The Effects of Direct and Associative Brand Extension Strategies on Consumer Response to Brand Extensions

C. Whan Park, University of Pittsburgh
Michael S. McCarthy, University of Pittsburgh
Sandra J. Milberg, Georgetown University
ABSTRACT - This study was conducted in order to examine a number of important issues related to brand extension strategies. The results revealed that while the use of a direct brand extension strategy may lead to undesirable negative reciprocity effects, the use of an associative brand extension strategy may mitigate these effects while simultaneously preserving the desired effects on the evaluation of the brand extension.
[ to cite ]:
C. Whan Park, Michael S. McCarthy, and Sandra J. Milberg (1993) ,"The Effects of Direct and Associative Brand Extension Strategies on Consumer Response to Brand Extensions", in NA - Advances in Consumer Research Volume 20, eds. Leigh McAlister and Michael L. Rothschild, Provo, UT : Association for Consumer Research, Pages: 28-33.

Advances in Consumer Research Volume 20, 1993      Pages 28-33

THE EFFECTS OF DIRECT AND ASSOCIATIVE BRAND EXTENSION STRATEGIES ON CONSUMER RESPONSE TO BRAND EXTENSIONS

C. Whan Park, University of Pittsburgh

Michael S. McCarthy, University of Pittsburgh

Sandra J. Milberg, Georgetown University

ABSTRACT -

This study was conducted in order to examine a number of important issues related to brand extension strategies. The results revealed that while the use of a direct brand extension strategy may lead to undesirable negative reciprocity effects, the use of an associative brand extension strategy may mitigate these effects while simultaneously preserving the desired effects on the evaluation of the brand extension.

INTRODUCTION

A number of researchers (Keller and Aaker 1992; Park, Milberg, and Lawson 1991; Roedder-John and Loken 1990; Romeo 1991) have expressed concern that the use of brand extensions may also lead to brand image dilution or negative reciprocity effects. Reciprocity effects are generally defined as changes in the consumer's original brand attitude and beliefs caused by brand extensions. Although it is possible that reciprocity effects can be both positive (reinforcing the brand attitude and beliefs) or negative (diluting the brand attitude or beliefs), researchers and practitioners express the most concern for negative reciprocity effects. Since brand extensions attempt to leverage the strengths of the brand name, managers need to consider: (1) the effects of the original brand on the evaluation of the brand extension and (2) the reciprocity effects of the brand extension on the subsequent evaluation of the original brand. Unless reciprocity effects are carefully examined together with extension effects, the use of a brand extension strategy may not create the intended synergy between the original brand and the brand extension.

While previous research has primarily examined how the direct brand extension strategy facilitates evaluations of brand extensions, this study explores two other issues: (1) the extent to which negative reciprocity effects occur due to extending the original brand into new product categories, and (2) how the use of an associative brand extension strategy may facilitate favorable evaluations of brand extensions while protecting the original brand from negative reciprocity effects.

THEORY AND HYPOTHESES DEVELOPMENT

Since previous studies (Park et al. 1991; Roedder-John and Loken 1991) have shown that categorization and schema theories appear to provide a valid theoretical base for brand extension research, we base our predictions for reciprocity effects on the same theoretical ground. Specifically, we base our predictions upon theories of stereotypic belief change [see Crocker, Fiske, and Taylor (1984) for a review] which are also grounded in categorization and schema theory. While several models of the mechanisms underlying attitude and belief change have been proposed (e.g., schema + tag, bookkeeping, conversion and subtyping), they all agree that: (1) attitudes and beliefs change in response to new instances that vary in terms of the degree of congruence with the person's existing category or schema, and (2) that these changes occur via the processes of assimilation and accommodation (Crocker et al. 1984).

Assimilation occurs when the new instance is not very different from the existing schema. The relatively high degree of fit allows the instance to be integrated into the existing schema leaving it essentially unchanged by the incorporation of the new instance. Conversely, accommodation occurs when the new instance is very different from the existing schema, thereby requiring the schema to be altered to accommodate the new instance. The implications of assimilation and accommodation for brand extensions are quite straightforward. As the congruence, or fit, between the brand extension and the original brand schema increases, the degree of observed negative reciprocity effects should decrease.

