Brand Extension Research: a Review

ABSTRACT - Research publications about brand extension have increased dramatically since 1990. Therefore, it is very useful to structure this field of research. This article presents a state of the art review that relates past studies to the paradigms from cognitive psychology theories they are derived from and to the variables taken into account and to the major results they provide.

Citation:

Elyette Roux and Frederic Lorange (1993) ,"Brand Extension Research: a Review", 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: 492-500.

European Advances in Consumer Research Volume 1, 1993      Pages 492-500

BRAND EXTENSION RESEARCH: A REVIEW

Elyette Roux, ESSEC, Cergy-Pontoise, France

Frederic Lorange, ESSEC, Cergy-Pontoise, France

ABSTRACT -

Research publications about brand extension have increased dramatically since 1990. Therefore, it is very useful to structure this field of research. This article presents a state of the art review that relates past studies to the paradigms from cognitive psychology theories they are derived from and to the variables taken into account and to the major results they provide.

It raises some unresolved issues and proposes areas of studies to be put on the research agenda on consumer evaluation of brand extension especially for non-functional products such as luxury goods.

INTRODUCTION

Research on brand extension has studied a diffusion process which includes different stages. From 1967 (Fry, Kerby) to 1987 (Boush et al.) the basis for future developments were set : emphasis was put on semantic generalization (Osgood, 1963) and whether or not affect transfer occurred between a brand and its possible extension based on the level of perceived similarity between the brand and the new product. Boush et al's. contribution (1987) seems to be a turning-point in this perspective, since it builds on previous research and introduced a categorization-based approach (Rosch and Mervis, 1975 ; Fiske, 1982) derived from cognitive psychology. The authors demonstrated that the higher the similarity between the current and new product, the greater the transfer of negative or positive affect to that new product.

The next stage started with Aaker and Keller's (1990) important contribution. They tested three basic dimensions of perceived similarity (complementarity, substitutability and transferability) combined with the overall quality of the original brand and the perceived difficulty in making the extension. Those factors are typical of similarity-based approach in categorization theory (Rosch and Mervis, 1975 ; Tversky, 1977). They also contribute to put into perspective alternative approaches (theory-based approach : Barsalou 1982 ; Cohen and Basu, 1987 ; Murphy and Medin, 1985 ; Sujan 1985) by including context effects in the evaluation process .

As a result, given that on one side brand extension evaluation from a consumer's standpoint is strategic for the firm (Tauber, 1981, 1988 ; Park et al, 1986), and that on the other side research has raised some unresolved questions and opened a broad avenue for research embedded in different cognitive psychology paradigms, studies on brand extension seem to have received a boost from 1990 to 1992.

The brand extension literature has already been reviewed (see for example Hartman, Price and Ducan, 1990; Rangaswany et al., 1990). However, due to the increase in publications, such contributions are now outdated. Consequently, the objective of this paper is to propose a typology that attempts to relate brand extension researches to their different paradigms, to the variables they measure and to the product category taken into account.

A special focus will be put on studies that investigate the psychological processes underlying consumer evaluation of brand extensions, since they represent most of the recent publications. This analysis will raise the question of the research paradigms and methods most adapted to consumer evaluation of luxury brand extensions.

A PROPOSED TYPOLOGY OF RESEARCHES ON BRAND EXTENSION

Our attempt here is to classify most papers published since 1967 in homogeneous groups. Rangaswamy et al. (1990) have done a systematic scanning of brand extension literature that shows the diversity of the researcher's interests and objectives. Therefore we adapted their classification and distinguish three types of contribution that are fairly independent and separate from one another (see table 1) :

1. Studies that attempt to articulate an economic or strategic rationale for extensions (Group 1 ).

2. Studies that develop models and measurement approaches to design optimal brand extensions (Group 2).

3. Research trying to explain psychological processes underlying consumer brand extension evaluations (Group 3)

We shall at first devote the following section to group 1 and group 2 type of studies which have not been deserved much attention. We then put the emphasis on group 1 type of research as the main focus of our review. Our presentation will be restricted to significant studies since there have been a lot of papers published on the topic in the past ten years that we cannot, for space reasons, summarize here.

Studies that attempt to articulate an economic or strategic rationale for extensions (Group 1 ):

Early in the 80's, Tauber clarified the role of brand extension strategies, by defining first in which a brand extension strategy was different from other opportunities. Then, he designed the conditions which enabled implementation of such a strategy ie: - the parent brand has to provide leverage in the new category,Bconsumers have to perceive new item to be consistent with the parent brand, and finallyBthe company must enter the new category with comparable or superior production / merchandising / distribution and/or advertising capability. In 1988, he then definitively confirmed the strategic role of brand leverage in a context of a cost-control world. Similarly, Aaker (1990, 1991) noted the advantages and potential pitfalls of brand extensions strategies. He emphasized the conditions under which brand extensions are profitable, or inversely, damaging for the parent brand. According to him, the brand has to add value to the new product considered and must trigger strong and quality associations. Besides, both products (the existing and the new one) have to share two fit dimensions : transferability of both skill and assets and complementarity.

