Memory Structure of Brand Names

C. Whan Park, University of Pittsburgh
Robert Lawson, University of Pittsburgh
Sandra Milberg, University of Pittsburgh
ABSTRACT - This study proposes that the transfer of brand name memory associations to new products is determined by categorization judgments, and such judgments are influenced by the nature of the memory structure associated with the brand name. Two experiments were conducted to test these propositions with three different brand name concepts, functional, symbolic and usage situation-based brand name concepts. The results suggest that functional and usage-based brand name concepts are understood and organized according to the product characteristics or usage situations of the products while symbolic brand name concepts are organized in terms of their direct linkage to the superordinate concept. Several implications on these differences in memory structure 'of different brand name concepts are offered.
[ to cite ]:
C. Whan Park, Robert Lawson, and Sandra Milberg (1989) ,"Memory Structure of Brand Names", in NA - Advances in Consumer Research Volume 16, eds. Thomas K. Srull, Provo, UT : Association for Consumer Research, Pages: 726-731.

Advances in Consumer Research Volume 16, 1989      Pages 726-731

MEMORY STRUCTURE OF BRAND NAMES

C. Whan Park, University of Pittsburgh

Robert Lawson, University of Pittsburgh

Sandra Milberg, University of Pittsburgh

ABSTRACT -

This study proposes that the transfer of brand name memory associations to new products is determined by categorization judgments, and such judgments are influenced by the nature of the memory structure associated with the brand name. Two experiments were conducted to test these propositions with three different brand name concepts, functional, symbolic and usage situation-based brand name concepts. The results suggest that functional and usage-based brand name concepts are understood and organized according to the product characteristics or usage situations of the products while symbolic brand name concepts are organized in terms of their direct linkage to the superordinate concept. Several implications on these differences in memory structure 'of different brand name concepts are offered.

One of the most valuable resources a firm has the marketplace is the reputation of its brand name. Consumers often rely on a brand name as an important information cue in evaluating products (see Park and Lessig 1981, for example). This would be particular true in situations where consumers are attempting to evaluate new products which have been introduced using an existing brand name (brand name extension Although brand name extension strategies are becoming more prevalent (Nielson 1985, Kesler 1987), little is known about consumers' responses to these strategies (Aaker and Keller 1987).

One assumption underlying the use of brand extension strategies seems to be that consumers' responses to a brand extension will be based, in part on existing brand name associations stored in memo and these associations, including affect, will transfer to the new products. The relevant question here is when and how these associations will be transferred? We propose that 1) the transfer of associations is determined by categorization judgments, in particular assessing category membership in terms of goodness of fit and 2) that such judgments are influenced by the nature of the memory structure associated with the brand name.

In other words, memory associations and affect will transfer when consumers identify a new instance in this case a brand extension, as a member of a particular category (Cohen 1982, Fiske 1982, Hutchinson and Alba 1985, Sujan 1985). The processes of these judgments, however, differ, depending on the type of brand name and the memory structure associated with it.

One can think of brand names as both helping to define category membership (Cohen 1982) and as organizing product-category associations. Some bran names are represented in memory based on individual features and cues (feature-based) while others are represented in memory based on a more abstract holistic array of cues (concept-based). With feature based brand names consumers judge "goodness of fit" based on whether the extension shares some critical feature or features (e.g., physical attributes, functions, usage context). For example, the brand name Honda is promoted and understood in terms of products that are related to small gasoline engines, Arm and Hammer is understood in terms of products that are related to deodorizing functions and Coleman in terms of products related to outdoor camping. In the case of Honda and Arm and Hammer, "goodness of fit" judgments would be based on whether the extensions were perceived as sharing physical attributes (small gasoline engine) or as performing similar functions (deodorizing function). For extensions introduced with the Coleman brand name "goodness of fit" would be based on whether consumers perceive the extensions as sharing the same age context.

On the other hand, for brand names promoted and represented in memory based on a more abstract array of cues, category membership is assessed in terms of some overall criterion of fit and not by the presence or absence of readily identifiable common features. For example, the brand names Rolex, Jaguar and Gucci are understood in terms of abstract concepts such as high status, luxury and wealth. Consumers determine how well the extension fits with the concept implied by the category label (Cohen and Basu 1987) and not with individual features.

