Role of Product Knowledge in Evaluation of Brand Extension

ABSTRACT - This paper examines the role of product knowledge in consumer evaluation of brand extensions. Specifically, we give a set of hypotheses on the moderating effect of the variables of product knowledge and type of similarity on similarity judgment between original and new product categories as well as on attitude extension. An experiment was conducted to test these hypotheses and the results of this experiment support some of our predictions.

Citation:

A. V. Muthukrishnan and Barton A. Weitz (1991) ,"Role of Product Knowledge in Evaluation of Brand Extension", in NA - Advances in Consumer Research Volume 18, eds. Rebecca H. Holman and Michael R. Solomon, Provo, UT : Association for Consumer Research, Pages: 407-413.

Advances in Consumer Research Volume 18, 1991      Pages 407-413

ROLE OF PRODUCT KNOWLEDGE IN EVALUATION OF BRAND EXTENSION

A. V. Muthukrishnan, University of Florida

Barton A. Weitz, University of Florida

[We thank John Lynch, Jr., and Rich Lutz for their suggestions at various stages of this project. We also thank Joe Alba, S. Ratneshwar, Barb Bickart and Susan Broniarczyk for their comments on the earlier versions of the manuscript.]

ABSTRACT -

This paper examines the role of product knowledge in consumer evaluation of brand extensions. Specifically, we give a set of hypotheses on the moderating effect of the variables of product knowledge and type of similarity on similarity judgment between original and new product categories as well as on attitude extension. An experiment was conducted to test these hypotheses and the results of this experiment support some of our predictions.

INTRODUCTION

In recent years there has been an increased interest in research on brand extension. The thrust of this research has been on identifying a set of factors determining the success of brand extension (for example Aaker and Keller, 1990; Boush et al, 1987; Smith, 1990). The studies examining brand extension from a micro perspective concentrated on the product and brand characteristics that account for the judgement of similarity between the original and new product with the same brand name. However, one of the issues not addressed by these studies is the interaction of individual characteristics with brand or product characteristics in similarity judgment as well as in attitude extension. This paper examines the ,ole of an important individual characteristic, product knowledge, in similarity judgment as well as in attitude extension.

Previous Research on Brand Extension

Though there is much evidence for the success of brand extensions (Tauber, 1981; 1988), there are many instances in which brand extensions resulted in diluting the value of original brand equity (Leuthesser, 1988). Recent research on branding has concluded that being a well known brand is not sufficient for extending the brand name to other product categories. As Tauber (1988) states, few consumers would want Jell-O shoe laces or Tide frozen entrees. The key to brand extension then is the logical fit or perceived similarity between the original and the new product bearing the same brand name.

The role of similarity in product evaluation has been examined by a few authors. The earliest studies in this area concluded that brand extension will be successful only when the consumers perceive a high degree of similarity between the original and extended product categories (for example, Boush et al 1987). The research that followed investigated the possible bases of similarity. Tauber (1988) after analyzing actual brand extensions involving 115 different brands in consumer durable and non durable categories suggests seven types of leverages a company should consider when seeking to extend its brands. These are 1) same products in different forms, 2) distinctive taste/ingredient/ component in the new item, 3) companion products, 4) same customer franchise 5) technical expertise 6) benefit/attribute/feature owned and 7) designer image/status. He terms the first two types of leverages as line extension opportunity and the rest as brand extension opportunity. The assumption behind the types of possible leverages listed above may be that a consumer sees some form of consistency between the original product and the new product. More recently some consumer researchers identified the types of perceived consistency or similarity used in product evaluation and their effect on the success of brand extension.

Although not specifically investigating the issue of brand extension, a study by MacInnis and Nakamoto (1989) identified the possible bases of similarity between pairs of products. While product attributes were the most recalled features, benefits and usage situations played the most important role in similarity judgment. The other factors in this study like manufacturing requirements and marketing factors did not have a significant role in the similarity judgment.