This type of contribution can be compared with that of Smith (1991). He empirically examined the effects of brand strategy on sales growth and advertising costs, and how these effects are moderated by characteristics of the product and the market in which it competes. The findings show that brand extensions have higher sales growth rates and lower advertising costs than individual brands.

More technically, Heath and Chatterjee (1991) found that adding an asymmetrically dominated entrant similar to the target, will increase the target's market share, but as the entrant becomes weaker and increasingly similar to the nontarget brand, the improvement in the target's market share weakens. Sullivan (1991), suggested that brand extensions should enter the markets later than new-name brands when the markets are more competitive and less risky. Finally, she demonstrated that brand extensions that enter early are less likely to survive than both new-name brands entering at the same time and brand extensions entering later.

TABLE 1

TYPOLOGY OF RESEARCH ON BRAND EXTENSION

Other authors contributed to that field by studying other strategic aspects of brand extensions. Day (1989) focused on the synergies in the marketplace or in production as a basis for forming joint ventures that can lead to brand extensions. Varadarajan and Rajaratnam (1986) studied the symbiotic relationships between firms and their consequences for the company names used to enter markets with a new product.

Studies that develop models and measurement approaches to design optimal brand extensions (Group 2) :

A few authors have tried to model brand extensions strategies and to develop measurement approaches to design optimal brand extensions. Sappington and Wernerfelt (1985) tested a model that predicts to what extent a multiproduct firm will brand a new product with the established company name and where the new product will be located (positioned) in relation to the reputation established by the firm in other markets. Green and Krieger (1987) described and applied a consumer-based approach to finding "optimal" line extensions from an analysis of consumer trade-off data. They developed a set of reference products, that come very close to being the highest utility product for each consumer, given the available range of product design level ; then they selected an optimal product, with regards to the firm's current line and the presence of competitive brands. Wernerfelt (1988) analytically demonstrated that family branding is a way of signalling the quality of a new "experience" product. Sullivan (1990) provides a methodology to analyse spillover effects in unbrella-branded products and uses it to measure a positive and a negative spillover effect. Finally, Rangaswamy et al. (1990) used a linear additive model of consumer utility for experimentally and successfully testing the hypothese that the extendibility of a brand is a function of three types of utility : utility for the brand, utility for the product attributes and utility for interaction between the brand name and product-specific attributes, which plays a key role in determining brand extendability.

Studies trying to explain psychological processes underlying consumer brand extension evaluations (Group 3) :

Whereas in the introduction studies have been presented in chronological order, we distinguish here sub-groups of research according to the main paradigm they are referring to. Such an attempt is not an easy one since concepts and corresponding theories are not systematically clearly defined and /or differentiated from one another. We propose to group the different types of research as follows:

1. Studies refering to the principle of semantic generalization and to the process of affect transfer from the core brand to the extension

2. Studies refering to the structure of the representation of the concepts stored in human memory driven from (a) similarity-based approach and from (b) theory-based approach.

1. Studies refering to the principle of semantic generalization and to the process of affect transfer from the core brand to the extension.

Semantic generalization was first dealt with by Osgood (1963) and is derived from the conception of human long term memory that differentiate semantic from episodic memory. According to him, semantic generalization occurs when there is a transfer of meaning from an object (significate) to a sign of that object (such as a verbal representation), and then from that sign to either another significate or to a sign of that significate. This transfer does not depend on physical similarity between objects (which is the case in physical generalization) but when there is a common meaning shared by the objects (i.e. a brand name) despite their possible different physical characteristics. Three out of the five contributions listed in table 1 based their studies on semantic generalization (Kerby, 1967; Roman, 1969; Neuhaus and Taylor, 1972). Fry (1967) did not speak of semantic generalization per se but rather of preference generalization within a family brand. In consumer purchasing behavior, a generalized preference will be said to exist when there is a positive association between various product categories belonging to the same brand. It is therefore close to the concept of semantic generalization.

Fry (1967) demonstrates that preference generalization is operative in consumer behavior for various purchases of products belonging to family brands. He also finds that the level of generalized preference is positively correlated with the level of similarity, for the products considered, of competitive brand sets. But, he partially fails to show that this level of generalized preference is correlated with price similarity : Fry makes the hypothethis that the level of generalized preference could be related to the level of price variation between two product categories for a brand and to the level of price of one branded product compared to the rest of the category. For some brands, it appears to be true and for other brands regression fails to support the hypothesis. Kerby (1967) and Roman (1969) test the same items but differently : they wanted to assess semantic generalization in the (transformation) formation of consumer attitudes for family-branded products. Roman succeeds where Kerby fails because the former distinguishes two levels of attitude,Bthe attitude-object (product area, itself) and the attitude-situation (brand of the product), while the latter does not isolate those levels. As a consequence, Kerby and Roman's findings are opposite. Capitalizing on those findings, Neuhaus and Taylor (1972) demonstrate that generalization of attitude for a family brand is positively related to the retail shelf arrangement (in family brand) and to the constancy of price differentials among brands over similar products within a product group and finally to the product categories itself.

