An Examination of the Effects of Multiple Brand Extensions on the Brand Concept

ABSTRACT - This study examines the effects of multiple extensions on two types of brand equity: equity of the original parent brand or anchor product, and a global "family brand" image. Unlike most previous work stressing equity as brand affect, equity is construed in terms of its impact on brand concept accessibility, evaluations and accessibility of brand specific beliefs. The results indicate that consistent extensions lead to higher brand concept accuracy and accessibility. Independent extensions are associated with higher accuracy and do not appear to detrimentally affect family brand equity, but can lead to decreased accessibility of brand beliefs.


Sandy D. Jap (1993) ,"An Examination of the Effects of Multiple Brand Extensions on the Brand Concept", in NA - Advances in Consumer Research Volume 20, eds. Leigh McAlister and Michael L. Rothschild, Provo, UT : Association for Consumer Research, Pages: 607-611.

Advances in Consumer Research Volume 20, 1993      Pages 607-611


Sandy D. Jap, University of Florida

[This is a condensed version of the originally submitted manuscript. A longer, more readable version is available from the author; her address is: University of Florida, Dept. of Marketing, 212 Bryan Hall, Gainesville, FL 32611.]

[This manuscript has benefitted greatly from the insightful advice and comments of Joe Alba, Rich Lutz, and especially John Lynch.]


This study examines the effects of multiple extensions on two types of brand equity: equity of the original parent brand or anchor product, and a global "family brand" image. Unlike most previous work stressing equity as brand affect, equity is construed in terms of its impact on brand concept accessibility, evaluations and accessibility of brand specific beliefs. The results indicate that consistent extensions lead to higher brand concept accuracy and accessibility. Independent extensions are associated with higher accuracy and do not appear to detrimentally affect family brand equity, but can lead to decreased accessibility of brand beliefs.


Brand extensions have been defined as the use of a brand to introduce products in different categories outside of the parent brand category as a means of achieving higher sales growth rates, higher ROI, and advertising and promotion efficiencies (Baldinger 1990). More recent research has examined the effects of multiple brand extensions on perceptions and evaluations of the proposed extension as well as the original core brand (Keller & Aaker 1992). The present research takes an additional step toward better understanding the effects of multiple extensions with respect to the original core brand and its anchor product.

The purpose of this study is to examine how multiple extensions might affect the brand at the level of the original parent product and at a more global, family level. More specifically, it is hypothesized that extensions can enhance and reinforce the parent brand concept, when they are made in a manner consistent with brand-specific associations. However, when extensions are made in a manner independent of the brand specific associations, the brand's global, family concept may be detrimentally affected.


The brand extension literature has found that consumers tend to search for a common "link" or degree of similarity that explains the extension in relation to the brand (See Keller 1991 for a review). Although similarity is often defined as similarity of product categories, the fit or consistency of a brand extension to the parent brand in the present research is defined as a function of the brand concept resulting from brand-specific features or associations (Broniarczyck 1992).

The brand concept consists of the primary brand-specific association, either at the attribute or product image level, including abstract meanings (e.g. "durability," "quality," "status") or concrete attributes (e.g. "pure ingredients," "dandruff control"). Although brands often have multiple associations, this study assumes that a dominant association can be identified that is able to explain or link the original product and all its extensions as a group. Brand concepts serve to position the product within its category, often referring to the distinguishing feature that differentiates the brand from its competitors. According to this conceptualization, similarity is a search for a meaningful commonality among what might otherwise be a dissimilar product grouping under a common brand name.


Just as "Smith" refers to a family name, and "Tom Smith" refers to a specific member, brand concepts can exist at two levels: the family level and the parent product level. When brand names such as Kelloggs, and Kodak are associated with multiple products, the family brand concept (e.g. "breakfast foods," "photography") encompasses the entire product line. The parent brand concept is almost always the same as the family brand concept, since the parent product is the brand's anchor product. However, this does not preclude the possibility that the brand concept may be one that is independent of the family brand concept. [In other instances, firms may introduce independent brand concepts by creating independent brand names in conjunction with the family name (e.g. Black & Decker Dust Buster, Nabisco Cheez-its). The use of "double" brand names is an interesting direction for future research in brand equity, but is beyond the scope of the present study.] Brand equity should be strengthened at the family or parent level when a concept is strongly associated to a particular brand, and diluted when the concept is weakly associated with the brand. [Brand dilution is defined as a decrease in the strength of beliefs about the parent brand (Roedder John & Loke 1990).]