While this seems to be a reasonable expectation, previous research on negative reciprocity effects (Roedder-John and Loken 1990; Romeo 1991; Keller and Aaker 1992) have not demonstrated consistent evidence of such effects. One possible explanation is that these studies manipulated the fit between the original brand and the extension along only one dimension such as typicality (Roedder-John and Loken 1990) or similarity (Romeo 1991, Keller and Aaker 1992). However, Park et al. (1991) demonstrated that different types of brand extension fit, specifically brand concept consistency and product level similarity, moderate the evaluation of extension products utilizing the direct brand extension strategy. Therefore, it is possible that different types of brand extension fit may also have different effects on the degree to which negative reciprocity effects occur. Given this premise, the present study examines how the brand concept consistency and product level similarity of a brand extension affects the extent of negative reciprocity effects caused by that extension.

In the event that brand extensions do lead to negative reciprocity effects, it would be desirable to identify an alternate branding approach which would allow marketers to realize the desirable synergistic effect of brand name leveraging, without the undesirable negative reciprocity effects. We propose the use of an associative brand strategy (hereafter referred to as associative branding), wherein a new brand name is used in combination with an existing brand name in the introduction of a new product For example, Marriott has used associative branding twice in the introduction of two new motel chains: Courtyard by Marriott and Fairfield Inn by Marriott, offering a different level of service and accommodations than those generally associated with the Marriott hotel chain.

The use of associative branding by practitioners appears to be based upon the assumption that the combination of a new brand name with a familiar brand name allows the consumer to selectively transfer existing brand attitudes and beliefs from the existing brand memory to the associated brand. This allows the consumer to differentiate the extension from the other products associated with the brand name, thereby reducing the likelihood of negative reciprocity effects. This explanation is consistent with the schema+tag model of assimilation (Graesser, Gordon, and Sawyer 1979) which would consider the associative brand name to act as a schematic "tag", allowing the consumer to segregate the discrepancies between the brand extension and the existing brand schema.

The above discussion suggests that when brand extensions exhibit high levels of fit, there should be no difference between the direct and associative branding strategies with regards to negative reciprocity effects. However, when the level of brand extension fit is low the associative branding strategy should result in lower levels of negative reciprocity effects than the direct extension strategy. Since it is difficult to know a priori precisely how poor the degree of fit must be to obtain negative reciprocity effects, we define low fit as existing when either product level similarity and/or brand concept consistency is low. Based upon this definition, we hypothesize:

H1: Compared to the direct brand extension strategy, an associative branding strategy mitigates negative reciprocity effects for brand extensions exhibiting low product similarity and/or low brand concept consistency.

TABLE 1

BRAND EXTENSIONS USED IN EXPERIMENT 1 BY BRAND CONCEPT AND PRODUCT SIMILARITY

Although we expect the associative brand strategy to mitigate negative reciprocity effects where they occur, we can not conclude that associative branding will be preferred to direct brand extensions unless we can determine that associative branding does not also mitigate the positive effects of the brand name association on the evaluations of the brand extension. In other words, we may not be inclined to use associative branding if it were to reduce the desired extension side effects. Therefore, we must consider how the use of associative branding affects brand extension evaluations.

Prior brand extension research has proposed that, depending upon the degree of brand extension fit, existing brand attitudes and beliefs transfer from the brand to the extension. However, since we would not expect consumers to have either attitudes or beliefs toward a unfamiliar associative brand name, we would not expect any attitudes or beliefs to transfer from the associative brand to the brand extension. Therefore, under both the direct and associative branding strategies, the existing brand name should be the only source from which brand attitudes and beliefs can transfer to the brand extension. Based upon this reasoning, we expect that both the direct and associative branding strategies will lead to similar brand extension evaluations, and we advance the following hypothesis:

H2: Compared to the direct brand extension strategy, an associative branding strategy has no differential effect on the evaluations of brand extension.