Boush et al. (1987) add two other theories to address the problem of similarity hardly treated by the former authors. They introduce the concept of categorization and schema congruence to test the role of similarity between products in affect transfer. Boush et al. (1987) focus their study on the notion of similarity between the original product category of a brand (i.e. existing branded products) and the proposed extension. They conclude that the higher the similarity between the current and new product, the greater the transfer of negative or positive affect to that new product. They also suggest that brand positive image in one product category could lead to negative evaluation of a new product in an unrelated category.

From that point of view, their study can be considered as the major turning-point in the brand extension literature. Almost all studies after this one base their hypotheses on categorization theory. In other words, the previous studies test whether or not affect transfer occurs in family branding, while Boush et al try to explain psychological variables intervening when there is affect transfer.

2. Studies refering to the structure of the representation of the concepts stored in human memory and processing of a new information driven from (a) similarity-based approach and from (b) theory-based approach.

Numerous research on brand extension evaluation have looked for the relevant theories that attempted to characterize the structure of knowledge stored in human memory. A first distinction can be made between memory based approches (i.e. when prior knowledge is necessary for an evaluation task ) and stimulus based approches (all information beeing present) (Lynch and Srull, 1982). In the former case, prior knowlege "is assumed to be comprised of relatively static and well bounded packets of information" (Barsalou and Medin, 1986). Those packets of information are defined according to the most quoted theories, as categories (Rosh and Mervis, 1975 ; Mervis and Rosch, 1981), schemas (Bartlett, 1932 ; Fiske, 1982 ; Fiske and Taylor, 1984), nodes in networks (Anderson and Bower, 1973 ; Collins and Loftus, 1975) or frames (Barsalou, 1991, 1992).

Categorization theories have drawn much attention from researchers. Smith and Medin (1981) divide theories of category representation into three basic classes : the classical view, the exemplar view and the probabilistic view (see also Cohen and Basu, 1987).

In the classical view, the categories are defined by singly necessary and jointly sufficient features. In the exemplar view, categories are represented by individual previously experienced category members called "exemplars" which are in other words : good known exemples. In the probabilistic view, categories are represented in terms of features that are typical of the category rather than defining the category considered (as opposed to the classical view). The prototype model is drawn from the probabilistic view : a prototype is an "abstract ideal mean" category member or the central tendency of the category (as opposed to previously experienced instances in the exemplar view). Given these three basic classes of concept representation in memory, two kinds of psychological processes may occur : categorization may first be based on attribute matching on the basis of feature list similarities : this process is called "similarity based approches". The second type of processing, which is called "theory based approaches", will be discussed in another section.

(a) similarity-based approach studies. According to the logic of similarity-based approaches, Aaker and Keller (1990) major contribution is centered on the definition of similarity between the brand and the proposed extension. The authors test three basic dimensions of similarity : complementarity, substitutability and transferability. They include two other main measures in their design : the perceived quality of the original brand and the perceived difficulty of making the extension. Transferability is the only dimension of similarity which has a direct impact on brand extension evaluation. Complementarity and substitutability therefore are not significant and have only an indirect interaction effect with brand perceived quality. Perceived quality turns out to have a positive effect on evaluation only when there is already a basis of fit between the two product classes. Finally, subjects' perceptions of the difficulty of making the extension has a positive effect on evaluation. In a separate study, they manipulate information cues on brand quality and on extended product attributes. They find out that elaboration at the product level is the only significant influence on evaluation. Therefore, they conclude that it is more effective from a strategic view point to protect the extension from negative association rather to attempt to reinforce positive association about the original brand.

Following those tracks, Romeo (1991) studies the effects of negative information about extended products on evaluation of both the proposed extension and the overall brand image. Based on the schema conception of human memory, he concentrates further on the notion of similarity : he distinguishes two levels of perceived similarity, similarity vs dissimilarity at the product category level and similarity vs dissimilarity at the attribute level. Product category had the only significant main effect on extension evaluation, which means that negative information is more influential on evaluation of extensions that are closely related to the family brand in term of product category only. The results show also that the overall family brand image is lowered by negative information when the extension is in a similar product category to the brand, whereas when the extension is in dissimilar product category, this leads to an increase in family brand image. This increase may be considered as inconsistent and due to sample or measurement effects. Therefore such an issue should be further tested. In conclusion, Romeo (1991) confirms the effects of negative information about extended product attributes on the evaluation of the extension (Aaker and Keller, 1990) and also the reciprocal impact on the core brand image.

