Product Category Perceptions, Elaborative Processing and Brand Name Extension Strategies

Dipankar Chakravarti, University of Arizona
Deborah J. MacInnis, University of Arizona
Kent Nakamoto, University of Arizona
ABSTRACT - This paper argues that consumers' judgments of fit between established brand names and new product extensions are determined by the associations that are activated and elaborated upon in a given situation. A pilot study explored the elaborative processing that occurred, both independently and when prompted by a cue, when consumers considered the extension of some established brand names into hypothetical product categories. The methodological and managerial implications of the results are discussed.
[ to cite ]:
Dipankar Chakravarti, Deborah J. MacInnis, and Kent Nakamoto (1990) ,"Product Category Perceptions, Elaborative Processing and Brand Name Extension Strategies", 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: 910-916.

Advances in Consumer Research Volume 17, 1990      Pages 910-916

PRODUCT CATEGORY PERCEPTIONS, ELABORATIVE PROCESSING AND BRAND NAME EXTENSION STRATEGIES

Dipankar Chakravarti, University of Arizona

Deborah J. MacInnis, University of Arizona

Kent Nakamoto, University of Arizona

ABSTRACT -

This paper argues that consumers' judgments of fit between established brand names and new product extensions are determined by the associations that are activated and elaborated upon in a given situation. A pilot study explored the elaborative processing that occurred, both independently and when prompted by a cue, when consumers considered the extension of some established brand names into hypothetical product categories. The methodological and managerial implications of the results are discussed.

INTRODUCTION

When a firm attempts to penetrate a new product category, considerable managerial attention is devoted to the naming of the new brand. The importance of this decision is underscored by the substantial resources that firms are willing to invest in the process of naming a brand (Mamis 1986, Leuthesser 1988). One common recommendation is to extend a well-established brand name to the new product via a family branding or similar strategy. It is argued that the franchise enjoyed by the established brand name will transfer to the new product. Such a transfer may provide several potential benefits, e.g., easing the task of generating awareness for the new product, providing an instant image for the new product, enhancing acceptance of the new product by the trade, and allowing potential promotional efficiencies. These arguments partially explain why firms have increasingly tended to adopt a brand extension strategy when introducing new products (Erickson and Dagnoli 1988, Kesler 1987, Tauber 1981).

The prevailing wisdom is that the above benefits of brand extensions should be most pronounced when the original and new product categories are similar (Kotler 1988). However, little is known about the nature of these similarity assessments. Indeed, recent psychological research on the organization and structure of human memory (Mervis and Rosch 1981, Smith and Medin 1981) suggests that this prevailing view may be overly simplistic. Whether or not a brand name transfers as intended from an old to a new product category may depend on the associations that are activated and elaborated either naturally or by a marketing cue.

In this paper, we discuss and empirically explore some factors that may influence the cognitive processes underlying consumers' judgments of "fit" between the original product category and new extension. Specifically, we consider (1) the degree to which the associations between the old and new categories are shared versus unique; (2) the implications of the relative salience or non-salience of these associations in memory and (3) the impact of the presence/absence of marketing communications cueing specific associations. We then describe a pilot study designed to provide some initial insights on these aspects of the brand name transfer mechanism. Finally, we discuss the methodological and managerial implications of the study findings.

FACTORS INFLUENCING BRAND NAME TRANSFER

Recent discussions of the nature of product knowledge in consumer memory draw upon the cognitive psychology literature (Anderson 1983, Wyer and Srull 1986) and argue that knowledge of an established brand is represented in memory as a series of associations (Lynch and Srull 1982). Some associations may be generic to the product category whereas others may be brand-specific. Thus, associations for Haagen Dazs ice-cream may include generic features of ice-cream such as "cold", "sweet" and "high in calories", as well as features specific to Haagen-Dazs, such as "rich" and "expensive". Beyond product features, the associative knowledge structure for a brand or product category may also comprise of benefits, usage situations and skills pertaining to manufacture. We now consider the aspects of these associations that may influence the judged St and evaluation of a new product that has been given an established brand name, i.e., the determinants of brand name transfer (Aaker and Keller 1988).