METHOD

Overview

The hypotheses were tested using methods and procedures based upon Park et al. (1991). The stimuli consisted of two brands, Timex and Rolex, and eight extension products. Timex and Rolex were chosen as the parent brands for a number of reasons. First of all, Timex and Rolex respectively represent a functional and prestige brand concept. Since Park et al. (1991) showed that subjects differed in their responses to the same extension of a functional and prestige brand we wanted to explore if these types of differences would be replicated with respect to negative reciprocity effects. Moreover, these brands exhibited two additional characteristics important to the objectives of this study. First, both brands were real (not fictitious) brand names that were highly familiar to subjects, thereby increasing the external validity of our findings. Second, neither brand had been extended to other product categories at the time of data collection. This allowed us to limit our product similarity manipulations to reflect only the wristwatch product category. The eight brand extension products used in this study (see Table 1) are a subset of the twelve extensions used by Park et al., and were pretested to vary in terms of product similarity (high vs. low) and product concept dominance (functional vs. prestige).

Design and Procedure

The design is a 2 x 2 x (2 x 2) mixed design with brand name (Timex or Rolex) and brand extension naming strategy (direct vs. associative branding) as the between-subjects factors and product similarity (high or low) and brand concept consistency (high vs. low) as the two within-subjects factors. Brand concept consistency was manipulated by the combination of the brand name (Timex or Rolex) with the product concept dominance of the extension (functional vs. prestige). For example, when the brand name was Timex (Rolex), functional products such as a calculator or smoke detector were considered to be high (low) in brand concept consistency. However, products such as necktie and cologne were considered to be low (high) in brand concept consistency with Timex (Rolex).

The subjects were 96 students enrolled in the full-time MBA program at a large eastern university and participated during their regular class meeting. Subjects were randomly assigned to one of the four between-subjects conditions. In order to avoid subject fatigue, each subject evaluated only four of the eight products in Table 1, with each product representing one of the four within-subjects conditions. The eight products were randomly assigned to one of two groups: (batteries, garage door opener, ring, and necktie or calculator, smoke detector, bracelet, and cologne). Therefore, an additional between-subjects factor, serving as a replicate, was the group of products evaluated by the subject. However, since this between-subjects factor had no effect on the results, it will not be discussed further.

In order to to identify an unfamiliar brand name that could be used as an associative brand name with Timex and Rolex, a pretest was conducted with 28 part-time MBA students. Based upon the results of the pretest, the brand name of Syntax was selected because its mean familiarity rating was only 1.71 on a seven-point scale (1=not at all familiar, 7=very familiar with the brand name) while Timex and Rolex rated a 6.02 and a 6.39 familiarity rating respectively.

The procedure was as follows. Subjects were advised that they would be evaluating a number of new product ideas for Timex (or Rolex). Subjects responded by answering a series of seven-point scale items which contained the manipulation checks and the dependent measures which are described with the results. The entire procedure took 15 to 20 minutes and subjects were debriefed at a later date.

RESULTS

Manipulation Checks

Consistent with the findings of Park et al. (1991) the results indicated successful manipulations of product similarity and product concept dominance. Product similarity to a watch was measured by indicating the degree to which each product went together with a watch (1=not at all, 7=very much). As expected, products initially assigned to the high similarity condition were rated higher in similarity to a watch (=4.19) than those assigned to the low similarity condition (=2.38, t(95)=12.6, p=.000).

Functional (prestige) product concept dominance was measured by indicating how important the characteristics of reliability and durability (luxury and status) were when considering the purchase of the product. As expected, the products assigned to the functional-orientation condition were judged to have higher functional importance (=6.21) than those assigned to the prestige-orientation condition (=3.43, t(95)=17.1, p=.000). Likewise, the products assigned to the prestige-orientation condition were judged to have higher prestige importance (=6.04) than those assigned to the functional-orientation condition (=2.76, t(95)=22.6, p=.000).

Hypothesis Testing

The tests of Hypothesis 1 were based upon the change in the subject's attitude toward the brand, which was operationalized as the difference between the pre- and post-extension measures of overall brand attitude in the following manner. First, pre- and post-extension brand attitudes were based upon measures of brand liking (1=disliked very much, 7=liked very much) and brand feelings (1=feel very bad about the brand, 7=feel very good about the brand) which were combined into a single brand attitude measure due to their high correlation (r=.782, p<.01). Then, the pre-extension brand evaluation was subtracted from the post-extension brand evaluation with negative differences indicating negative reciprocity effects. Therefore, the dependent measure for the reciprocity effect caused by a particular extension was computed by subtracting the brand evaluation prior to the consideration of any brand extensions from the brand evaluation given after the consideration of the brand extension of interest.

