Don’t try to manage brand loyalty Abstract The objective of this paper is to review some 50 years of buyer behavior knowledge, leading to a discussion of what it might mean for marketing (and research) going forward, specifically addressing whether brand loyalty should be dispensed with as a marketing goal. The methods involve a review of academic literature on brand loyalty focusing on patterns of brand buying in markets where consumers make repeated choices between alternative offerings. It is well understood that over time, the incidence of 100%-loyal buyers for any brand is small (and predictable). Most buyers are polygamously loyal: they have a habitual portfolio of a few brands, buying some more often than others, and sampling their portfolio “as-if at random”. As a consequence, from a brand perspective, most of the volume comes from many occasional purchasers who are buying other brands more often. Results take the form of empirical generalizations. It might seem reasonable to try to target occasional buyers of a brand to make them buy the brand more often (increase their loyalty). But in general the evidence on brands that do succeed in growing shows that they do so whilst attracting even more occasional buyers! Growth almost never comes just from increasing the loyalty of existing buyers. It’s not that loyalty isn’t important, but that it takes a particular structural form, which it seems, cannot be manipulated directly. These results raise important questions about whether marketers waste time, effort and money on hapless tactics such as loyalty programs and relationship marketing directed at small subgroups of users. The conclusions address marketers' efforts to develop loyalty, as opposed to efforts to attract as many buyers as possible regardless of whether they are loyal. They also explore implications for targeting, and communication strategy. Key Words: Brand loyalty, Habitual behavior, Stochastic 1 Introduction and Objectives: Brand Loyalty in the Literature Over the past 50 years, a great deal has been written about brand loyalty. Early work defining the concept of loyalty was followed by a variety of methods to measure, manipulate, improve and capitalize on it. Modeling based on individual-level-data was developed using brandloyalty as an independent construct in discrete-choice models (Guadagni and Little 1983). The purpose of this paper is to go briefly through some of the history, but without great detail because much of the ground has been previously covered—see for example Jacoby and Chestnut’s “Brand Loyalty: Measurement and Management” (1978), Mellens et al’s, 1996 “A Review of Brand-Loyalty Measures in Marketing” (1996), or Bennett and Rundle-Thiele’s “The brand loyalty life cycle: Implications for marketers” (2005). All three are admirable assemblages with insightful commentaries upon the ever-growing literature on brand loyalty. We briefly review the academic literature on brand loyalty in Part 1 and then focus on patterns of brand buying in markets and determine whether there is any indication that brand loyalty levels are predictable by anything other than a simple stochastic model. In other words, we do not question that brand loyalty exists and is an important concept that marketers would do well to understand as a metric, however, we strongly doubt that brand loyalty is subject in any meaningful degree to marketing interventions. Rather, brand loyalty exists in categories as a function of the marketplace in which levels of competition and the size of the various brands dictates the overall market level of brand loyalty. Large share brands have loyalty levels that are slightly above the average market level, while smaller brands have loyalty rates that are slightly below average. There may be some exceptions to these rules, but they are mostly small and explainable. These findings are laid out in section 2. 2 Part 1. Brand Loyalty The standard form for beginning a paper on brand loyalty is something along the lines: The long term success of a firm rests on the ability of its brands to attract and retain customers. To put this into context, for the candy company Cadbury’s a single share point in the UK alone is worth about £18 million per year—that is about 60 million purchases, and for a company like Cadbury’s with about 35% market share, that amounts to about 1 billion purchases each year in the UK alone, Clearly there is a need to continually generate repeat buying, or loyalty. In competitive markets, both attracting and retaining customers are critical tasks, but retention attracts more attention because it seems more manageable. Marketers are naturally interested in finding ways to get their customers to buy more and for longer, and presumably the converse is true as well--making their competitors’ customers buy less or switch. And from a financial perspective loyal customers may also be more rewarding while also being more efficient in marketing terms. Some of the reasons put forward to concentrate on loyal customers include: • Keeping an existing customer costs about one sixth of attaining a new one. (Rosenberg and Czepiel 1983) • Brand loyal customers are less price-sensitive and therefore have higher customer value than occasional buyers (see e.g. Krishnamurthi and Raj 1991; Reichheld and Sasser 1990; Kumar and Shah 2004). • Positive word of mouth from loyal customers is another potential mechanism for firms to save on marketing costs (East et al 2008). • Loyalty rates tend to rise with market share, and high market share in turn is associated with higher rates of return on investment (Buzzell et al 1975) • Brand loyalty feeds into firm profitability (Reichheld and Sasser 1990, Reichheld 1996) and brand equity (Srivastava 2002) and can be exploited through ‘customer-driven marketing’, relationship marketing, direct marketing, and loyalty programs. 3 One of the core issues in debates around brand loyalty is whether purchasing patterns or behavioral measures capture the true meaning of loyalty. Many authors over the years have argued that behavioral measures of loyalty on their own are inadequate because ‘true loyalty’ must include an attitudinal (e.g. preference) component (Jacoby and Kyner 1973, Dick and Basu 1994, Day 1969, Mellens et al. 1996). In a standard work on loyalty Jacoby and Chestnut (1978) surveyed over two hundred research projects and identified dozens of different ways to describe loyalty. They reached the almost rueful conclusion that, “…it is extremely interesting to find, upon reviewing this literature that no one quite agrees on exactly what…loyalty is.” Their own conceptual (as opposed to operational) definition includes both attitudinal and behavioral components. Brand loyalty is: The (a) biased, (b) behavioral response (c) expressed over time (d) by some decision-making unit, (e) with respect to one or more alternative brands out of a set of such brands, and (f) is a function of psychological (decision-making, evaluative) processes (Jacoby and Chestnut 1978, p. 80). This definition assumes that behavior is habitual and based on rational or rationalizing processes of experienced buyers. It also makes a necessary simplification that attitude has relevance through its effect on behavior. The definition is worth examining in some detail: The fact that the definition specifies ‘some decision making unit’ means that it is specified at the “individual” level (whether that is a person, or a household or whatever). However, it says nothing about how loyalty sums over individuals to give brand loyalty. ‘Biased behavioural response’ implies that some consumers buy at different levels from others, or from a norm. The natural tendency might be to think of this as a positive bias. But any response that differs from, say, the mean tendency (brand share) would be biased. This could include non-buyers (zero tendency). Hence every person could be said to have a loyalty to every brand (which is how the Dirichlet model is specified). ‘Behavioural response expressed over time’ implies that consumers have an underlying and systematic tendency to buy a certain brand or group of brands. It doesn’t specify whether this is a zero order process (i.e. fixed ongoing propensity, where the choice of brand on one purchase occasion has no affect on the probability of buying that brand or any other brand at the next occasion). If this 4 were true, then brand loyalty would be beyond the influence of marketing interventions. Brand loyalty also entails actual purchases of a brand; intention to purchase is not sufficient in this definition. ‘A set of such brands’ means both that loyalty only arises in a choice situation, and that they may actually be loyal to more than one brand, a phenomenon observed by many marketers and researchers (e.g. Ehrenberg 1972; Jacoby 1971). This is especially true for low involvement goods, in which the consumer often does not evaluate brands on a continuous scale, but classifies them as ‘brands I buy’ and ‘brands I don’t.’ Similar behavior has also been observed for infrequently bought goods such as televisions and mobile phones (Bennett, 2009). If more than one brand is acceptable, an individual might be indifferent between them and exhibit loyalty to a group of brands rather than to a single one. The obvious problem that multi-brand loyalty presents for marketers is that it makes it difficult to distinguish purchasing within a portfolio from brand switching. The fifth condition, ‘Function of a psychological (decision-making, evaluative) process’ assumes that brand loyalty involves mental activity, at least at some point in time. It assumes that consumers process information and reach decisions to buy a particular brand. This is in line with standard views of consumer behavior that will be discussed further below. It also recognizes that consumers do not always seek information actively, but over time may develop a commitment towards a brand and become brand loyal, with loyalty being shown as consistent repurchase of a brand. Here Jacoby and Chestnut’s composite definition bridges the divide between those who feel that commitment is an essential component of brand loyalty (e.g. Mellens, Dekimpe and Steenkamp 1996, Dick and Basu, 1996, Aaker and Joachimsthaler 2002), and those who view loyalty as a measure of behavior (e.g. Foxall 1987, Ehrenberg, Uncles and Goodhardt 2004). The former feel that for consumers to be considered as truly loyal they must not only act loyally by repeatedly purchasing a brand, but must also have strong, enduring positive attitudes towards it (see Amine 1998, Samuelson and Sandvik 1997). In this view ‘True loyalty’ may be distinguished from e.g. ‘spurious loyalty’ (those who buy a brand without holding favorable attitudes (Dick and Basu, 1994)), and this is important because these customers may be vulnerable to competitors’ marketing efforts. Moreover, a 5 purely behavioral definition of loyalty fails to explain the causes of loyal behavior. A long-running debate: Attitudinal Loyalty The debate over attitude and behavior goes back to the 1940s when Guest (1944) defined attitudinal loyalty as consistent preference over time (Guest 1955). This definition derives from the personal or relational view of loyalty as expressed between people (as distinct from expressions towards goods, as in buying them). People have relationships with each other and loyalty implies a notion of sticking with a relationship even in hard times, through thick and thin, of being supportive even with misgivings. But in commercial or buying contexts it is perhaps a stretch to talk about being loyal to a brand of toothpaste or breakfast cereal through good times and bad. After all consumers may behave loyally without strong feelings or attitudes of attachment. Aside from brand managers whose career depends on a brand, few consumers will admit to strong emotional connections to a toothpaste or cereal, nor that they have any meaningful relationship with these products. Even so, these same consumers would admit to buying the brand repeatedly over the years. They exhibit loyal behavior. The attempt to explain why people like and buy particular brands hinges on affective response models of decision-making, in which a central idea is the concept of attitude. The classic definition of attitude is given by Allport (1935): “The mental process by which an individual -- on the basis of past experience and stored information – organizes his perceptions, beliefs and feelings about a particular object and orientates his future behavior.” In attitudinal research, loyalty is usually based on the assumption that consumer beliefs and feelings are antecedent to purchase. As a result such research is designed to understand the involvement and commitment of consumers - why the consumer has positive or negative attitudes towards a brand. The main areas of interest in this field are the cognitive and affective processes that underlie purchasing behavior (Fishbein & Ajzen, 1975) and the reasons behind the individual consumer’s attitudinal loyalty. 6 Attitudinal measures are based on the stated preferences, commitment or purchase intentions of consumers. These measures make it possible to separate out the relevant decision-makers and give insight into the motivations for the consumer’s choice behavior. It is also true that people are capable of rationalizing and that situations arise in shopping (sales, promotional offers, stock-outs, etc.) so that attitudinal measures might not always be accurate reflections of reality. Hence the validity of the attitudinal measures depends on the strength of the relationship between attitude and behavior. An implied question about habits and attitudes is whether they are changeable. Baldinger and Rubinson (1996) studied steadfastness in consumer loyalty and whether brand loyalty can be predicted from attitude. While they were not clear about how they operationalize attitudes towards FMCG brands, the results showed a strong correlation between behavior and attitude: the few buyers with strong attitude mostly showed high brand loyalty, while most of those with weak attitude showed low loyalty. But the vast majority (70-80%) of each brand’s buyers had weak or medium attitudes towards it and bought it at low to medium levels. And durability of both attitudes and buying loyalty was poor; of those who had high loyalty in year 1, only about half remained highly loyal in year 2, thus revealing that the vast majority of buyers are neither behaviorally nor attitudinally very loyal towards single brands. Subsequent studies have confirmed this finding (e.g. Dall’Olmo Riley et al 1997). Behavioral Loyalty Behavioral loyalty was defined in the 1950s and measured in a variety of ways, such as the portion of total purchases devoted to one or two brands (Cunningham 1956). This approach was developed further by e.g. Ehrenberg (1972) and Bass (1974) who advocated a stochastic view of buying behavior that holds that consumers purchase goods in random patterns and do not employ much rational thought in doing so (Bass 1974, Hoyer 1984). This work was enabled by the growth of the panel data industry that tracked individual purchases made by thousands of households over many years. These data were originally concerned with consumer goods purchasing in steady-state markets. The wealth of panel data enabled the seminal contributions of the Negative Binomial Distribution (NBD) and the Dirichlet 7 (Ehrenberg, 1972, Goodhardt, Ehrenberg and Chatfield 1984). These models were used in many replication and extension studies to describe purchasing behavior and predict repeat purchase rates (behavioral loyalty) with great accuracy over a wide range of product categories, countries and market conditions (see Ehrenberg and Uncles 2000). The use of stochastic modeling helped marketers to understand how people buy brands in well-established or steady-state markets (Uncles et al 1995) and later work into developing markets (Bennett 2007), subscription markets (Sharp and Wright 2000) and into less frequently bought categories of goods such as cars (Terech et al 2003, Bennett 2005) and TVs (Bennett 2009) revealed