Consistent Extensions

Brand concepts can act as explanations or theories, that simplify reality, making a category of knowledge increasingly informative, useful and efficient (cf. Murphy & Medin 1985). When multiple extensions are made on a basis consistent with the family brand concept, each extension emphasizes the same brand concept as the original product, making the family brand concept increasingly coherent. Consequently, when consumers are exposed to the family brand name or any of its products, the family brand concept should be quickly and accurately recalled.

Consistent extensions increase the opportunity for cognitive elaboration (relative to a control group) as exposure to the family's products strengthens the family concept and causes consumers to make the association between the original brand concept and the new product. As the family and product level brand concepts are repeatedly strengthened, beliefs concerning its nature should be enhanced. This reduction in uncertainty can lead to positive affect (Nuttin & Greenwald 1968; Obermiller 1985). Changes in evaluation are unlikely to be dramatic, such as positive evaluations becoming negative; however, preexisting positive evaluations may become more positive.

Given the prior discussion, if the original family brand concept is maintained over an increasing number of product categories, the brand concept becomes increasingly accessible because extensions are made in accordance with the consumer's prior knowledge/expectations of the brand and are continually fortifying the brand's original family concept. Constant repetition of the brand concept through multiple extensions should also strengthen existing beliefs about the nature of the brand concept and increase liking for the brand via reduced uncertainty. Additionally, there is no reason to believe that this might differentially affect the parent brand at the family or parent brand level.

H1: Consistent extensions lead to increased family and product brand concept accessibility, increased belief strength, and more positive evaluations of the original product and the family brand name.

This hypothesis further assumes that the brand concept is positively valued and ceiling effects on brand concept accessibility, belief strength, and evaluations are nonexistent.

Independent Extensions

Brand dilution research is sometimes examined as the result of inconsistent extensions, where an inconsistent extension basis represents the opposite end of the brand concept continuum (i.e., gentle-harsh, high quality-low quality). This is a convenient assumption, and one that lends itself well to hypotheses similar to those found in the stereotype and schema revision literature. In reality, few firms would introduce brand extensions that blatantly contradict the original brand concept. Firms that want to broaden their customer base to include new segments are likely to use an independent brand name with an association that does not conflict with their current brand concept. Therefore, this study is based on the premise that firms are more likely to introduce brand extensions on bases independent of the family brand concept, as opposed to extension bases directly opposed to the parent brand concept. An independent brand extension does not necessarily imply that the extension basis is irrelevant in the product class; rather, it means that the extension association is made on a dimension that has no necessary implication for the parent brand association, or an association that is subjectively uncorrelated with the parent brand concept. If Neutrogena's family brand concept is "pure ingredients," independent extensions may emphasize other associations such as "deep cleaning," "invigorating scent," or "pricey."

Family Level. Independent extensions are more likely to result in equity dilution at the family level than at the parent brand level because as the association set size linked to a brand name increases, the strength of a "link" or "retrieval path" is weakened (Anderson 1976; Raaijmakers & Shiffrin 1981), and competing associations inhibit and interfere with retrieval (Anderson 1976; Meyers-Levy 1989; Sujan & Bettman 1989). Over time, the brand may lose its definitive meaning or association. Additionally, the presence of new independent associations may lead to decreases in the strength or accessibility of the consumer's beliefs about the family brand concept.

Decreases in brand concept accessibility and brand beliefs are likely to impact evaluation of the parent brand, as evaluations are subject to change as an individual learns more about a stimulus. If a consumer is unable to integrate the independent associations into a coherent brand concept, uncertainty and aversive tension may arise leading to decreased affect (i.e., Biehal & Chakravarti 1986; Johnson & Levin 1985; Obermiller 1985; Meyer 1981; Simmons & Lynch 1991). Thus, it is expected that brand evaluations will also be lower at the family level than at the parent brand level.

