The Effect of Brand Typology on Brand Extension Fit: Commercial and Academic Research Findings

ABSTRACT - Based on an exploratory qualitative research study among French and British marketing practitioners with recent brand extension experience, this paper presents a brand extension research agenda which looks at brands in terms of a series of typologies. Three brand typologies are identified which both practitioners and recent academic research findings suggest may have an effect on how consumers evaluate brand extensions. The typologies are based on: the level of involvement which the category and/or brand evokes; a distribution of brand equity; and the nature of the brand in terms of its abstract or functional meaning. The research agenda identifies those areas where academic research can play a useful and relevant role in developing our understanding of consumer evaluation and decision making for brand extensions.

Citation:

Gil McWilliam (1993) ,"The Effect of Brand Typology on Brand Extension Fit: Commercial and Academic Research Findings", in E - European Advances in Consumer Research Volume 1, eds. W. Fred Van Raaij and Gary J. Bamossy, Provo, UT : Association for Consumer Research, Pages: 485-491.

European Advances in Consumer Research Volume 1, 1993      Pages 485-491

THE EFFECT OF BRAND TYPOLOGY ON BRAND EXTENSION FIT: COMMERCIAL AND ACADEMIC RESEARCH FINDINGS

Gil McWilliam, London Business School, UK

ABSTRACT -

Based on an exploratory qualitative research study among French and British marketing practitioners with recent brand extension experience, this paper presents a brand extension research agenda which looks at brands in terms of a series of typologies. Three brand typologies are identified which both practitioners and recent academic research findings suggest may have an effect on how consumers evaluate brand extensions. The typologies are based on: the level of involvement which the category and/or brand evokes; a distribution of brand equity; and the nature of the brand in terms of its abstract or functional meaning. The research agenda identifies those areas where academic research can play a useful and relevant role in developing our understanding of consumer evaluation and decision making for brand extensions.

INTRODUCTION

In response to practitioner interest in brand extensions, academics have been assiduously trying to fill the gaps in our understanding about how consumers evaluate brand extensions, (eg. Aaker & Keller (1990), Farquhar, Herr & Fazio (1990), Park, Milberg & Lawson (1991), Roedder John & Loken (1991)). Most of the research to date has concerned itself with the issue of "fit" as a generalised concept which is applicable across all types of product category. Fit has been looked at in terms of the extent to which the relationships between consumer perceptions of the parent brand, its product category, the extension, and the new product category, pre-determine the ease with which the parent brand can extend (Farquhar, Herr & Fazio 1990, Boush & Loken 1991, Park, Milberg & Lawson 1991). Roedder John & Loken (1991) have concentrated on the relationship between the extension and the parent brand in terms of harmful "feedback" effects.

Other research has examined the elements which comprise the concept of fit. Aaker & Keller (1990) examined extensions in terms of whether the two product classes were complements, substitutes, the perception of transfer of skills (eg in manufacturing), and the perception of how "trivial" the extension was ("difficult" in their terms). Bridges (1991) examined extensions in terms of product features and image associations and their role in preserving schema cohesion. Meyers-Levy & Tybout (1989) have examined the optimal degree of fit, or schema (in)congruity. Most of this academic research has been based on the responses of US consumers, or students, usually to hypothetical extensions presented in an experimental setting. There has been little or no empirical research into the views of marketing practitioners, based on their experience of real-life brand extensions.

As a first step towards filling this gap, this paper reports an exploratory study of French and British practitioners' beliefs and practices in researching actual brand extensions. The study explored areas of brand extension research which had proved difficult, and areas of extension activity which were still plagued by ignorance. Specific attention was devoted to the concept of fit, so that practitioners' field experience of this concept could be compared with academic findings, and new areas for research developed.

The paper examines the concept and research into the various aspects of fit between parent and brand extension, both as it has been explored in the academic literature to date, and as it has been experienced by marketing managers and commercial marketing researchers. The design of the research into the practitioner-related issues is explained, and the key findings as they relate to the issues of fit are detailed. In particular we discuss the finding that practitioners do not operate from a single homogeneous view of branding, but rather utilize various branding typologies to help them understand brand extension evaluations. A research agenda is proposed which highlights several areas where more research effort needs to be directed if it is to be responsive to the needs of practitioners.