Shared and Unique Associations

The degree to which the old and new product categories have shared and unique associations is expected to be a factor that determines the likelihood of brand name transfer. Figure 1 represents this schematically. A brand name would be more likely to transfer when an old and a new product category share more associations (see the solid lines in Figure 1). Thus, Haagen Dazs sherbet may be a logical extension from a consumer's perspective, and may be seen as a good fit with Haagen Dazs name because the two products have common associations or shared features.

However the overlap between two products is seldom complete. Thus, for any given extension, the established and extended brand are likely to have knowledge associations that are not shared (see the dotted lines in Figure 1). For example, while Haagen Dazs ice-cream is made with butterfat, sherbet is not. Moreover, the two products often come in quite different flavors. Thus, the old and new product categories are also likely to have unique (distinctive or dissimilar) features. The greater the extent of such unique associations, the lower may be the transferability of brand names across the two categories.

FIGURE 1

ASSOCIATIONS BETWEEN AN ESTABLISHED AND EXTENDED BRAND

These ideas on brand name transferability are closely related to theories of categorization which suggest that the greater the feature overlap between items, the greater the likelihood that such items will be perceived to belong to the same cognitive category (Rosch and Mervis 1975, Rosch 1978, Mervis and Rosch 1981, Smith and Medin 1981). Other theories of similarity judgments (Tversky 1977) also propose corresponding ideas. Thus, Haagen Dazs sherbet may seem like a natural extension since ice-cream and sherbet share many physical features. However, to the extent dissimilarities become salient during the fit judgment, the extension may be viewed as less appropriate.

Shared benefits may be another basis for judging the appropriateness of a brand extension. Barsalou (1982, 1983) suggests that, in considering items linked by a common goal (e.g., things to eat on a diet), it is shared benefits rather than physical features that guides category judgments. Thus, although cake and ice-cream share few physical features, both may be linked to the same goal (i.e., things to eat for dessert). This may explain why a brand extension such as Haagen Dazs cake may be judged appropriate. By a similar logic, the Arm and Hammer brand name may transfer readily from baking soda to air freshener because the two products provide similar deodorizing benefits.

Usage complementarity among product categories may provide a basis for assessing the appropriateness of a brand name transfer. This type of complementarity between product categories is a special case of shared benefits in that it makes the joint benefits of the two products higher than the sum of their benefits viewed individually. Thus, the Haagen Dazs brand name may readily transfer from ice-cream to chocolate syrup because the two products are complementary in usage.

From a firm's perspective, the decision to produce an extended brand is often determined by marketing or manufacturing synergies between the two products. In some cases, these technological or manufacturing synergies are also obvious and meaningful to consumers and may provide the basis for a perceived similarity between the old and the new product categories. Thus, skills associated with the molding of plastic products makes Rubbermaid a credible brand name across a number of disparate product categories such as floor mats and storage boxes. Thus, manufacturing synergies that are readily perceived by the consumer may also be a basis for judged fit of an extended brand name.

Salient and Nonsalient Associations

Associations among products may be differentially salient to consumers. Just as some associations come to mind very easily, others are less accessible in memory or more difficult to perceive in the environment. For example, it may make sense for Haagen Dazs to produce frozen vegetables because it already commands a developed technology for distributing frozen foods. Likewise, Heinz may be capable of producing high quality apple juice because the production technology for ketchup is akin to that used for making apple juice (extraction). However, despite marketing and manufacturing similarities between the categories, consumers may not focus on these similarities. Thus, Haagen Dazs frozen foods or Heinz apple juice may be viewed as inappropriate extensions of the brand names. Thus, in addition to considering the type and nature of associations between the established and the extended products, (i.e., shared vs. unique features), it is important to consider the salience of the associations.

