 7 6 4 Monzo 4 26 22 18 24 15 19 23 9 7 8 4 Average - 24 21 22 20 13 14 13 14 6 7 5 5 All competing brands will share their customers with other brands and will do so broadly in-line with their brand’s size (penetration) in the category. Data Source: YouGov BrandIndex UK 2023 © All rights reserved. Table 3 Duplication of retail banking customers, UK, 2022. % banking with % also banking with … Nation Halifax Santan Barclays Natwest HSBC UK Lloyds Tesco Monzo First D RBS Co-op Nationwide 21 20 21 18 16 13 13 12 8 8 5 5 Santander 19 23 19 19 15 14 13 12 10 7 5 4 Halifax 18 23 20 20 16 14 15 14 8 8 6 4 Barclays 17 22 20 20 15 15 14 13 9 7 4 4 Natwest 15 22 20 19 18 13 11 11 11 5 4 4 Lloyds 14 21 20 18 18 12 11 12 8 6 4 3 HSBC 13 22 20 20 19 14 12 11 11 7 3 3 Tesco Bank 8 31 30 27 27 19 17 20 8 13 7 5 Monzo 8 23 19 25 21 21 18 14 8 8 4 4 First Direct 6 27 24 23 19 13 15 14 18 11 4 4 RBS 4 22 22 21 16 12 9 12 13 7 5 5 Co-op 4 25 19 17 17 13 11 11 10 8 6 5 Average - 23 21 21 19 15 14 14 12 9 7 5 4 Note: The row and column order are different in Table 2 (compared to Table 1) due to some banking brands growing or declining from 2019 to 2022. Data Source: YouGov BrandIndex UK 2023 © All rights reserved. 2.2. Observation #2: competing brands sell to similar customers studies highlighting the value in niche segments (see, Dibb and Simkin, 1991). Another well-established empirical generalisation, that conflicts It should be said that a large body of work exists on the topic of with STP theory, is that the brand user profiles of competing brands tend segmentation – much of it concerned with how to do segmentation. A not to differ. In other words, the buyers of competing brands within a plethora of segmentation schemes and approaches have been examined category look strikingly similar in terms of potential segmentation in scholarly literature, such as segmenting by loyalty (Frank, 1967), criteria such as demographics and lifestyle (e.g., gender, age, income). values and lifestyles (Novak and MacEvoy, 1990), product involvement This empirical regularity has been repeatedly documented, across a (Lockshin et al., 1997), benefits (Cermak et al., 1994), even astrological range of markets and segmentation variables. For example, a lack of signs (Mitchell and Haggett, 1997). However, even segmentation pro­ brand-level segmentation has been observed across many consumer ponents find results that are consistent with the finding that brand user goods categories (e.g., Anesbury et al., 2017; Hammond et al., 1996; profiles differ little. As Fennell et al. (2003, p. 223) wrote, after ana­ Uncles et al., 2012), as well as durables and services (Kennedy and lysing 52 product categories looking for segments using dozens of po­ Ehrenberg, 2001b). In relation to retailing, United Kingdom grocery tential segmentation variables, ‘[d]emographic and general retailers were found to sell to consumers with extremely similar de­ psychographic variables … are not useful for predicting relative brand mographics, attitudes and media use (Kennedy and Ehrenberg, 2001a). preference’. Sportswear brands were shown to have little differences in appeal across The finding that competing brands’ customers look similar, while consumer segments in the UK, using both simple demographics and the surprising for many marketers, can be seen as good news. It suggests sophisticated ACORN geodemographic targeting scheme (Dawes, 2009). there are many people potentially available to purchase one’s own Lynn (2007) found no meaningful segmentation differences between brand, and an overwhelming majority of brands do not have to competing hotel and cruise ship brands. Lynn (2013, p. 92) later pigeonhole their activities and sell to a specific target segment. Instead, concluded in a study of quick service restaurants, ‘most … time, energy ‘competition ultimately means selling successfully to the same potential and money should be devoted to mass marketing and not targeting customers’ (Hammond et al., 1996, p. 48). This is not a bleak story for subsets of consumers’. marketers, but rather an enabling one. Pairing this finding with The similarity of brand user profiles appears to hold in consumer knowledge about the key difference in brand performance metrics be­ panel data (e.g., Hammond et al., 1996) as well as survey data (Kennedy tween large and small share brands is the size of the customer base (see and Ehrenberg, 2001b). The cumulative empirical evidence, including later in paper), it becomes clear that the route to brand growth is to those about retailers and other service providers, is in contrast to case acquire people who purchase the category. This is the sustainable path 4 B. Sharp et al. Journal of Retailing and Consumer Services 76 (2024) 103566 to brand growth, including for retailing and consumer services brands. adequately build physical and mental availability. Indeed, recent evi­ Kotler (2004, p. 10) asserts that not understanding one’s target dence points to the importance of distribution (a component of physical customers is a ‘deadly sin’ in marketing. The often-overlooked similar­ availability) in new product success (e.g., Sinapuelas et al., 2015). ities in competing brands’ customer bases, along with the regularities Together, these