Hypothesis 1 was tested by comparing the significance of the negative reciprocity effects observed under the direct brand extension strategy with the the significance of the negative reciprocity effects observed under the associative branding strategy. If the difference score was significantly different from zero in the direct extension condition, then negative reciprocity effects were said to have been obtained. Evidence of mitigation of negative reciprocity effects was obtained if the difference score was not significantly different from zero in the corresponding associative branding condition.

Hypothesis 1 predicted that associative branding would mitigate negative reciprocity effects for low product similarity and/or low brand concept consistency extensions. Figure 1 presents the findings which, interestingly, supported this prediction in the Timex condition but not in the Rolex condition.

In every case, directly extending the Timex brand led to significant negative reciprocity effects on the Timex brand attitude while using an associative branding strategy did not. When the Timex extensions exhibited only low product similarity, the negative reciprocity effects were significantly different from zero under the direct strategy (functional=-.40, t(24)=2.49, p<.01), but not so under the associative branding strategy (functional=-.21, t(20)=1.14, p=n.s.). Likewise, when the Timex extensions exhibited only low brand concept consistency the negative reciprocity effects were significantly different from zero under the direct strategy (prestige=-.36, t(24)=2.26, p<.05), but not under the associative branding strategy (prestige=-.05, t(20)=0.26, p=n.s.). Finally,when the Timex extensions exhibited both low product similarity and low brand concept consistency the negative reciprocity effects were significantly different from zero under the direct strategy (prestige=-.88, t(24)=4.21, p<.001), but not under the associative branding strategy (prestige=-.45, t(20)=1.85, p=n.s.). These results strongly support H1 for the Timex case. In addition, when Timex extensions exhibited both high product similarity and high brand concept consistency the negative reciprocity effects were significantly different from zero under the direct strategy (prestige=-.32, t(24)=2.03, p<.06), but not under the associative branding strategy (prestige=.02, t(20)=0.14, p=n.s.). Although this result was unexpected, it indicates that even high-fitting extensions may lead to undesirable negative reciprocity effects which can, in turn, be mitigated by the use of an associative branding strategy.

An examination of the means for Rolex indicates that although there was directional support for the predictions of H1, there was no significant effect of associative branding for any of the Rolex brand extensions. Taken together, the results of the Timex and Rolex conditions seem to indicate that associative branding may be effective at mitigating negative reciprocity effects for functional brands, but not for prestige brands.

Hypothesis 2 predicted that there would be no differential effect of associative branding on the evaluation of brand extensions for either the product-related or prestige brands. This hypothesis was tested by using the overall evaluation of the brand extension which was computed by averaging the two measures of the attitude toward the brand extension (feel good/bad, like/dislike) which were highly correlated (r=.795, p<.01).

A 2 (Timex vs. Rolex) x 2 (direct vs. associative branding strategy) x 2 (high vs. low product similarity) x 2 (functional or prestige-orientation) mixed design ANOVA was used to test Hypothesis 2. Figure 2 presents the results of this analysis which strongly supported Hypothesis 2. First of all, none of the main or interaction effects of strategy on brand extension evaluations achieved significance at the a=.05 level. Furthermore, an additional analysis comparing the means of the direct and associative branding strategies showed that in only one case did the means of two strategies differ significantly. When Timex extensions exhibited high product similarity and low brand concept consistency, using an associative branding strategy improved the evaluation of the extension (Direct=2.0, Associative brand=2.7, t(45)=1.99, p=.055). The combination of these findings supports the expectation that using the associative branding strategy should not adversely affect brand extension evaluations.

In addition, an analysis of subjects in the direct brand extension condition provided a replication of Park et al. (1991). Specifically, a separate 2 (Timex vs. Rolex) x 2 (high vs. low product similarity) x 2 (functional or prestige-orientation) mixed design ANOVA was performed on only the subjects in the direct branding condition. This MANOVA revealed: (1) a main effect of product similarity (High=3.55, Low=2.95, F(1,49)=28.03, p=.000), (2) a brand by concept consistency interaction (F(1,49)=99.6, p=.000), and (3) a three way interaction of brand, concept consistency, and product similarity (F(1,49)=7.46, p=.009). The three-way interaction of brand, concept consistency, and product similarity can be seen by examining the solid lines in Figure 2. This pattern of results is virtually identical to those obtained by Park et al. (1991).

