Consumer Evaluations of Line Extensions: a Conjoint Approach

Moonkyu Lee, Yonsei University
Jonathan Lee, University of Pittsburgh
Wagner A. Kamakura, University of Pittsburgh
ABSTRACT - Line extensions have been a basis for strategic growth for many firms during the past decade. The viability of line extensions largely depends on how consumers perceive the new features of the extensions. This study examined how consumers differ in their evaluations of original products and line extensions in terms of behavioral characteristics, using an individual-level conjoint analysis. Of particular interest was to compare and contrast the effects of consumer characteristics on evaluations of line extensions with those of brand extensions. A within-subjects conjoint experiment was conducted which involved rankings of full profiles of original products and line extensions. The results showed that subjects' evaluations of the new features of line extensions were influenced by three behavioral factors: perceived brand strength, perceived typicality, and product usage. Theoretical and managerial implications of the results are discussed.
[ to cite ]:
Moonkyu Lee, Jonathan Lee, and Wagner A. Kamakura (1996) ,"Consumer Evaluations of Line Extensions: a Conjoint Approach", in NA - Advances in Consumer Research Volume 23, eds. Kim P. Corfman and John G. Lynch Jr., Provo, UT : Association for Consumer Research, Pages: 289-295.

Advances in Consumer Research Volume 23, 1996      Pages 289-295

CONSUMER EVALUATIONS OF LINE EXTENSIONS: A CONJOINT APPROACH

Moonkyu Lee, Yonsei University

Jonathan Lee, University of Pittsburgh

Wagner A. Kamakura, University of Pittsburgh

ABSTRACT -

Line extensions have been a basis for strategic growth for many firms during the past decade. The viability of line extensions largely depends on how consumers perceive the new features of the extensions. This study examined how consumers differ in their evaluations of original products and line extensions in terms of behavioral characteristics, using an individual-level conjoint analysis. Of particular interest was to compare and contrast the effects of consumer characteristics on evaluations of line extensions with those of brand extensions. A within-subjects conjoint experiment was conducted which involved rankings of full profiles of original products and line extensions. The results showed that subjects' evaluations of the new features of line extensions were influenced by three behavioral factors: perceived brand strength, perceived typicality, and product usage. Theoretical and managerial implications of the results are discussed.

INTRODUCTION

Launching a new product is a risky endeavor because of the high costs of introduction and the low probability of success. One increasingly popular strategy for lowering the costs and improving the odds is to extend a well-known brand name to a new product. The use of existing brand names involves at least two strategic options: line and brand extensions (Aaker and Keller 1990; Farquhar 1989; Tauber 1981). Line extensions occur when the original brand name is extended by modifying features (such as flavors, sizes, or varieties) within the existing product category (e.g., Diet Pepsi or Liquid Tide), whereas brand extensions take place when the brand name is used to enter a completely different product category (e.g., Clorox laundry detergent or Zenith computers). Thus, these two strategies, although conceptually related, are quite different from each other and are expected to work in different ways.

Line extensions, the focus of the present study, are prevalent in the marketplace. According to Gorman's New Product News, 6,125 new products were accepted by groceries in the first five months of 1991. Of these, 89% were line extensions, 6% were brand extensions, and only 5% bore new brand names (Dornblaser 1992). From a company's point of view, it may be less risky and costly to use an existing brand name in introducing a new product with relatively minor changes. Although intuitively appealing and widely used, this line extension strategy does not necessarily guarantee success. Another study conducted by the Association of National Advertisers reports that 27% of line extensions fail (ANA 1984). As an example, in 1993, PepsiCo. introduced Crystal Pepsi, a clear cola with the impression of purity, to tap into a New Age mentality among young consumers. But the sales figures of this product have not been encouraging thus far (Business Week 1994). Industry analysts speculate that the company failed to impress consumers with any real benefits of the product other than the pure and natural image (Marketing News 1994). Then what determines the success or failure of a line extension? The Crystal Pepsi example provides insights into this question. It suggests that critical to the success of a line extension is whether or not consumers value the newly added features of the extension. If so, a question remains as to what makes consumers appreciate or reject those features.