results downplay the importance of building a position seen in the Duplication of Purchase Law (discussed above), are useful based on differentiation and superiority, and instead support the analysis approaches that help brands understand their customers. importance of building market-based assets, even when it comes to new Additionally, they highlight how STP theory misleadingly says that products. brands should sell to different sorts of customers. Because competing brands share their customers with other brands, and these customers 2.4. Observation #4: competing brands have similar satisfaction scores tend to have similar profiles, this logically calls into question the belief that customers need to perceive brands as being different in order to Kotler and Keller (2021) state that offering products that are superior purchase them and develop loyalties. We next look at the evidence on to competitors leads to satisfied customers, who then repeat-purchase, perceived brand differentiation and how it contrasts to the rationale ensuring the ongoing success of the firm or brand. It seems a truism proposed from STP theory. that offering great products or services should lead to customer satis­ faction, and indeed having highly satisfied customers is desirable. 2.3. Observation #3: competing brands are weakly differentiated However, the relationship between customer satisfaction, loyalty and brand competition is far from the simple picture painted in STP theory. Past research investigating the perceived differentiation of In this respect, STP theory is at odds with one of the best-known competing brands has produced the finding that even loyal (regular) empirical laws of marketing: the Double Jeopardy Law. This law-like buyers of a brand are unlikely to perceive it as very different from pattern shows bigger brands (with more customers) get a bit more loy­ competing brands. Even for the most successful (high share) brands in a alty, and small brands (with fewer customers) get a bit less. This pattern market, very few brand users, only around one in ten, state that the is described in more detail in the next section. Despite the clear patterns brand they use is either ‘different’ or ‘unique’ (Romaniuk et al., 2007). between big and small brands in terms of size and loyalty, there is This research included brands operating in packaged goods and durable generally not a clear relationship between a brand’s satisfaction levels categories (e.g., cars), in addition to retailers (e.g., supermarkets) and and its market share (e.g., Fornell, 1995). consumer service brands (e.g., banking). This discovery was originally Take the example of the satisfaction of retail banking brands in the made using Young & Rubicam’s Brand Asset Valuator data (see, United Kingdom (see Table 4). Data in Table 4 was supplied by a com­ Romaniuk et al., 2007), and was later independently replicated in mercial research project. The banks have around a 6-fold difference in Kantar’s BrandZ tracker (see, Hollis, 2011). The latter study tracked over the size of their customer bases (i.e., 32%–5%). The banks with more 6000 brands over a 10 year period and revealed ‘the proportion of customers also get similar, but slightly more loyalty, as shown by the people willing to endorse any brand as “different from other brands of (a average number of products customers have with the brand. However, specific category)” is low’ (Hollis, 2011, p. 2). the level of customer satisfaction these competing service brands receive This finding has been corroborated in other brand image data, which has no relationship to either their size or loyalty. Instead, all brands have again shows striking similarities in the way that those who are familiar a score that is around 8. We conclude that (a) bigger brands do enjoy a with a brand see their brand, and those that are familiar with another bit higher loyalty, but also that (b) bigger brands don’t get higher brand see theirs (Collins, 2002). Other research using brand image data satisfaction, therefore (c) it is difficult to see how higher satisfaction shows that many customers do not uniquely associate the brands they scores can be assumed to lead to higher loyalty. buy, even the highly successful ones, with any particular image Furthermore, individual customers’ satisfaction ratings are not perception (Romaniuk and Gaillard, 2007). Indeed, extensive evidence consistent over time. Around half of a brand’s customers provide a shows that competing brands tend to share brand attributes with other different satisfaction score six week after the initial measurement, brands far more than they stand out on any one particular image attri­ despite not having another encounter with the brand over that time bute (Collins, 2011; Dawes, 2011; Romaniuk, 2001). Moreover, buyer’s (Dawes et al., 2020). Similar instability in repeat rates for brand image brand image associations for brands appear to generally not be strongly attributes is seen in past research (Dall’Olmo Riley et al., 1997). held, as only around 50% of respondents attach the same attribute to a There is no doubt that satisfied customers are good to have, but brand when they are re-interviewed (Castleberry et al., 1994; Dall’Olmo satisfaction is not a metric that