It is important to note that for brand names represented by feature-based memory structures (functional and usage) judgments of category membership is related to the identifiable or inherent relationships among the products (e.g., shared physical attributes, functions, usage context). Whereas, for those brand names represented by concept-based memory structures (symbolic) judgments of fit are independent of any identifiable or inherent relationships among the products. It is the concept of the brand name that provides a link between products that would not otherwise exist.

Although previous research has demonstrated that category membership (fit) and liking for extensions is based on perceptions of shared physical and functional characteristics (University of Minnesota Consumer Behavior Seminar, 1987), no research has demonstrated that category membership can also be usage-based and concept-based as well. The purpose of the present research then is to: 1) demonstrate that different brand name concepts (symbolic, functional, usage) are associated with different memory structures (Experiment 1) and 2) demonstrate that different memory structures associated with brand name concepts affect categorization judgments for existing products and by implication for brand extensions as well (Experiment 2). More specifically, we seek empirical data supporting the idea that categorization judgments are based not only on shared physical or functional attributes among products but can be usage or concept based as well, depending on the nature of the memory structures associated with a particular brand name. We believe that this line of research will help marketers better understand how consumers evaluate brand name extensions.

EXPERIMENT 1

Rationale

The purpose of Experiment 1 is to investigate differences in brand name memory structure that may exist among different types of brand names (Park, Jaworski and MacInnis 1986) with respect to whether or not a linkage exists between brand names and superordinate concepts. We expect that brand names associated with the symbolic concept differ from those brand names associated with functional or usage concepts in that a superordinate linkage exists in the former concept but not in the latter concepts when products do not share common features.

More specifically, when individual brand names are understood on the basis of the functional features or usage context, perceptions of commonality among such brand names will also be based on such characteristics or cues. However, these features or cues associated with each brand name are not often shared among the brand names. Relating different brand names to one another based on common features or cues would, therefore, be difficult. Consumers in this situation are expected to engage in the reconstruction process of internal memory in an attempt to identify the common characteristics that serve as a superordinate concept. To perform this type of problem solving task will take time. Due to the absence of an organizing concept category among these brand names that facilitates retrieval. recall of these brand names is also expected to be difficult.

When a brand name is promoted with concepts difficult to describe with individual features (symbolic), they are best understood by the Gestalt arrangement of multiple cues. Concepts like wealth or high status may best be understood through associations with unique combinations of product, person and usage situational setting. For example, consumers come to understand the brand name Rolex when the product (watch) is promoted using persons (e.g. Donald Trump) and situations (e.g. yacht party) associated with the concepts of luxury and status. Following the same process, when a brand name like Gucci successfully elicits one of these unique combinations the name Gucci will also be understood as the company promoting status consciousness. Since these unique combinations all belong to the same concept category (luxury, status), consumers are expected to identify the commonality across these brand names readily. When there is a recognition of a common superordinate concept among these brand names, that concept itself serves as a retrieval cue for brand name recall.

In Experiment 1, in which subjects are given the task of thinking of what reasons sets of symbolic, functional and usage-based brand names have in common and then are given an unexpected recall task, we hypothesize a greater rate of brand name concept identification and better recall performance for the symbolic list than for the other two. This pattern of results would occur if consumers first realized the basis of commonality for the symbolic brands and then used this knowledge as a retrieval cue to aid their recall of the list members.

Method

Subjects consisted of a class of 35 MBA students from a large Eastern University. At the beginning of the session, they were given a test booklet which contained all the materials necessary for completing their tasks. The materials were three lists of 11 brand names each of which varied on the dimension of brand name concept. One list contained symbol c brand name concepts with each brand name conveying the idea of luxury and high status. Another list contained functional brand name concepts in which the products associated with each brand name shared certain features such as General Electric products all having to do with electronics. A final list contained usage-based brand name concepts in which products associated with each brand name shared a common usage context, such as Coppertone products being used for skin care in "sun-related" activities. [The items comprising each list were chosen on the basis of experimenter judgment and confirmed in a separate study in which an independent group of subjects assigned brand names to one of these three categories.]

These three lists, which are shown in Table l(a), were presented to subjects in counterbalanced order on successive pages of the test booklet. Upon inspecting these lists their task was to list the reason or reasons that these particular brand names were included as members of the list. In addition, subjects were asked to note the time to the nearest second at three separate points for each list: when they started reading each list, when they thought of their first reason and when they thought of their last reason.

Following this search for common bases, subjects were unexpectedly asked to recall the brand names in each of the three lists. They were given as much time as needed; Finally, subjects were asked to turn back to each list and rate their familiarity with each brand name on a seven-point scale with greater familiarity being indicated with a higher number.