Parent Brand Level. Independent extensions are likely to lead to minimal dilution effects of the original parent brand, because these associations are the most well-learned over time, and because the category context in which the information was originally encoded may trigger encoding specificities (Tulving & Thomson 1973) that enhance recall when the category is cued. Therefore, it is hypothesized that independent extensions differentially dilute brand equity at the family and original product level. This occurs because independent associations in various product categories make it increasingly difficult to cognitively process all the information into a coherent family brand concept. Encoding specificities that occur at the time of learning, in conjunction with the presence of a category cue, may account for the apparent lack of dilution, and changes in brand beliefs and evaluation that are observed with respect to the original parent product.

H2: Independent extensions have little effect on parent brand concept accessibility, brand beliefs, or brand evaluation at the product level. However, independent extensions reduce brand concept coherence, brand beliefs, and brand evaluation at the family level.

Recent research on brand equity dilution has reported seemingly conflicting findings. Roedder John and Loken (1990) examined conditions that lead to dilution at the brand's global image level; they found that dilution occurred when moderately similar brand extensions were made. In contrast, Keller and Aaker (1992) reported that unsuccessful extensions in dissimilar product categories do not adversely impact the parent brand. This apparent difference might be explained by better understanding the procedures that were used. Roedder John and Loken (1990) cued subjects by the parent brand name and reported their beliefs and associations about the parent brand at the family brand level.

However, in the Keller and Aaker (1992) study, subjects were presented with the original parent brand product and then the dependent parent brand measures were taken. Encoding specificities with regard to the original product could likely have been triggered as a result of exposure to the product, inhibiting the occurrence of dilution. The difference in the results of the two studies might thus be attributed to whether or not a specific product category was cued. It could be that merely asking, "does Brand A mean gentleness?" leads to qualitatively different results than when the same question is posed in the presence of a specific product category cue (i.e. "does Brand A product mean gentleness?"). This study will account for both of these situations by using category cues when dependent measures are taken at the parent brand level, and taking family level dependent measures in the absence of product cues.



Subjects were students enrolled in an introductory marketing class at a major university in the Southeast United States. Forty subjects were shown sixteen health and beauty aid advertisements. The group was given five minutes to study an ad book and product display, and were told that although the ads were relatively standardized, there was enough information to form some opinions about each product. The ad for the focal parent brand, Coast soap, was always the third ad in the booklet. Following this ad were ads for extensions of Coast soap. Ten subjects saw shampoo, anti-perspirant, and sunblock extensions that made a consistent "clean scent" claim. Ten subjects saw extension ads that made independent claims ("body building" shampoo, "dry" anti-perspirant, and "maximum sun protection" sunblock). Twenty subjects saw no extension advertisements. Subjects were allocated in such a manner so as to provide a stable baseline estimate for the control condition, since many of the focal contrasts would be made with respect to this group. All other filler brands were randomized within the book. Each condition had four randomized versions of the ad books. The product display contained the high and low market share brands and high and low quality brands (arranged by product category) used to anchor the affect ratings administered later in the procedure.

A five-minute filler task was administered, and then subjects were given an additional five minutes to review the ads and product display a second time to insure sufficient opportunity to elaborate upon the product associations and form opinions. This was followed by another five-minute filler task after which subjects were given sheets of paper with brand names. They were asked to take into consideration all that they knew about the brand and then list five associations (e.g. product qualities, images, characteristics and features) that came to mind whenever they saw the brand name.

Subjects were then seated at computers and presented with ten true/false statements in random order. Each subject's response time and accuracy were unobtrusively recorded by the computer. Subjects then responded to sixteen randomly ordered brand belief questions; they were told to take into account all that they knew about the brands. The final portion of the computerized questionnaire asked for evaluative ratings. Subjects were first presented with the brand names displayed on the product table, in order to "anchor" their rating scales equivalently across conditions before making their evaluations of the treatment brands. At this point, subjects completed a 30-minute filler task, the free association measures, and the computerized measures at the other level of measurement (family or parent brand). A post-experimental questionnaire was administered at the conclusion of the experiment.