predicts brand loyalty, nor does it Riley et al., 1997; Rungie et al., 2005). These facts contrast with the idea distinguish big brands from small brands. This is certainly at odds with that it is essential for a brand to build ‘strong, favourable and unique the sentiment that ‘satisfaction is the key to building customer loyalty’ brand associations’ to be bought and to enjoy high loyalty (Kotler and (Kotler and Keller, 2021, p. 448). This finding also undermines the STP Keller, 2015, p. 311), as per STP theory and related work (John et al., 2006; Keller, 1993). Keller (2014, p. 706) does discuss the idea that a Table 4 brand can emphasise points of parity not just points of difference, but Penetration and satisfaction of United Kingdom retail banking customers. adds that to be in a strong position, a brand should achieve an advantage % Banking with Avg. # products Satisfaction score on those points of difference. A corroborating picture about the role of differentiation is seen in Halifax 32 1.7 7.9 Barclays Bank 29 1.8 7.8 new product research. A traditional saying is that “good marketing can’t Nationwide 26 1.8 8.3 save a bad product” but nowadays most brands do not launch ‘bad’ Santander 24 1.6 7.7 products, their launches are well made and are pre-tested for consumer Lloyds Bank 20 1.6 7.9 acceptance. Furthermore, new launches have often had the opportunity NatWest 15 1.6 8.1 to lie in wait and learn how to create ‘superior value’ over the options HSBC 14 1.8 7.6 TSB Bank 8 1.3 8.2 available in the market. Some new launches might have this opportu­ Bank of Scotland 7 1.5 8.1 nity, but there are still many new product failures (Victory et al., 2021), First Direct 6 1.5 8.7 including new private label products (Salnikova et al., 2020). Even Royal Bank of Scotland 5 1.6 7.9 many new launches voted ‘Product of the Year’, which are innovative in Average - 1.6 8.0 some way and are consumer-voted successes, end up being withdrawn Brands that have more customers have little variation in either their loyalty or (Victory and Tanusondjaja, 2023). These are seemingly not to do with satisfaction scores. These results suggest that greater customer satisfaction is not offering poor value to consumers but rather likely due to their failure to a primarily driver in explaining how brands compete and grow. 5 B. Sharp et al. Journal of Retailing and Consumer Services 76 (2024) 103566 viewpoint as to how brand growth and competition occurs. In this sec­ in the size of the customer base, rather than the loyal competing brands tion we touched on the fact that firm’s satisfaction scores often do not receives, is also demonstrated in Table 4. There are some instances correlate with their size, or their brand loyalty levels. In the next section where brands have higher loyalty than is expected, but where this does we further contrast the STP view, that sees loyalty as an outcome of happen the deviations tend to be predictable. For example, private labels brand positioning or superior value, with further extensive evidence on often exhibit higher loyalty for their penetration level (e.g., Bound and how buyer and brand loyalty develop. Ehrenberg, 1997; Dawes, 2022), which is a symptom of the brand having exclusive but restricted distribution (i.e., only people who shop that 2.5. Observation #5: competing brands have predictable loyalty store can buy the brand). In turn, these findings show that sustainable brand growth comes STP theory strongly implies that loyalty has to be earned (i.e., if some from greatly enlarging the size of the customer base, with commensu­ brands do better at earning it, they will be rewarded with high loyalty); rately much smaller gains in loyalty. This is certainly a much different to which has a corollary that brands should be able to vary considerably in the story about how loyalties emerge, as told through STP theory. loyalty. Moreover, that in order for buyers to show loyalty towards a brand they must perceive it to offer superior value. The empirical evi­ 3. Moving to the Market-Based Asset Theory of Brand dence, however, again paints a different picture. Competition Psychological experiments and field studies have documented how quickly buyers adopt loyal behaviours, even to identical offerings. In summary, Kotler along with colleagues developed a model of how Tucker (1964) found buyers developed repeat-purchase loyalty for brands compete that is largely based around making competition indi­ identical loaves of bread with the only difference between them being a rect and gaining perceived value and loyalty (repeat-purchasing) ad­ letter such as L, M, P or H. Similarly, McConnell (1968) found consumers vantages. STP theory encourages firms to find segments of customers, quickly developed loyalty towards three ‘brands’ of identical beer research what they want, selectively target, and make their brand labelled only with alphabet letters. Other work has documented people appear different or unique at least partly via advertising. develop loyalty towards university lecture seats and even toilet roll In this paper, we have outlined an array of empirical discoveries and orientation (Sharp, 2017a). Therefore, not only do few people perceive scientific laws (which are well documented