Loken and Roedder John (1991) address specifically the issue of negative impact of brand extension on core brand image. They provide the first strong empirical evidence that negative brand extensions have a negative dilution effect on brand names beliefs. This is due to the fact that instead of measuring overall brand image (Romeo, 1991 ; Aaker and Keller , 1990), they measure it at its specific attribute level. More precisely, they show that subjects who perceive the extension as moderately typical of the existing branded products dillute their brand beliefs. At the opposite, brand evaluation remains non affected for subjects that perceive the extension as highly or lowly typical of the brand. In the first case, the extension may lead to confusion in consumer's perception, which may explain the dillution of brand beliefs. In the second case, for an extension rated as highly typical, negative information may be judged by consumers as inconsistent with their prior brand beliefs based on their own brand experience. Therefore the negative information is discounted. For a low typical rated extension, the very distance at the attribute level does not affect the original brand attitude. It should be noted that for Loken and Roedder John (1991), the fit between the original brand and the extended product is not considered in term of similarity (Aaker and Keller, 1990 ; Romeo, 1991) but in term of typicality. According to the probabilistic view, typicality is usually defined as the degree to which an item is perceived to represent a category (Rosch and Mervis, 1975). [The construct is usually measured by asking subjects to rate items along scales with endpoints of "good example/poor example," "representative/unrepresentative," or simply "typical/atypical".] This can be explained by the fact that Aaker and Keller (1990) refer to the classical view of the categorization, that Romeo (1991) bases his study on the classical view and on the schema, while Loken and Roedder John (1991) make reference to the probabilistic view. They however have in common the use of a similarity-based approach to test that fit. Following this process, Boush and Loken (1991) study more specifically the relationship between brand extension typicality (defined as the extension's similarity to the brand's current products) and the process by which a brand extension is evaluated. Their goal is to assess to what extent this psychological process is influenced by the typicality of the extension and the brand's breadth, i.e. variation among a brand's current products. For that purpose, they use a time variable measure as is the case in psychological categorization research, to identify the model used by consumers to process information. For typicality, the results show that moderately typical extensions are evaluated by a two-stage process including an attempted categorization followed by piecemeal processing. For extremely typical or extremely atypical extensions, a direct categorization process seems to be prevalent. This is true in particuliar for broad brands. Results concerning the effects of brand breadth on the nature of evaluating processing are inconclusive. Finally, they confirm the assertion that the transfer of positive attitude from a brand to its extension is positively associated with the degree of extension typicality.

Typicality in family branding is also expored by Boush (1991). He bases his study on the prototype model and on Tversky's (1977) findings on asymmetries in similarity judgements. His results reveal first, that less typical product will be perceived as more similar to the typical product than the reverse, second that brand names can reverse assymetries in similarity judgements and finally that a common brand name will increase perceptions of similarity between two products. This suggests that in brand extension strategy, it is usefull to select prototypes that are more likely to influence consumers' evaluations, given the crucial role of similarity in affect transfer and assymetries in similarity judgements implying assymetries in affect transfer.

Assymetry in relations between a brand and its products is also been investigated by Farquhar et al. (1990). They distinguish three types of relations : - brand to category associations, i.e the likelihood that the parent brand activates features that caracterize the basic category. In this regard brand to category associative strenght is defined as typicality. - Category to brand associations, that is, the "ease with the brand is retrieved from memory given exposure to the category". This category to brand associative strength is called cognitive dominance. And finallyBcategory to category relatedness is dependent upon the strength and acessibility of superordinate concepts linking the two categories. They take as granted that the relations of dominance and typicality are asymmetrical. They advocate that typical brands in a product category are more easily extended to closely related target categories than to distant ones.

We summarize the various contributions discussed above in this section in table 2.

Compared to the affect transfer approach, similarity-based approaches have shed a new light on the crucial role of similarity / fit / typicality. They point three main conclusions :

- for functional products, negative or positive information has more effect when it concerns the product category level,

- brand effects on extension evaluation is only significant when there is already a basis of fit between product categories (transferability).

- to optimize the basis of fit , it is more efficient to extend from a prototype which gives more basis of perceived similarity than every other product.

"Similarity based approches" have been criticized by Medin and his various co-authors who propose an alternative conceptualization of categorization called "theory based approches" [Theory based approaches (Murphy and Medin, 1985) are also referred to as: Knowledge based approaches or explanation based (Wattenmaker, Nakamura, Medin, 1988); or context based approaches (Barsalou, Medin, 1986; Medin 1983).] . A second group of research on brand extension is therefore derived from this view. Their main goal is to investigate the determinant of that crucial fit.

TABLE 2

SUMMARY OF SIMILARITY-BASED CONTRIBUTIONS

(b) theory-based approach studies. According to Murphy and Medin (1985) a theory is defined as "a host of mental explanations rather than a complete, organized, scientific account" p. 280. Therefore they consider that "people theories of the world embody conceptual knowledge". Instead of attributes matching, categorization is based here on an inference process supplied by underlying principle on information concerning operations, transformations and relations among attributes

According to this view, people have different theories about the world and, as a result have different representations of categories. As a consequence, the importance of individual experience is critical (Murphy and Medin, 1985). Research on consumer's evaluation of brand extension that follows a theory based approach includes consumer's knowledge in the product category and / or the brand, whereas similarity based approach studies do not.