Barsalou (1983) suggests that some attributes of items in a category are salient regardless of context, while the salience of others are context-dependent. Context-independent attributes tend to be those central to the meaning of the category. One may speculate that the basic features of product categories, and those central to a product's ability to deliver its core benefit, would be context-independent and hence naturally salient. On the other hand, if consumers typically think of products based on their benefits or physical features, potential technological or commercial synergies between the products are less likely to be salient.

Figure 1 depicts associational patterns between an established and an extended brand, classified on the basis of whether they are shared/unique or salient/nonsalient. Thus, the two products may be (1) similar on both salient and nonsalient dimensions, (2) similar on salient dimensions but dissimilar on nonsalient dimensions, (3) dissimilar on salient dimensions but similar on nonsalient ones, and (4) dissimilar on both salient and nonsalient dimensions. This fourfold classification scheme permits us to consider the nature of the fit judgments and also the managerial relevance of each classificatory cell.

Cases 1 and 4 are straightforward and imply strong and weak fit judgments respectively. Cases 2 and 3 have significant managerial relevance. First, consider Case 2 in which an established and extended category pair have dissimilar salient associations but are similar on nonsalient dimensions (e.g., Heinz ketchup and apple juice). Because the two brands are dissimilar on salient features, benefits or usage, consumers may perceive Heinz apple juice as a poor fit and evaluate it negatively. The manufacturer then would have to ensure that previously nonsalient manufacturing synergies become salient to the consumer. Likewise, consider Case 3 where an established and extended product pair share salient associations, but there exist nonsalient issues that may raise legitimate questions regarding a firm's competence to manufacture the extension. For example, though Heinz relish may seem like a natural extension due to the perceived complementarity between ketchup and relish, the manufacture of the two products involve quite different processes (extraction versus pickling). A competitor pointing out these differences would lower the credibility of Heinz relish. These examples illustrate why one must consider both salient and nonsalient associations in judgments of fit between an established and an extended category.

Cueing Effects

The salience of associations is likely to be subject to a number of external and situational factors. One factor investigated in this study is the effect of a general, externally-provided cue that is designed to prompt elaboration of previously nonsalient associations. Advertising is particularly important in this regard. Advertising claims could alter the salience of particular associations or, more generally, shift the mental frame or set in which judgments take place, changing the set of associations that are salient.

HYPOTHESES

Given the above, one would expect that without elaboration, nonsalient similarities or dissimilarities would have no effect on the perceived fit of the extended brand with the established brand. Salient cues would drive judgments of the goodness of a brand extension. Thus:

H1: When elaboration is not cued, judgments of the goodness of an extension are best when the established and the extended products have similar features, benefits or usage, and are worst when the two products have dissimilar features, benefits or usage.

However, when an externally provided cue prompts greater elaboration, previously non-salient cues may begin to matter. Thus, two products which initially seem dissimilar may be more favorably evaluated as consumers come to find similarities that were not immediately salient. On the other hand, upon elaboration, a consumer may also find dissimilarities between two products on non-salient dimensions, even though the products initially seemed to be quite naturally associated. The identification of these distinctive features may change how well the established brand name is perceived to fit the extended product and may also change perceptions of the goodness of the extended brand.

H2: Cued similarities on nonsalient dimensions may reduce the impact of salient dissimilarities on the perceived goodness of the extended brand, while cued dissimilarities should reduce the impact of salient similarities on perceived goodness judgments.

In summary, we suggest that the basis for evaluative judgments regarding a brand extension lies in the associations that link the existing branded product and the new product class. Associations may be shared or unique (i.e., similar or dissimilar) and they may be either salient or nonsalient. Salient associations will affect judgment regardless of external cues. Other, initially nonsalient associations may become salient if they are cued. In the next section, we describe a pilot study that examined how the sharing and the salience of associations influences consumers' evaluation of brand extensions.

