Introduction
The topic of brand extensions has been an important focus for current research in the marketing field, with several influential researchers making their mark on the extant literature: Aaker (1990), Aaker and Keller (1990), Voelckner and Sattler (2006), Ahluwalia (2008), Batra, Lenk and Wedel (2009), Boush and Locken (1991), Bottomley and Doyle (1996), Buil, de Chernatony and Hem (2009). Defined as “the use of established brand names to launch new products” by Voelckner and Sattler (2006, p.18), brand extensions are present in almost every market and product category, as the ability to stretch a brand by introducing an assortment of new products or services can have tremendous financial benefits in today’s marketplace through trade promotions, lower entry prices, lower advertising spendings or better retailer acceptance (Collins-Dodd and Louviere 1999, Tauber 1988, Simon and Sullivan 1993).
Keller, Apéria and Georgson (2008) describe two main strategies for brand extensions, namely line extensions and category extensions. Whereas line extensions refer to extensions that target a new market segment with an existing product class, category extensions are more complicated as they require the brand to enter a new product class, thus giving rise to various problems, such as the suitability of the new class, the transferability of the brand’s attributes to the new product based on the new category characteristics, or the potential lack of fit between the two categories. Line extensions are seen as harmless and have a similar customer base, whereas category extensions due to their higher stretchability are considered more dangerous.
The popularity of brand extensions as a new product introduction strategy has increased, with over 80% of the new product launches in 2004 despite their high failure rates according to Batra, Lenk & Wedel (2009, pg. 3). Usually considered a cheap and easy new product introduction strategy, brand extensions are seen in a highly positive light by managers across the world.
The general argumentation behind using brand extension strategies lies in the existing brand associations that consumers hold in their memories, which can be used in the evaluation of the extension Morrin (1999). We should also not ignore the fact that due to increased international trade, and the plentitude of new brands, products and services offered, the minds of the consumers are constantly subjected to new information. Thus using a familiar brand can diminish the amount of information search that consumers use to make purchasing decisions (Brucks,1995) and decreases the level of perceived risk (Roselius, 1971).
Although many studies have shown the benefits of using brand extensions (Smith & Park, 1992; Tauber, 1991), and a significant amount of companies use the strategy to introduce a growing number of sub-brands, their strategies are not always successful, with countless extensions failing to meet expectations. However, at the other extreme, there seem to be brands that can extend into almost any category and still be successful: Virgin has managed to extend into categories such as travel and tourism, leisure, shopping, media and communications, finances, health and even social and environment; similarly, countless luxury brands, including Gucci or Prada due to their high status and intangible brand promises seem to be able to stretch farther than normal brands. Thus, the question of what influences brand stretchability – the extent to which an existing brand name can be used to introduce products in different categories (Ahluwalia, 2008) – came to be questioned.
From Aaker and Keller’s (1990) study, several researchers have focused in identifying the key success and failure factors of brand extensions, mostly confirming the role of brand fit (similarity) between the initial product and the newly developed one (Boush and Loken 1991, de Ruyter and Wetzels 2000, Voelckner and Sattler 2006). Although a strict conceptualization of “fit” does not exist, what is known that the fit of an extension can vary from very high fit (e.g. line extensions: Pepsi Light), to moderate extensions (e.g. similar category/usage situation: Mars icecream) to low fit (e.g. dissimilar category and difficult-to-uncover similarities: Cosmopolitan yoghurts). The most widely accepted differentiation of “fit” was defined by Aaker and Keller (1990) in terms of complementarity, substitution and transferability.
Techniques to establish the fit between of a parent brand and its extension have also been subjected to scrutiny, focusing mostly on communication methods: advertising exposure (Lane 2000, Klink and Smith 2001), priming (Boush 1993, Pryor and Brodie 1998), explanatory links (Bridges et all. 2000), emphasizing the salience of association (Bridges et al. 2000), or the amount of information sharing (Klink and Smith 2001).