In a recently published article, Aaker and Keller (1990) present the results of two studies that explored the role of a set of moderator variables on attitude extension. In the first study the authors proposed that affect transfer to hypothetical brand extensions may depend on 1) the quality perceptions of the brand in the original product category, 2) the extent to which the new product is perceived as a "substitute" for or "complement" to the original product, 3) the extent to which the firm's manufacturing expertise could be transferred to the new product and 4) perceived difficulty of extension. The authors found that of the proposed set of factors only two,"transfer" and ''difficulty of extension", had a significant effect on attitude toward the extended brands. In the second study, they manipulated positive quality cues about the brand (cue present or absent) and opportunity to elaborate (a neutral description of the extension attributes alone or in combination with positive quality cues). It was found that providing positive quality cues did not affect attitude extension while opportunity to elaborate did.

The studies discussed above have certainly improved our understanding of the factors accounting for the success of brand extension. However, there is a need to incorporate individual differences such as product knowledge as explanatory variables. As product evaluation may be a function of familiarity with and expertise in a particular product category, these variables may play a significant role in attitude extension. More interestingly, these knowledge related variables may interact with product characteristics to affect one's similarity perceptions as well as attitude extension. The following section gives the theoretical rationale for including the variable of product knowledge in our study.

Product Knowledge and Attitude Extension

In recent years there has been a substantial amount of research on the role of product knowledge on various stages of consumer behavior (for example Bettman and Park, 1980; Brucks, 1985; Johnson and Russo, 1984; Rao and Monroe, 1987 and Sujan, 1985). These studies conclude that the decision processes and strategies of consumers who are high on product knowledge differ from those who are low. Based on an extensive conceptual analysis, Alba and Hutchinson (1987) suggest that familiarity (the number of product related experiences that have been accumulated by the consumer) and expertise (the ability to perform product related tasks successfully) are two separate components of product knowledge. Furthermore, these authors suggest that experts are superior to novices in terms of their cognitive structure, analytic capabilities, ability to make elaborate inferences and memory capabilities. On the basis of these conclusions, we suggest that the basis of similarity or fit judgment in an attitude extension may not be uniform across all segments of consumers; it may vary between experts and novices.

We extrapolate the following propositions of Alba and Hutchinson (1987) to the specific context of attitude extension. Because of their deep, richly intertwined category structure, experts may be able to comprehend similarity between two different classes of products although surface perceptual cues may not suggest any obvious similarity between those products. When novices process information selectively, they are more likely than experts to select peripheral surface cues for judging a brand extension processing (Alba and Hutchinson, 1987)

Furthermore, experts may more often use elaborate inferences to find fit between two product categories. For example, Gillette once test marketed a line of blank tape cassettes. Though, the Gillette name is not associated with sound reproduction, knowledge of the commonalities between the production of cassette tapes and shaving equipments made the association between the brand and the product seem less incongruent (Alba and Hutchinson, 1985).

Thus experts may find similarity between the original product and the new product with the same brand name on the basis of deep cues. We define deep cues as those factors that may account for the performance of the product in the original category and may also be related to the performance in the new category. At the attribute level, these may include similarity in terms of technology, design and fabrication and materials and components used in the manufacturing process. These may also include abstract benefits that require elaborate inference making.

Novices, on the other hand, may tend to relate the original and new products on the basis of surface level cues. These factors are not related to the performance of products. Rather these may include perceptual cues like package, shape, color, size, etc., perception of two products being either substitutes or complements, similar retail outlets and similar promotional techniques used. Also benefits that require no elaborate inference making may come under this category.

The findings of several streams of research in the areas of attitude formation and change support our propositions. Our proposition that similarity between two products may be judged on the basis of either surface level factors or deeper level factors is similar to the propositions of Petty and Cacioppo (1981) and Chaiken (1980). In both these models, ability is one of the factors that decide the type of processing a person engages in. In a framework on relationship among belief,attitude,intention and behavior, Feldman and Lynch (1988) propose that prior knowledge may be one of the factors that determine the "perceived diagnosticity" of an input.

As previous studies have shown, attitude extension is a function of perceived similarity between two products. Based on the logic given above we suggest that experts are more likely to transfer their attitude when the original and new product categories are similar in terms of deep properties. Conversely, novices are more likely to transfer their attitude when categories are similar only in terms of surface level factors. It has to be noted, however, that when two products are related at surface level, experts as well as novices may identify-the similarity; Unlike novices, experts may not perceive such a similarity diagnostic and hence may not transfer their attitude from the original product to the new product bearing the same brand name.