that purchasing behavior patterns generalize even under highly varied conditions. The behavioral camp views attitude as a consequence rather than an antecedent of purchase behavior. It holds that buying behavior is a result of instrumental conditioning and that buying behavior alone is capable of explaining brand loyalty, and indeed is better at predicting future loyalty than is attitude (Bhattacharya 1997, Sharp, Sharp and Wright, 2000). In this view, buying behavior is essentially habitual, routinized, and not generally subject to much ongoing thought or decision-making. This behavior, plus experience with the product or the brand, gives rise to attitudes about the brand, in the process influencing commitment, preferences and a propensity to repurchase. As a result attitudes are treated as of secondary importance because they are in the main, consequences of behavior, (e.g. Cannon, Ehrenberg and Goodhardt 1970, Franzen 1994). This notion of behavior harks back to the initial debate around rationality and problem solving. Dewey (1930) discussed the role of habits in human behavior concluding that: “Habits are conditions of intellectual efficiency…Outside the scope of habits, thought works gropingly, fumbling in confused uncertainty… This view of habits was advanced by Katona (1953) who proceeded from psychological principles to develop six propositions that he claimed were to some extent, findings or empirical generalizations. His first four propositions were: 8 • Problem solving behavior is a relatively rare occurrence…. • The main alternative to problem solving behavior is not whimsical or impulsive behavior. When genuine decision-making does not take place, habitual behavior is the most usual occurrence: people act as they have acted before under similar circumstances, without deliberate choosing. • Problem-solving behavior is commonly recognized as a deviation from habitual behavior. • Strong motivational forces--stronger than those which elicit habitual behavior—must be present to call forth problem-solving behavior. Being in a “cross-road” situation, facing “choice points” or perceiving that something new has occurred…. Thus for the vast majority of purchase occasions, decisions have long since been made, and consumers behave (buy) without much thought or involvement. This applies to frequently bought goods (FMCGs and convenience goods) but it may also have application to durable goods: the fifth car that a person buys, or the third TV because these repeated behaviors are generally not presenting totally new problems for solution. 9 Composite Loyalty By the 1960s, many researchers, notably Day (1969) proposed that loyalty should be viewed as a composite of both behavioral and attitudinal concepts. Day described loyal behavior as “Consisting of repeated purchases prompted by strong internal disposition.” The rationale was that including both attitudinal and behavioral components might improve construct validity--the behavioral elements described what people bought, and the attitudinal elements helped to explain why they did so. Dick and Basu (1994) advanced the case for composite measures by describing loyalty as the relationship between a consumer’s attitude towards an entity (brand, service, or shop) and patronage behavior. In this work, each consumer’s relative attitude had two dimensions: the strength of the attitude and the degree of attitudinal differentiation among competing brands. This theory allowed marketers to distinguish between various types of loyalty (true loyalty, latent loyalty, spurious loyalty) and to identify possible drivers of loyalty. Further research has explored loyalty not just to brands, but also to the shops that sell them, e.g. (Cunningham 1956, Cunningham and Ross 1961, Dunn, Reader and Wrigley 1983, Kau and Ehrenberg 1984, Leszcyc and Timmermans 1997, East et al, 1995, Brewis-Levie and Harris 2000). East et al (2000) suggest customer loyalty represents a customer’s commitment to a brand, shop or supplier, based on a strong favorable attitude manifested in consistent patronage. In doing so they criticize the Dick and Basu model, saying that while it is widely cited, it is seldom tested. With this in mind, researchers have continued to broaden loyalty by introducing new dimensions and metrics (Athiyaman 1999, Bloemer and Kasper 1995, Butcher et al 2001). It has also been argued that different definitions should be used depending on the category, e.g. behavioral measures for stable consumer goods markets and composite measures for less stable, higher involvement categories (Rundle-Thiele and Bennett 2001). In addition to research to describe different patterns of loyalty, and differing ways of being loyal, research has also been directed at understanding the factors that promote loyalty. Two recurrent themes in the attitudinal side of this research are satisfaction and trust. Many authors claim these factors are related (e.g. Bolton 1998, Cronin et al 2000) but some have also criticized this stream by showing that a high degree of satisfaction does not always translate 10 into high loyalty (Mittal and Lassar 1998). On the behavioral side of the debate, the factors that most affect loyalty positively are the stability of the category (Bennett 2007), and size of the largest brand(s) or double jeopardy (Ehrenberg, Uncles and Goodhardt 2004). However, the view that attitudes are vital to any definition of loyalty has been challenged as being at odds with basic epistemological principles and empirical evidence regarding the stability of measures of attitude (Sharp Sharp and Wright 2000). These measures may be very steady at the aggregate level, but changeable at the individual level—attitudinal response rates towards brands show average repeat-response rates of only 50% (Baldinger and Rubinson 1996, Barnard et al 1986, Dall’Olmo Riley et al 1997) which implies that consumers are inconsistent or fickle, and that attitudes are no more stable or enduring than situational variables. Indeed this points to why studies that use attitudes to predict future behavior generally have very poor results (see Kraus 1995, Wright and Klÿn 1998). Despite huge amounts of time, effort and money poured into research into brand loyalty, managers are still not clear on how to build and maintain it (Gounaris and Stathakopolous 2004). This is partly because of changes in the marketing environment. In the early days of brands in the late 1800s, the appeal of new brands grew rapidly, peaking in the consumer boom that followed the Second World War. From then on, growing choice, globalization and steadily rising quality of available alternatives led to increasing product parity and the ubiquity of multi-brand loyalty (Dekimpe et al 1997, Ehrenberg et al 2004). At the same time, by the 1980s the baby boom generation, who had had broader product, brand and consumption experiences than their parents seemed much more willing to switch between brands (Zaltman 2003). Over the past two decades, generation X, Y and so on, long used to brand, product, marketing and media proliferation, have shown themselves to be rather blasé about brands (O’Loughlin and Szmigin 2006) to the point where research effort and managerial attention has shifted away from brand loyalty because efforts to influence it have failed (Divett, et al 2003), towards sales and market share (Dekimpe et al 1997, Zaltman 2003). This idea of declining brand loyalty is usefully summarized in Table 1 (Bennett and Rundle-Thiele 2005). 