This study investigates how consumers evaluate original brands and their line extensions using an individual-level conjoint analysis. Two brands and their extensions are examined. The objectives are: (a) to trace the changes in attribute importance from original products to line extensions, (b) to determine the importance weights of the new features of line extensions, and (c) to assess the relative influence of several consumer behavioral characteristics on the importance weights of those new features. These individual characteristics include (a) consumers' perception of similarity between the original and extended brand, (b) brand attitude, (c) brand knowledge, (d) product experience, and (e) product usage.

THEORETICAL BACKGROUND

Consumer Evaluations of Brand and Line Extensions

Recent research attention has predominantly focused on consumer responses to brand extensions. It has been shown that consumers engage in a categorization process when evaluating a brand extension, and that they use their affect associated with the original brand when they find a good "fit" or similarity between the parent brand and the extension (Aaker and Keller 1990; Boush and Loken 1991; Roux and Lorange 1993). However, as mentioned earlier, the distinction between line and brand extensions is a matter of type, not a matter of degree. A line extension typically involves modification of an attribute(s) of a current product. Therefore, it creates at most a slight or moderate level of inconsistency from its parent brand. From the standpoint of the categorization framework, such a level of inconsistency is likely to be filtered or ignored (Neisser 1976), and thus, the line extension is perceived to be "typical" of the original brand and triggers affect transfer from the original to the extended brand. If so, according to the categorization model, all line extensions should be successful. In reality, however, this is not the case as there are numerous examples of failure in the market. Thus, consumer evaluations of line extensions should be considered from a different perspective.

The Schema Pointer Plus Tag Model

The schema pointer plus tag (SP+T) model provides a useful framework for understanding how consumers evaluate line extensions. The SP+T model makes specific predictions about how information is encoded, comprehended, and recalled (Graesser 1981; Graesser, Gordon, and Sawyer 1979; Woll and Graesser 1982). It suggests that information that is congruent with schematic expectations is stored in memory along with a "pointer" to the generic schema that best matches the information. On the other hand, schema-incongruent information is encoded in a separate memory location and marked with a unique "tag." The model predicts that schema-incongruent information is particularly well recognized and better recalled in comparison with congruent information, since it occupies a unique space in memory.

Line extensions generally share most attributes in common with parent brands, with the exception of one or more features. Thus, from the standpoint of the SP+T model, these common attributes are encoded within the original brand schema with new, modified features tagged with them. It is expected that consumers pay more attention to and remember the new features better, and thus, when they are asked to make their evaluations of a line extension, they rely more heavily on these features than on their affect associated with the original brand. These features become a focal point of the evaluations of the line extension. Therefore, the success of a line extension hinges on whether these new features are perceived as important and appealing.

TABLE 1

DESCRIPTION OF CONJOINT DESIGN

The Effects of Consumer Characteristics on Evaluations of Line Extensions

Little research to date has dealt with consumer characteristics as potential determinants of line extension success. Research on brand extensions offers some insights into how individual characteristics affect the way consumers react to extended brands in general.

First, as mentioned earlier, many studies found that the "fit" or similarity between the parent and extended product has a positive impact on consumer evaluations of extensions (Aaker and Keller 1990; Boush and Loken 1991; Park, Milberg, and Lawson 1991). Second, Smith and Park (1992) empirically demonstrated that the strength of the parent brand is related positively to the market share of the brand extension. Brand strength or the equity built up in the name of an existing brand is operationalized by consumer attitude toward the brand (Aaker and Keller 1990; Marketing Science Institute 1988) or brand familiarity (Keller 1993). These two factors are also examined as potential determinants of line extension success. Finally, it has been indicated that consumer knowledge about the original and new product categories positively influences consumer evaluations because experts, compared to novices, tend to use more elaborate inferences to find a fit between the two categories (Muthukrishnan and Weitz 1991).

Two knowledge constructs have been conceptualized: subjective and objective knowledge (Brucks 1985; Park and Lessig 1981; Park, Mothersbaugh, and Feick 1994). In the present research, in addition to consumers' self-assessed knowledge, product usage is included as an objective measure of knowledge.