Priming is generally understood as “the fact that recently and frequently activated ideas come to mind more easily than ideas that are not recently or frequently activated” (Fiske and Taylor, 1984, p. 231). In brand extension research, the concept has been proven “to activate dormant attributes which the parent and extension category have in common” (Pryor and Brodie, 1998, p. 497) as reported by Chakravarti et al. (1990) who investigated brand name extension strategies and Johnson (1986) who looked at consumers’ similarity judgements. The work of Pryor and Brodie (1998) and Boush (1993) has clearly indicated that priming slogans can directly influence the evaluation of a brand extension when emphasizing the similarities between the two. However, due to the use of fictitious brand and the general use of the “fit” construct, there is stil little know about the effect of priming on real brands and on which fit dimensions can be influenced via this technique.
The main research question of this paper will focus on how priming can help improve the stretchability of a brand by appealing to various fit dimensions between the parent brand and its extension. Thus, we will ask:
RQ: What is the effect of priming on the perceived fit of brand extension for various fit levels?
H1: For high fit extensions, priming will not influence the evaluation of the brand extension (due to the already existing similarity between the parent brand and the extension).
Thus, we can conclude that:
H1a: For high fit extensions, priming category similarities will have no effect on the brand extension evaluation.
H1b: For high fit extensions, priming similar attribute characteristics will have no effect on the brand extension evaluation.
H1c: For high fit extensions, priming similar usage situations will have no effect on the brand extension evaluation.
H2: For moderately fit extensions, priming will lead to better brand evaluations (due to the activation of dormant brand associations).
Thus, we can conclude that:
H2a: For moderately fit extensions, priming category similarities will have a strong positive effect on the brand extension evaluation.
H2b: For moderately fit extensions, priming attribute similarities will have a strong positive effect on the brand extension evaluation.
H2c: For moderately fit extensions, priming similar usage situations will have a strong positive effect on the brand extension evaluation.
H3: For low fit extensions, priming will moderately positively affect the evaluation of the brand extension (due to the increased difficulty of making similarity associations).
Thus, we can conclude that:
H3a: For low fit extensions, priming category similarities will have a moderate positive effect on the brand extension evaluation.
H3b: For low fit extensions, priming attribute similarities will have a moderate positive effect on the brand extension evaluation.
H3c: For low fit extensions, priming similar usage situations will have a moderate positive effect on the brand extension evaluation.
Due to the current economic crisis, marketing communications investments have decreased, and it is becoming increasingly important to determine what constitutes a successful communication campaign for brand extensions that would determine consumers to better evaluate brand extensions.
Consequently, this research will focus on understanding how to design marketing communication strategies are efficient in promoting brand fit of brand extensions by using priming.
This study contributes to research on brand extension in several ways. First, it highlights the importance of communication techniques in establishing brand fit for brand extensions. Second, by looking at the various fit dimensions, we can identify which ones can be subjected to change via the use of priming. Third, no other study on the effects of priming on brand extensions has used real brands. Thus this will be an opportunity to put existing theories to the test of the market place and existing brand associations. Fourth, from a managerial perspective, our study shows which type of advertising techniques are suitable for establishing brand fit in the minds of the consumers. Finally, the data can be used to apply this persuasion model to individual business contexts.
II
Literature Review
Brand Extensions
The benefits of brand extensions were noticed as back as 1981, with Tauber being one of the first researchers to confirm that new products do indeed benefit from existing brand names; after 1990 research in the field of brand extensions exploded, with Aaker and Keller (1990) leading the way, by looking at how consumers evaluate extensions; Boush and Loken (1991) conducted a process-tracing study on brand extension evaluations; Keller and Aaker (1992) also looked at the sequential introductions of brand extensions. Loken and John (1993) were one of the first to look at the potential negative effect of brand extensions on parent brands; Broniarczyk and Alba (1994) continued evaluating the relationship between the parent brand and its extension by establishing the role of the former on making judgements about the latter.
Bottomley and Holden (1996) followed with research on how attitudes for brand extensions are formed; Bridges et al. (2000) looked at improving brand extension evaluations via communication strategies; de Ruyter and Wetzels (2000) applied the concept of similarity in service extensions; Lane (2000) also looked at ad repetition effects on improving brand fit for extensions. More recent research, such as Buil et al. (2009) has focused on cultural differences in brand extension evaluations.