11 Table 1, the Brand Loyalty Life Cycle Era Characteristics Implications for brand loyalty Birth of brand Loyalty 1870-1914 • Introduction of Quaker Oats, Gillette, Pears, Coca-Cola • Brands offered consistency in a period of varying product quality • Branded goods offered growth to organisations • High initial resistance to brands from retailers and consumers • Advertising assists organizations to increase both their respectability and market share • Branding assisted customers to distinguish between the variety of products on offer • Branding reduced risk • Customers were more likely to re-buy brands that had proved satisfactory in the past Golden age of brand loyalty 1915-1929 • Customers were grateful for improved quality brands offered • Value of brands recognized by retailer • High levels of brand awareness • Highly creative advertising campaigns • Cynicism towards advertising emerges towards the end of the golden age of brands • High incidence of sole loyalty • Functional brand loyalty • Trust in brands reduced consumer resistance to brands, thus increasing loyalty • As availability of branded products increased, customer loyalty increased Latent brand loyalty 1930-1945 • Depression of 1929 and World War II saw a reduction in brand availability • Brand loyalty hit by situational factors • Lack of availability altered consumer habits • Consumer preferences towards brands increased despite inability to purchase The birth of multi-brand loyalty 19461970 • Comeback of manufacturer brands after WW II • Explosion of new products and in retail outlets • Increased threat from generic brands and discount stores • With improved in product quality from competition, brand differentiation was reduced • Baby boomers begin to drive consumer behaviour trends • Beginning of multi-brand buying • At the start of the era, after the war, consumers returned to their preferred brand when it became available again • As differentiation declined and choice increased, consumers were increasingly multibrand loyal • Increasing price sensitivity Decline in loyalty 1971present • Multi-brand loyalty dominant • Intense competition between an increasing array of brands and alternatives • Very low levels of differentiation • Generic brands increasing market share • Majority of new products offer incremental changes and minor product modifications • Brand communities (Harley-Davidson, Apple) • Lower risk in brand switching • Some brands are bought to convey self-identity, rather than just to guarantee consistency of quality • Consumers demanding experiences not just a product • Increased expectations from consumers • Brand loyalty levels in fast-moving consumer goods has declined • The incidence of inertia increases • As consumers become more demanding, dissatisfaction increases • Brands that are functional and low involvement may have reduced loyalty • Brands that convey image and self-identity may have higher loyalty A recurring theme in the literature bemoans the steady loss or erosion of brand loyalty (see Dekimpe et al 1997, Capizzi and Ferguson 2005, Divett, et al 2003, Trout and Rivkin 2000). While the decline of brand loyalty is debatable, these articles reflect the fact that the management of brand loyalty has not progressed very far. That is, marketing interventions have generally not had much effect on consumers in so far as affecting their levels of brand repurchasing. 12 One potential explanation for the stubbornness of loyalty is that by and large, many categories are remarkably stable in terms of underlying category structure. That is, brand shares remain remarkably stable for years at a time. Even though marketing practitioners invest very substantial sums to influence consumer behavior through communications, sales promotions, loyalty schemes, etc., in efforts to grow their brands, the data reveal that even over the course of a year or two brand shares in most FMCG categories remain very nearly stationary (Dekimpe and Hansens 1995), with any gains or losses being only temporary (Srinivasan and Bass 2000). New research reveals that brand share stationarity is the norm over the long term as well--seventy five percent of brands (n = 106) remained within 3 share points of their starting average over six years (Graham 2009), Moreover, the five percent of brands that managed to grow by six share points or more over as many years, achieved their growth not through manipulation of the promotional mix, but through changing their brand architecture or some discontinuous innovation. This implies that tactical marketing efforts make very little difference to brand growth. At best, ongoing marketing mix programs may be described as brand share maintenance. In the exceptional cases when brands do grow, they do so not by increasing the loyalty of existing buyers, but by adding new buyers, or increasing the penetration (Baldinger et al 2002, Graham 2009). When such growth does occur, it does not alter the underlying structure of the market, including such patterns as double jeopardy, rather it simply raises all the brand performance measures of the growing brand in line with that brand’s share, including its repeat buying rate. The, the only empirically verified way to increase loyalty is to increase brand size and move further up the double jeopardy line (Habel et al 2005). Brand loyalty is an effect of brand size, and not the other way round. 13 Part 2: Empirical Generalizations or the art of the real Andrew Ehrenberg and his colleagues have followed an alternative data-driven approach to the analysis of brand buying behavior for the past several decades. They have described the behavior of consumers in competitive markets and established that how often people buy a product and the brands or products that they buy is largely habitual, with individual behavior aggregating to measures of brand performance which follow regular law-like patterns (e.g. Ehrenberg, Uncles and Goodhardt, 2004). This approach stems from the well-known NBDDirichlet model of purchase incidence and brand choice in established competitive markets. The finding that most markets behave in a predictable ‘Dirichlet’ manner led to the conclusions that: • Loyalty (the propensity to purchase) at the individual consumer level has multiple causes. However, it produces a common effect at the brand level, which is captured by many different measures, • Competing brands differ little in the levels of loyalty they enjoy Brand loyalty and switching studies have generally been based on purchase sequences over time (e.g. Kau and Ehrenberg 1984, Colombo and Morrison 1989, Keaveney 1995, Leszcyc and Timmermans 1997) and much thought has been devoted to the difference between buying two brands in the sequence ABABAB or in the sequence AABB. In both cases A and B are bought 50% of the time, but the first customer looks like a serial switcher, while the second seems to have had a complete change of preference. In fact however, neither pattern is very common. Brown (1953), Cunningham (1956) and Ehrenberg (1988) have shown repeatedly that most households are loyal to more than one brand and buy regularly from a portfolio or repertoire of brands within a category, typically buying one brand more than others. Brand portfolios are implicit in buyer classifications (Aaker 1996). At one extreme are noncustomers who do not buy a brand at all and at the other, committed customers who buy only one brand. All those in between are multiple brand buyers, moving between brands perhaps in response to availability, promotions, changes in price, advertising, or a desire for variety, with the result that they spread their loyalty amongst several brands. 