RESEARCH DESIGN

Because of the exploratory nature of this research, it was decided to conduct qualitative in-depth interviews amongst marketing practitioners with experience of brand extensions. Interviewees were encouraged to talk about recent brand extension projects with which they had been involved. They were then asked to comment on some hypothetical extensions (the same or localized versions of the hypothetical extensions employed by Aaker & Keller (1990)) with a view to exploring further their concept of fit, and also to determine the degree to which they were conscious of generalisable determinants of fit such as transferability, complementarity, substitutability, and degree of difficulty as discussed by Aaker & Keller (1990).

In order to capture some of the range of European knowledge and experience about brand extensions, including possible differences in emphasis between countries, it was decided to carry out interviews in two European countries, France and Britain. Each of these countries has a well-established infrastructure of branding specialists. Moreover, practitioners in these countries tend to believe that their experience and understanding of brands differ from those of practitioners in North America (Lannon & Cooper 1983). Both British and French advertising agencies feel that their early concentration on the intangible, more emotional aspects of brands as a means of differentiation, rather than on tangible product-feature attributes, has contributed to the often noted differences in European and US television advertising.

However, there are also differences between the French and British approaches to branding and brand extensions. For example, the field of semiotics is well-established in France, and some brand research projects will employ semioticians as a matter of course. Semiotics is hardly used at all in the UK (there is only one registered company offering semiotic services in the British Market Research Society). Conversely British advertising agencies make heavy use of Account Planners, whose job it is to understand, through market research, brand-related consumer behaviour. French agencies tend not to use such planners. Britain also has a longer tradition of specialist consultancies in branding and in new product development .

The sample was chosen to reflect those differences between the two countries. Semi-structured interviews were held with senior marketing managers, senior brand managers, and market researchers in both countries; and in Britain with advertising account planners, and executives from brand consultancies, (four interviews), and in France with semioticians (two interviews). Eight interviews were held with executives in manufacturing companies (four in France and four in the UK). The remainder of the sample comprised market researchers (both qualitative and quantitative specialists, and some who carried out both types of research). The interviewees' experience of branding ranged from ten to twenty years. Twenty-six interviews were conducted (13 in London, and 13 in Paris) in June and July 1991. Each interview lasted about an hour, and was conducted in the interviewee's language. Given that the interviews centred on recent and on-going brand extension projects, we cannot discuss actual extensions for reasons of confidentiality. In consequence this paper concentrates on the generalized findings. Where necessary we use hypothetical examples, or examples which are already in the public domain, to illustrate the points made by the interviewees.

THE EFFECT OF BRAND TYPOLOGY ON BRAND EXTENSION EVALUATION

Given the differing infrastructure and research emphasis between the two countries, the French and British practitioners' beliefs about what constituted a good brand extension were surprisingly similar. In particular, practitioners in both countries shared a strongly held belief in the importance of "fit" between the parent and extension brands and their respective product categories. However the differences and similarities between the findings of academic research and practitioner experience of brand extension activity provides several areas for future research.

Of particular note was the practitioner view that treating the brand as a single homogeneous concept might lead to misinterpretations of how consumers evaluate extensions, especially in relation to the concept of fit. Based on the interviews, and building on academic research in this area, we have identified three brand typologies which affect how consumers evaluate brand extensions. These typologies are based on:

1. The level of involvement which the product category and/or brand evokes;

2. A schema based on the distribution of equity; and

3. The degree of abstract meaning carried by the brand.

The rest of this paper explores these typologies as they have been experienced and researched by practitioners and academics.

Typology 1: High and Low Involvement in Consumer Evaluations of Fit

Brand extensions, according to the interviewees, were frequently a cheaper, quicker, and less risky alternative to launching a new brand in a new product category. In these instances, using an established brand to enter a new market was a decision based more on considerations of short term profitability than as a result of a prolonged search for the target market which would optimally fit with the qualities and values of the parent brand.