across many categories, competing brands to be differentiated (Romaniuk et al., 2007), they do countries and other conditions), including many examples from retailing not have to see them as differentiated to buy them, and re-buy them and consumer services contexts, that clash with the STP theory of brand repetitively. competition. The collective interpretation of these empirical laws has The other, very well-established fact about brand loyalty is that it is led to a new theory of brand competition called the market-based asset highly predicable from the size of the brand (as briefly mentioned in theory. We now explain this theory, how it emerged, and how as any Observation 4). From as far back as the 1960’s, researchers discovered scientific theory must do, it fits with the known empirical laws. that more popular TV presenters, newspapers, products or brands also The idea of market-based assets entered the marketing literature enjoyed slightly higher loyalty, while the less popular alternatives also almost 30 years ago (Sharp, 1995; Srivastava et al., 1998). This concept suffered from somewhat lower loyalty (Martin, 1973; McPhee, 1963). was in line with the emerging strategy literature at the time, that views From these studies the term ‘Double Jeopardy’ was coined. the firm as a bundle of resources and capabilities (Amit and Schoemaker, This finding has been generalised over numerous markets including 1993; Wernerfelt, 1984, 1995). It was proposed that ‘the role of mar­ consumer packaged goods, durables, services, and for loyalty in terms of keting was concerned with the task of developing and managing purchasing as well as brand attitudes (e.g., Ehrenberg and Goodhardt, market-based assets’ (Srivastava et al., 1998, p. 16). That such intan­ 2002; Ehrenberg et al., 1990; Graham et al., 2017). An illustrative list of gible assets could have considerable financial value was uncontrover­ recent publications demonstrating the Double Jeopardy Law in diverse sial, but much of the literature at the time had a different focus, such as contexts is also shown later in the paper. Importantly for this special how to lower the risk of launching new brands. For example, the 1980s issue, Double Jeopardy has been documented in retailing and consumer marketing literature on brand equity near exclusively focussed on hy­ services contexts, including in fast food retailing (e.g., Pleshko and pothetical brand extension experiments. Heiens, 2022), convenience stores (e.g., Pleshko and Souiden, 2007), Sharp (2010) then introduced a market-based assets theory of brand retail banking (e.g., Mundt et al., 2006), and grocery retailers (e.g., competition. Importantly, this theory was based on the known empirical Uncles and Ehrenberg, 1990; Uncles and Kwok, 2008). laws about consumer behaviour and brand performance. This theory The Double Jeopardy Law, probably the most famous of all empirical was developed using an empirical first, empirical-then-theoretical laws in marketing, can be expressed algebraically as: approach (see, Barwise (1995); Bass (1995); Golder et al. (2022). In W = W0/(1− b). other words, Sharp’s (2010) market-based assets theory is grounded in Where W is brand purchase frequency, W0 is the constant estimated empirical laws and describes the mechanism of how brands compete and as the average of W(1-b) for all brands, and b is brand penetration grow: (Ehrenberg et al., 1990). ‘In the long run, brands essentially compete in terms of mental and Double Jeopardy stands in stark contrast with STP theory that sug­ physical availability. Even product innovation largely works (when it works) gests brands can carefully accommodate a chosen target segment, satisfy by enhancing mental availability and gaining further physical distribution. it with superior value and enjoy high buyer loyalty - from a small portion Building mental availability requires distinctiveness and clear branding, while of the market. By contrast, the Double Jeopardy Law (with its wide­ brands seldom compete on meaningful differentiation. This means that spread empirical support) says that small brands will have predictably marketing attention should be focused on building these assets so that a brand lower loyalty. is easier to buy, for more people, and in more buying situations. No marketing Market share is informed by the number of customers a brand has, activity, including innovation, should be seen as a goal in itself, its goal is to and the behavioural loyalty these customers devote to the brand. The hold on to or improve mental and physical availability’ (p. 196). Double Jeopardy Law explains that higher market share is due over­ Mental availability is defined as the propensity for a brand to be whelmingly the size of the brand’s customer base. For example, Dries­ noticed, recognised and/or thought of in buying situations (Romaniuk ener et al. (2017) show that toothpaste brands vary 20-fold in their and Sharp, 2004). Drawing on the Associative Network Theory of penetration but less than two-fold in loyalty. Similarly, Colombo et al. memory (e.g., Anderson and Bower, 1979; Teichert and Schontag, (2000) examined brand-switching for cars, showing much larger dif­ 2010), mental availability requires developing a breadth of relevant ferences