STUDY DESIGN AND PROCEDURES

A set of seven well-known brands were selected (Table 1). We expected that product category characteristics would be strongly associated with them (Carpenter and Nakamoto 1989). Each brand served as a replicate of the basic design. Five types of extensions were developed for each brand (Table 1). One extension was a control product actually produced under the brand name. Others were extensions to hypothetical categories that varied in similarity or dissimilarity to the established brand. Some extension categories were picked such that the similarities or dissimilarities on dimensions that we presumed would be naturally salient. Operationally, these tended to be physical features, benefits, or usage characteristics. Other extensions concerned dimensions that we presumed would be non-salient. These mainly concerned technology or manufacturing dimensions.

TABLE 1

STIMULUS PRODUCT CATEGORIES

Finally the study manipulated whether an external cue (elaboration instruction) was present or absent. If present, the cue was embedded in the study instructions. It suggested that in evaluating the extensions, subjects should beware of first impressions, and that technological synergies and shared marketing efficiencies could allow the company to produce a quality product at a reasonable price.

The study used a 2 (salient dimensions similar or dissimilar) by 2 (nonsalient dimensions similar or dissimilar) x 2 (cueing present or absent) experimental design. Since, for this pilot study, we had a limited number of students (97) in a convenience sample, each subject judged only five of the seven extensions. The extensions were assigned-using a balanced incomplete block design. In the no cue condition, each subject saw all four treatment combinations plus a control. In the cued condition, only the mixed similarity extensions were used, i.e., the salient similarity - nonsalient dissimilarity or the salient dissimilarity nonsalient similarity cases. The subject judged two of one type and three of the other type.

For each of the five brand extensions, subjects provided four judgments on 7-point scales anchored by high and low. The judgments, (1) similarity to other products made by the brand, (2) fit with the brand's image, (3) quality of the new product, and (4) the extended brand's expected sales, were assumed to reflect the perceived goodness of the extension. After providing judgments for all five products, subjects provided some qualitative reasoning for their judgments of fit and quality.

TABLE 2

JUDGMENTS AS A FUNCTION OF SIMILARITY/DISSIMILARITY ON SALIENT AND NONSALIENT DIMENSIONS

RESULTS

The judgments for the extended brand were analyzed in a repeated measures analysis of variance using the brand extension as the unit of analysis and the similarity manipulations on salient and nonsalient dimensions as factors. Our first hypothesis (H1) was that, in the absence of an external elaboration cue, judgments would be driven by naturally salient associations. The results were somewhat consistent with expectations (Table 2). For associations deemed to be salient, (physical features, benefit, and usage), similarity had a strong main effect (p < 0.001) on judgments. Unexpectedly however, similarity had a significant main effect (p < 0.001) for nonsalient associations as well. Moreover, as indicated by a marginally significant salient x nonsalient interaction, (p < 0.06), the effect of similarity varied by whether the associations were salient or nonsalient. For space reasons, Table 2 reports the means for only the fit judgments. However, the data for the similarity, quality and sales judgments showed similar patterns.

The interaction suggests that the impact of nonsalient similarities (technological synergy) on judgments was greater when physical feature or usage similarity was low. Thus, if consumers initially perceive that the established and extended products are dissimilar on salient dimensions, they may engage in greater elaboration to find similarities that may not be immediately obvious. These results suggest that in some circumstances, consumers may look for and ''find'' similarities on nonsalient dimensions even without an external cue.

It was also hypothesized (H2) that the presence of an elaboration cue increases the impact of nonsalient similarities and reduces the impact of salient dissimilarities on judgments. In an ANOVA on the four types of judgment, (considering only the mixed similarity extensions judged), the cue by extension type interaction was marginally significant (p < 0.06). However, the means were not ordered as expected (Table 3). When brands were similar on physical features, benefits or usage, the cue had no impact. Even when cued, the lack of technological synergy did not adversely affect judgments regarding these extensions. Only when technological synergy was present did the cue have the expected effect of increasing fit judgments. These results are consistent with the idea of a confirmation bias (Einhorn 1982). Specifically, these analyses, pooled over the seven brands used in the study, suggest that consumers primarily search for evidence of similarity on salient features. If brands are similar on these salient features, the impact of dissimilarities on nonsalient dimensions is minimal.