The desire to launch a brand in the first place (whether a new brand or an extension of an established brand) can also be motivated by a desire to enter a new and growing market; to make an aggressive move against a competitor in a second market; to reap economies of scale from an established distribution network; to move away from a product category which had a finite life, (eg because of changes in fashion, in technology, or public opinion and pressure groups; to enter markets with higher added value; or as the result of a global rationalisation of brands.

All of these reasons suggest that the companies had a predetermined view of the product category into which they wanted to extend the parent brand. The research question in these instances is not about where to extend but rather how to extend into a predetermined category. This has important implications both for the research methodology employed, and also the notion of "fit".

The observation that practitioners often decide on the target product category for strategic reasons, and leave "how to get there", the execution of that strategy, to reflect consumer sensitivities, may be the result of their making prima facie judgements about brand extension fit at a subconscious level, before research programs are even considered. However, several interviewees implied that they had a model in their minds about brand extensions which, crudely stated, says the consumer will let you/your brand extend into virtually whichever product categories you want, as long as you do it sensitively/properly.

In this conceptualisation, consumers act as rather passive observers of brand extension activity. Based on an assumption that the company "knows what it's doing" or "must have a good reason for doing this", consumers may infer a link from the observation of the activity - assuming they think about it at all. This is a process which is not too dissimilar from Bem's (1970) concept of self-perception. Consumers infer what they like (their attitudes) from an observation of their own behaviour ("I eat a lot of brown bread so I guess I like it"). In the instance of brand extension, they would see an extension and seek to rationalize its existence (for example, "I guess Mars introduced Mars ice-cream because they discovered that lots of people like me were keeping Mars Bars in their refrigerators"). As long as the parent brand is sufficiently highly regarded and trusted, and the explanation is sufficiently plausible, consumers are willing to try the brand extension. Once it has been tried however, it will be judged in its own right, and perhaps with little reference back to the parent brand for positioning support. Although there may be some cognitive activity in the process of rationalizing the linkage, this cannot be taken for granted. Many interviewees gave the impression that their consumers, on the whole, gave less thought to extensions than, say, new brands. Indeed this was even cited as one of the reasons why brand extensions are so popular with marketing companies.

This can be regarded as a "low involvement" view of brand extension. That is, the consumers are fairly indifferent to the extension; they do not generate counter-arguments, or take positions relative to it; they may not even go as far as consciously rationalizing its existence (unless asked to do so by a researcher!). They see it, they try it, then on the basis of actual experience they decide whether to purchase it again or not. This is very similar to Ehrenberg's (1974) Awareness-Trial-Reinforcement model.

Given the amount of research and interest over the years about consumer involvement, it is perhaps time to apply what we have learned about involvement to the area of brand extension. Generalized concepts, such as fit, need to be examined in terms of their applicability to product categories, brands, and the relationship between categories and brands across different levels of involvement (McWilliam 1991). Some laboratory research evidence suggests that there may be differences in the evaluation processes of extensions in different product categories. For instance, Boush and Loken (1991) found differences in the time respondents took to evaluate extensions in grocery products and electronics which might suggest different cognitive processes for extensions in low and high involvement categories.

However to date, most academic and practitioner research into brand extensions has implicitly taken a high involvement approach. It has tended to operate with the underlying assumption that consumers do care about the parent brand and therefore its extension. This assumption is partly a reflection of the research methodologies employed. Whether qualitative or quantitative, these tend to ask respondents to reflect more or less consciously and verbalize their feelings about the parent and the extension. While this raising of consciousness may be essential for the execution of a sensitively developed extension or its advertising platform, it cannot be assumed that the same thoughtfulness is brought into play in the supermarket. Indeed the results of twenty-five years of research and theorizing about consumer involvement suggest the opposite, especially in the context of the frequently-purchased goods and services that form the basis of much of the research.