in the number of buyers the car brands achieved but dramati­ memories linked to a brand, to increase the size of the brand-related cally lower variation in the loyalty to those makes. The larger difference network in people’s memories (Romaniuk, 2013). These memories 6 B. Sharp et al. Journal of Retailing and Consumer Services 76 (2024) 103566 need to be anchored to the brand in memory (via direct branding and Table 5 through a brand’s Distinctive Assets). This conceptualisation repre­ Illustrative summary of academic studies published since (2010) sented a substantial pivot from brand equity theory, which stressed positive differentiating memories (i.e., asking what memories the brand Competing Brands Share Customers (Duplication of Purchase Law) elicits, rather than what elicits the brand (Keller, 2001)). Non-profit/charity brands (Faulkner et al., 2022) Music listening (Anesbury et al., 2022) The importance of mental availability to brand competitiveness is Shopping baskets and department stores (Tanusondjaja et al., 2016; Tanusondjaja also underscored by the fact that a variety of situations bring consumers et al., 2022) into a category to purchase (Romaniuk, 2016; Vaughan et al., 2021). Iranian e-brands (Naami et al., 2021) This fact makes it important for brands to link to multiple, relevant Luxury brand competition (Romaniuk and Sharp, 2016) category situations. Brands that are linked to more of these relevant Cross-category brand purchasing (Grasby et al., 2022) Customer mindset metrics (Mecredy et al., 2021) purchase cues in buyer’s memory have a higher chance of being pur­ Healthy vs unhealthy food (Anesbury et al., 2018b) chased than brands with fewer people having links to fewer cues. This Fresh produce purchasing (Anesbury et al., 2020) conceptualisation also deviates away from simpler brand awareness Consumer goods categories with expected partitions (Anesbury et al., 2021) metrics where only the link of the brand to the category name cue is Cigarette purchases (Dawes, 2014) Consumer goods in Russia (Kennedy and McColl, 2012) measured (see, Bergkvist and Taylor, 2022). Moreover, consumers are Online and offline purchases (Dawes and Nenycz-Thiel, 2014) ‘cognitive misers’, evoking and considering a far smaller subset of Private labels (Dawes and Nenycz-Thiel, 2013) brands than exist and compete in most categories. It is reported that Competing Brands Have Similar User Profiles (Brand User Profiles Seldom consideration sets are often small, two brands or less (Lapersonne et al., Differ) 1995; Shocker et al., 1991). This pattern is also demonstrated in CPG/grocery categories (Anesbury et al., 2017; Uncles et al., 2012) International brands vs local brands (Tanusondjaja et al., 2015) retailing and services contexts like retail banking (e.g., Dawes et al., Healthy vs unhealthy food (Anesbury et al., 2018b) 2009; Honka et al., 2017). This means that even financial services Fresh fruit purchasing (Anesbury et al., 2018a) brands (traditionally seen as being in highly ‘rational’ categories) with Competing Brands Have Predictable Loyalty (Double Jeopardy Law) higher mental availability are thought of, considered, and bought, by B2B (Romaniuk et al., 2021) Fashion online auctions (Chowdhury et al., 2021) more people. Using choice experiments to find Double Jeopardy patterns (Greenacre et al., 2015) Physical availability, on the other hand, is about how easy it is for Brand associations (Romaniuk, 2013; Stocchi, Driesener and Nenycz-Thiel, 2015; people to purchase the brand (Romaniuk and Sharp, 2021; Sharp, 2010). Stocchi et al., 2017) Marketers build physical availability by winning distribution in stores Average spend per buyer (Dawes et al., 2017) and websites, gaining listings on menus, paying for search and display Customer mindset metrics (Mecredy et al., 2021) Stents for surgical procedures (i.e., industrial market) (McCabe et al., 2013) (e.g., on Google and Amazon), offering credit and various ways of Wine varietals (Cohen et al., 2012) paying, providing car parks, increasing opening hours and so on. The Cultural venue/event attendance (Trinh and Lam, 2016) market-based asset of physical availability has a quality dimension as Wine/butter purchase by country of origin (Trinh et al., 2019) well as pertaining to quantity, that is, how easy is the brand to buy, and Purchases in online supermarkets (Trinh et al., 2017) Healthy vs unhealthy food (Anesbury et al., 2018b) for how many buyers. This means in addition to having a relevant Fresh produce purchasing (Anesbury et al., 2020a) portfolio and gaining presence in places where people shop, it also in­ CPG brands in China (Kennedy and McColl, 2012) cludes gaining prominence in these locations (Nenycz-Thiel et al., Private label brand image data (Nenycz-Thiel and Romaniuk, 2014) 2016). Yavorsky et al. (2021) provide an illustrative example of how Online and offline purchases (Dawes and Nenycz-Thiel, 2014) physical availability ‘works’ in the context of retail auto dealerships, Competing Brands Have Predictable Loyalty (NBD/Ehrenberg’s Law of Buying Frequencies) whereby consumers were found to search only a limited number of Blood donations (Faulkner et al., 2016) dealers, and were very likely to buy a vehicle from the dealer Shopping