The benefits of using brand extensions to introduce new products have been subjected to the test of research. It seems that brand extensions might benefit both consumers and managers. Consumers benefit from integrating new product information into their existing schemas, decreasing information search time (Roselius 1973), being promised a constant quality due to the use of similar skills and technology (Aaker and Keller 1990). Managers also benefit from brand extensions as consumers require reduced amounts of information (Smith and Park 1992), develop awareness and trust at a faster pace (Smith and Park 1992) and advocate trial (Smith and Park 1992) due to spillover effects (Simonin and Ruth 1998), decrease advertising spendings which are “moderated by intrinsic similarity”(Smith and Park, 1992, p.308) and even market share when the parent brands have already strong associations (Smith and Park 1992).
However, as with any strategy, brand extensions also have potentially negative effects. Many researchers looked at the potential negative impact brand extensions might have on parent name familiarity, dilution of the parent brand,
Fit and establishing successful brand extensions
Most research in the field of brand extensions agrees that the fit between the parent brand and its extension plays an important role in determining the success rate of the extension. First identified by Aaker and Keller (1990) as a success determinant, the role of brand fit has been confirmed by various research such as Boush and Locken (1991), Bottomley and Doyle (1996), Bottomley and Holden (2001), Völckner and Sattler (2006), Buil et al. (2009). The results were validated for both product and service categories, with Boush et al. (1987), Aaker and Keller (1990), Park et al. (1991) focusing on the product brand extensions, while de Ruyter and Wetzels (2000), van Riel et al. (2001), Hem et al. (2003) and Lei et al. (2004) extending the brand extension research to the product category.
The reasons the concept of fit has received so much attention is because it has been proven to lead to a better overall evaluation of the brand extension (Aaker and Keller, 1990; Boush and Loken, 1991; de Ruyter and Wetzels, 2000; Völckner and Sattler, 2006), reinforce the parent brand’s image (Zimmer and Bhat, 2004; Buil et al., 2009), require smaller investments for new product introductions (Smith and Park, 1992), increase market share (Smith and Park, 1992). Moreover, Smith and Park (1992) agree that when there is a strong fit, brand extensions become a very successful market entry strategy. Although initially it was believed that the role of brand fit in extant research might have been influenced by the predominant use of hypothetical brand extensions (Klink and Smith, 2001), Völckner and Sattler (2006) noticed the same strong effect of fit in determining the success of a brand extension, although their model only used real extensions.
Although many conceptualizations of the concept do exist, it seems that the exact role of “fit” is still rather vague in the literature, with Völckner and Sattler (2006) asking themselves what fit actually means after noticing its strong effect on the evaluation of brand extensions, as no exact conceptualizations exists to date (p. 31). Most of the definitions have a common base, as identified by Buil et al. (2009): “the degree of proximity between parent brand and extension that consumers perceive.” (p. 1303) Most researchers have considered fit as the similarity between the parent brand and the extension by looking at the product category and attributes that the two share (Broniarczyk and Alba, 1994; Keller, 2002; Park, Milberg, and Lawson, 1991).
Although research proves that cultural differences might influence brand extension evaluations due to different thinking and categorization styles, as proposed by the self-construal theory elaborated upon by Ahluwalia (2008). Although differences might occur, it seems that there is a clear trend in evaluating extensions, as can be observed in Graph 1 : under low or moderate processing motivation, brand extensions with a high perceived fit are considered to have a low stretch from the parent brand, whereas small perceived fit is congruous with a higher stretchability. It seems that the relationship is not direct, but rather has a S-curve appeal.
Graph 1: Expected pattern of effect for high-versus low-interdependence consumers for varying levels of brand stretch under moderate to low processing motivation
(Ahluwalia, 2008)
The fit between the parent brand and its extensions has potential consequences for both the parent brand and the extension. A high perceived fit can lead to a positive spillover effect on the parent brand (Keller and Aaker 1992), reinforcing the original associations of the parent brand (Aaker 1990), increased brand equity (Swaminathan et al., 2001), or stronger parent brand awareness (Aaker 1991). At the same time, a strong fit has been proven to have a strong positive effect on the quality perception of the new product, (Aaker and Keller 1990) due to the transfer of knowledge.
The general opinion is that perceived fit “can contribute to more positive evaluations when combined with other dimensions such as parent brand quality.” (Aaker & Keller, 1990, p. 38) Broniarczyk and Alba (1994) also observed that when brand specific associations are at a constant level, brand fit can account for the transfer of affect from the parent brand to the extension when certain other associations are held constant and that when consumers cannot judge brands via affect or other associations, they resort to fit as the main evaluation criterion.