The above discussion leads us to the following set of hypotheses:

H1(a): When similarity between the original product and the new product with the same brand name is based on surface factors, experts as well as novices are likely to identify the similarity.

H1(b): When similarity between the original product and the new product with the same brand name is based on deep factors, experts are more likely than novices to identify this similarity more accurately.

H2(a): When similarity between the original product and the new product with the same brand name is based on surface factors, novices are more likely than experts to transfer their positive or negative attitude toward the original product to the new product.

H2(b): When similarity between the original product and the new product with the same brand name is based on deep factors, experts are more likely than novices to transfer their positive or negative attitude toward the original product to the new product.

Although these hypotheses were proposed explicitly for the variable of expertise, we expected the same pattern of results for the variable of familiarity.

METHOD

Subjects

Subjects in this study were undergraduate students enrolled for Marketing Management and Marketing Research courses at a South-eastern university. There were 106 subjects (52% female and 48% male). The subjects participated in the study as a part of a course requirement.

Procedure

The subjects were told that the study was being conducted by the university for a sporting goods manufacturer. To increase the response involvement, subjects were told that their responses would be used as input for the product and promotional strategy of the company. Then they were asked to fill out a four section questionnaire. Each section of the questionnaire was distributed and collected separately. The sections of the questionnaire asked for information in the following order: a. attitude toward the brands in the original product category; b. attitude toward hypothetical brand extensions; c. judgment of similarity between original and new products; d. the level of familiarity with and expertise in the original as well as new categories. At the end of the session subjects were debriefed.

Design

There were three factors in this study product knowledge at two levels (expert vs novice and high familiarity vs low familiarity), basis of similarity at two levels (deep vs surface), and attitude toward brands in the original category at two levels (positive vs negative). Product knowledge was a between subjects factor and the other two were with in subject factors. There were two brands nested in each combination of the two with in subjects factors and "brands" was treated as a fixed factor. Finally the set of subjects was treated as random factor nested in product knowledge.

Independent Variables

The study was conducted with tennis racquet as the proposed extension category using brand names of existing tennis shoes or golf clubs. Tennis rackets bear a surface similarity to tennis shoes. Probably the most obvious relationships between these products are that both of them are tennis products and thus may be viewed as complements and both are sold at the same retail outlets.

Tennis rackets were considered to be similar to golf clubs at a deeper level. It was learned through lengthy discussions with sporting goods sellers that same materials were used to manufacture components of these two products. Further the design and fabrication techniques for these products were quite similar.

The second factor in the study was the initial attitude toward the brand in the original category. The brands of golf clubs and tennis shoes considered by sellers of sporting goods to be high quality were used as "positive brands" and the ones they identified as low quality were used as "negative brands". Their classification highly correlated with the sales figures for these brands, suggesting that consumers may share their perceptions of quality.

The third factor was the level of subjects' knowledge in tennis and golf products. -Both familiarity and expertise were measured, in order to ascertain whether differently moderated similarity judgments and attitude extension. Expertise was measured through two item self rating. In previous studies this measure of subjective knowledge was found to be correlated with objective knowledge (for example, Brucks, 1985). Familiarity was measured through a nine item scale. Some of the items were similar to those used by Sujan(1985) to measure familiarity in another product domain. Others were constructed on the lines of the components of familiarity identified by Alba and Hutchinson (1987). Appendix- A gives the items used to measure expertise and the contents of familiarity scale. Cronbach's alpha for-subjective expertise was 0.969 for tennis shoes, 0.985 for golf club and 0.981 for tennis rackets. Median split was used to separate high and low knowledge groups in each of the product categories. Subjects falling in the upper half of all the three categories were classified as "high familiar group" and the others were classified as "low familiar group". A similar procedure was used to classify subjects as experts and novices.

Dependent Variables

Attitude toward brands in the original category, attitude toward the new product with the same brand names and judgment of similarity between the original and new product categories were used as dependent variables. Attitude toward brands in the original categories was included as a check for our manipulation of positive and negative brands. The manipulation check revealed that the ratings of the subjects did not agree with those of the sellers in the case of two brands each in tennis shoes and golf club categories. Hence although we collected data for six brands in each category, only data for four brands in each category were included in the final analysis.