METHOD

An individual-level conjoint analysis was used in this study to estimate partworths of each attribute level across respondents using rank-ordered data obtained from multiple conjoint tasks. The experiment examined two product categories, laundry detergents and soft drinks, where line extensions are commonly found. A brand was selected for each category: Tide for laundry detergents and Coke for soft drinks. Then, line extensions were chosen or created for each brand; the extension for Tide was Tide with Bleach, an existing extension, while the extension for Coke was Coke Clear, a hypothetical case. Finally, based on Consumer Reports (1992), several relevant attributes from each product category were selected for the conjoint task.

An orthogonal fractional-factorial design was used in the study. For each product category, the experimental task involved the ranking of two sets of full profiles, that is, 12 profiles for each original product and line extension. Therefore, each subject completed four conjoint tasks, which resulted in two sets of within-subjects rankings. In addition to the conjoint data, subjects provided information about their individual characteristics which were mentioned earlier. The profiles of the line extensions included marginal or additional attributes which were not present in those of the original products. Subjects for this study were 188 undergraduate and graduate students at a major state university. They were offered an incentive of four dollars to complete the questionnaire. Table 1 shows the design of the conjoint experiment for the two product categories described above.

TABLE 2

PARTWORTH ESTIMATES OF COKE AND COKE CLEAR

DATA ANALYSIS AND RESULTS

Based on the rank-ordered data obtained from the conjoint experiment, the following procedure was used to test the effects of behavioral factors on the changes in attribute importance. First, individual-level conjoint models were fit separately to the rankings of the full profiles of the original products and line extensions. With rankings transformed into ratings, which then were rescaled as deviations from the mean, a model was estimated by ordinary least squares (OLS). The model yielded unbiased estimates of individual-level parameters. Individual-level estimation allows arbitrary heterogeneity in the coefficients across respondents (Elrod, Louviere, and Davey 1992). Therefore, for each subject i within each choice set, the individual-level regression model of preference ratings, Rij, of the alternatives j was given as:

(1) EQUATION

For both product categories, the effect of price was assumed to be linear and modeled based on a single variable, while the effects of other variables were assumed to be non-linear and modeled based on two effect-type dummy variables.

Second, the importance of each attribute, Iik, was computed using partworths of different attribute levels. Then, the changes in attribute importance from original products to line extensions, DIik, were obtained from the differences between them for each individual. That is, for each attribute in both the original products and line extensions, the importance of attributes for subject i was obtained by:

(2) EQUATION

And the changes in attribute importance between the original products and line extensions were given as:

(3) EQUATION

The changes in attribute importance obtained from equation (3) served as dependent variables in testing the effects of behavioral factors. Table 2 shows the partworth estimates and importance of each attribute obtained by averaging estimates over all subjects for Coke and Coke Clear respectively; Table 3 reports those for Tide and Tide with Bleach.

TABLE 3

PARTWORTH ESTIMATES OF TIDE AND TIDE WITH BLEACH

Finally, five individual difference variables were measured: (a) perceived similarity between the parent and extended brands, (b) brand attitude, (c) brand knowledge, (d) product experience, and (e) product usage. Except for product usage, all variables were measured with multi-item scales. Measurement items used in this study were excerpted from the previous research on brand extensions mentioned earlier.

Several principal components were identified based on the behavioral variables. The results of the factor analysis in Table 4 and Table 5 show that four rotated factors provided a good summary of the data, accounting for 96 percent of the standardized variance. The first component in both product categories was a measure of brand familiarity since it showed approximately equal loadings on brand knowledge and product experience. Product usage and perceived typicality were based on single items, while perceived brand strength was a combination of brand attitude, knowledge, and product experience.

Then, the regressions of DIik were run on the rotated factor scores of those principal components to examine the influence of the behavioral factors on any changes in preferences between the original products and line extensions. For both product categories, four common factors were found: (a) brand familiarity, (b) perceived brand strength, (c) perceived typicality and (d) product usage. Table 4 present the results of the factor analyses for the soft drink and for the laundry detergent categories, respectively. Based on the factor scores of each individual, SiF, the regression model was given as:

(4) EQUATION

Pooled regressions were run across subjects with changes in attribute importance as dependent variables and four behavioral factors as independent variables.