Brand concept consistency has also been proven to indicate a fit between an extension and a parent brand, with prestige or luxury brands benefiting more from this type of fit (Park et al., 1991), partially due to the increased accessibility of concept characteristics that luxury brands enjoy. The first advocates to including brand stretchability as a measure of brand equity, Park et al. (1991) concluded that prestige brands might be more extendable than regular brands under the condition that the extension category or products have a consistency with the existing concept, thus accounting for the success of brands such as Prada or Jaguar offering phones, Ferrari offering laptop cases or high-class strollers.
The most widely accepted categorization of “fit” is the one developed by Aaker and Keller (1990) who distinguished between three types of fit: complementarity, transferability and substitution.
Complementarity refers to product that “are consumed jointly to satisfy some particular need” (Aaker and Keller, 1990, p. 30). In the study, complementarity has not been found as a factor contributing to better brand extension evaluations. Substitution refers to products that “tend to have a common application and use context such that one product can replace the other in usage and satisfy the same needs” (Aaker and Keller, 1990, p. 30). Similar to complementarity, substitution also had no effect on brand extension evaluations in the study. The only dimension of fit that improved the evaluation of a brand extension was transferability, “the perceived ability of any firm operating in the first product class to make a product in the second product class” (Aaker and Keller, 1990, p. 30).
Although two fit dimensions did not seem to have a direct effect on the evaluation of brand extensions, they cannot be dismissed as unimportant for several reasons, including possible indirect effects, or potential differences between the use of fictitious versus non-fictitious brands.
When looking at the effect of brand extensions on parent brand memory structures and retrieval processes, Morrin (1999) noticed that the degree of affective transfer from parent to its extension is directly proportional to the level of fit between the two, an effect also noticed by Keller and Aaker in 1992.
Although most researchers agree that the extent of fit between the parent brand and an extension can affect its overall success, Ahluwahlia (2008) recognizes that there is still a lack of understanding on the drivers of fit and admits that there are no clear guidelines helping managers understand the stretchability of a brand. The findings of this research shed some light on how consumers evaluate brand fit, and suggest that consistency levels seem to be higher when there is a stronger similarity in terms of category and attributes; that is not always the case for usage situations and target market overlaps (Ahluwalia 2008).
Park, Milberg and Lawson (1991) conclude that the type of brand also influences the potential success levels of an extension, with luxury brands having a higher stretchability due to their abstract rather than functional concepts, such as status symbols, comfort or craftsmanship which allows them to “accommodate a more diverse set of objects that share fewer features” (p. 192)
Categorization theory and perceived fit
In order to understand how consumers assign different products to different product categories, we need to have a closer look at categorization theory. The most widely accepted theory about how our brains organize long-term information states that we use associative network theories to retrieve information, and that we store information in hierarchical networks which contain various nodes, or concepts interconnected through various links or associations. According to this theory, family brands, as described by Loken & John (1993) can become categories in themselves, and are thus associated with a number of specific attributes based on the attributes of the compounding products.
The way we assign different products to certain categories has been the subject of various research throughout the years. The theory suggests that when confronted with a brand extension, consumers can evaluate it in two possible ways, either by inferring attribute beliefs and their evaluative importance consistent with piecemeal processing, or by transferring the overall attitude of the parent brand to the extension when there is a perceived category fit between the two (Aaker and Keller 1990). For moderate fit, piecemeal processing seems to be greater than for very similar or dissimilar extensions, as discovered by Boush & Loken (1991). The same study proved that attitude transfer from an existing brand to its extension has a linear relationship with the typicality of the relationship, as can be observed in Graph 2 and Graph 3.
Graph 2: Variations in Piecemeal Responses by Typicality Rating for Grocery Products
(Boush & Loken, 1991)
Graph 3: Variations in Piecemeal Responses by Typicality Rating for Elecronic Products
(Boush & Loken, 1991)
Morrin (1999) proved that for well established brands, the type of extension fit does not influence the categorization speed, but it does so only for non-dominant parent brands (brands with little awareness).