Brand-Level Analyses

In the preceding analyses, the effects of brand and of type of judgment were also significant and interacted with the salient and nonsalient similarity/dissimilarity manipulations. Because of our focus on fit perceptions, judgments of the fit of the brand extension with the brand's current image were of particular interest. We therefore considered these judgments in more detail at the brand level.

TABLE 3

JUDGMENTS AS A FUNCTION OF CUEING AND EXTENSION TYPE

These analyses provide further insight on the reasons for the pattern; noted above.

We summarize these results without going into detailed descriptions of the data. In essence, the data for one brand and its extensions (Kellogg's) showed the pattern originally expected. In the absence of cueing, physical and usage similarity drove fit judgments. Thus, oatmeal and bread were both rated very high on fit, while birdseed and beer were rated poorly. The cue had the effect of increasing ratings slightly for birdseed (because of the shared production technology), and decreased fit ratings for bread. The qualitative reasoning explaining the fit judgments were also consistent with our expectations.

However, the remaining six brands reflected idiosyncratic patterns. In some cases, (e.g., Kleenex Tree Nurseries), products that were expected to be seen as naturally dissimilar were not seen that way. The presence of the cue seemed to intensify the impact of the physical feature/usage similarities. In other cases, (Haagen Dazs chocolate syrup), product-specific factors such as usage complementarity appeared to have strong effects, regardless of the cue. In some cases, (Alpo canned tuna), even the recognition of technological synergies did not change evaluations because of other salient inferences. Thus, although Alpo tuna uses many of the same manufacturing and packaging technologies as Alpo dogfood, consumers made inferences about the extension (i.e., would taste like dogfood) lowering fit perceptions. Moreover, though subjects recognized the potential sharing of equipment used to make both pet food and human food products, this was perceived as disgusting by some.

DISCUSSION

While the study results did not conform to our expectations, they provided a number of insights on the varied inference processes that may underlie judgments about brand extensions. These suggest both methodological caveats and managerial implications for future research.

First, as may be expected, salient associations have a strong impact on judgments about brand extensions and similarity between the established and extended brands seems to have a positive impact on fit judgments. However, it is clear that perceptions of similarity can be based upon multiple dimensions and that the specific context may drive what is salient. Physical feature and usage similarity are two generally effective bases, but other bases (e.g., product complementarity) emerged as well, sometimes dominating the effects of physical disparity.

Cues to nonsalient dimensions had mixed effects. In the absence of other sources of similarity between the established and extension categories, the cue seemed to have had an asymmetric effect increasing ratings if it cued a basis for similarity, no effect if it did not. However, if other more obvious sources of similarity were present, the cue, when effective, seemed to cause increased elaboration of these obvious sources. This increased their impact, but did not enhance the salience of the intended nonsalient attributes.

The first implication of these findings concerns the identification of "good" and "bad" extensions. A complex cluster of associations may exist between an established brand name and an extension product category. Hence, it is difficult to specify a priori which of these are likely to be accessed and elaborated in a judgment situation. Thus, to the extent that a priori judgments of salience/nonsalience were inaccurate, it is difficult to predict fit judgments. Thus, for future research examining these effects, it may be necessary to develop a baseline set of associations for each brand pair. In theoretically-motivated empirical work, this may be done at the individual subject level and used to assign subjects to experimental groups. In other, managerial applications, one may look for segments of individuals who depict similar, stable clusters of salient and nonsalient associations for a product category.

Second, the effect of cues to nonsalient associations seems highly variable. In the present study, the cue was rather global and provided a general invitation to consider certain bases for comparison. Based on the verbal reports, the cue appears to have been successful only in some cases. Thus, self-driven elaboration took rather idiosyncratic directions. By contrast, advertising cues tend to be much more specific. Hence, associations cued by well-executed advertising messages may have a more uniform impact.