Based on attitude scales used in a number of previous studies, attitudes toward brands in original as well as extended category was measured through a scale of four 7 point semantic differential items. The items in the scale were positive-negative, favorable-unfavorable, good-bad and like-dislike. Cronbach's alpha for scale items for attitude toward the original brand was 0 754 and it was 0.868 for scale items for attitude toward the extended brand. The overall attitude was the average of these four items.

The perceived similarity between the original category and new category was measured through two items (1= not at all similar to 7= highly similar and 1=not at all related to 7 extremely related). Cronbach's alpha for these items was 0.911 for the pair of tennis shoes-tennis racquet and 0.930 for the pair of golf club-tennis racquet. The subjects were also asked to give reasons for their similarity judgments which served as an indicator of the possible bases of similarity. Later these responses were coded by one of the researchers as deep and surface level factors.

Plan of Analysis

To test Hypotheses 1a. and 1b. two models were used. In the first model either expertise or familiarity and type of similarity (manipulated) were the independent variables and subjects' rating of similarity was the dependent variable. Expertise/familiarity was a between subjects variable and type of similarity was a two levels (surface and deep) repeated factor. A mixed factor analysis of variance was used to test these hypotheses. To support these hypotheses there should be a significant simple effect of product knowledge at deep level of similarity and there should be no such effect at surface level of similarity. In other words, there should be a significant expertise/familiarity * Type of similarity interaction.

In the second model, the reasons for similarity judgment for the pairs of tennis shoes, tennis racquet and golf club-tennis racquet,coded as accurate and inaccurate was used as the dependent variable and familiarity/expertise was the independent variable. This model was an additional test of hypothesis lb. If our assertion that experts would identify the basis of similarity more accurately when two products are related at a deeper level is correct, then there should be a significant effect of expertise/familiarity. Categorical linear model (CATMOD) was used as the technique of analysis.

RESULTS AND DISCUSSION

Manipulation Checks

As mentioned earlier, data on some brands had to be excluded from analysis as the initial attitude expressed by a majority of subjects did not agree with the opinions of sellers. We conducted manipulation checks to ensure that the eight brands included in final analysis were perceived as positive or negative by subjects. We were interested in two effects. First, to show that our manipulations worked, there should be a significant main effect of the factor of initial attitude. Second, we tested whether the difference in liking between the positive and negative brands of tennis shoes was equal for experts and novices and whether the analogous difference in liking for positive and negative brands of golf clubs was equal for experts and novices. Such expert novice difference in attitudes toward the original brands could artificially produce differences in liking for the extended brands. This explanation could be ruled out by finding no significant three way interaction of product knowledge*initial attitude*type of similarity.

The results we obtained confirmed that the initial attitude manipulation was strong (F (1,104)=97.16, p < 0.0001) [In this section we are only discussing results for analyses using expertise as an independent variable. We did similar analyses using familiarity as an independent variable. The results supported Hypotheses 2(a) and 2(b) but not hypotheses 1(a) and 1(b). The results are given in the footnote.]. The three way interaction was not significant (F (1,104)= 2.35, p > 0.128). This shows that there was no expert-novice differences in the liking for brands in the original product categories.

Similarity Judgments of Experts and Novices

The model using-expertise as a between subjects variable, type of similarity as repeated factor and similarity judgment as dependent variable yielded partial support for our hypotheses l(a) and 1(b). Table 1 gives the type of similarity x expertise mean scores.

The type of similarity x expertise interaction was highly significant ( F (1,104) =23.9, p < 0.0001). The cell means for expertise x type of similarity were in the predicted direction. However, contrary to our prediction that the effect of expertise would be significant only at deeper level of similarity,we found that expertise had a significant simple effect at surface level (F = 8.36 p < 0.004) as well as at deep level of similarity (F = 16.9 p < 0.0001). These results suggest an interesting possibility that experts may perceive two product categories less similar when these categories are related only on the basis of surface factors.