14 Disloyalty is a younger field and tends to focus on the ending of relationships. This is discussed in various contexts such as switching (Keaveney 1995, Mittal and Lassar 1998, Roos 1999), customer exit (e.g. Bolton 1998), termination (Hocutt 1998), customer defection (Colgate et al 1996, East et al 1997, Garland 2002), erosion (East and Hammond, 1996), demotion (East and Hammond 2003) attrition (Szymigin and Carrigan 2001) and churn (Keaveney 1995, Sharp et al 2002). Each of these terms means something slightly different because the industries in which the studies were based use their own particular descriptions for the ending of buyer--seller relationships. Much of this literature analyzes services or business-to-business where relationships are contractual and therefore defection easier to spot (Heskett 1995) and customers typically have “a repertoire of one” i.e. they subscribe to one brand for all their category requirements until they defect to another (Sharp and Wright 2000). With contractual relationships, there are also generally high barriers to switching, hence higher costs and both customers and suppliers view break-ups as significant events. Generalizations in marketing The developments in research in marketing have been much influenced by explanations of different types of phenomena that can be generalized despite apparent differences in conditions. These phenomena are known as empirical generalizations and are now considered integral to our understanding of buyer behavior. Such generalizations form a basis for reusable knowledge in marketing. The problem of the impossibility of conclusive proofs of universal statements or theories, was discussed by Popper (Thornton, 2002) who said the first criterion for any empirical theory is that it be falsifiable; that is, it must be possible to conceive of an observation which would contradict the theory. Popper claimed Falsifiability was the key point of demarcation between science and non-science and advocated that rather than trying to prove that a theory is true, we should instead try to show that it is false by subjecting it, along with competing theories, to the strongest possible tests to determine which provides the best results and has the fewest serious falsifying instances. For a theory to be falsifiable it must be capable of making testable predictions which can be compared to observations and also to the predictions of other theories. The best 15 theories will have: (i) fewer serious falsifying instances than competing theories; (ii.) make a wider variety of predictions; and (iii) successfully predict some outcomes which are at odds with the predictions of other theories. Furthermore, if we cannot conceive of some evidence to refute a theory, we must say that it is unscientific, and provides no basis for progress. Falsification and boundary conditions do not render theory useless (though that may happen), rather they lay the basis for development of new theories. Without the possibility of falsification, a theory can never be contradicted, so there is no possibility for empirical testing or failure, and therefore no need to improve the theory. A criticism of falsification and other empirical approaches is that the observations on which they depend are subjective and uncertain and so cannot be reliably used to either falsify or justify theories. This subjectivity and uncertainty arises because of such factors as errors of measurement (sampling error, fallible instruments, data collection procedures), confounding influences (extraneous activity, seasonality, economic cycles, etc.) and subjectivity on the part of the observer (bias, error, incompetence, dishonesty, etc.). An elegant solution to these issues is available through varied replication (Wright and Kearns 1998) in which: a. further samples reduce sampling error, b. different data collection methods make observations somewhat independent of bias arising from any one particular method, c. different times or places reduce the effects of confounding influences, d. additional observers reduce subjectivity, bias, error or dishonesty of single observers. e. meta-analysis establishes norms and an understanding of the contributory factors and possible exceptions to theoretical application. Marketing has few replicated or extended studies and when they have been done, they generally offer little support for the original study (Armstrong 2000). This emphasizes the need for replications to provide an effective method for ensuring that research results are meaningful and reliable. 16 Core empirical generalizations of brand buying behavior Bass (1995) defined empirical generalizations as a pattern or regularity that repeats over different circumstances and that can be described simply by mathematical, graphic or symbolic methods. This pattern repeats, but need not be universal over all circumstances. Ehrenberg (1982) notes that: The law-like relationships of science are descriptive generalizations, often at quite low level. But the variables which do not appear in the equation greatly aid our understanding (e.g. That the type of gas, the type of apparatus, etc., do not matter). They are also the building blocks of higher level theory and explanation. Ehrenberg characterizes law-like relationships as having the following properties: - they are of limited generality, rather than universal - they are approximate, rather than exact - they are not necessarily derived from theory - they are broadly descriptive rather than causal Thus empirical generalizations have the potential to lead to a robust marketing research tradition (Morrison and Schmittlein 1988). The concept has advanced by various techniques to investigate different aspects of consumer behavior through studies with varying methodologies and goals. Blattberg et al (1995) for example analyzed empirical generalizations related to sales promotion by looking at the effects found consistently in a large series of research projects. Dekimpe and Hanssens (1995) also derived several empirical generalizations about conditions under which markets are likely to evolve. They conclude that generalizations provide a foundation to study the long-term effectiveness of marketing programs. Farley et al (1995) illustrated how empirical generalizations can be produced through the use of meta-analysis, for example concerning parameters in models of advertising, pricing, diffusion, and consumer behavior. While the above studies and a few others (e.g. Reibstein and Farris 1995, Kaul and Wittink 1995) attempt to find results that are generalizable, these are the exceptions. Most studies and attempts to build analytical models are isolated cases without follow up 17 (Leeflang and Wittink 2000). While the lack of generalizations is a problem, the studies above give some encouragement by showing the wide range of topics (e.g. strategy, market response, competition, diffusion, innovation, relationships) for which empirically testable results may be generated. In this line Barwise (1995) advocates replication studies across multiple sets of data over a range of conditions which then form the foundations of scientific knowledge. His characteristics of good empirical generalizations are: - Scope - Precision - Parsimony - Usefulness - Link with theory Generalized empirical research findings have a strong link with theory. For example the NBD-Dirichlet model (Dirichlet for short) of choice behavior in competitive marketing situations (Goodhardt, Ehrenberg, Chatfield, 1984) provides one such link that has been confirmed, verified and extended in over 30 years of research across dozens of categories and markets. The model is simple in that an item’s market performance measures— penetration, repeat-buying, category share, and switching patterns--are all determined simply by the item’s market share. This in effect means that there is no impact of specific marketing–mix factors or product attributes on the market performance measures. In theory, these factors and attributes simply affect the item’s market share and through that its other performance measures. In the model, each consumer has certain propensities or probabilities to buy the available brands. These probabilities are assumed to be individually steady (at least for the time being), but very heterogeneous, i.e. differing greatly across consumers. The model itself is defined for markets that are both stationary and non-partitioned (i.e. with steady marketshares and no clustering of particular brands). The model only purports to describe what markets are like when they are approximately steady and non-partitioned. The Dirichlet is a descriptive model, which aims to describe patterns that are observable in the data, and from these develop empirically grounded benchmarks and insights for evaluating marketing action. Like other models of consumer buying behavior (e.g. 18 Hendry, first order Markov, NBD etc.) it has received coverage in the modeling literature (Guadagni and Little 1983, Leeflang and Wittink 2000), but with little detailed discussion. Brand Buying Behavior