Third, the brand stimuli used here were established names that were perhaps prototypical and may have anchored the cluster of associations in their original categories. Relative to names that were less prototypical, the set of associations evoked by these brand names may have been difficult to overcome even with the help of a cue. Thus, with less stable associations and a weaker product category member, external cues may have stronger effects.

Finally, the judgments reflected a difference between the noting or comprehending of an association and the adjustment of a fit evaluation on that basis. In the Alpo Tuna case, it was clear that the cued associations (technological synergies in food processing and canning) were clear to many subjects, but were not compelling. The pattern of evaluations suggested that it was impossible for subjects to discount the incongruity between dog food and human food.

Thus, if a taxonomy of brand extensions is to be based on the idea of salient associations, it will be necessary first to develop some more detailed understanding of the classes of associations that may arise in relating an established brand and a novel product category. We are currently studying the structure of brand knowledge as a basis for predicting the natural or baseline perceptions likely to emerge when an extension is introduced. Despite the somewhat mixed findings here, it remains an intriguing possibility that ultimately, a brand name may enjoy greater credibility in a product category if it permits priming of deeper and more meaningful structural associations, even if the surface similarity is not judged to be great.

REFERENCES

Aaker, David and Kevin Keller (1988), "Consumer Evaluations of Brand Extensions", Working paper, Stanford University, Stanford, CA.

Anderson, John R. (1983), The Architecture of Cognition, Cambridge, MA: Harvard University Press.

Barsalou, Lawrence (1982), "Context-independent and Context-dependent Information in Concepts", Memory and Cognition, 10, 1, 82-93.

Barsalou, Lawrence (1983), "Ad Hoc Categories", Memory and Cognition, 11, 3, 211-227.

Carpenter, Gregory and Kent Nakamoto (1989), "Consumer Preference Formation and Pioneering Advantage", Journal of Marketing Research, forthcoming.

Einhorn, Hillel (1982), "Learning From Experience and Suboptimal Rules in Decision Making," in D. Kahneman, P. Slovic and A. Tversky, eds., Judgment Under Uncertainty: Heuristics and Biases, Cambridge, UK, Cambridge University Press.

Erickson, Julie and Judann Dagnoli (1988), "Brand Frame Spreading: Line Extensions are-Marketers' Lifelines", Advertising Age, February 22, 3 & 76.

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

Kotler, Philip (1988), Marketing Management: Analysis, Planning, Implementation and Control, 6th Edition, Englewood Cliffs, NJ: Prentice Hall.

Leuthesser, Lance (1988), "Defining, Measuring and Managing Brand Equity", Summary of the Marketing Science Institute Conference on Brand Equity, Report #88-104, May 1988.

Lynch, John and Thomas Srull (1982), "Memory and Attentional Factors in Consumer Choice: Concepts and Research Methods," Journal of Consumer Research, 9, 1, 18-37.

Mamis, Roben A. (1985), "Name Calling", Inc, July 1984, 57-72.

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

Mervis, Carolyn B. and Eleanor Rosch (1981), "Categorization of Natural Objects", Annual Review of Psychology, 32, 89-115.

Rosch, Eleanor (1978) "Principles of Categorization", in Eleanor Rosch and Barbara Lloyd (eds.), Cognition and Categorization, Hillsdale, NJ: Lawrence Erlbaum Associates.

Rosch, Eleanor and Carolyn B. Mervis (1975), "Family Resemblances: Studies in the Internal Structure of Categories", Cognitive Psychology, 7, 573605.

Tauber, Edward M. (1981), "Brand Franchise Extensions: New Product Benefit from Existing Brand Names", Business Horizons, 24 (2), 3641.

Tversky A. (1977), "Features of Similarity", Psychological Review, 84, 327-352.

Wyer, Robert S. Jr. and Thomas K. Srull (1986), "Human Cognition in its Social Context", Psychological Review, 93 (July), 322-359.

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