Table 2 gives the results of categorical linear analysis done with bases of similarity rating (coded as relevant or irrelevant) given for the pairs of tennis shoes-tennis racquet ("surface" pair) and golf club - tennis racket ("deep" pair) as the dependent variable and expertise as independent variable. This is an additional test for Hypotheses 1(a) and l(b).

We obtained strong support for our hypothesis 1 b(chi-square (1)=36.9, p < 0.0001). Experts gave deeper level reasons for similarity between the product categories of tennis rackets and golf clubs. Novices either judged the similarity to be low or gave surface level reasons for their judgment. We combined these two categories of responses and called it as inaccurate basis of similarity. Table 3 gives the frequencies for expertise by type of reasons for similarity for the pair of tennis shoes-tennis rackets. No one gave a reason that could be identified as deep. However, a few subjects perceived no similarity between these two product categories. Hence, the levels of the factor basis of similarity are appropriate (surface level) and inappropriate (no similarity between these two product categories).

TABLE 1

SIMILARITY JUDGMENT OF EXPERTS AND NOVICES

TABLE 2

FREQUENCIES OF REASONS FOR SIMILARITY X EXPERTISE FOR DEEP SIMILARITY PAIR

TABLE 3

FREQUENCIES OF REASONS FOR SIMILARITY X EXPERTISE FOR SURFACE SIMILARITY PAIR

As there are very few frequencies in the "inappropriate" cells, no further analysis is done on this data. An examination of frequencies in the cells appears to support hypothesis 1 (a).

Test of Hypotheses on Attitude extension

Table 4 gives the means of Expert x initial Attitude x Type of Similarity.

The three way interaction was significant, (F (1,104) = 9.58, p < 0.003). Since the interaction is significant, we conducted a follow up test. In this test, we compared type of similarity x initial attitude interaction at each level of expertise. For the experts this interaction was moderately significant (F (1, 18) = 3.68, p < 0.08). For novices this interaction was highly significant (F (1, 86) = 8.92, p < 0.004. The directions were as predicted. Since the experts were smaller in number, power to detect this interaction effect may be low. This may explain the lower statistical significance for experts.

Limitations and Future Research Directions

Although we found evidence for most of our hypotheses, there are certain limitations in this study. First of all, we used only one product in each of the categories relevant to our research objectives. Secondly, the most important variable in this research, product knowledge, was measured rather than manipulated. However we recognize that manipulating product knowledge may be an extremely difficult task. A more important point is that when brands were treated as random factor nested in initial attitude and type of similarity, appropriate error terms changed, and the observed treatment effects sometimes failed to attain statistical significance. Hence the findings may not be generalizable to products or brands other than the ones used in this study. It may be necessary to replicate this study using a number of products/brands. Finally, in the process of categorization, we suggested that our novice subjects' evaluation may be based on a similarity judgment. They may identify similarity between tennis shoes and tennis rackets as they are used in the same sport (complements). However it is also possible that they just went by the category label 'Tennis" as this label was missing in the other pair. [We thank Susan Broniarczyk for suggesting this possibility.] Hence it is possible that novices' evaluation simply may be on the basis of the category label and not on the basis of perceptual fit of surface level factors. Future research could examine these competing explanations on the process of attitude extension.

TABLE 4

ATTITUDE EXTENSION BY EXPERTS AND NOVICES

GENERAL DISCUSSION AND CONCLUSION

In this research we suggested that experts and novices may differ in their reactions to brand extension on the basis of the type of relationship between the original product and the new product with the same brand name. Our results may have implications in the areas of branding and promotion. When a firm extends its brand name to another product category on the basis of commonality in technology, experts may appreciate such an extension while novices may not. Similarly, novices may find surface factors diagnostic while novices may not.In such a case, it will be advantageous for the firm to educate consumers on the technical/manufacturing commonalities and convince them that the extension is logical. The expert - novice differences may also be useful in positioning of the extended brand. However, when the global image of the firm is very high then there may not be any expert - novice difference in product evaluation. In such an instance experts as well as novices may have high attitude toward extended brand.