FIGURE 1

THE EFFECT OF DIRECT AND ASSOCIATIVE BRANDING STRATEGIES ON NEGATIVE RECIPROCITY EFFECTS FOR TIMEX AND ROLEX

DISCUSSION

The results of the present study suggest that not only can poor fitting brand extensions cause negative reciprocity effects but that negative reciprocity effects can occur even for high fitting brand extensions. This implies that negative reciprocity effects can occur even when consumer evaluations of the brand extension are favorable. The results of this study also raise an important issue as to why the associative branding strategy effectively mitigated negative reciprocity effects for the Timex brand extensions, but not for the extensions of the Rolex brand. One possible explanation is the possibility of brand memory structure differences between the Timex and Rolex brands (Park, Lawson,and Milberg 1989). Assuming that Timex is primarily associated with the wristwatch product class, consumers might have difficulty assimilating non-wristwatch brand extensions into their memory structure of Timex. The associative branding strategy might facilitate this type of assimilation process by offering a basis for schema+tag processing (Graesser, Gordon, and Sawyer 1979), thereby mitigating negative reciprocity effects. However, assuming that the Rolex brand is primarily associated with the concepts of status and prestige, when the Rolex brand is extended via associative branding to products that do not share a prestige or status concept, consumers still perceive generally high levels of brand schema incongruity. Therefore, even though associative branding still appears to help to some degree, it is unable to compensate for the perceived brand category incongruity enough to reduce the negative reciprocity effects to insignificant levels. While this explanation seems plausible, more research focusing on this issue is necessary.

FIGURE 2

THE EFFECT OF DIRECT ASSOCIATIVE BRANDING STRATEGIES ON BRAND EXTENSION EVALUATIONS

It should be noted that while the results of the present study offer several important findings, they also need to be understood within the study's potential limitations. Specifically, the findings of the present study are the result of within-subjects measurement of pre- and post-extension brand attitudes. Although this type of design allows us to measure the change in brand attitude due to the brand extensions, it also leaves open the possibility of demand effects. However, given the discrepancy between the Timex and Rolex results, the likelihood of demand effects seems remote since subjects should have been similarly affected by experimental demand. This explanation notwithstanding, the results of this study need to be replicated in a between-subject environment where the possibility of demand effects is much smaller. Furthermore, the limited nature of the stimuli used in this study (i.e., two brands and eight extensions) makes the generalizability of these results somewhat tenuous.

REFERENCES

Crocker, Jennifer, Susan T. Fiske, and Shelly E. Taylor (1984), "Schematic Bases of Belief Change," in Attitudinal Judgment, J. Richard Eiser, ed. New York, NY: Springer-Verlag, 197-226.

Graesser, Arthur C., Sallie E. Gordon, and John D. Sawyer (1979), "Recognition Memory for Typical and Atypical Actions in Scripted Activities: Tests of a Script Pointer + Tag Hypothesis," Journal of Verbal Learning and Verbal Behavior, 18 (June), 319-332.

Keller, Kevin Lane, and David A. Aaker (1992), "The Effects of Sequential Introduction of Brand Extensions", Journal of Marketing Research, 29 (February), 35-50.

Park, C. Whan, Robert Lawson, and Sandra Milberg (1989), "Memory Structure of Brand Names," in Advances in Consumer Research, Vol 16, ed. Thomas Srull, Provo, UT: Association for Consumer Research, 726-731.

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

Roedder-John, Deborah and Barbara Loken (1990), "Diluting Brand Equity: The Negative Impact of Brand Extensions", working paper, Carlson School of Management, University of Minnesota.

Romeo, Jean B. (1991), "The Effect of Negative Information on the Evaluations of Brand Extensions and the Family Brand" in Advances in Consumer Research, Vol. 18, Rebecca H. Holman and Michael R. Solomon, eds. Provo UT: Association for Consumer Research, 399-406.

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