Two related conjectures emerge from the observation that consumers often seem to be fairly indifferent to the choice of product category into which a brand extends. First, we can conjecture that consumers evaluate brand extensions from low involvement product categories into other low involvement categories, in a different way from those in which the parent and/or target category is high involvement. Here, it may be that a low/low involvement extension is evaluated in much the same way as a low involvement advertisement (Krugman 1965). Consumers do not engage in counter-arguments, or indeed much cognitive processing, because there are no resistant attitudes to overcome. Over time, with exposure to the extension through advertising or distribution visibility, a reorganization of the perceptual structure will take place to include the extension. Alternatively they may process a low/low extension in a manner closer to that outlined in Petty, Cacioppo & Schumann's (1983) peripheral elaboration likelihood model. Conversely, brand extensions from, and/or to, a high involvement product category may induce a higher level of overall involvement, and the evaluation process in consequence take either a different route or be based on a high degree of cognitive processing which may involve counter-arguments. A second related conjecture would be that consumers require a different type of persuasive message when being confronted with extensions from and into low involvement categories. This brand typology may therefore have direct practical implications for the marketing communication of extensions.

The Role of Involvement in Communicating Fit

The notion that consumer involvement may moderate the concept of fit (and the communication of that fit), is further suggested in an examination of both practitioner and academic findings on the role of communication in brand extensions, especially when it comes to communicating the positioning of the new extension.

Practitioners see the task of communicating an extension's positioning as complex since it involves managing a number of incongruities or "tensions". The brand extension must remain "true" to the values of its parent; it must draw on links which are made obvious, but the advertising for it must be sufficiently different from the parent; to avoid the risk of being miscoded as being for the parent. Similarly, the new extension has to operate according to the dynamics and rules of its new category, but at the same time be perceived as sufficiently different (and better) than the brands which already inhabit that category. These tensions, whereby the extension has to be similar but different in two directions (parent brand and new product category), are not easy to execute and require a good deal of costly advertising and product development research.

The idea that extensions must be similar but different, both in terms of their parent brand and in terms of their new "home", finds some echoes in the research of Roedder John & Loken (1991) and Meyers-Levy and Tybout (1989). The latter found that a moderate degree of incongruity between a schema (or parent brand) and an object (or extension) improved the likelihood of a favourable evaluation. They argued that favourable evaluation depended on a satisfactory resolution of the incongruity in the mind of the consumer; this was possible with a moderate degree of incongruity but less easy with an extreme degree. Scenarios where there was perfect congruity were not very noteworthy, and therefore unlikely to prompt extensive cognitive elaboration. These findings have implications not only for the type of communication which is necessary to introduce a brand extension but perhaps also for the direction of those extensions.

There is a sense in which Myers-Levy & Tybout also draw on notions of involvement; whereby communication which presents the schema and object as so similar that it does not evoke any cognitive elaboration (low involvement) results in a less favourable evaluation than that which does evoke cognitive processing (higher involvement), but not so much as to cross the fine line between moderate and extreme incongruity. However the findings of Myers-Tybout (1989) are in part contradicted by those of Loken & Ward's (1990) and Boush & Loken's (1991) laboratory tests which found a direct correlation with increases in the level of typicality (congruity) between parent and target category and positive evaluation. However it is hard to compare these two sets of studies because of the very different methodologies.

We might conjecture that extensions from and into low involvement product categories require a moderate degree of incongruity in their communication, (ie. require a communication which induces a moderately higher level of response involvement) while extensions from and/or to high involvement categories may require greater congruity and typicality between the parent and extension categories and brands.

Typology Two: Fit and an Equity Based Typology

The practitioners interviewed expressed a good deal of confusion and uncertainty about whether it was easier to extend from a brand which was tightly focussed in its meaning (most typically a single-line brand) or from one which had a less focussed, broader-based set of attributes. Examples of focussed single-line brands might be Marmite, for an English audience, or Orangina for a French audience.

Several examples were given where brand extensions were more to do with a desire to build on a strong established single-line brand than with the attractiveness of entering a particular new market. The justification for extending a single-line brand might be to recoup the premium paid on acquisition of the brand by doing something "different" with it; to capitalize on the established brand's equity by making it work harder; to increase the in-store presence and feeling of stature and familiarity of the parent brand; or to take advantage of a realization that the company owned a brand capable of extension.