There are two mainstream theories about how our brains react to inconsistent information, such as the information resulting from a farther brand extension and a potential misfit between the category established by the initial parent brand and an extension that lacks a connection to it or does not share the same attributes. Loken and John (1993) describe several models. One of them is the “bookkeeping model” which assumes that consumer beliefs “change incrementally as new information is received” (p. 72). A second alternative is the “typicality-based model”, which states that the impact consumers attribute to the inconsistent information depends on how typicality of the product to the category and is based on the belief that typical members would share more attributes with other category members than non-typical products (Loken & John, 1993, p. 73). Both models were supported by their research, with a few significant differences: when the researchers measured family brand beliefs first, evidence supported the bookkeeping model, whereas when typicality judgments were assessed first, the typicality-based model proved more reliable. At the same time, Loken and John (1993) inferred that “typicality is more likely to affect beliefs that relate to core or distinctive attributes.” (p. 81)
Communication methods for increasing fit
According to the categorization theory, in order to retrieve a certain concept, or node, an activation process has to occur. Achieving this activation can be done through various techniques, one of them being that of priming, as discussed by Morrin (1999).
Priming, “the fact that recently and frequently activated ideas come to mind more easily than ideas that are not recently or frequently activated” (Fiske and Taylor, 1984, p. 231) would lower the threshold levels of activation for a certain node, thus facilitating the retrieval process. At the same time, due to increased network expansion (over-stretching a brand), a concept might also be inhibited because of the expansion of its initial network, which can be a consequence of for instance adding a new product category to the existing ones.
At the same time, Morrin (1999) proved that brand extension advertising does not facilitate parent brand recognition in real-world situations if consumers have more than four product choices on the shelf, thus proving that it might be more beneficial to emphasize the elements of the original brand and not those of the extension for crowded product categories.
When looking at the effect of advertising slogans, Boush (1993) observed a clear effect that slogans have not only on improving the perceived fit between the parent brand and the extension, but also on the overall evaluation of the brand extensions by “drawing attention to attributes that the new product either has in common with existing products or that conflict with existing products” (Boush, 1993, p. 67). The main reasoning for using priming has been that when processing new information, by being offered specific cues on which to focus is more likely to emphasize the importance of those characteristics (Boush 1993).
Repetition also plays a role in establishing a fit between the parent brand and the extension, as confirmed by Lane (2000) especially for moderate and farther extensions. Repeated exposure to an ad stimulus improved the overall evaluations of the extension (Lane 2000). The study emphasized the role of elaborative processing as a means of influencing fit (Aaker and Keller 1990, Broniarczyck and Alba 1994).
It is generally believed that when confronted with moderate extensions, independent self-construals have a lower evaluation of the perceived fit between the two and that they “tend to decontextualize the object from its field and process it analytically in an attempt to establish its category membership” (p. 340) Independent self-construals tend to poorly evaluate moderate and farther brand extensions and preferred being offered with explicit information about the fit between the extension and its parent brand (Ahluwalia, 2008).
Based on previous knowledge, we proposed the following hypotheses:
H1: For high fit extensions, priming will not influence the evaluation of the brand extension (due to the already existing similarity between the parent brand and the extension).
Thus, we can conclude that:
H1a: For high fit extensions, priming category similarities will have no effect on the brand extension evaluation.
H1b: For high fit extensions, priming similar attribute characteristics will have no effect on the brand extension evaluation.
H1c: For high fit extensions, priming similar usage situations will have no effect on the brand extension evaluation.
H2: For moderately fit extensions, priming will lead to better brand evaluations (due to the activation of dormant brand associations).
Thus, we can conclude that:
H2a: For moderately fit extensions, priming category similarities will have a strong positive effect on the brand extension evaluation.
H2b: For moderately fit extensions, priming attribute similarities will have a strong positive effect on the brand extension evaluation.
H2c: For moderately fit extensions, priming similar usage situations will have a strong positive effect on the brand extension evaluation.
H3: For low fit extensions, priming will moderately positively affect the evaluation of the brand extension (due to the increased difficulty of making similarity associations).
Thus, we can conclude that:
H3a: For low fit extensions, priming category similarities will have a moderate positive effect on the brand extension evaluation.
H3b: For low fit extensions, priming attribute similarities will have a moderate positive effect on the brand extension evaluation.
H3c: For low fit extensions, priming similar usage situations will have a moderate positive effect on the brand extension evaluation.