A 2x3x2 factorial design was used, with measurement level (family level; parent brand level) as a within subject variable, and extension type (consistent, independent, control) as a between subject variable. An order variable of measurement level (parent brand first-family second, family-first-parent brand second) was also introduced, to test for order effects; this variable was not significant, reducing the overall design to a 2x3 factorial.

Pretesting of Stimuli. In order to avoid any possible effects that might result from category expertise, three health and beauty aid product categories were chosen; differences in knowledge were expected to be minimal. Each product category was pretested to insure that product familiarity did not differ significantly by gender and all brands used were moderately familiar. This was done to avoid ceiling effects and dependent measure ratings that might be biased upward.

The extension positionings were pretested to insure that they were similarly perceived across individuals and were relevant to their product categories. In order to reduce demand artifacts and hypothesis guessing, the extensions were not emphasized as extensions of the original product category; they were merely seen as products with the same brand name in the ad book.

Dependent variables. The dependent measures emphasized the brand's primary image positioning in the advertisement and were comprised of four tasks. Brand concept accessibility was measured by its effects on recall order in free association listings and response times for association verification. Accessible brand concepts should be one of the first associations recalled upon exposure to the brand name. This measure did not yield significant findings in any condition and subsequently will not be addressed further. The association verification task at the family level required the subject to answer true/false to ten statements such as, "Coast means clean scent." At the parent brand level, subjects saw a product category name before responding to the true/false statements.

Brand beliefs were measured by questions such as, "To what degree do you believe that Coast actually possesses a clean scent?" At the parent brand level, the question was phrased, "To what degree do you believe that Coast soap actually possesses a clean scent?" Responses were indicated on a 7-point Likert scale (1=I don't believe the brand has it; 7=I do believe the brand has it).

Brand evaluation was measured by presenting the subject with a brand name and asking the subject to indicate the extent to which he/she felt the brand represented an inferior/superior product. Individuals in the parent brand level condition were given the product category before making their responses. Responses were indicated on a 7-point Likert scale (1=inferior product; 7=superior product).


Only two subjects indicated an awareness of Coast as the treatment brand in the post-experimental questionnaire. Analysis of the data without their responses indicated no significant differences, so their responses were retained in the analysis. The questionnaire further indicated that demand effects were not a threat. The results are displayed in Table 1.

Consistent Extensions vs. Controls

Consistent extensions were not expected to differentially affect family and product level equity. As predicted, there were no significant interactions of extension type (consistent vs. control) and level on any of the dependent variables. Response times to the verification statement in the consistent extension condition (m=1528) were significantly lower (F1,31=7.22; p<.05) than response times for the controls (m=2310). Subjects in the consistent condition also displayed a significantly higher accuracy rate (m=1.0) than the control group (m=.78), (F1,31=8.24; p<.01).

Brand beliefs were higher following consistent extensions (m=6.5) then following no extensions (m=5.9), although this was not significant (F1,31=2.51). The accessibility of these beliefs in the consistent condition (m=3958) did not differ significantly from the controls (m=4167), (F1,31=.035). Additionally, subjects' evaluations in the consistent condition (m=6.0) became more positive than the controls (m=5.5) although the difference was not significant (F1,31=1.74).

In sum, there is partial support for hypothesis 1, consistent extensions appear to lead to more accessible brand concepts and higher accuracy. The results for the brand beliefs and evaluation prediction are directionally correct, but did not reach significance. These results might be due to the fact that some of the contrasts were underpowered. The power was .56 (w2=.049) for the brand belief contrast, and .50 (w2=.042) for the evaluation contrast. With larger cell sizes, these differences may approach statistical significance.

Independent Extensions vs. Controls

The predicted interaction of measurement level and extension type on the response times of the verification task was not significant (F1,31=.035), although reaction times in the independent condition at the family level (m=2172) were higher than at the product level (m=1929). Response times in the independent condition did not differ significantly from the controls at the family level (F1,31=.796). As predicted, for parent brand response times, the difference between control subjects and those exposed to independent extensions was not significant (F1,31=.21).