Results

Reasons for Grouping. In general, subjects were able to comprehend the experimenter-defined underlying basis for the symbolic list, but not for the functional or usage-based list. Two judges independently found that of 35 subjects, 30 supplied satisfactory reasons for the symbolic list, only one was able to identify the underlying functional basis, and no subject identified the usage-based rationale. In addition, subjects required significantly less time to think of their first reason for the symbolic brand names (X=26.44 sec.) than for the functional (X=80.63 sec.) or usage-based (X=87.19 sec.) brand names (F(2,30)=6.95, p<01). A common strategy for dealing with the functional or usage-based lists was to find subsets of common product classes (e.g., Wilson, Nike, Adidas - sports equipment; Tide, Clorox, Ivory cleaning supplies) and identify the function of those subsets.

TABLE 1

LIST OF BRAND NAMES (EXPERIMENT 1) AND PRODUCTS (EXPERIMENT 2)

List Recall. The hypothesis that memory for the symbolic brand names (X=6.60) would be superior to the functional (X=5.77) and usage-based (X=4.83) brand names was confirmed (1(2,68)=9.24, P<.001). This memory advantage for the symbolic list occurred in spite of indications that subjects spent less total time (F(2,28)=3.05; P=.06) pondering the symbolic list (X=139 sec.) than the functional (X=217 sec.) or usage-based (X=230 sec.) lists.

Also, the familiarity ratings indicate that the memory advantage for the symbolic lists cannot be explained by subjects' greater familiarity with the symbolic brand names. To the contrary, familiarity ratings were lower for the symbolic brand names (X=5.33) than for the functional (X=5.72) or usage-based (X=5.44) brands (F(2,5 )=2.75; P=.07).

Discussion

The results of the first experiment indicate the concept of symbolic brand name has a degree of psychological reality for consumers that is not present for functional or usage-based brand name concepts. When brand names from each of these three categories are grouped together and presented for consumers' inspection, they were easily able to identify the underlying basis of the symbolic list but not of the other two lists. In addition, the superior recall of the symbolic list points to a memory structure for symbolic brand names such that a linkage exists between the individual brand name concepts and the superordinate level of "high-status" brands.

EXPERIMENT 2

Rationale

Insofar as Experiment 1 provided evidence of the "upward" memory links that exist between brand names and their associated superordinate concepts, at least in the case of the symbolic category of brand name, Experiment 2 seeks to investigate the nature of the "downward" linkages that exist between brand names and their associated products. In particular we propose that products associated with functional and usage-based brand name concepts are bound together by similarities that are inherent to the product lists themselves, either in sharing the same features or usage situations; whereas products that are associated with a symbolic brand name are not directly linked, but rather depend upon the brand name itself to supply the underlying common concept to which the various products are perceived as fitting together.

To test this hypothesis, we presented lists of products associated with each of the three brand name concepts either with or without the brand name as an organizing label for the list, and measured judgments of similarity and fit.

Starting with the assumption that the similarity judgment task is based on shared product features (Tversky 1977), we reasoned that this judgment would not be affected by the presence of the brand name label. This is expected since the perception of similarity of the symbolic, functional and usage-based products is based on direct product to product comparisons, and not on the presence of brand name labels. On the other hand, the fit judgment encouraged other processes to occur in addition to the assessment of shared product features.

Specifically, the fit judgment is expected to occur in two different ways. First, when the products share the common features, the features may serve directly as the superordinate concept which, in turn, brings the products together, thus facilitating the fit judgment. Second, when the products are linked to the superordinate concept directly, they may also be judged as fitting together. Therefore, even if products do not share the common features (are not perceived as similar), they may be judged as fitting together as long as they are directly linked to the superordinate concept. With this reasoning we hypothesize that the fit judgment for the symbolic brand name list would be positively affected by the presence of the brand name because it supplies the underlying abstract concept that ties the products together. By contrast, this judgment for the functional and usage-based brand name lists should not be affected by the presence of the brand name label, because as noted earlier, the inherent similarity of these products to each other is based on the common features among them.

Method

As in Experiment 1, subjects were asked to find bases on commonality in lists associated with symbolic, functional and usage-based brand name concepts. This time, however, the lists which are shown in Table l(b), consisted of eleven names of products associated with three different brands, chosen to represent the three different categories of brand name concepts: Gucci (symbolic), Sony (functional) and Coleman (usage-based). The lists were presented under one of two brand name conditions: with the brand name present and with it absent. This resulted in a mixed experimental design with brand name presence as the between subjects factor and lists as the within subjects factor.