FOOTNOTE

The analyses with familiarity instead of expertise as an independent variable yielded results somewhat different from those reported in the results and discussion section. The type of similarity x familiarity was not significant at 0.05 level but was significant at 041 level (F (1,104) = 2.93, p < 0.1). For the pair of tennis racquet and golf club basis of judgment x familiarity interaction was significant (chi-square (1)= 12g, P < .0005. The three way interaction among familiarity, initial attitude and type of similarity was significant (F (1,104)=4.38, p < 0.05). However, the mean in positive-surface condition (3.24) was higher than the mean in positive-deep condition (2.76) for High familiar group. This is contrary to our prediction. This shows that as Alba and Hutchinson (1987) argue, familiarity and expertise are different constructs and treating them as same may cause serious construct validity problems.

APPENDIX - A

The contents of Familiarity Scale

The items included in the familiarity scale sought information on (1) How long the subject been playing the sport? (2) How often does she play? (3) How often does she visit stores exclusively selling these products? (4) How many times has she bought tennis rackets, tennis shoes and golf clubs for herself or for others? (5) How often and how much is she consulted by her friends and relatives in the purchase of these products? (6) How much of attention does she pay to the ads for these products? (7) How many brands in each of these product categories could she recall? (8) What magazines pertaining to these product categories does she read? and (9) How much of time does she spend watching TV programs on these sports?

The subjective expertise scale had two items 1) I consider myself least knowledgeable = 1 to highly knowledgeable = 7 and 2) I consider myself novice = 1 to expert = 7. These were measured for each product category separately.

REFERENCES

Aaker, D and K. Keller (1990), "Consumer Evaluations of Brand Extensions", Journal of Marketing, 54, 27-41.

Alba, J. W and J. W. Hutchinson (1985), "A Framework for Understanding Consumer Knowledge II. Comparison and Inference Process", Working Paper, University of Florida, Center for Consumer Research.

Alba, J. W. (1987), "Dimensions of Consumer Expertise", Journal of Consumer Research, 13, 411 -454.

Bettman, J. R and C. W. Park (1980), "Effects of Prior Knowledge and Experience and Phase of the Choice Process on Consumer Decision Processes: A Protocol Analysis", Journal of Consumer Research, 7, 234-248.

Boush, D et al.[ University of Minnesota Seminar ] (1987), "Affect Generalization to Similar and Dissimilar Brand Extensions", Psychology and Marketing, 4, 225-237.

Brucks, M (1985), "The Effects of Product Class Knowledge on Information Search Behavior", Journal of Consumer Research, 12, 1-16.

Chaiken, S (1980), "Heuristic versus Systematic Information Processing and the Use of Source versus Message Cues in Persuasion", Journal of Personality and Social Psychology, 39, 752-766.

Feldman, J. M and J. G. Lynch, Jr. (1988), "Self-Generated Validity and Other Effects of Measurement on Belief, Attitude, Intention and Behavior", Journal of Applied Psychology, 73, 421 -435.

Johnson, E. J and E. Russo (1984), "Product Familiarity and Learning New Information", Journal of Consumer Research, 11, 253-263.

Leuthesser, L (1988), "Defining, Measuring, and Managing Brand Equity", MSI Report No. 88104.

MacInnis, D and K. Nakamoto (1989), "Cognitive Associations and Product Category Comparisons": The Role of Knowledge Structures and Context", Working paper, University of Arizona.

Petty, R. E and J. T. Cacioppo (1981), "Attitudes and Persuasion: Classic and Contemporary Approaches", Dubuque, IA, William C. Brown.

Rao, A and K. B. Monroe (1988), 'The Moderating Effect of Prior Knowledge in Cue Utilization in Product Evaluation", Journal of Consumer Research, 15, 253-263.

Smith, D (1990), "An Examination of Product and Market Conditions That Affect the Financial Outcome of Brand Extensions", Working paper, University of Wisconsin.

Sujan, M (1985), "Consumer Knowledge Effects on Evaluation Strategies Mediating Consumer Judgments", Journal of Consumer Research, 12, 31-46.

Tauber, E. M (1981), "Brand Franchise Extension: New Product Benefits from Existing Brand Names", Business Horizons, 24 (2), 36-41.

Tauber, E. M. (1988), "Brand Leverage: Strategy For Growth in a Cost-Control World", Journal of Advertising Research, 28, (August/September), 26-30.

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