The interaction of measurement level and extension type on accuracy was not significant (F1,31=1.7). However, a main effect of extension type on accuracy was significant (F1,31=4.9, p<.05); accuracy in the independent condition (m=.95) was higher than the control group (m=.78). This was an unexpected finding; it could be that exposure to the treatment brand "reminds" the subjects of the original parent brand concept such that accuracy is maintained, even in the presence of newly learned, independent associations.



The hypothesized interaction of measurement level and extension type on brand beliefs was also not significant (F1,31=0). At the product level, the brand belief difference between the independent (m=6.4) and control group (m=6.0) was not significant, as predicted (F1,31=.629).

The interaction of measurement level and extension type on response times to the brand belief statements was not significant (F1,31=.74). However, there was a significant main effect of extension type on response time (F1,31=7.25; p<.05); brand belief accessibility in the control condition (m=4167) was higher than the independent group (m=5666). Additionally, the independent condition response time was significantly higher (F1,31=9.40; p<.01) than the consistent condition (m=3959).

The interaction of measurement level and extension type on evaluation was not significant (F1,31=1.07). Contrary to expectations, independent extensions increased family brand evaluations (m=5.8) while having no effect on parent brand evaluations (m=5.5); however, this difference was not significant.

Taken together, these data partially support hypothesis two in that no significant differences were found at the parent brand level. The results suggest that learning new independent associations inhibits brand belief accessibility without affecting brand concept accuracy.



There are several modifications that can improve the study. First, larger sample sizes in the consistent and independent conditions should improve statistical power for key contrasts. Second, multiple measures of brand evaluation would strengthen confidence in the reliability of the measures. Finally, the theory might be better tested with dependent measures that are less "noisy" so that effects are easily detected. Response times are characterized by a high error component resulting in the need for large sample sizes in order to bring the power of the study to a satisfactory level.


One purpose of this study was to show that brand extensions can enhance and reinforce the original parent brand concept when extensions are made in a manner consistent with the family brand's primary brand-specific associations. The results seem to lend partial support to hypothesis 1. Consistent extensions lead to significant increases in brand concept accuracy and accessibility. Although consistent extensions did not lead to significant increases in brand beliefs, they did become higher, suggesting that future research may be worthwhile to better understand the process by which this occurs.

There was mixed support found for hypothesis 2; parent brand equity in the independent condition did not differ significantly from the controls. However, independent extensions improved brand concept accuracy at both the family and product level. A plausible explanation for this result might be that exposure to the treatment brand reminds subjects about the original brand concept, enabling them to maintain a high degree of accuracy. Although subjects were exposed to independent associations, family equity appeared resistant to dilution effects; there were virtually no significant decreases in category accessibility, brand beliefs and evaluations at the family level.

The limitations of a laboratory study may account for why dilution effects were not observed. First, in the real world, it is likely that dilution effects occur over time, and so the half-hour filler task delay was not long enough for memory deficits to occur. Second, brands often have several strong associations as opposed to a single, dominant brand concept. It could be that new extensions that emphasize only one of a family brand's multiple associations may result in dilution of the non-emphasized associations. Finally, since independent associations did not introduce information that disconfirmed or contradicted the original brand concept, there may have been no reason to modify original beliefs that "Coast possesses a clean scent."

Independent associations had a detrimental effect on brand belief accessibility, holding serious implications for brand equity. If the product and brand concept association is not accessible, then the product is less likely to enter the consumer's evoked set in a memory based choice situation. Even if the brand were to enter the consumer's evoked set, an inaccessible brand concept might result in the brand being evaluated on its competitors' strengths instead of its own strengths.

Nevertheless, the decomposition of brand equity at the family and product level is conceptually useful, holding strategic implications for brand management. Certain types of extensions may be instrumental in strengthening the global brand meaning in the consumer's mind. Alternatively, management might deliberately use extensions to weaken a brand's family image, perhaps with the goal of providing a broader platform for launching future new products.

Future Research

Future avenues of research that have been highlighted by this study include further examination of the consequences of accessibility of the brand concept and beliefs. Inaccessible beliefs can prevent a brand from being evoked and considered, as well as prevent the brand from being evaluated on its dimensions of strength.

