A Relational Model For Category Extensions of Brands

Peter H. Farquhar, Carnegie Mellon University
Paul M. Herr, Indiana University
Russell H. Fazio, Indiana University
ABSTRACT - This paper introduces a relational model for studying the effects of brand-category, category-brand, and category-category associations on the potential success of extending a brand to a new category. The paper describes the managerial significance of category extensions, considers methods for measuring the different associations, and discusses the implications of recent research on category extensions.
[ to cite ]:
Peter H. Farquhar, Paul M. Herr, and Russell H. Fazio (1990) ,"A Relational Model For Category Extensions of Brands", in NA - Advances in Consumer Research Volume 17, eds. Marvin E. Goldberg, Gerald Gorn, and Richard W. Pollay, Provo, UT : Association for Consumer Research, Pages: 856-860.

Advances in Consumer Research Volume 17, 1990      Pages 856-860


Peter H. Farquhar, Carnegie Mellon University

Paul M. Herr, Indiana University

Russell H. Fazio, Indiana University


This paper introduces a relational model for studying the effects of brand-category, category-brand, and category-category associations on the potential success of extending a brand to a new category. The paper describes the managerial significance of category extensions, considers methods for measuring the different associations, and discusses the implications of recent research on category extensions.


Many firms capitalize on brand names that are well-known in one category by extending them to products in other categories that are new to the firm. For example, Tauber (1981) cites Jello Pudding Pops, Bic Disposable Lighters, Woolite Rug Cleaner, and Clairol Hair Blow-Dryer as successful extensions of familiar brands to new categories.

Farquhar (1989) describes two types of "brand extensions". A "line extension" applies an existing brand name to a product in one of the firm's existing categories. On the other hand, a "category extension" applies an existing brand name to a category that is new for the firm (Tauber, 1981). Our interest here is the study of category extensions of brands.

Potential benefits of category extensions include immediate name recognition and the transference of benefits associated with a familiar brand. For example, Chipman-Union's introduction of branded deodorizing athletic socks in 1980 was helped by licensing the "Odor-Eater's" name used originally by Combe, Inc. for its successful line of deodorizing foot-pads.

Moreover, category extensions eliminate the high costs of establishing a new brand and often reduce the costs of gaining distribution (e.g., Morein, 1975; Kesler, 1987). Such advantages have even led some companies to resurrect "has-been" brands; Pfizer revived the Barbasol brand and extended it first to a new line of aerosol-foam shaving creams (Saporito, 1986) and then to shaving gels and anti-perspirants.

A well-known brand name does not guarantee a successful extension. Abrams (1981) points to Arm & Hammer anti-perspirant, Life Savers gum, and Welch's prune juice as extensions that failed. Indeed, an extension can weaken the parent brand in the original category (Tauber, 1981).

Therefore, leveraging a brand's equity through extensions to new categories carries both opportunities and risks. The opportunities are represented by the firm's growth potential in the new category. Risk comes not only from a new product failure, as noted above, but also from the success of a category extension. Success carries the risk that the entire brand franchise can be diluted. Ries and Trout (1986) describe many of the dangers of over-extending a brand name, and generally advise against category extensions.


Farquhar (1989) defines brand equity as the "added value" that a brand endows a product; this I added value can be viewed from the perspective of the firm, the trade, or the consumer. Brand equity from a consumer perspective is reflected by the increase in the strength of associations an individual has for a product using the brand.

This paper specifies more precisely the structure of these associations and seeks to identify the processes which mediate consumers' judgments of a category extension of a brand. Herr, Farquhar, and Fazio (1989) implement the relational model by I measuring (a) the strength of associations between j brands and product categories, (b) the relatedness of W different product categories, and (c) the [ transferability of features between a parent brand and t a new item in a different product category. Some of their key findings are summarized here.

Figure 1 describes the relational model for f investigating product categorization and brand extensions. The "parent brand" is presumably associated with only one product category, called the "parent category". Thus the basic model considers neither "family brands" used across several categories, such as IBM or Kraft, nor "ambiguous brands" that evoke multiple, unrelated categories, such as Cadillac (automobiles and pet food) or United (airlines and gasoline). The relational model can be modified later to accommodate these generalizations if necessary.

The brand-to-category association in (la) > measures the likelihood that observation of the parent brand activates various features that characterize the basic category. For example, Crest is strongly associated with the features of toothpaste, while for many consumers Close-up is weakly associated with the toothpaste category. We refer to the "brand-to-category associative strength" as typicality (Smith and Medin, 1981). Barsalou (1985) discusses possible determinants of typicality, while Loken and Ward (1987), Ward and Loken (1988), and others describe recent consumer studies of typicality.