Consumer knowledge has two major components: familiarity and expertise (Alba and Hutchinson, 1987). They define familiarity as "the number of product-related experiences that have been accumulated by consumer" p. 411. Expertise is "the ability to perform product-related tasks sucessfully". As product familiarity increases,Bthe cognitive effort required is reduced, -the cognitive structures used to differentiate products become more refined and complete,Bthe ability to isolate which is more important and task relevant increases,Bthe ability to generate accurate knowledge improves, and finally,Bthe ability to remember product information improves also (Alba and Hutchinson, 1987). Sujan (1985) demonstrates that "experts" generate more product cognitions than " novices" even if the information given does not match the category.

Muthukrishnan and Weitz (1991) show that brand attitude beeing equal among groups, "experts" give deeper level reasons for similarity between products categories than "novices". They conclude that "experts" might judge an extension differently than "novices".

Mac Innis, Nakamoto and Nami (1991) analyze the role of knowledge structures and context in similarity judgements. They suggest that self related bases with regard to the product have the strongest influence on overall similarity ratings (i.e. similarity in how, when, why the products are used). They claim that according to Barsalou (1982) that asssociations that are likely to be salient for a product depend on the context in which the product is embedded. They show that in making judgement about category relationships, concrete elements as physical attributes tend to be salient, but self-related elements associated with product consumption tend to be most important. Some of those results are shared by Chakravarti et al. (1990) who demonstrate that judgements are driven by salient associations such as physical features, benefits and usage. Non salient associations, such as technological synergy, are also relevant. Therefore, perceptions of similarity may be based upon multiple dimensions that are context dependent.

Context effects on brand extension evaluation have been also studied by Schmitt and Dube (1991) within the theory based approach of categories. They suggest that brand extension evaluations are influenced by information that is internally accessible to consumers when they judge the extension (see also Duncan and Nelson, 1986). They show that brand extensions are subject to context effects and seem to be best represented by contextual and unique characteristics, rather than by feature list similarities. Based on Barsalou's recent conceptualization (1991) they propose that frame theory should be looked upon in extension research. According to them, a frame is made of three basis components: attributes value sets, relations between attributes, and relations between values called constraints. The core of a frame is composed of a set of cooccurring characteristics called attributes which have different values as a frame represent different exemplars. Values are subordinate concepts of an attribute. Attributes in a frame are not independant but often related, as well as values. This representation of consumer knowledge of brand extension needs however to be tested against the more traditional views.

Following Murphy and Medin's view (1985) of concept relationships among objects, Park et al (1991) propose that evaluations of brand extensions depend on the degree of perceived fit between the extension product and the brand name. According to them, the degree of perceived fit is a function of both product-feature-similarity perceptions and brand-concept-consistency perceptions. Product-feature-similarity perception depend on identifying the relationships between product extension and the brand's existing products, whether concrete (feature correlations, attribute matching) or abstract (shared usage situations). Concept consistency perceptions rely on the extension product's ability to accommodate the brand concept. Therefore they control in their research design for subjects brand familiarity, and brand image that might affect brand concept consistency judgements in order to test both aspects of brand extension evaluation. In the same vein, Bridges (1991) proposes a schema unification model. She defines a brand schema as "a mental category which represents everything a consumer knows, believes or infers about a brand"(p.4). She selects variables which determine perceived fit: [Perceived fit is dependant upon the type of brand schema and on the nature between the brand and the extension and types of informations provided to the consumer. It is measured by five different seven-point scales: 1) appropriateness of the extension for the company, 2) degree to which the extension makes sense, 3) fit between company and product, 4) understanding of connections, and 5) confidence in explaining why the firm is planning to introduce the extension.] - the type of consumer's brand schema depending on dominant type of associations consumers have for the brand (concrete / product-related vs abstract / product-related vs image-related),Bthe type of relationship between a brand and its extension (shared physical product attributes or no-product-level relationship) andBthe type of information provided to consumers about the extension (aiming at raisng the salience of explanatory links or raising salience and relevance of those explanatory links). Share physical product attributes means that the brand and the extension have common attributes that are important to the functioning of each product.. No product-level relationship implies that the brand and the extension do not share physical attributes and also that they are not complements or substitutes. Her results support the hypothesis that the fit rating is higher for a product-related schema than for an image-related schema when the existing and the extension product share physical attributes. When the brand and the extension do not share product level relationship, fit rating is higher for an image related schema than a product-related schema.

Keller and Aaker (1992) consider that extension evaluation depend on three factors: -how salient or accessible the core brand associations are in the extension context,Bhow relevant consumer perceive this information to their extension evaluations,Bhow favorable inferred associations are in the extension context (see also Keller, 1991). They seem to stick to the similarity based approach , by defining similarity in terms of the salient attributes shared by the core brand and the extension product class (see also Aaker and Keller, 1990); however they move to a more theory based approach by integrating consumer knowledge of the core brand as relevant context. Such knowledge includes brand perceived quality, company credibility, and more specifically previous brand extensions that might affect brand image and extension evaluations bases. They also measure as covariates the consumer's perceived expertise or knowledgeability with the brand and the extension product category, and product usage rate.