The following conceptual model will be tested for validity:
We expect the following phenomena to occur:
III
Methodology
Procedure
Stimuli
This study will assess the differences in brand extension evaluations for real brands and potential extensions under different priming conditions aimed at improving the level of fit between the parent brand and the extension. In order to conduct the study, we will use the stimuli introduced by Ahluwalia (2008) due to their already proven different stretch levels. The results obtained by Ahluwalia (2008) can be observed in Table 1:
Table 1: Brand extension stimuli
Extension |
Category Similarity |
Attribute Similarity |
Usage Situation Similarity |
Perceived Fit* |
Extension Evaluation* |
|---|---|---|---|---|---|
|
Sony high-definition television |
6,80 |
6,64 |
6,76 |
5,85 |
6,09 |
|
J&J lotion |
6,74 |
6,08 |
6,49 |
5,07 |
5,00 |
|
J&J stuffed toys |
2,84 |
6,12 |
2,69 |
3,38 |
4,80 |
|
Adidas brown shoes |
4,16 |
4,38 |
3,35 |
3,10 |
4,26 |
|
Sony vacation resort |
1,98 |
3,52 |
5,18 |
2,56 |
4,15 |
|
Adidas calculator |
1,68 |
2,82 |
2,43 |
1,79 |
3,53 |
|
J&J instant noodles |
1,24 |
1,76 |
1,16 |
1,84 |
3,52 |
the results for low interdependence consumers were used as our study will be conducted in Europe, which is known to score low on interdependence self-construal
Pretest
A pre-test with approximately 50 students will be carried out in order to confirm Ahluwalia’s (2008) findings and their applicability to our subject pool. Unlike Ahluwalia (2008), our subjects will be informed that the purpose of the survey is to test different slogans for their efficiency in promoting the new extensions. They will first be provided with two paragraphs on the company statements of the four parents brands and another paragraph describing the seven extensions. Afterwards, the participants will be asked to rate the extensions in terms of the three fit categories identified by Aaker and Keller (1990): transferability, complementarity and supplementarity, using similar scales as the original study. An overall evaluation of the similarity of the extension with the parent brand will also be included. Respondents will be asked to rate the extensions on a 7-point Likert scale anchored at “strongly disagree” and “strongly agree”:
Fit Dimension |
Question |
|---|---|
|
Substitute |
The new product can replace the original one in usage. |
|
Complement |
The original and the new product can be used simultaneously. |
|
Transfer |
The skills and the resources used to make the original product can also be used to manufacture the new product. |
|
Overall fit |
Overall, the original product and the extension are similar. |
Respondents will also be asked to fill in several open-ended questions aimed at identifying possible examples for the three fit categories:
Fit Dimension |
Question |
|---|---|
|
Substitute |
When do you think the two products could be used to replace one other? |
|
Complement |
When do you think these two products can be used at the same time / in the same place? |
|
Transfer |
How would you use the skills needed to create the first product to create the new one? |
Main Experiment
A sample of approximately 300 respondents will be asked to fill in an online questionnaire as part of a 4 (priming conditions) x 7 (brand extension replicates) between-subject factorial design. Using 7-point Likert scales, the respondents will evaluate the seven extensions in terms of:
- product concept (anchored at: “very unfavorable” and “very favorable”)
- the perceived fit of the extensions on the four categories as described in the pre-test after being presented with one of the priming conditions.
- report of brand familiarity
- report on desirability, favorability and quality for the parent brand as manipulation checks.
- report on agreement with the primed fit dimensions.
- overall evaluation of the brand extension: purchase intention/ likability/ desirability/ quality
The participants will be asked to read a brief advertising copy containing one paragraph on the original brand and one on the extension. The priming will be conducted via different slogan manipulations, similar to the studies conducted by Boush (1993) and Pryor and Brodie (1998) but focused on the fit categories.
Similar to Ahluwalia (2008), brand familiarity will be used as a covariate in our study. Planned contrasts will be used to find statistically different evaluations of the brand extensions in the presence of the priming conditions. The results will be interpreted using ANOVA and ANCOVA in order to identify differences between the four priming conditions (no priming, priming substitute, priming complement and priming transfer). The results will be analyzed using PASW Statistics.