baskets (Martin et al., 2020) geographically closest to them. B2B/industrial purchases (Wilkinson et al., 2016) Put together, the market-based assets view presents a far less Cultural venue/event attendance (Trinh and Lam, 2016) romantic view of the power of brands. Brands that are easier to buy, get Sporting event attendance (Trinh, 2018) Comparing buying for different ethnicities (Trinh et al., 2020) bought by more people, more often (Sharp, 2017b). That is, they are Wine/butter purchase by country of origin (Trinh et al., 2019) known for more situations (mental availability), by more people, and are Fresh food category purchases (Anesbury et al., 2020b) more widely available (physical availability). Big brands have greater Cigarette purchases (Dawes, 2014) physical and mental availability (Romaniuk, 2013), and these brands Several studies over the last decade continue to find the empirical regularities have larger marketing budgets to support their assets. In contrast to STP described in this paper in new contexts. This includes in durables, emerging theory, the market-based asset theory says brands do sell to similar types markets, experiential purchases, and in attitude image data. of buyers and all brands compete head on and share buyers with each other as if they are direct substitutes. produce), luxury brands, wine, business-to-business and industrial Since the publication of How Brands Grow (Sharp, 2010), many goods, charities, music listening, and event attendance. studies have tested and extended the market-based assets theory’s un­ This is unusual in marketing, with an analysis showing that the vast derlying laws about how brands compete in new contexts. Some re­ majority of marketing theories are rarely exposed to more than a single searchers have conducted direct replications of the empirical laws that test, usually only in terms of qualitative direction, and usually by the corroborate the existence of the market-based asset theory for brands in authors who propose the theory and hypotheses (see Kenworthy and consumer packaged goods categories. For example, Steenkamp (2017) Sparks, 2016). concluded: “Never in my 35 years of research have I encountered such a strong 4. Testing market-based asset theory: explaining variation in relation between two marketing metrics. Brands with a large market share brand size have far more buyers than brands with a low market share. Market share increases depend on substantially growing the size of your customer base”. We now show how market-based asset theory helps us to better An illustrative list of studies published since 2010 that test the un­ understand why some brands are far bigger than others, using several derlying law-like patterns for market-based asset theory introduced empirical examples. First, we use market-based asset theory to explain above, are shown in Table 5. The patterns have been investigated in variation in brand size among competing brands within a market. Sec­ many contexts, including in consumer goods, retailers and stores, ond, we investigate how market-based asset theory explains variation in shopping baskets, private label brands, unbranded goods (e.g., fresh 7 B. Sharp et al. Journal of Retailing and Consumer Services 76 (2024) 103566 brand size across markets. brands are large: when non-buyers purchase from the category, they are more likely to buy brands they are aware of (and whose advertising they are aware of). Data were provided by YouGov (2023). Note, we use 4.1. Variation in brand size within a market advertising awareness, not brand awareness for this analysis because it is not possible to split out non-brand users who are aware or not brand In this section we present evidence on the role of mental and physical aware in the YouGov dataset. The number of stores for each retailer were availability in underpinning brand size for two types of retailers. First, sourced from ScrapeHero (2023a). we examine the association between brand usage, awareness and There is again a strong, positive association between the size of a advertising awareness for the largest 20 high street fashion retailers in QSR brand and advertising awareness, as shown in Table 7. The corre­ the United Kingdom. Data were provided by YouGov (2023) for the lation among brand size and ad awareness among each of the three analysis. We tabulate the proportion of respondents in 2022 who re­ customer groups is very high (r = 0.90, r = 0.93, and r = 0.89 among ported they were current customers of these fashion retailers, the pro­ current, former and non-customers respectively). There is also a portions who were simply aware of the brand, and the proportion who noticeable difference in the advertising awareness that larger brands were aware of advertising for the brand. While advertising awareness is achieve among former and non-customers compared to the smaller not a mental availability measure per se (see, Romaniuk, 2013), brands (e.g., McDonalds still receives 26% ad awareness among non- advertising is a primary mechanism for building mental availability. customers compared to 1% for Taco Bell). Awareness across infre­ Therefore, the higher awareness of a brand’s advertising in the popu­ quent (former customers)3 and non-buyer groups is important