Farquhar, Herr & Fazio (1990) talk of "typical" and "dominant" brands. Typical brands are those where the brand name triggers an immediate association with the product category, as would be the case with say Xerox, Filofax, or Hoover. Conversely, brands which are easily triggered by the mention of the category, (say chocolate and Cadburys, coffee and NescafT or baked beans and Heinz) are referred to as dominant brands. Preliminary findings using a time-latency model in a laboratory setting (Farquhar, Herr & Fazio 1990) suggested that neither typical nor dominant brands were easily extended into distant categories, but were more easily extended into related categories than were weaker brands. Since the strength of category-brand associations (in both directions) is strongly correlated with market share, these results seem to suggest that extending brands with a high market share anywhere other than a closely related product category is difficult as well as risky, and whether the brand is typical or dominant is irrelevant. However it is those brands with a high market share in one single market which brand owners are most seeking to extend. Indeed the reason they may wish to extend them well beyond the confines of their parent category is precisely because of the fear that they will become overly identified with the category and become "generic".

FIGURE 1

A BRAND TYPOLOGY BASED ON DISTRIBUTION OF EQUITY

The notion of typicality in brands (Farquhar et al 1990, Loken and Ward 1990) does link however with an implicit model, held by several of the practitioners interviewed in this study, and which may be a more robust platform from which to judge whether some brands will be easier to extend than others. This model is based on an evaluation of the distribution of equity or meanings in brands.

Implicit in this understanding is an acknowledgement that the history and more important the typology of the parent brand provide a basic set of generic rules and opportunities with respect to brand extensions. Boush & Loken's (1991) study of "brand breadth" differences in the evaluation of extensions is interesting in this respect. They characterize single-line brands as "narrow" (eg Cambell soups) and brands which embrace several product categories as "broad" (eg Heinz food products). Their findings suggest the notion that brand extension evaluation is influenced by the degree of breadth which the parent brand has. Narrow brands appeared to have advantages for extensions that were highly similar to the parent category, and broad brands appeared to have advantages for extensions that were only moderately similar.

From all the interviews and detailed discussions on individual brands and their extensions, and building on the research findings of Boush & Loken (1991), it has been possible to build a second brand typology, as seen by practitioners. This typology is based on the distribution of the brand equity across different types of brand (see Fig 1). A particular brand may hold 100% of the equity (as would be the case with single-line brands, monolithic corporate brands such as IBM, or luxury brands such as Vuitton or Dunhill), or a percentage less than 100% depending upon a pre-existing relationship with a sub-brand or umbrella (range) brand.

The notion of a distribution of equity can perhaps be illustrated by the following (hypothetical) extensions. The Swiss breakfast muesli brand, Alpen, is a single-line brand existing on its own, without any extensions, and no corporate or umbrella brand endorsement. All of the brand equity inherent in the name Alpen is contained within that name. If its management wanted to extend it into other categories, the extension would be constrained only by that which is inherent in the brand name and its history. The brand name Lean Cuisine, on the other hand, is distributed (in Britain) by the company Findus (part of NestlT) with both names on the packaging. Findus carries with it significant brand equity in Britain as a respected purveyor of a wide variety of frozen foods. At the time of writing, it is likely that any equity which Lean Cuisine enjoys, partly, accrues from its endorsement by Findus.

At the time of the launch of Lean Cuisine, the Findus endorsement was no doubt extremely helpful in gaining listing in the major supermarkets and consumer trial. Since then, the Lean Cuisine name has no doubt over time given the Findus brand a degree of focus and status that it might not have achieved on its own. However the Findus endorsement may also have endowed the Lean Cuisine brand with associations of frozen foods, which Lean Cuisine, had it been a brand with 100% of the equity, or had it been associated with some other endorsement, say Heinz or NestlT, might not have had.

The manoeuvrability of Lean Cuisine in terms of brand extensions could well be constrained by the manoeuvrability and limits of the Findus brand name, and if it were ever decided to extend the Lean Cuisine name out of the freezer cabinet, the relationship between endorser and sub-brand might have to change first, ie the distribution of the equity shift more in favour of Lean Cuisine (eg through packaging redesign etc.).

Further research which analyses the extendability of brands according to this typology would be of considerable benefit to practitioners since it acknowledges that brands do not exist in a history-free, context-free limbo. The advantage of this typology over that of Boush & Loken (1991), is that it takes account of the interplay between multiple brand meanings which have been brought together for various strategic reasons, eg a corporate brand used to endorse a single-line brand.