Thirty-eight subjects, who were MBA students from the same institution as the Experiment 1 subjects, were first asked to inspect the lists of products for the purpose of stating reasons why "the products may or may not belong together." They were then asked to rate these products on a seven-point scales according to (a) how the products "fit together as a group," and (b) the extent to which they were similar to each other. The reasons and rating scale items were completed together for one set of products at time before proceeding to the next set. The three lists were presented in counterbalanced order.

While doing these tasks, half the subjects were simply presented with the list of products, while the other half received the associated brand name as a label heading for the lists and as an adjective for each product. This redundancy was intended to insure that subjects processed the brand name.

Results

Ratings of Similarity and Fit. The mean similarity and fit ratings are presented in Table 2. Overall, subjects' ratings of similarity and fit indicate that perceptions of product belongingness are greater for the functional (Sony) and usage-based (Coleman) brand names than for the symbolic (Gucci) brand name. Using seven-point scales, subjects gave mean "similarity" ratings of 5.16, 5.05 and 3.53 for the Sony, Coleman and Gucci products, respectively across the two brand name conditions (F(2,72)=29.55, P<.001), and mean "fit" ratings of 5.84, 6.29 and 4.21 for those three brands respectively (F(2,72)=36.10, P<.001). Both similarity and fit ratings were significantly higher for the functional and usage-based brand names than for the symbolic brand name. This appears to be due to the inherent product similarity present in the functional (Sony) and usage-based (Coleman) brand names when compared to those in the symbolic brand name (Gucci). However, neither rating task produced a significant main effect of brand name presence condition, with perceptions of similarity being slightly higher for the brand name absent condition (F(1,36)=.51) and perceptions of fit being slightly higher for the brand name present condition (F(1,36)=2.08, P<.15). As expected, the presence or absence of brand name did not affect the similarity judgment of products for each type of brand name.

The most interesting aspect of the data concerns the Lists by Brand Name Condition interactions. Here the prediction was that symbolic brand show higher fit (but not similarity) ratings when the brand name was present while the other brands would not be affected by the brand name presence. This pattern of results did occur. For the fit ratings, the interaction effect reached borderline significance (F(2,72)=3.01, P=.05) which reflects the fact that subjects perceived Gucci products as fitting together more (t(36)=2.07, P<.05) when the brand name was present (X=4.74) than when it was absent (X=3.68).

Such an increase in similarity ratings did not occur for the Gucci list when the brand name was present (X=3.37 vs. 3.68 for the brand name absent and brand name present conditions, respectively). It was not expected, however, that subjects perceived products as more similar when the brand name Sony was absent (X=5.74) than when it was present (X=4.58). This produced the significant interaction of lists and brand name conditions (F(2, 72) = 5.33, P<0.01) for the similarity measure. It appears that some of the products listed under the name Sony such as cassette tapes, video tapes, car stereo and turntable might not have been typically associated with the Sony brand name. This may have negatively influenced subjects' judgments of product similarity under the presence of Sony brand name. Subjects appear to be highly sensitive in judging the similarity among products under the functional brand name. This result suggests that for the functional concept the brand name extension has a very limited boundary. In summary, as predicted, the presence of the brand name enhanced fit ratings, but not similarity ratings, for the symbolic brand; and failed to increase either the fit or similarity ratings for the functional and usage-based brands.

TABLE 2

MEAN SIMILARITY AND FIT RATINGS IN EXPERIMENT 2

GENERAL DISCUSSION

The results of these two studies together suggest that different categories of brand names are structured differently in consumers' minds. Symbolic brand names, when presented to consumers in a list, are understood as representatives of a superordinate concept with a meaning akin to "products that are luxurious, high-status and expensive." By contrast, functional or usage-based brand name concepts are concepts that are not spontaneously articulated by subjects when presented with lists of brand names. Rather, brand names classified under these concepts appear to be understood in terms of their associated products without being understood in terms of the superordinate concept. Subjects' ratings of similarity and fit showed the same patterns for the functional and usage-based brand names; they were not increased by the presence of the brand name. On the other hand, subjects' judgment of how well the products fit together was enhanced by presenting the symbolic brand name.