for Individual Brands In the past much research has been carried out to generate empirical generalizations in the area of loyalty to brands. Some of the main characteristics of buying behavior for brands are discussed below. It has been repeatedly observed that the (5 or 10) leading brands in packaged goods (FMCG) product categories usually have the following characteristics: • The market shares of individual brands differ greatly • Brands have very different numbers of buyers. This is also in line with their market shares, i.e. brands with bigger market shares have more buyers (higher percentage penetration) • In contrast, the average purchase frequencies for brands are much more similar (that is the number of times a brand was bought over the analysis period does not vary much), • There is a small downward “double jeopardy” trend with market share, i.e. smaller brands not only have fewer buyers, but those who buy them do so slightly less often, • The average amount bought per purchase occasion varies little from brand to brand, • Most buyers of a brand buy it very infrequently. Most buyers have a repertoire of brands from which they habitually choose, spreading their purchases across several brands, • Period-by-period repeat buying is much the same for different brands, and it tends to be low. 19 Buying Competitive Brands Most consumers tend to buy more than one brand over a period of time: e.g. McDonald’s and Burger King, Ford and Vauxhall, and so on. The patterns are again very much the same for different brands and products, and there are several key reasons for this are: • Few consumers of a brand are 100% loyal to that one brand over any extended series of purchases (and there are even fewer 100% loyals for smaller brands (following the double jeopardy effect)). • 100% loyal buyers usually do not buy the brand heavily • A brand’s customers buy other brands in total far more often in a period like a year or over many purchases than they buy the brand itself • Which other brands a brand’s customers also buy is mostly much the same from brand to brand • The dominant factor for the purchase duplications between brands is the penetration of each brand, in a near constant proportion (i.e. the Duplication of Purchase law) • Some clusters of brands or sub-markets with higher or lower duplications can occur as systematic deviations from the Dirichlet pattern. Sub-markets arise particularly for functionally different types of product, e.g. regular or diet soda, regular or unleaded petrol, but partitioning is often remarkably slight and is a second-order effect • Brand shares are much the same for the lighter and heavier category buyers, i.e. brand choice and purchase frequency are largely independent. Taken altogether, the main effects of the above generalizations are that: • The different brand performance measures tend to vary together • This correlation is largely a matter of market share, i.e. big brands score higher than smaller ones. • The loyalty measures of different brands tend otherwise to be very similar Major exceptions to these effects and to the more detailed regularities above are rare, and usually explainable. The detailed patterns of consumers’ buying behavior are much the 20 same for different brands, products, and services. Thus there are heavy, light and nonbuyers of any category and they choose between brands which in a competitive market are mostly similar. In practice, repeat-buying and brand-switching patterns are dominated by how big each brand is (its market share), and not, or hardly, by any idiosyncratic attributes or values of the brands (Chatfield and Goodhardt 1975, Uncles, et al, 2003). An important reason for this is that consumers themselves are mostly very experienced in terms of buyer propensities. Consequently buying the product and brand in question is largely habitual. This is a condition known as repertoire market buying where people form personal repertoires of three or four brands from which they habitually choose one brand more often than others. Within such a framework of mostly steady but divided loyalties, individual purchases are made in an apparently irregular or even “as-if random” manner. Portfolios of such habitual brands can differ greatly from one person or household to the next. But this heterogeneous behavior aggregates to various brand performance measures that follow much the same regular or law-like patterns for different brands and products. These patterns have been found in over 50 product and service categories ranging from grocery products to prescription drugs (Stern 1995) and motorcars, as summarized in Table 2. These patterns hold under varied conditions and are found across different datasets, they are also empirically generalizable and provide usable norms or benchmarks for testing the patterns from new sets of purchasing data. Table 2 Varied conditions for Dirichlet-Type Patterns Products and Services Time Space and People Food, Drink, cleaners and personal care OTC medicines, Prescription drugs Petrol, Aviation fuel, Cars, PCs Retail Shops, Chains TV episodes, Programs and Channels Different points in time, 1950-2004 Different-length analysis periods Britain, USA, Europe, Australia, etc. Light and heavy buyers, subgroups Household or individual purchases Brand and Product Variants Market Conditions Large and small brands Pack-sizes; flavors, forms, formats Private labels Price bands Near steady-state markets Dynamic markets (for loyalty measures) Non-partitioned markets Partitioned sub-markets Source: Ehrenberg, Uncles and Goodhardt (2004) 21 Conclusions There is no doubt that brand loyalty exists. Some consumers exhibit and express ongoing preferences for a brand and buy it more frequently, although rarely exclusively. However, despite decades of effort to massage, improve and heighten loyalty for individual brands, it remains stubbornly connected to each brand’s market share. There is little or no evidence that brands have idiosyncratic levels of loyalty, nor that loyalty is related to specific characteristics of a brand, other than its size. The implication is that attempts to manipulate loyalty directly are doomed to fail. All brands sell to a mixture of more and less loyal consumers. The balance between these is very similar from brand to brand and highly predictable. Growth does not come from increases in the loyalty of existing regular or heavy customers. When brands do grow, they mostly do so by increasing the propensity to purchase across the broad range of their potential customer base, including many light and non-buyers as well as a few heavier ones (McDonald and Ehrenberg 2003). This usually manifests itself by an increase in penetration, with possibly a small increase in frequency. Implications for Marketing Management Although there is no prescribed mechanism for reliably achieving growth, the broad implications for marketing are clear. To sell more, a brand has to get more customers to buy, many of whom will only be very occasional and not highly loyal. It is not enough simply to work on the more loyal customers. This in turn means strategies that emphasize reach, making the brand more widely available, communicating to more people and so on. Mass marketing methods succeeded in the 20th Century because they were the right thing to do, or fit to the purpose of making brands grow, not just because the media and retail landscape made them easy. Trying to restrict marketing efforts to highly refined targets, and the associated processes involved in trying to find the right target and measure the effects of targeted activity are wasteful of time and money and ultimately unlikely to succeed. 22 Even for maintaining an existing position, the same principles apply. Loyalty cannot be manipulated in a meaningful way by loyalty programs, CRM or other approaches that focus on existing customers in attempts to get them to alter their behavior and buy more. Brands that restrict their activity to a narrow base of consumers will find that their many but occasional consumers will be eroded by competitive activity or simply lack of salience. 