Kardes and Allen (1991) address the question of perceived variability of a parent brand's offerings as a determinant of brand extension evaluation. Brand extension can have a negative impact on the parent brand image : the overall image decreases with a favorably evaluated new entrant, and especially for umbrella brands which are affected by a dilution effect. When perceived variability in the target category is high, generalization happens to be difficult. When it is low and when the parent brand as a niche positionning, there seems to be some opportunity for brand leverage.

Hartman, Price and Duncan (1990) include consumers' prior knowledge for the brand name and the product category as one of the five elements of their conceptual model of franchise evaluation. The other elements are: / degree of match or perceived similarity between the franchise extension and prior knowledge; / motivation for processing of the franchise extension; / extended processing ; / and moderating influences from individual factors and situationnal factors (i.e. product involvement). Their reasoning is within the similarity based approach but the knowledge variable they integrate in their model stems from a theory based approach (that is not referred to by the authors).

In related vein, Kim et al. (1991) propose an analogical reasoning perspective as an alternative theory to semantic generalization and category-based affect transfer. We decided to put this contribution in this section, because the results show that analogical reasoning prevails when subjects have no prior knowledge about the brand extension. Inversely, when subjects are given the brand extension together with an existing brand, no analogical reasoning is observed. Finally, the results support that analogical reasoning is a function of personal characteristics (i.e. it occurs when subjects are more self-confident).

Finally, the main contribution of the theory-based approaches studies is to investigate the underlying determinants of fit, or the salience and relevance of associations on which fit is based on. Table 3 summarizes the major determinants of that fit. This table emphasizes that in brand extension strategies, fit is very contextual and dependent on consumers'knowledge about the brand.

CONCLUSION

The different stages in brand extension research may be summed up as follows : first, it focused on the need for fit (semantic generalization and affect transfer), second it shed a new light on the role of fit (similarity-based approach) and finally it investigated the determinants of fit (theory-based approach).

TABLE 3

SUMMARY OF THEORY-BASED CONTRIBUTIONS

Given the importance of context, future research should integrate in their design the relevant variables measuring those contextual effects, be they linked to consumers prior knowledge or to the brand prior extensions. Consumers' perceived similarity is driven by salient associations that might be concrete and / or abstract. Most brand extension evalution research has been conducted with so called functional products. With the exception of Park et al. (1986, 1991), echoed by Hartman et al. (1990) and to some extent by Van Raaij and Schoonderbeeck (1992), the issue of symbolic products is not specifically raised. In fact, evaluation of brand extensions for a luxury product implies a more abstract and complex task (categorization process) for the consumer, and a more risky decision on the firm's side. A poorly evaluated extension may have more consequences for a brand bought on the basis of its name, (i.e. its inherent values, status or self image identification) rather than on strictly functional differential attributes. The initial brand image may suffer more from inadequate extension decisions and brand equity can be clearly damaged in that case (Ries and Trout, 1981 ; Loken and Roedder, 1991 ; Kardes and Allen, 1991 ; Keller 1991). Therefore, future research should be conducted on symbolic products to assess to what extend research results on functional products may or may not generalize to that specific category. Similarity-based approaches seems to be more relevant for a basic level categorization process. For a more abstract process involved in the evaluation of luxury brand extension, theory-based approaches may to be more appropriate. Evaluation of extension will depend on what comes to the consumer's mind about the brand in the extension context. Leading French luxury "brands" with exclusivity and a prestigious image are derived from French "Haute Couture" such as Christian Dior, Chanel or Saint Laurent. They have been already heavily extended in both women's and men's ready-to-wear markets , perfumes, cosmetics, beauty care, accessories, etc... Other brands such as Hermes are based on an historical handcraft skills. The strategic reality of such brands is wherelse the brand may be extented, in order to stick to the brand legitimate territory in the consumer's mind and in order to reinforce its strenght. This reality implies that one should take into account contextual effects dependent upon the breadth of the parent brand's previous extension [The authors have a research program currently on going on that issue.], to determine in those cases the specific determinants of fit. Moreover those brands are international by definition, therefore consumer's evaluation of brand extension should be conducted cross culturally to define to what extent what comes to minds in the extension context is culturally bound or not. [For example the replication of Aaker's and Keller's (1990) results is subject to variations when tested in a different country see for example: Brodie, Roderick and Lorraine Sunde (1992) "Consumer evaluation of brand extension: further empirical results", Proceedings of the Annual Conference of the European Marketing Association, p.121-133.]

REFERENCES

Aaker, David A. (1990), "Brand Extensions : The Good, The Bad, and the Ugly", Sloan Management Review, Summer, 47-56.

Aaker, David A. (1991), Managing Brand Equity, The Free Press.

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

Alba, Joseph W. and Wesley Hutchinson (1987), "Dimensions of Consumer Expertise", Journal of Consumer Research, 13, March, 411-54.

Anderson, John, R. and Gordon H. Bower (1980), Human Associative Memory : A Brief Edition, Hillsdale, N. Y.

Barsalou, Lawrence, (1982), "Context-independent and Context-dependent Information in Concepts", Memory and Cognition, 10, 82-93.