Additional research is also needed that examines the role of uncertainty in the development of beliefs and evaluations toward the parent brand or any of its products. Uncertainty was not explicitly manipulated or measured in this study, but a better understanding of its effects would contribute to our knowledge of brand equity.

There is still much to be learned about how brand extensions might be used to effectively implement long term corporate strategy goals. This study attempted to take a step in this direction by better understanding how consistent and independent extension types might affect consumer perceptions of the corporate family brand and parent brand.


Anderson, John R. (1976), Language, Memory, and Thought, Hillsdale, NJ: Lawrence Erlbaum Associates.

Baldinger, Allan L. (1990), "Defining and Applying the Brand Equity Concept: Why the Researcher Should Care," Journal of Advertising Research, 30, RC2-RC5.

Biehal, Gabriel, Dipankar Chakravarti (1986), "Consumers' Use of Memory and External Information in Choice: Macro and Micro Perspectives," Journal of Consumer Research, 12, (March), 382-405.

Broniarczyck, Susan M. (1992), unpublished dissertation, Department of Marketing, University of Florida.

Johnson, Richard D., Irwin P. Levin (1985), "More than Meets the Eye: The Effect of Missing Information on Purchase Evaluations", Journal of Consumer Research, 12, (September), 169-177.

Keller, Kevin Lane (1991), "Conceptualizing, Measuring, and Managing Customer-Based Brand Equity," Working paper, Stanford University.

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

Loken, Barbara, James Ward (1990), "Alternative Approaches to Understanding the Determinants of Typicality," Journal of Consumer Research, 17, (September), 111-26.

Meyer, Robert J. (1981), "A Model of Multiattribute Judgments Under Attribute Uncertainty and Informational Constraint," Journal of Marketing Research, 18, (November), 428-441.

Meyers-Levy, Joan (1989), "The Influence of a Brand Name's Association Set Size and Word Frequency on Brand Memory," Journal of Consumer Research, 16, (September), 197-207.

Murphy, Gregory L., Douglas L. Medin (1985), "The Role of Theories in Conceptual Coherence," Psychological Review, 92, (3), July, 289-316.

Nuttin, Joseph and Anthony G. Greenwald (1968), Reward and Punishment in Human Learning, New York: Academic Press.

Obermiller, Carl (1985), "Varieties of Mere Exposure: The Effects of Processing Style and Repetition on Affective Response," Journal of Consumer Research, 12, (June), 17-29.

Raaijmakers, Jeroen G. W., Richard M. Shiffrin (1981), "Search of Associative Memory," Psychological Review, 88, (March), 93-134.

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

Simmons, Carolyn J., John G. Lynch, Jr. (1991), "Inference Effects without Inference Making? Effects of Missing Information on Discounting and Use of Presented Information," Journal of Consumer Research, 17, (March), 477-491.

Sujan, Mita, James R. Bettman (1989), "The Effects of Brand Positioning Strategies on Consumer's Brand and Category: Some Insights from Schema Research," Journal of Marketing Research, 26, (November), 454-467.

Tulving, Endel and Donald M. Thomson (1973), "Encoding Specificity and Retrieval Processes in Episodic Memory," Psychological Review, 80, (September), 352-373.



Sandy D. Jap, University of Florida


NA - Advances in Consumer Research Volume 20 | 1993

Share Proceeding

Featured papers

See More


N10. How Does It Make You Feel? Emotional Reasoning and Consumer Decisions

Andrea Rochelle Bennett, University of North Texas
Blair Kidwell, University of North Texas
Jonathan Hasford, University of Central Florida, USA
David Hardesty, University of Kentucky, USA
Molly Burchett, University of Kentucky, USA

Read More


O9. The Role of Numerical Identification in Customer Reaction toward Service Failure

JIEXIAN (Chloe) HUANG, Hong Kong Polytechic University
Yuwei Jiang, Hong Kong Polytechic University

Read More


Always Trust in Your Friends? Cross-cultural Effects of Review Source and Incentives on Trustworthiness

Dionysius Ang, Leeds University Business School

Read More

Engage with Us

Becoming an Association for Consumer Research member is simple. Membership in ACR is relatively inexpensive, but brings significant benefits to its members.