Table 5 reports the results of the regression based on equation (4) for Coke and Coke Clear. Analysis of Variance (ANOVA) results show that the importance changes in fizziness and color were significant. Furthermore, in both regressions, perceived brand strength and perceived typicality were significant in evaluating those attributes. However, it should be noted that the direction of the main effects was negative (cf. Aaker and Keller 1990; Boush and Loken 1991; Smith and Park 1992).

TABLE 4

CONSUMER BEHAVIORAL VARIABLES AND FACTOR PATTERN

TABLE 5

REGRESSION RESULTS (COKE) INFLUENCE OF CONSUMER BEHAVIORAL FACTORS UPON CHANGES IN ATTRIBUTE IMPORTANCE

TABLE 6

REGRESSION RESULTS (TIDE) INFLUENCE OF CONSUMER BEHAVIORAL FACTORS UPON CHANGES IN ATTRIBUTE IMPORTANCE

Table 6 shows the regression results for Tide and Tide with Bleach. The effect of perceived typicality was significant for changes in price importance. For the other attributes, there were no significant effects of the behavioral factors which were common to the original product and the line extension. Unlike the soft drink category, neither perceived brand strength nor perceived typicality was found to be a significant behavioral factor in evaluating the marginal attribute, extra bleach. Only product usage was significant. ANOVA results show that the significant effects of the behavioral factors were found only for changes in price importance between Tide and Tide with Bleach.

DISCUSSION

The results of this study generally support the intriguing notion that the additional features of line extensions are not always valued by consumers, that is, they are valued or ignored depending on several behavioral factors. For Coke Clear, it was found that two factors had a negative influence on the valuation of the "clear" color: perceived brand strength and perceived typicality. This is somewhat counter-intuitive and appears to be inconsistent with the results of the previous studies on brand extensions. For instance, Smith and Park (1992) found that brand strength is positively related to the market performance of brand extensions, and Boush and Loken (1991) revealed that perceived fit between parent brands and brand extensions enhances consumer evaluations of extensions. Although these studies are not directly comparable to the present research due to differences in methodology and dependent measures, the results of this study imply that the consumer evaluation process for line extensions might be quite different from that for brand extensions. For Tide with Bleach, product usage had a positive impact on the valuation of the added attribute, "extra bleach." The results from the two extension conditions, considered in combination, suggest that there are brand differences in consumer valuations of new features of line extensions. Thus, marketing managers should first understand how their brands are perceived by consumers when planning on a line extension rather than simply relying on the strength of the reputation of their brands.

Although the additional features of the line extensions were not selected on the basis of any theoretical taxonomy, some interpretations are possible by understanding the basic characteristics of those features. For many consumers, the clearness of cola, although not very meaningful or relevant, has a unique, pure, and natural image. The results of this study, however, indicate that such an attribute is not valued even by consumers who have a favorable impression about the brand name, Coke, and those who perceive the extension, Coke Clear, to be consistent with the image of original Coke. It is speculated that since Coca Cola has been in the market for a long time, consumers seem to have built up a strong emotional attachment or inertia about the brand, that is, they do not like to see any change in it. This is why Coca Cola company had to abandon its New Coke line and introduce Coca Cola Classic years ago. Consumers liked the taste of New Coke but did not consider it important (McCarthy and Pereault 1993). Carpenter, Glazer, and Nakamoto (1994) found in their experiments that consumers often value distinguishable, unique, but irrelevant attributes. However, specific brand names were not given in their studies. Their results might have been different if subjects were given brand names.

The bleach content of laundry detergent is a utilitarian or functional feature. It is also valuable and relevant to creating an actual benefit for consumers who use the product. Therefore, it is quite natural that this attribute is more valued by those who use more laundry detergent. The name Tide seems to be associated with function and performance, rather than any symbolic values. The results of this study are consistent with Park et al. (1991) and Broniarczyk and Alba (1994) in the sense that these studies underscore the importance of brand-specific images and associations for the successful introduction of brand extensions. Marketing practitioners planning on line extensions should make sure they add to or modify the feature(s) that match the image of their current brand. More importantly, they should keep in mind that the strength of a brand name does not always guarantee the success of a line extension since the evaluation of the extended product is determined primarily by consumer perceptions of the new feature(s).