One measure of brand-category associative strength or typicality is the time subjects take to respond correctly to the query, "Is a product with brand X sold in category Y?" By comparing response latencies within a given product category, we can establish the relative typicality of a given pair of brands.

On the other hand, the category-to-brand association in (lb) considers the ease with which the brand will be retrieved from memory given exposure to the category (Shiffrin and Dumais, 1981; Fazio, 1989). We refer to the category-to-brand associative strength as cognitive dominance. Thus, a "heavier" or "more dominant" brand is more likely to be recalled in a category than a "lighter" brand.



Fazio (1987) compared different measures of cognitive dominance for various brands in 25 product categories and found high convergent validity among the measures. One method measures response latencies to correctly identify a brand as belonging to a product category. The query is, "In category Y is a product sold with brand X?" Faster correct responses indicate stronger category-brand associations.

One distinguishing characteristic of the relational model is the directional nature of associations between a brand and a category. A strong brand-to-category association (e.g., the "Ferrari" brand with the category "automobiles") does not necessarily imply a strong category-to-brand association (e.g., the "automobile" category does not usually activate "Ferrari"). Thus a "typical" brand, in the sense defined, need not be a "dominant" brand in a given product category. Conversely, a "dominant" brand in a category need not be a "typical" brand (e.g., "Haagen-Daz" is a super-premium ice cream). The asymmetric relationship between a category and a member is discussed further in Tversky and Gati (1978) and Smith and Medin (1981). An analogous pair of relationships exists for (3a) and (3b).

The category-category relatedness in the relational model in (2) depends upon the strength and accessibility of superordinate concepts linking the two categories. For example, Aaker and Keller (1990) discuss three factors that contribute to relatedness from a longer list of factors given by Albion (1985): complementarity (e.g., a common usage situation - tennis rackets and tennis balls), substitutability (e.g., a common function - wine and beer), and similarity (e.g., common features automobile and motorcycle). There are obviously other factors that relate two product categories from a consumer's perspective.


The likelihood of activating a memorial representation of the product category and its defining features when encountering a brand name presumably increases with the typicality of the brand. We hypothesize that typical brands in a product-category are more easily extended (in the sense of learning new associations) to closely related target categories than to distant target categories. Furthermore, we conjecture that dominant brands will be difficult to extend to unrelated target categories, because of the exemplary nature of such brands in their original product categories.

Fazio's (1987) investigation of cognitive dominance provides a standard set of stimulus materials that are well calibrated on category-brand associative strength. For each of 25 different product categories, Fazio identified brands that vary in the strength of their association with the product category.

Herr, Farquhar, and Fazio (1989) used a subset of these brands and categories as stimulus materials in testing various aspects of the relational model. One study measured typicality by examining subjects' latencies to respond correctly to queries about a given brand belonging to a parent category. Comparisons with Fazio's (1987) category-brand queries for dominance confirmed the hypothesized asymmetric relation between typicality and dominance.

Another study tested the hypothesis that relatedness influences the time subjects need to identify a product category as not being sold by a particular brand. The results provide overwhelming support that the less related the target category and the parent category, the faster subjects could identify the target category as not being associated with the brand.

Herr, Farquhar, and Fazio (1989) further demonstrated that the learning of new brand to target category associations depends upon the effects of the brand's cognitive dominance and the parent to target category relatedness. For example, subjects memorized one of several lists of brand-category pairs. Subjects were then given a cued-recall test where a brand was presented and they were required to retrieve the target category with which it had been paired in the list. Subjects were able to retrieve significantly more target category terms when those categories were closely related to the parent category than when the target categories were distantly related.

Tauber (1981) notes that feature transfer from the parent brand to the category extension is a critical element in the success of a newly introduced item. Herr, Farquhar, and Fazio (1989) also addressed this issue using the relational model. The focus, however, shifts from the typicality and dominance in (la) and (lb) to the analogous relations in (3a) and (3b). Nedungadi and Hutchinson (1985), Ward and Loken (1988), and others have conducted preliminary studies showing positive correlations between typicality and preference for a brand, but did not identify the underlying processes for transferability of attitudes from the parent brand to an extension.