given the lation is likely an indicator of higher mental availability for the brand. role acquiring new and infrequent buyers in sustainable brand growth. We see in Table 6 a clear association between brand size and brand In addition to larger QSR brands having higher advertising aware­ awareness. Overall, the correlation between awareness and brand size is ness than smaller brands, we also see a strong positive correlation be­ positive and significant r = 0.44 (p < 0.05). For advertising awareness, tween brand size and the number of stores, with a correlation of r = we also see an overall positive association with this metric and brand 0.44. We do note a marked deviation, whereby Subway appears to have size, again with some apparent exceptions such as Primark and Next many more stores than we would expect given its size (e.g., 2227 stores which have low ad awareness for their size, however, the overall cor­ for 7% buying, compared to McDonalds’ 1391 stores for 30% buying). relation between brand size and ad awareness is also positive and sig­ This pattern occurs for Subway in other markets, and is attributable to it nificant at r = 0.78 (p < 0.05). There are some exceptions to the broad having smaller stores, with only a small proportion of stores with drive- pattern, such as River Island and Levi’s having quite high awareness for through. Overall, this QSR data is again consistent with the market- their size, however Levi’s has the benefit of being a long-standing based asset theory, that the principal underpinnings of brand size and clothing brand, sold in multiple stores, before it was a retailer. While growth are mental and physical availability. this analysis between brand size and awareness is not causal, it is consistent with the concept that the market-based asset of mental 4.2. Variation in brand size across markets availability underpins market share. We now consider a second analysis using data for eight of the most The market-based asset theory of brand competition not only ex­ popular Quick Service Restaurants (QSRs) in the United Kingdom. We plains the variation in market shares within a context (as we demon­ again examine the association between brand size and advertising strated above for fashion and QSR retailers), but also explains why awareness. In this example, we use advertising awareness among three global brands can have vastly different market shares in different groups: current customers, former customers and those who have not countries. This is a phenomenon that looks rather odd from an STP bought the brand. The rationale for this approach is to identify if larger theory and global brand management perspective. brands have more mental availability among all these three consumer groups. If larger brands have more mental availability even among those Table 7 who do not buy them, this explicates the explanation for why those large Store coverage and ad awareness of quick service restaurants, UK. Brand % Current # Ad Awareness Table 6 Customer Stores % Current % Former % Non- Brand size, awareness and ad awareness of fashion retailers, UK. Customers Customers Customers Brand % Current Customer % Brand Aware % Ad Aware McDonalds 30 1391 60 40 26 Marks & Spencer 23 95 26 KFC 11 1006 48 26 15 Primark 19 95 7 Dominos 8 1215 47 27 14 Next 14 93 9 Subway 7 2227 29 13 8 Tu Clothing 11 70 10 Burger 6 1092 30 12 7 George 10 86 8 King H&M 10 91 10 Pizza 3 470 35 8 5 TK Maxx 9 93 12 Express Matalan 8 90 9 Papa 2 484 26 9 4 F&F Clothing 8 71 8 John’s ASOS 8 77 8 Taco Bell 1 132 20 5 1 New Look 7 87 5 Average 9 1002 37 18 10 JD Sports 5 92 11 Zara 5 81 3 Brands with more customers have tend to have more stores and have ads aware Clarks 4 89 4 by more people. In other words, bigger brands have higher physical and mental Fat Face 3 73 3 availability than smaller sized brands in the market. Data Source: YouGov Boohoo 3 73 7 BrandIndex UK 2023 © All rights reserved. River Island 3 91 3 Levi’s 3 91 4 Joules 2 50 2 Peacocks 2 78 1 Average 8 83 8 3 The former customers of a QSR brand have likely not ‘left’ it or defected in Brands with more customers are known by more people. There is a strong pos­ that year, they simply have not purchased it for some time. For example, itive correlation between the number of customers a brand has and brand size. although not in the QSR category, past research of consumer goods brands Data Source: YouGov BrandIndex UK 2023 © All rights reserved. shows 80% of brand buyers buy less than once a year (Dawes et al., 2022). 