Thus it may be conjectured that brand names which do not "own" 100% of their equity are harder to extend into areas which are untypical of their "branding partner" (but typical in their own terms) than they would if they "owned" 100% of their equity. Moreover the nature of the partnering brand may have an additional effect. For example, there may well be differences in consumers' evaluations of a "partnership" of two hitherto focussed brands (such Findus and Lean Cuisine) and a partnership based on a less focussed corporate brand, such as ICI with a focussed brand such as Dulux. We may conjecture that in the latter case any range extensions would be evaluated only in terms of the focussed brand Dulux (with ICI acting as a quality endorsement), whereas in the former case of Findus and Lean Cuisine, both parent brands would influence consumers' evaluation of an extension.

Typology 3: Brands with Abstract or Symbolic Meanings vs Brands with Functional Meanings

The interviewees also tended to distinguish brands according to the degree of their dislocation or abstraction from their product category. That is, they saw some brands as being heavily identified with the product category and its features, and others being able to embrace a wide number of very different product categories: the luxury brands such as Gucci, or Dunhill would fall into this latter category.

Indeed many practitioners saw this state of dislocation as being the ultimate stage in the development of the brand. This view was held particularly in France, where many interviewees referred to the work of J-N Kapferer (1991), who explicitly portrays a state of dislocation as an aspiration "c'est lorsqu'un nom perd totalement son statut de designateur rigide qu'il devient marque au sens plein du terme." (p 131). A brand moves through a stage where it is the product, to being a way of doing things ("la recette"), to conveying an expertise or quality ("savoir-faire") through levels where it acquires more abstract meanings through a development of its central focus to reach what Kapferer refers to as its core philosophy. It is only when a brand achieves this level of dislocation that it can embrace a wide portfolio of product categories:

" ... pour rendre cohTrents des extensions de plus en plus distantes, la marque s'Tloigne du physique et accFde nTcessairement au titre de source s'inspiration, de systFme de valeurs pouvant s'enraciner dans des mTtiers diffTrents." (p 140)

The idea that the more abstract the brand meaning, the easier it is to extend to distant product categories is an idea which has received recent research attention in the work of Park, Lawson & Milberg (1989) and Park, Milberg & Lawson (1991). The Park, Lawson & Milberg (1989) categorization of brands into symbolic, functional or usage based brands would be familiar to the practitioners interviewed, although only a few practitioners explicitly categorised brands in this way. The notion of the importance of "concept consistency" (Park, Milberg & Lawson 1991) as a dimension of perceived fit between extended brands, their new product categories, and the parent brand and its product category was supported by the interviews amongst the practitioners.

The problem for the practitioners was (i) to decide which of several, sometimes competing, abstract brand values to choose to exploit further, and (ii) the knowledge that on occasion it was hard to tell whether the abstract value in fact existed within the parent brand's set of meanings, or whether it had been "retro-fitted" from the proposed extension back to the parent (as a means of rationalising a decision which had been taken for others reasons perhaps: hindsight operating as the greatest source of wisdom in brand extensions as it does in other areas of marketing!).

The problem of choosing the "right" abstract value was made manifest when the interviewees were asked to respond to a series of hypothetical extensions based on well-known parent brands. One aim of this exercise was to explore whether practitioners, at least implicitly, used the concepts of complementarity, substitutability, and transferability in their evaluations of brand extension fit (Aaker & Keller 1990). To this end several of Aaker & Keller's hypothetical extensions were included, eg. McDonalds into frozen french fries, theme park or photo-processing. The concepts of complementarity, substitutability, and transferability did indeed seem to underlie many of the evaluations, but at an implicit level. This finding strongly suggests that these concepts are rarely systematically researched or explored in practitioner extension research, at least in France and Britain.