Barsalou, Lawrence W. and Douglas L. Medin (1986), "Concepts : Static Definitions or Context-Dependent Representations ?", Cahiers de Psychologie Cognitive, 6, 2, 187-202.

Barsalou, Lawrence, (1991), "Components of Conceptual Representation : from Feature Lists to Complex Frames", in Categories and Concepts : Theoretical Vioews and Inductive Data Analysis , Van Mechelen et al. (Edts), san Diego, Academic Press.

Barsalou, Lawrence, (1992), " Frames, Concepts and Conceptual Fields", in Frames, Fields, and Contrasts : New Essays in Semantic and Lexical Organization, E. Kittag and A. Lehrer (Edts), Hillsdale, N.Y., 21-74.

Barlett, F.C. (1932), Remembering : A Study in Experimental and Social Psychology, London Cambridge University Press.

Boush, David M. (1991), "Brand Name Influences on Product Similarity Judgments", Paper Presented at the Association for Consumer Research Conference, October, 17-20.

Boush, David and AL. (1987), "Affect Generalization to Similar and Dissimilar Brand Extensions", Psychology & Marketing, 4, 3, 225-237.

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

Bridges, Sheri (1991) "A Schema Unification Model of Brand Extension", Paper Presented at the Association for Consumer Research Conference, October, 17-20.

Brodie, Roderick J. and Lorraine Sunde (1992), "Consumer Evaluation of Brand Extensions : Further Empirical Results", Proceedings of the Annual Conference of the European Marketing Association, 121-133.

Chakravarti, Dipankar, Deborah Mac Innis and Kent Nakamoto (1990), "Product Category Perceptions, Elaborative Processing and Brand Name Extension Strategies", Advances in Consumer Research, 17, 910-916.

Cohen, Joel B. and Kunal Basu (1987), "Alternative Models of Categorization : Toward a Contingent Processing Framework", Journal of Consumer Research, 13, March, 455-472.

Collins, Allan, M. and Elizabeth Loftus (1975), "A Speading-Activation Theory of Semantic processing", Psychological Review, 82, 6, 407-28.

Day, Diana (1989), "An empirical test of the Synergistics Effects of Relatedness", Working Paper, University of Pennsylvania.

Farquhar, Peter H. , Paul M. Herr and Russel H. Fazio (1990), "A Relational Model for Category Extensions of Brands", Advances in Consumer Research, 17, 856-860.

Fiske, Susan (1982), "Schema Triggered Affect : Applications to Social Perception", in Affect and Cognition : the 17 th Carnegyie Symposium on Cognition , M. Clark and S. Fiske (Edts) Hillsdale, 171-90.

Fiske, Susan and Shelley Taylor (1984), Social Cognition, AddisonBWesley Publishing Company.

Fry, Joseph N. (1967), "Family Branding and Consumer Brand Choice", Journal of Marketing Research, 4, August, 237-47.

Green, Paul E. and Abba M. Krieger (1987), "A Consumer-Based Approach To Designing Product Line Extensions", Journal of Product Innovation Management, 4, 21-32.

Hartman, Cathy L. , Linda L. Price and Calvin P. Duncan (1990), "Consumer Evaluation of Franchise Extension Products : A Categorization Processing Perspective", Advances in Consumer Research, 17, 120-127.

Heath, Timothy B. and Subimal Chatterjee (1991), "How Entrants Affect Multiple Brands : A Dual Attraction Mechanism", Advances in Consumer Research, 18, 768-772.

Kardes, Frank R. and Chris T. Allen (1991), "Perceived Variability and Inferences about Brand Extensions", Advances in Consumer Research, 18, 392-398.

Keller, Kevin L. (1991), "Conceptualizing, Measuring, and Managing Customer-Based Brand Equity", Research Paper 1140, September, Stanford University Graduate School of Business.

Keller, Kevin L. (1991), " A Conceptual Model of Cusumer Evaluation of Brand Extension", Working Paper, Stanford University Graduate School of Business.

Keller, Kevin L. and David A. Aaker (1991), "The Effects of Sequential Introduction of Brand Extensions", Paper 2002, June, Stanford University Graduate School of Business, (Forthcoming JMR Feb.92).

Kerby, Joe K. (1967), "Semantic Generalization in the Formation of Consumer Attitudes", Journal of Marketing Research, 4, August, 314-317.

Kim, Kunbea, Arvind Rangaswamy and Raymond Burke R. (1991), "Consumer Evaluation of Brand Extension : An Analogical Reasoning Perspective", Paper Presented at the Association for Consumer Research Conference, Oct. 17-20.

Loken, Barbara and Deborah Roedder John (1991), "Diluting Brand Beliefs : The Negative Impact of Brand Extensions", Paper Presented at the Association for Consumer Research Conference, Oct. 17-20.

Lynch, John G. and Thomas K. Srull (1982), "Memory and Attentional Factors in Consumer Choice : Concepts and Research Methods", Journal of Consumer Research, 9, June, 18-37.

Mac Innis, Deborah J. , Kent Nakamoto and Gayathri Mani (1991), "Cognitive Associations and Product Category Comparisons : The Role of Product Knowledge Structure and Context", Paper Presented at the Association for Consumer Research Conference, Oct. 17-20.