8 B. Sharp et al. Journal of Retailing and Consumer Services 76 (2024) 103566 Consider the case of car brands, which are sold through dealers, where they have invested the most money in building mental and therefore an excellent example of a retailing and consumer services physical availability over decades. We see in Table 8 that in Australia, context. We obtained 2022 market share information in three markets Toyota is the market leader (21% share), it has far more dealers (290) for three well-known car brands with global teams in charge of their than any other, Kia has fewer (177) and Subaru less again (128). In the positioning and brand strategy: Toyota, Kia and Subaru. We also sourced United States, Toyota has the largest market share of the three (13%) information on the number of dealerships each brand has in the three and has far more dealerships than Kia (781) or Subaru (639). Whereas in countries, being a measure of physical availability.4 Metrics pertaining the United Kingdom, Toyota has far lower market share (6%) than in the to mental availability were not publicly available, but some information other two countries, this is reflected in the fact that its dealership on advertising spend was identified and is discussed later in this section. numbers are similar to Kia’s (177 for Toyota, 187 for Kia). And Subaru The results are shown in Table 8. In Australia, Toyota is the market has only 1% share in the UK, with less than half the dealers of the other leader with 21%, whereas in the United States, it has far lower market two brands (73) whereas in the US and Australia it has not that many share at 12% and in the UK it has only 6%. Similarly, Kia is considerably fewer than Kia. larger in Australia than in the UK (7% vs. 4%) and Subaru in the UK has In relation to building mental availability, in Australia (in 2020) only a fraction of the market share it has in the other two markets. Toyota was the number one spender in the category, Kia sixth and How can these across-country differences in market share occur if Subaru was not in the top ten spenders (Nielsen, 2020). In the United strong brands are ‘superior’ as per STP theory? The cars are essentially States, Toyota is the fourth biggest spender in the category, Kia is ninth, the same from one country to another. Toyota, Kia and Subaru have Subaru tenth (Statista, 2023a). Although there is less publicly available global model platforms. We verified that each brand offers approxi­ information for these car brands in the United Kingdom, we see Toyota is mately the same number of models in each country (for 2022 or as close not a major spender there (Digital Intelligence, 2019) unlike the other a year as could be identified),5 so the differences are not due to broader two markets, which is consistent with it having lower market share. or narrower range in one country compared to another. Also, we verified To sum up, where these brands have more mental availability and that the relative prices of these car brands are quite similar in the three physical availability, they have much higher market share. This is a far countries.6 Therefore, these differences in market share across countries better explanation than the STP theory of unique perceived brand cannot be explained by the rationale of STP theory that brands will win differences. if they provide consumers superior value. Why do some brands have vastly different market shares across 5. Conclusion markets? Could the answer pertain to consumer preferences for locally manufactured product? Toyota did make vehicles in Australia, but ‘It doesn’t matter how beautiful your theory is, it doesn’t matter how drastically reduced its manufacturing operation in 2012 and shut it smart you are. If it doesn’t agree with [the] experiment, it’s wrong.’ - Pro­ down in 2017. Subaru and Toyota do manufacturer in the United States, fessor Richard Feynman, Nobel Prize Winner. but this does not explain why one of them has four or five times the In this paper we presented several known, law-like patterns in how market share of the other in that country. Kia cars are all made in Korea. brands compete and challenge that their existence would not be possible So, these market share differences are not due to ‘buy local’ preferences. if STP theory is to be believed. The views that come from STP theory Could Toyota’s large share in Australia be because the big players, GM have been challenged in the past due to the lack of compelling evidence and Ford left Australia? Yes, that did help their competitors, but there (Wright and Esslemont, 1994) and there has been no convincing evi­ are many other brands such as Hyundai, and Volkswagen, and Ford still dence in the three decades since, despite many academics and marketers has 15% market share. And this does not help us to understand why the still accepting and practicing the mantra to segment, target and market share differences occur between the other two markets. (differentiate their) position. The market-based asset theory presented Instead, the answer is that these brands are bigger in the countries in this paper challenges traditional marketing thinking and can provide marketing practitioners a clear framework about how brands compete and grow, based on empirical evidence. Table 8 Scientific theory is retained until it does not fit the known evidence. New vehicle sales car brand market share & number of dealerships. Over the past 50 years a considerable amount of evidence, from an array Brand % Market Share # Dealerships of sources/perspectives, has coalesced into a non-intuitive but coherent picture of how brands compete. In this article we have only had space to Toyota Kia Subaru Toyota Kia Subaru give a brief overview of this evidence and introduce the current theory Australia 21 7 3 290 148 128 that fits with this evidence. No doubt the market-based assets theory United States 13 5 4 1270 781 639 presented in this paper will be further adapted in light of evidence, and United Kingdom 6 4