The "rightness" of the hypothetical extension and the ability to rationalize the linkage not surprisingly depended upon the practitioner's view of the brand, and in particular on their view of its abstract meaning. Thus those who saw McDonalds as essentially about consumption activities, would reject the notion of a theme-park as too frivolous, or insufficiently "nutritional" or "foody" in its values, (while more likely favouring the notion of the frozen french fries); whereas others who saw McDonalds as being essentially about family entertainment, and a treat for the children, saw the theme-park as the "right" extension, and frozen fries to be eaten at home as "definitely wrong". Yet others thought both would work equally well. Large-scale research might provide an answer, but it a further finding of this research was that (given the frequent intention to save money by extending a brand, rather than launching a new one), expensive research projects were not that common.

Practitioners also expressed concern about diluting the parent brand's equity by extending the brand (the brand"stretch" of Ries and Trout (1986) . Thus, suppose that both values of nutrition and entertainment are equally present in the McDonald's brand personality as it currently stands (ie. that McDonalds uniquely combines nutritional values and entertainment). To extend in either direction might result in a successful brand extension, but the extension might then have a significant influence on what the brand would come to mean, since in the longer term it would have pulled its meaning in one direction.

Not only was it difficult to find the "right" abstract value, but many of the tales recounted of failed extensions indicated that important key elements in the brand were not discovered at the time of the research, but only guessed at when the failures occurred. (These were post-rationalizations about failures). For example, a brand of fragrance extended into skin cream had failed, it was thought, because fragrances lie on the skin, whereas creams are intended to penetrate the skin, and the consumer therefore saw these two applications are incompatible. And yet, this "fact" was not picked up during an extensive programme of pre-launch qualitative and quantitative attitudinal research. A stock-cube, which was a household name in gravy (meat sauce) making, failed when it was extended to embrace ready-to-pour gravy because, it was felt, consumers saw the brand as an ingredient in the gravy-making process, not as an end-product. Again, this subtle distinction was not picked up during extensive pre-launch research.

The point here is not that these post-rationalizations may be wrong (although many of them may be), but that apparently obvious elements (with hindsight) of the brand or its product category associations, were not picked up in the prior research. Moreover in each of these cases the over-arching abstract value (the superordinate concept of eg Farquhar 1989) which linked the two product categories together clearly made sense and "worked" in the hot-house of the focus group, but was apparently found wanting in the cold light of day.

It may be conjectured from the findings of this exploratory study, that conceptual linkages which operate at an abstract level (for example, beauty, self-indulgence, caring cook, quality) may require an additional detailed product level linkage, a linkage which may have more to do with understanding the usage similarities between the parent and target product categories, than simple product feature similarities. Alternatively, it may be as Bridges (1991) has suggested that abstract associations may obscure important product-level links, and raising the salience of these links in advertising for example, may help improve the evaluation of the extension. Future research which builds on the work of Park, Milberg, and Lawson (1989, 1991) and Bridges (1991), examining in greater detail the need for an explicit relationship between the abstract concept and the product based concept (whether product feature similarity or usage contexts), could considerably aid practitioners in this difficult area.

CONCLUSION

Research amongst these French and British practitioners has shown that despite several decades of brand extension activity, and a perceived increase in the amount of recent extension activity, there is still a good deal of ignorance and anxiety about brand extension. Extending an existing brand may be seen as easier than launching a new brand, but is no less complex in terms of consumer and market behaviour.

This study identified three brand typologies which may influence how consumers perceive an evaluate brand extensions, which were based on: the level of involvement which the category and/or brand evokes; a distribution of brand equity; and the nature of the brand in terms of its abstract or functional meaning.

Additionally, four broad areas for future research are suggested which might help practitioners in their task by throwing yet further light on how consumers evaluate brand extensions:

1. Does involvement influence the way consumers evaluate brand extensions? A series of conjectures which have emerged from this piece of practitioner research would suggest that consumers might evaluate brand extensions from and into low involvement categories in a different way from extensions from and/or to high involvement categories. Involvement research would suggest differences would be seen in either the processing routes, or the presence of cognitive activity if an involvement typology is used to examine brand extension evaluation. Research based on such a typology would have a significant effect not only on the methodology of extension research but also potentially on the choice and communication of brand extensions.

2. Additionally, the level of involvement of the parent and target category may affect the degree of congruity which is portrayed in any extension announcement. The work of Meyers-Levy and Tybout (1989) suggests a level of moderate incongruity is best. This work requires further extension a)in the light of Loken & Ward's (1990) findings; b) in the context of high and low product category and brand involvement; and c) in the context of high and low involvement communication, that is, the need to be different yet the same in communication terms, especially in advertising.