From a modeling point of view, it should be pointed out that the estimation method used in this research does not allow the test of statistical significance of the changes in attribute importance. Thus, an immediate extension of the current study would be to develop a latent class model of line extension evaluations that can infer schema characteristics from the significant changes in attribute importance using multiple choice sets and associate the schema pointer plus tag model with consumer behavioral factors.

REFERENCES

Aaker, David A. and Kevin L. Keller (1990), "Consumer Evaluations of Brand Extensions," Journal of Marketing, 54, 27-41.

Association of National Advertisers, Inc. (1984), Prescription for New Product Success, New York.

Broniarczyk, Susan M. and Joseph W. Alba (1994), "The Importance of the Brand in Brand Extension," Journal of Marketing Research, 31 (May), 214-228.

Boush, David M. and Barbara Loken (1991), "A Process-Tracing Study of Brand Extension Evaluation," Journal of Marketing Research, 28, 16-28.

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

Business Week (1994), "Does Pepsi Have Too Many Products?" February 14, p.64.

Carpenter, Gregory S., Rashi Glazer, and Kent Nakamoto (1994), "Meaningful Brands from Meaningless Differentiation: The Dependence on Irrelevant Attributes," Journal of Marketing Research, 31 (August), 339-350.

Consumer Reports (1992), Buying Guide Issue, 56, 12, New York: Consumers Union of United Sates, Inc.

Dornblaser, L. (1992), "New Product Totals for 1991 Soar Past 16,000 Level," New Product News, Gorman Publishing Co.

Elrod, Terry, Jordan J. Louviere, and Krishnakumar S. Davey (1992), "An Empirical Comparison of Ratings-Based and Choice-Based Conjoint Models," Journal of Marketing Research, 29 (August), 368-377

Farquhar, Peter H. (1989), "Managing Brand Equity," Marketing Research, 1, 24-35.

Graesser, Arthur C. (1981), Prose Comprehension Beyond the Word, New York: Springer-Verlag.

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, 319-332.

Green, Paul E. and Abba M. Krieger (1987), "A Consumer-Based Approach to Designing Product Line Extensions," Journal of Product Innovation Management, 4, 21-32.

Keller, Kevin L. (1993), "Conceptualizing, Measuring, and Managing Customer-Based Brand Equity," Journal of Marketing, 57 (January), 1-22.

McCarthy,E. Jerome and William D. Perreault (1993), Basic Marketing, 11th ed., Homewood, IL: Irwin.

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

Reddy, Srinivas K., Susan L. Holak, and Subodh Bhat (1994), "To Extend or Not to Extend: Success Determinants of Line Extensions," Journal of Marketing Research, 31 (May), 243-262.

Roux, Elyette and Frederic Lorange (1993), "Brand Extension Research: A Review," in European Advances in Consumer Research, ed. W. Fred Van Raaij and Gary J. Bamossy, Provo, UT: Association for Consumer Research, Vol. 1, 492-500.

Marketing News (1994), "Consumers Show Little Taste for Clear Beverages," 28, 11 (May 23), p.1.

Marketing Science Institute (1988), Research Topics 1988-1990, Cambridge, MA.

Neisser, U. (1976), Cognition and Reality: Principles and Implications of Cognitive Psychology, San Francisco: W. H. Freeman.

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

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, 185-193.

Park, C. Whan, David L. Motherbaugh, and Lawrence Feick (1994), "Consumer Knowledge Assessment," Journal of Consumer Research, 21, 1, 83-99.

Smith, Daniel C. and C. Whan Park (1992), "The Effects of Brand Extensions on Market Share and Advertising Efficiency," Journal of Marketing Research, 29 (August), 296-313.

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

Woll, Stanley B. and Arthur C. Graesser (1982), "Memory Discrimination for Information Typical or Atypical of Person Schemata," Social Cognition, 1, 4, 287-310.

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