Herr, Farquhar, and Fazio (1989) found that subjects expressed greater liking for brands extended to closely related rather than distantly related categories. Moreover, dominant brand extensions were liked significantly more than their weak counterparts when extensions were to closely related categories. Similar results were found with "intention to buy" as the dependent measure. Thus, "evaluative" associations of the brand were transferred to the category extension as well. Recent research by Fazio (1986, 1989), Herr (1989), and Herr and Fazio (1989) describes the cognitive processes involved in such feature transfers.

Additional studies have investigated the relationship between basic components of the relational model and market level variables, such as market share. Preliminary analysis of the Fazio (1987) brands reveals a very strong correlation between the strength of category-brand association and the market share held by a brand. For these product categories, "cognitive dominance" is indeed reflected by "market dominance". Data were also collected on levels of advertising expenditures per brand as a function of cognitive dominance. In every case, the "strong" category-brand associate spent more on advertising than did the "weak" category-brand associate.


Several researchers are developing somewhat different frameworks that focus on consumer knowledge, memory, and inference in category extensions of brands. These approaches share a common theme that emphasizes the importance of memory content and structure in evaluating brand extension outcomes.

Aaker and Keller (1990) hypothesize that consumers may hold four types of associations in long-term memory that bear on category extensions. These associations occur between a brand and (1) the attributes of that brand, (2) consumer attitudes toward that brand. (3) the attributes of the product category, and (4) consumer attitudes toward the product category. Aaker and Keller suggest that the presence and strength of these associations affect consumer judgments of the "degree of fit" between the existing brand and its potential extension to a new category. These intervening judgments are important in predicting success of a given brand extension.

Park, Lawson, and Milberg (1988) propose that the success of a brand extension is mediated by a consumer's categorization of the new item. They suggest that categorization is driven primarily by similarity judgments that depend on the memory structure for the specific brand in question. Park et al. identify three distinct structures for memory of brands and draw managerial implications for each type of structure.

Bridges (1989) suggests that a brand schema ("the bundle of associations, beliefs, and expectations that consumers have for a brand", p. 7) affects consumers' evaluations of brand extensions. She conjectures that the outcome of judgments of "fit" depends upon the brand schema, the relationship between the existing and the extended products, and additional information about the extension.

Given the recent attention on brand equity by the Marketing Science Institute and other research organizations (Leuthesser, 1988), the number of related approaches is likely to grow.


The relational model for category extensions of brands includes three types of associations: (i) brand to category, (ii) category to brand, and (iii) category to category. The strengths of these associations are called typicality, dominance, and relatedness, respectively. The possible asymmetry in the relationship between typicality and dominance is a distinguishing characteristic of the model.

Early research provides support for the relational model. In particular, typical brands in a product category are more easily extended (in the sense of learning new associations) to closely related target categories than to distant target categories. Furthermore, dominant brands are not easily extended to distant target categories, because of the exemplary nature of such brands in their original product categories. Herr, Farquhar, and Fazio (1989) provide further details.

The relationship between typicality and dominance is of further practical interest for exploring the limitations of brand extensions. When the same brand has been extended to a wide variety of target categories, we do not expect the parent brand's dominance in the original category to diminish, but we would expect a dilution of typicality. The strength of association from the parent brand to the basic category should diminish as a direct function of the success, number, and variety of brand extensions. Thus, the relational model offers one way of measuring the potential for "brand confusion" from over-extension (see Ries and Trout, 1986).

The distinction between typicality and dominance in the relational model addresses another issue in brand extensions. Some brands are so typical of a particular target category that consumers mistakenly believe that the brand extension exists when it does not. For example, many respondents in aided awareness tests said they had heard of McDonald's onion rings and Wang pocket calculators. One firm reportedly found that a proposed brand extension was already thought to exist by about half of the customers surveyed (Goydon, 1984).

"Spurious awareness" of non-existent brand extensions is both a blessing and a curse for marketers (see Farquhar and Pratkanis, 1987). To the extent that a strong association already exists, a firm might well consider a brand extension to that target category. Less effort would be needed for creating awareness, and more could be spent on other activities. On the other hand, "spurious awareness" can lead to misinterpreting advertising effectiveness measures and mismanaging the introduction of the brand into a new category.

The relational model appears to be a useful conceptualization of the associations between brands and categories that influence the success of extensions from a consumer perspective. Continuing research will refine the basic model through further consumer experiments, field studies, and market analyses.


The authors gratefully acknowledge support for this research from the Marketing Science Institute, the Center for Product Research, and the National Science Foundation


Aaker, David A. and Kevin Lane Keller (1990). "Consumer Evaluations of Brand Extensions," Journal of Marketing Research.