According to our view, this occurred because subjects perceived the inherent feature-based similarity of the functional and usage-based product lists, and did not need the brand name as a cue to accomplish this. By contrast, the low similarity symbolic brand name product list was perceived as fitting together better when the brand name was presented because the brand name supplied the unifying superordinate symbolic concept that served as the basis for commonality. In other words, the presence of the Gucci brand name allowed subjects to see that such dissimilar products as belts, perfume, pens, etc. "fit together" because they could be luxury, high-status products.

We believe that these differences in memory structure for these different categories of brand name concept have definite implications for understanding how consumers perceive brand na ne extensions. Because functional and usage-based brand name concepts are understood and organized according to the product characteristics or usage situations of the products, a proposed extension will be judged according to the feature-based similarity of the new product to the existing set of products. Whatever beliefs and attitudes consumers have for the existing brand name and its products will transfer to the new product to the extent that it is similar to the features or uses of the existing products. The features or uses of the existing products will, therefore, serve as the extension boundary.

However, evaluations of extensions for symbolic brand names should follow a different process. Instead of referring to the existing products for a similarity check subjects will assess whether or not the new extension could ''fit'' that concept. One immediate implication of this distinction is that there exists a greater latitude of acceptability for extensions for symbolic brand names than for functional or usage-based brand names.

Another interesting distinction that has surfaced in this research is the one between judgments of similarity and judgments of fit. When a new instance is assessed for membership in some category, that process of categorization appears to vary depending upon the nature of the existing memory structure. One method of categorizing is based upon comparing features of the new instance to features of already categorized instances. However, the results of Experiment 2 suggest that such judgments of similarity may not be the only basis for categorization; under some circumstances, subjects make judgments of fit. In this case, the new instance is compared to an abstract concept. Further research on basis categorization processes is needed to fully understand brand name extension phenomena (cf. Cohen and Basu 1987).

Finally, the existence of different memory structures for different brand name concepts suggests that different processes are responsible for their formation. Functional or usage-based brand name concepts, whose associated products share common features, may be formed according to a "bottom-up" type of process, in which the brand name concept is abstracted as a result of promotion or being exposed to the brand name's different individual products. For example, consumers may come to understand Honda as "small, gasoline engine-based" products because of their exposure to Honda's motorcycle, compact car, lawn-mower, snow-blower, generator, etc. The formation of the symbolic brand name concept, however, may follow a different "top-down" course because the brand name is promoted with concepts that are difficult to describe with individual features. Symbolic concepts like wealth or high-status may come to be understood by consumers with unique combinations of product, person or usage situational setting. As long as products are positioned properly, establishing associations with the superordinate concept, they would be understood in terms of that concept. This may account for the formation of the symbolic brand name. Therefore, another direction for future research would be to explore how brand na ne concepts are initially formed.

REFERENCES

Aaker, David A. and Kevin Lane Keller (1987), "Consumer Response to Brand Extensions," Unpublished Working Paper, School of Business Administration, University of California, Berkeley, CA 94720.

Cohen, Joel B. (1982), 'The Role of Affect in Categorization: Towards a Reconsideration of the Concept of Attitude," in Advances in Consumer Research, Vol 9, ed. Andrew A. Mitchell, Ann Arbor, MI: Association for Consumer Research, 94-100.

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

Fiske, Susan T. (1982), "Schema-Triggered Affect: Application to Social Perception," in Affect and Cognition: The 17th Annual Carnegie Symposium on Cognition, eds. Margaret S. Clark and Susan T. Fiske, Hillsdale, NJ: Erlbaum. 55-78.

Hutchinson, J. Wesley and Joseph W. Alba (1985), "A Framework for Understanding Consumer Knowledge: I. Structural Aspects and Information Acquisition," working paper, Center for Consumer Research, University of Florida, Gainesville, FL 32611.

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

Nielson, A.C. (1985), "Testing Techniques," Vol. 1., No. 1.

Park, C. Whan, Bernard J. Jaworski and Deborah J. MacInnis (1986), "Strategic Brand Concept-Image Management," Journal of Marketing, 50, (October), 135-146.

Park, C. Whan and V. Parker Lessig (1981), "The Construct of Familiarity and Its Impact Upon Consumer Decision Process: Cognitive Biases and Heuristics," Journal of Consumer Research, 8, (September), 223-230.

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

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

University of Minnesota Consumer Behavior Seminar (1987), "Affect Generalization to Similar and Dissimilar Brand Extension," Psychology and Marketing, 4 (Fall), 225-237.

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