23 References Aaker D.A. (1996) “Building Strong Brands,” 1st Edition, Free Press Aaker, D.A., & Joachimsthaler, E. (2002) Brand Leadership, The Free Press, Division of Simon & Schuster, New York, New York Allport, G.W. (1935) Attitudes, in Murchison, C.A. (ed), A Handbook of Social Psychology, Worcester MA, Clark University Press, pp 798-844 Amine, A (1998) Consumers’ True Brand Loyalty: The Central Role of Commitment, Journal of Strategic Marketing, 6 305-319 Armstrong, J.S., (2000) “Discovery and Application of Marketing Principles: Major Academic Contributions from the 20th Century”, Presented at the Society for Marketing Advances Conference, November Athiyaman, A. (1999) A Closer look at brand loyalty in fast moving consumer goods, Australasian Journal of Market Research, Vol 11 p 1-20 Baldinger, A.L., Blair, E. & Echambadi, R. (2002) Why brands grow, Journal of Advertising Research, 42, p 7-14 Baldinger, A.L. and Rubinson, J. (1996), Brand Loyalty: the Link Between Attitude and Behaviour, Journal of Advertising Research, 36(6), 22-34 Barnard, N.R, Barwise, P. and Ehrenberg, A.S.C. (1986) Reinterviews in attitude research: Early results, in MRS Conference, Market Research Society, Brighton, UK p 1-15 Barwise, T.P (1995) “Good Empirical Generalizations,” Marketing Science, Vol 14, No 3 Barwise, T.P. and A.S.C. Ehrenberg (1985) Consumer Beliefs and Brand Usage, Journal of the Market Research Society, 27, 2, p81-93 Baumgartner, H. & SteenkampJ-B. (1996), “Exploratory Consumer Buying Behavior: Conceptualization and Measurement,” International Journal of Research in Marketing, 13 (2), 121-137. Bass F.M. (1995) Empirical Generalizations and Marketing Science: a Personal View, Marketing Science, 14(3) G6-G19 Bass F.M. (1974) The Theory of Stochastic Preference and Brand Switching, Journal of Marketing Research, 11, p1-20 Bennett, D.R. (2005) “Which Car will they buy next?” The R&D Initiative, No 19, Centre for Research in Marketing, South Bank University, London Bennett, D.R. 2007. Meta analysis of repeat buying loyalty. Proceedings of the Academy of Marketing, UK conference, Kingston University, UK, July 3-6 24 Bennett D.R. (2009) Brand Loyalty Dynamics – China’s Television Brands Come of Age, Australasian Marketing Journal, 16 (2) Bennett, R. and Rundle-Thiele, S. (2005) The brand loyalty life cycle: Implications for marketers, Brand Management, Vol 12, NO.4 p 250-263 Bhattacharya, C.B. (1997) Is your brand’s loyalty too much, too little, or just right?: Explaining deviations in loyalty from the Dirichlet norm, International Journal of Research in Marketing, 14, 421-435 Blattberg, R.C., Briesch, R. & Fox, E. (1995) How Promotions Work, Marketing Science, Vol 14 No 3, Part 2 of 2 Bloemer, J.M.M. & Kasper, H.D.P. (1995) The complex relationship between consumer satisfaction and brand loyalty, Journal of Economic Psychology, Vol. 16 p 311-329 Bolton, R.N. (1998) A Dynamic Model of the Duration of the customer’s Relationship with a Continuous Service Provider: The Role of Satisfaction, Marketing Science, 17 (1), p 45-65 Brown, G.H. (1953) “Brand Loyalty,” Advertising Age 24, Reproduced in Ehrenberg, A.S.C and Pyatt F.G. Consumer Behavior, Harmandsworth, Penguin books, 28-35 Brewis-Levie, M. & Harris, P. (2000) An empirical analysis of buying behavior in UK high street womenswear retailing using the Dirichlet model, International Review of Retail, Distribution and Consumer Research 10:1, pp. 41-57 Butcher, K., Sparks, B. & O’Callaghan, F. (2001) Evaluative and relational influences on service loyalty, International Journal of Services Marketing Vol 12, No. 4 p 310-27 Buzzell, R.D, Gale, B.T (1987), The PIMS Principles, Free Press, New York, NY, Buzzell, R.D, Gale, B.T, & Sultan, R.G.M (1975), Market share – a key to profitability, Harvard Business Review, Vol. 53 No. 1, January-February, pp.97-1 Cannon, T., Ehrenberg, A.S.C. & Goodhardt, G. (1970) Regularities in Sole Buying, British Journal of Marketing, 4 pp 80-86 Capizzi, M.T. and Ferguson, R. (2005) Loyalty trends for the twenty-first century, Journal of Consumer Marketing, 22/2 p 72-80 Chatfield, C. & Goodhardt, G. (1975) Results Concerning Brand Choice, Journal of Marketing Research, vol 12, p 110-113, Colgate, M., Stewart, K & Kinsella, R. (1996) Customer Defection: a study of the student market in Ireland, International Journal of Bank Marketing, 14/3 p 23-29 Colombo, R.A. and Morrison, D.G. (1989) A Brand-switching model with implications for marketing strategies, Marketing Science, 8, 1 pp 89-99 25 Cronin, J.J, Brady, M., Hult, G.T (2000) Assessing the effects of quality, value, and customer satisfaction on consumer behavioral intentions in service environments. Journal of Retailing, 76(2), 193–218. Cunningham, R.M. (1956) Brand Loyalty – What, Where, How Much? Harvard Business Review, 34 (Jan/Feb), 116-128 Cunningham, Ross M. (1961) Customer Loyalty to Store and Brand? Harvard Business Review, 39 (Nov/Dec), 127-137 Day, George (1969) A two dimensional concept of brand loyalty, Journal of Advertising Research, 9, 29-35 Dall’Olmo Riley, F., Ehrenberg A.S.C., Castleberry, S.B., Barwise, P. and Barnard, N.R. (1997) The Variability of Attitudinal Repeat-Rates, International Journal of Research in Marketing, 14, 437-450 Dekimpe, M. & Hanssens, D. (1995) Empirical Generalizations about Market Evolution and Stationarity, Marketing Science, Vol 14, No 3 part 2 of 2 Dekimpe, MG, Steenkamp, J-B, Mellens, M & Abeele, P. 1997. Decline and Variability in Brand Loyalty. International Journal of Research in Marketing, Vol 14, Issue 5, p405-420 Dewey, J. (1930) Human Nature and Conduct, New York: Modern Library, p31 (as quoted in Waller (1988)) Dick, A.S. & Basu, K. (1994) Customer Loyalty: Toward an integrated Conceptual Framework, Journal of the Academy of Marketing Science 22, 2, pp101 Divett, M., Crittenden, N., Henderson, R. (2003), Actively influencing consumer loyalty, Journal of Consumer Marketing, Vol. 20 No.2, pp.109-2 Dunn, R.S., Reader, S. & Wrigley N. (1983) An Investigation of the Assumptions of the NBD Model as Applied to Purchasing at Individual Stores, Applied Statistics, 32, p 249-259 East, R. & Hammond, K. (1996) The Erosion of Repeat-Purchase Loyalty, Marketing Letters: 2 163-171 East, R. & Hammond, K. (1997) The Relationship Between Share of Purchase and Repeat Purchase over Time, Kingston Business School Occasional Paper, No. 28 East, R., Hammond, K. and Lomax, W. (2008) Measuring the impact of positive and negative word of mouth on brand purchase probability, International Journal of Research in Marketing, 25 p 215-224 East, R. Harris P., Wilson, G., & Lomax, W., (1995) Loyalty to supermarkets, International Review of Retail, Distribution and Consumer Research 5:1, pp. 99-109 East, R., Sinclair, J. & Gendall, P., (2000) “Loyalty: Definition and Explanation,” published in conference proceedings, ANZMAC 2000, December 26 Ehrenberg A.S.C. (1972) Repeat Buying, North Holland Publishers, Amsterdam Ehrenberg A.S.C. (1982 reprinted 1994) “A Primer in Data Reduction,” John Wiley, Chichester Ehrenberg A.S.C. (1988) Repeat Buying, 2nd Edition Charles Griffin & Co. Ltd London Ehrenberg A.S.C., Uncles, M.D. & Goodhardt G. (2004), “Understanding Brand Performance Measures: Using Dirichlet Benchmarks,” Journal of Business Research Vol. 57 No.12, pp.1307-25. Ehrenberg, A.S.C. & Uncles, M.D. (2000), “Understanding Dirichlet-type Markets,” the R&D Initiative Research Report 1, South Bank University, London Farley, J.U., Lehmann, D.R. & Sawyer, A. (1995) Empirical Marketing Generalisation Using Meta-analysis, Marketing Science, vol 14, No 3, Part 2 of 2 Fishbein, M. & Ajzen, I. (1975) Belief, Attitude, Intention and Behavior, Reading Mass, Addison-Wesley Foxall, G. (1987) Radical Behaviorism and Consumer Research: Theoretical Promise and Empirical Problems, International Journal of Research in Marketing 4, 2 111-129 Franzen, G. (1994) Advertising Effectiveness: Findings from Empirical Research, Publications, Henley-on-Thames, UK, NTC Franzen, G. & Bouwman, M. (2001) The Mental World of