Abrams, Bill (1981), "Exploiting Proven Names Can Cut Risk of New Products," Wall Street Journal, January 22.

Albion, Mark S. (1985). "Consumer Perceptions of Related Products: A Taxonomy," working paper, Harvard Business School.

Barsalou, Lawrence W. (1985). "Ideals, Central Tendency, and Frequency of Instantiation as Determinants of Graded Structure in Categories," Journal of Experimental Psychology: Learning, Memory, and Cognition, 11, 629-649.

Bridges, Sherri (1989). "A Schema Unification Model of Brand Extensions," working paper, Stanford University.

Farquhar, Peter H. (1989). "Managing Brand Equity,' Marketing Research, 1, September, 24-33.

Farquhar, Peter H. and Anthony R. Pratkanis (1987). "Phantom Choices: The Effects of Unavailable Alternatives on Decision Making," working paper, Carnegie Mellon University.

Fazio, Russell H. (1986). "How Do Attitudes Guide Behavior?" In R. M. Sorrentino and E. T. Higgins (eds.), Handbook of Motivation and Cognition: Foundations of Social Behavior, Guilford Press, New York, 204-243.

Fazio, Russell H. (1987) "Category-Brand Associations and their Activation from Memory, unpublished report, Ogilvy Center for Research and Development.

Fazio, Russell, H. (1989). "On the Power and Functionality of Attitudes: The Role of Attitude Accessibility," In A. R. Pratkanis, S. J. Breckle and A. G. Greenwald (eds.) Attitude Structure and Function, Lawrence Erlbaum, Hillsdale, New Jersey.

Goydon, Raymond (1984). "Phantom Products," Forbes, May 21, 202-204.

Herr, Paul M. (1989). "Priming Price: Prior Knowledge and Context Effects," Journal of Consumer Research, 16, 67-75.

Herr, Paul M., Peter H. Farquhar, and Russell H. Fazio (1989). "Extending Brand Equity to New Categories," working paper, Indiana University.

Herr, Paul M. and Russell H. Fazio (1989). 'The Attitude-to-Behavior Process: Implications for Consumer Behavior." In A. A. Mitchell (ed.), Psychology and Advertising: Ad Exposure, Memory, and Choice, Lawrence Erlbaum, Hillsdale, New Jersey.

Kesler, Lori (1987). "Extensions Leave Brand in New Areas," Advertising Age, June 1, S1-S2.

Leuthesser, Lance (ed.) (1988). "Defining, Measuring, and Managing Brand Equity: A Conference Summary," Report 88-104, Marketing Science Institute, Cambridge, Massachusetts.

Loken, Barbara and James Ward (1987). "Measures of Attribute Structure underlying Product Typicality," Advances in Consumer Research, 1+ 22-26.

Morein, Joseph A. (1975). "Shift from Brand to Product Line Marketing, Harvard Business Review, 53, 56-64.

Nedungadi, Prakash and J. Wesley Hutchinson (1985). 'The Prototypicality of Brands: Relationships with Brand Awareness, Preference and Usage," Advances in Consumer Research, 1, 498-503.

Park, C. Whan, Robert Lawson and Sandra Milberg (1988)."Memory Structure of Brand Names," working paper, Graduate School of Business, University of Pittsburgh.

Ries, Al and Jack Trout (1986). Positioning: The Battle for Your Mind, 1st ed., rev. McGraw-Hill, New York.

Saporito, Bill (1986). "Has-Been Brands Go Back to Work," Fortune, April 28, 123 - 124.

Shiffrin, Richard M. and Susan T. Dumais (1981). "The Development of Automatism," in John R. Anderson (ed.), Cognitive Skills and Their Acquisition, Lawrence Erlbaum, Hillsdale, New Jersey, 111-140.

Smith, Edward E. and Douglas L. Medin (1981). Categories and Concepts, Harvard University Press, Cambridge, Massachusetts.

Tauber, Edward, M. (1981). "Brand Franchise Extension: New Product Benefits From Existing Brand Names," Business Horizons, 24, 3641.

Tversky, Amos and Itamar Gati (1978). "Studies of Similarity," in E. Rosch and B. B. Lloyd (eds.), Cognition and Categorization, Lawrence Erlbaum, Hillsdale New Jersey, 79-98.

Ward, James and Barbara Loken (1988). "The Generality of Typicality Effects on Preference and Comparison: An Explanatory Test," Advances in Consumer Research, 15, 55-61.