3. More work is required into the issue as to whether it is easier to extend from certain types of brand relationship. Brand extensions from different types of brands as diverse as corporate, umbrella, and single-line brands, and brands which exist in some combination of these types, might be compared using panel or diary data. Indeed longitudinal surveys using panel data might provide significant insight into these questions. More exploratory research needs to be conducted to evaluate the interaction effects between brand extensions and the parent brand when that parent does not "own" 100% of its equity but shares it with another brand name. It was conjectured in this paper that extension from brands which were in fact partnerships of two brand names would be more difficult than from brands which were single-line, or which owned "100%" of their brand equity, because good ness of fit had to be achieved for both names.

4. The work of Park, Milberg & Lawson (1991) could usefully be extended to shed further light on the inter-relationship between abstract or symbolic concept consistency and product feature and usage consistency. In particular, what happens in extension evaluations with brands which possess more than one strong abstract or symbolic meaning and do extensions based on abstract values need underpinning by linkage with usage similarities?

It is hoped that the above research agenda not only demonstrates that a great deal of interesting work is still to be done in the area of brand extensions, but also that it is useful from time to time to refer to practitioner practice to highlight and refine areas where academic research can play a useful and relevant role.

REFERENCES

Aaker, David A., and Kevin L. Keller (1990), "Consumer Evaluations of Brand Extensions", Journal of Marketing, 54, January, 27-41.

Bridges, Sheri (1991), "A Schema Unification Model of Brand Extensions", Paper given in Special Session at Association of Consumer Research Conference, Chicago.

Boush, David M and Barbara Loken (1991), "A Process-Tracing Study of Brand Extension Evaluation", Journal of Marketing Research, Vol 28 February, 16-28.

Ehrenberg A.S.C. (1974), Repetitive Advertising and the Consumer, Journal of Advertising Research, 14 (2), April, 25-33.

Farquhar, Peter H. (1989), "Managing Brand Equity", Marketing Research, September, 24-33.

Farquhar, Peter H., Paul M. Herr, and Russell H. Fazio (1990), "A Relational Model for Category Extensions of Brands", Association of Consumer Research, 17, 1-5.

Kapferer, Jean-Noel (1991), Les Marques, Capital de L'Entreprise, Les +ditions D'Organisation, Paris.

Krugman, Herbert E. (1965), "The Impact of Television Advertising: Learning without Involvement", Public Opinion Quarterly, 29, 349-356.

Lannon, Judy and Peter Cooper (1983), "Humanistic Advertising: A Holistic Cultural Perspective", International Journal of Advertising, 2, 195-213.

Loken, Barbara and James Ward (1990), "Alternative Approaches to Understanding the Determinants of Typicality", Journal of Consumer Research, 17, (September), 111-126.

McWilliam, Gil (1991), "Consumers' Involvement in Brands and Product Categories", Proceedings of EMAC 1991 Annual Conference, Dublin, 21-23 May, p 404-424.

Meyers-Levy, Joan and Alice M Tybout (1989), "Schema Congruity as a Basis for Product Evaluation", Journal of Consumer Research, 16, 39-54.

Park, C. Whan, Robert Lawson and Sandra Milberg (1989), "Memory Structure of Brand Names" Association of Consumer Research, 16, 726-731.

Park, C. Whan, Sandra Milberg and Robert Lawson (1991), "Evaluation of Brand Extensions: The Role of Product Feature Similarity and Brand Concept Consistency", Journal of Consumer Research, 18, September, 185-193.

Petty Richard E., John T Cacioppo, and David Schumann (1983), "Central and Peripheral Routes to Advertising Effectiveness: The Moderating Role of Involvement", Journal of Consumer Research, 10, September, 135-146.

Ries Al and Jack Trout, (1986), Positioning: The Battle For Your Mind, McGraw-Hill International.

Roedder John, Deborah and Barbara Loken (1991), "Diluting Brand Equity",Working Paper, University of Minnesota.

----------------------------------------