Brands: Mind Memory and Brand Success, World Advertising Research Centre, Henley-on-Thames, UK Garland, R. (2002) “Estimating customer defection in personal retail banking,” International Journal of Bank Marketing, 20/7 p 317-324 Goodhardt G., Ehrenberg A.S.C. & Chatfield C. (1984) The Dirichlet: A Comprehensive Model of Buying Behavior, Journal of the Royal Statistical Society, Series A, 147, Part 5, p621-655 Gounaris, S. & Stathakopolous, V. (2004) Antecedents and consequences of brand loyalty: An empirical study, Journal of Brand Management, Vol. 11, No. 4 pp 283-306 Graham, C. (2009) What's the Point of Marketing Anyway? New Findings on the Prevalence, Temporal Extent and Implications of Long-Term Market Share Equilibrium. Journal of Marketing Management, in press. Guadagni, P.M. & Little, D. (1983) A Logit Model of Brand Choice Calibrated on Scanner Data, Marketing Science, 2 (3) Summer, 203-238 Guest, L.P (1955) Brand Loyalty, 12 Years later, Journal of Applied Psychology 39, p405-408 Habel, C., Rungie, C.M. & Laurent, G. (2005) “Investigating Category Dynamics using the Double Jeopardy Line” proceedings of the Australia New Zealand Marketing Academy (ANZMAC) Conference 2005 27 Hammond, K. & East R., (2003) “Maximising Retention, Minimising Defection: Which Customers Should You Focus On?” Seminar at London Business School, December Heskett, J. (1995) “Strategic Service Management: Examining and Understanding It”, In (Eds) Glynn, W.J. and Bares, J.G. Understanding Services Management: integrating marketing, organisational behavior, operations and human resource management, John Wiley and Sons, Chichester Hocutt, M.A. (1998) Relationship dissolution model. Antecedents of relationship commitment and the likelihood of dissolving a relationship, International Journal of Service Industry Management, 9 (2) 189-200 Hoyer, W.D. (1984) An Examination of Consumer Decision Making for a Common Repeat Purchase Product, Journal of Consumer Research 11, 3, 822-829 Jacoby, J. (1971) A Model of Multi-Brand Loyalty, Journal of Advertising Research 11, 3, 25-31 Jacoby, J. & Kyner D.B. (1973) Brand Loyalty versus Repeat Purchasing Behavior, Journal of Marketing Research, 10, 1-19 Jacoby J. & Chestnut R. (1978) Brand Loyalty Measurement and Management, New York, Wiley Katona, G. (1953) Rational Behavior and Economic Behavior, Psychological Review (September, 1953) pp 307-318. Kau, A.K. & Ehrenberg, A.S.C (1984) Patterns of Store Choice, Journal of Marketing Research, 21, 399-409 Kaul, A. & Wittink, D. (1995) “Empirical Generalizations About the Impact of Advertising on Price Sensitivity and Price,” Marketing Science, Vol 14, No 3, Part 2 of 2 Keaveney, S.M. (1995) Customer Switching Behavior in Service Industries: An Exploratory Study, Journal of Marketing, Vol 59, April, p 71-82 Kraus, S.J. (1995) Attitudes and the Prediction of Behaviour: A Meta-Analysis of the Empirical Literature, Personality and Psychology Bulletin, 21 58-75 Krishnamurthi. L. & Raj. S.P. (1991) An Empirical Analysis of the Relationship between Brand Loyalty and Consumer Price Elasticity, Marketing Science, 10, 2. 172-183. Kumar, V. & Shah, D (2004) Building and sustaining profitable customer loyalty for the 21st century Journal of Retailing, Volume 80, Issue 4, Pages 317-32 Leeflang, P.S.H & Wittink, D.R. (2000) Building models for Marketing Decisions: Past Present and Future, International Journal of Research in Marketing, 17 (2-3) 105-126 28 Leszcyc, Peter T. & Timmermans, Harry J.P. (1997) Store Switching Behavior, Marketing Letters, 8:2, 193-204 McDonald, Colin & Ehrenberg, Andrew (2003) “What Happens When Brands Gain or Lose Share?” Report 33 for Members, Ehrenberg Bass Institute, University of South Australia Mellens, M., Dekimpe, G. & Steenkamp, J-B. (1996) A review of Brand Loyalty Measures in Marketing, Tidschrift voor Economie en Management Vol XLI, 4, 1996, p 507-533 Mittal, B. & Lassar, W.M. (1998) Why do customers switch? The dynamics of satisfaction versus loyalty, Journal of Services Marketing, Vol. 12, No. 3, p 177-194 Morrison, D. G. & Schmittlein, D.C. (1988) Generalizing the NBD Model for Customer Purchases: What are the Implications and is it Worth the Effort? Journal of Business And Economic Statistics, 6 (April), p 145-159 O’Loughlin, D. and Szmigin, I. (2006) Emerging perspectives on customer relationships, interactions and loyalty in Irish retail financial services, Journal of Consumer Behaviour, Vol 5, 2, p 117-129 Reibstein, D.J. & Farris, P.W. (1995) Market Share and Distribution: A Generalization, A Speculation, and Some Explanations, Marketing Science, 14 (3) Part 2 of 2 G190-G202 Reichheld, F.F (1996) Learning from Customer Defections, Harvard Business Review, March/April Reichheld, F.F. & Sasser, W.E. (1990) Zero Defections: Quality Comes to Services, Harvard Business Review, 68, p 105-111, Reichheld, F.F. & Sasser, W.E. (1996) The Loyalty Effect: The Hidden Force Behind Growth, Profits and Lasting Value, Harvard Business School Press, Boston Roos, I. (1999) “Switching paths in Customer Relationships,” Doctoral Dissertation No. 78, Publications of the Swedish School of Economics and Business Administration, Helsinki Rosenberg. L.J. & Czepiel, J.A. (1983) A Marketing Approach to Customer Retention; Journal of Consumer Marketing 2, 45-51 Rundle-Thiele, S. and Bennett, R. (2001) A brand for all seasons? A discussion of brand loyalty approaches and their applicability for different markets, Journal of Product and Brand Management, Vol 1(10): p25-37 Rust, R.T. & Oliver, R.W. (1994) Video Dial Tone: The New World of Services Marketing, Journal of Services Marketing, Vol. 8 No. 3, p 5-16 Samuelson, B.M & Sandvik K. (1997) In European Marketing Academy Conference, Vol. 3, (Eds. Arnott, D, Bridgewater, S.) Warwick Business School, Warwick, UK pp 1122-1140 Sharp, B., Wright, M. & Goodhardt, G.J. (2002) Purchase Loyalty is Polarized into either Repertoire or Subscription Patterns, Australasian Marketing Journal, 10(3), 7-20 29 Sharp, B., Sharp, A. and Wright, M. (2000) Questioning the value of the “true” brand loyalty distinction, for the Australia New Zealand Marketing Academy (ANZMAC) Conference 2000 Sharp, B, & Wright, M, (2000) “There are Two Types of Repeat Purchase Markets,” ANZMAC 2000, Gold Coast, November 29-December 1 Stern, P. (1995). Prescriptions for Branded and Generic Pharmaceuticals, Journal of Brand Management, 2, 177-183. Srinivasan, S. & Bass, F. (2000) Cointegration analysis of brand and category sales: Stationarity and long-run equilibrium in market shares, Applied Stochastic Models in Business and Industry, 16 p 159-177 Srivastava, R. (2002) Linking Customer Assets to Financial Performance, Journal of Services Research 5, 1, p 26-38 Szymigin, I. & Carrigan, M. (2001) Wherefore customer loyalty? Journal of Academy of Marketing Science, 15, Fall Terech, A, Bucklin, R.E. & Morrison, D.G. (2003) “Consideration, Choice and Classifying Loyalty,” from Andres Terech’s PhD, Anderson School at UCLA Thornton, S. (2002) “Karl Popper,” The Stanford Encyclopedia of Philosophy, Winter 2002 Trout, J. and Rivkin, S. (2000) Differentiate or Die—Survival in Our Era of Killer Competition, John Wiley & Sons, Inc, New York, NY Uncles, M D, Dowling, G & Hammond, K (2003) Customer Loyalty and Customer Loyalty Programs, Journal of Consumer Marketing Uncles, M D, Ehrenberg, ASC, Hammond, K (1995) Patterns of Buyer Behavior: Regularities, Models and Extensions, Marketing Science 14 (3) G71-78, Part 2 of 2 Wright, M. & Kearns, Z. (1998) Progress in Marketing Knowledge, Journal of Empirical Generalizations in Marketing Science, Volume Three Wright, M. and Klÿn, B. (1998) Environmental Attitude-Behaviour Correlations in 21 Countries, Journal of Empirical Generalisations in Marketing Science, 3 42-60 Wright, M. Sharp A & Sharp B. (2002) Market Statistics for the Dirichlet model: Using the Juster Scale to Replace Panel Data, International Journal of Research in Marketing Zaltman, Gerald. (2003) How Customers Think: Essential Insights into the Mind of the Markets. Boston: Harvard Business School Press, 30