Medin, Douglas L. (1983), "Structural Principles in Categorization", in Perception, Cognition and Development : International Analysis, T. J. Tigne and B. E. Shepp, (Eds.), Hillsdale, N. Y., 203-30.

Medin, Douglas L. , Mark W. Altom and Timothy D. Murphy (1984), "Given Versus Induced Category Representations : Use of Prototype and Exemplar Information in Classification", Journal of Experimental Psychology, 10, 3, July, 333-52.

Medin, Douglas L. , William D. Wattenmaker and Sarah E.Hampson (1986), "Family Resemblance, Conceptual Cohesiveness, and Category Construction", Cognitive Psychology , 19, 1-37.

Medin, Douglas L. and Lawrence W. Barsalou (1986), "Categorization Processes and Categorical Perception" in Categorical Perception, Harnad (ed), Cambridge University Press.

Medin, Douglas L. and Thomas J. Reynold (1983), "Cue-Context Interactions in Discrimation, Categorization, and Memory", in Perception, Cognition and Development : International Analysis, T. J. Tigne and B. E. Shepp, (Eds.), Hillsdale, N. Y., 323-55.

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

Murphy, Gregory and Douglas Medin, (1985), "The Role of Theories in Conceptual Coherence", Psychological Review, 92, 3, July, 289-316.

Muthukrishnan, A. V. and Barton, A. , Weitz (1991), "Role of Product Knowledge in Evaluation of Brand Extension", Advances in Consumer Research, 18, 407-413.

Neuhaus, Colin, F. and James R. Taylor (1972), "Variables Affecting Sales of Family-Branded Products", Journal of Marketing Research, 9, November, 419-22.

Osgood, Charles, (1963), "Psycholinguistics" in s.Koch, (ed.), Psychology : A Study of a Science, 6, Mac Graw-Hill , New York.

Park, C. Whan, Bernard J. Jaworski and Deborah J. Mac Innis (1986), "Strategic Brand Concept-Image Management", Journal of Marketing, 50, October, 135-145.

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, 185-93.

Rangaswamy, Arvind, Raymond Burke and Terence A. Oliva (1990), "Brand Equity and the Extendibility of Brand Names", Working Paper 90-019R, Wharton School , University of Pennsylvania.

Ries, Al, and Jack Trout (1981), Positionning : A Battle for your Mind, Mac Graw Hill.

Roman, Hope, S. (1969), "Semantic Generalization in Formation of Consumer Attitudes", Journal of Marketing Research, 6, August., 369-73.

Romeo, Jean, B. (1991), "The Effect of Negative Information on the Evaluations of Brand Extensions and the Family Brand", Advances in Consumer Research, 18, 399-406.

Rosch Eleonore and Caroline Mervis, (1975), "Family Resemblances : Studies in the Internal Structure of Categories", Cognitive Psychology, 7, 537-605.

Sappington, David E. , and Birger Wernerfelt (1985), "To Brand or Not to Brand? A Theoretical and Empirical Question", Journal of Business, 58, 3, 279-293.

Schmitt, Bernd H. and Laurette Dube (1991), "Contextualized Representations of Brand Extensions : Are Feature Lists or Frames The Basic Components of Consumer Cognition?", Paper Presented at the Association for Consumer Research Conference, Oct. 17-20.

Smith, Daniel, C. (1991), " An Examination of Product and Market Characteristics that Affect the the Financial Outcomes of Brand Extension", Report 91-103, February, Marketing Science Institute.

Sujan, Mita, (1985), "Consumer Knowledge : Effects on Evaluation Strategies Mediating Consumer Judgements", Journal of Consumer Research, 12, June, 31-46.

Sullivan, Mary (1990), "Measuring Image Spillovers in Umbrella-branded Products", Journal of Business, 63, 3, 309-329.

Sullivan, Mary (1991), "Brand Extension and Order of Entry", Report 91-105, March, Marketing Science Institute.

Tauber, Edward, M. (1981), "Brand Franchise Extension : New Product Benefits From Existing Brand Names", Business Horizons, 36-41.

Tauber, Edward, M. (1988), "Brand Leverage : Strategy for Growth in a Cost-Control World", Journal of Advertising Research, Aug. -Sept. , 26-30.

Tversky, Amos, (1977), "Features of Similarity", Psychological Review, 84, 4, 327-52.

Van Raaij, Fred and Wim Schoonderbeek, (1992), "Meaning Structure of Brand Names", in J. Olson and T. Reynolds, eds. , Understanding your Customers : Means-End Theory and Practice, Papers on Economic Psychology, 101, February, Erasmus University, Rotterdam.

Varadarajan, Rajan, and Daniel Rajaratnam (1986), Symbiotic Marketing Revisited", Journal of Marketing, 50, January, 7-17.

Vernerfelt, Birger (1988) , Umbrella Branding as a Signal of New Product Quality : an Example of Signalling by Postiong a Bond", Rand Journal of Economics, 19, automn, 458-66.

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