Creating stronger brands through effective brand consistency
An experimental investigation of the effect of brand consistency on customer-based brand equity, brand attitude and brand image and the role of brand familiarity
Name: Sil Leijdekkers Student number: 5894352 Master: Communication science Track: Persuasive communication Supervisor: D. Muntinga
Date: 29-01-2016 Word count: 12107
Index Introduction 2 Theoretical framework 6 Methodology 20 Results 31 Conclusion 38 Discussion 41 References 44
Appendix A: Questionnaire of the pretest 56
Appendix B: Questionnaire of the experiment 59
Appendix C: Stimulus material 66
ABSTRACT: The aim of this research is to find out if brand consistency has an effect on customer-based brand equity, brand attitude and brand image and if so, what level of brand consistency is more effective. Moreover, brand familiarity is measured to analyse its expected moderating influence on the mentioned effects. The research is based on the dominant perspectives in brand consistency research: inconsistent, personalized brand communications on the one hand, versus consistent, integrated brand communications on the other. The results of an experiment demonstrate not only that brand consistency has a significant effect on customer-based brand equity and brand attitude, but also that moderately consistent and consistent brand advertising are the most effective levels. Furthermore, no main effect of brand consistency on brand image was found, nor any effect of the moderator brand familiarity. Consequently, brand consistency should be considered a focus point when building customer-based brand equity and/or trying to influence brand attitude.
Introduction
In an ever more globalizing and advertising world, with growing numbers of brands, advertising and a further segmentation or increasing heterogeneity of smaller groups of target customers, it becomes harder for brands to communicate what the brand stands for and to maintain a consistent and a coherent brand image (Rosengren, Dahlén & Modig, 2013; Rotfeld, 2006). This can turn out to be disastrous, because not coherent brands, which do not communicate a clear brand image and brand values are more easily overlooked, forgotten and eventually neglected by customers. This, in turn, means lower sales levels and lower profits, if any. It is important for customers to know what a brand stands for and the brand image is one of only a few brand assets that have a possibility to do just that (Aaker, 1996). This might have the brand survive
in a overcrowded marketplace (Keller, 2008; Keller, 1993; Aaker, 1996), consisting of a huge amount of competitors for customers and market share and ever-growing levels of advertising and other marketing and branding activities.
Also, for customers, the huge amount of advertising clutter and brands fighting for a little piece of the customer’s attention could lead to several issues. Which brand is personally relevant or which brand should I buy and maybe keep on buying are becoming puzzles instead of questions one asks him or herself on a daily basis. Being confronted with more than one targeted advertisement of the same brand (because customers usually are part of more than one target group), which contain messages relating to different brand associations and brand features, makes these puzzles even harder to solve. It is that cultural differences between customers and inter-personal differing brand associations make it hard for brands to reach their customers in a relevant way, for example based on their personally relevant brand associations, but it does not justify that as many customers as possible are targeted with many brand advertisements. This way of communicating means that customers are increasingly unable to decipher clear brand images and find out about brand values and the right or relevant brand for them.
In this light, it is questionable whether brands do good when they segment their target groups further and try to reach niche groups of possible customers, even though putting customers central in the marketing and branding efforts is something that is said to be successful (Kotler & Armstrong, 2009). The aforementioned brand and customer issues call for a change in brand communications. Putting more effort in communicating a clear, consistent and coherent brand image instead might be a solution (Sharp, 2010; Keller, 1993). Still, this way of communicating has not been adopted as much, as it has not been proven successful, yet.
The brands’ core business, in most cases, is to sell, turn-over and in the end make money. This will be the case as long as the money coming in through selling products or services surmounts the money being spend on marketing- and branding-related activities. The brand that sells the most and spends the least will be the winner in terms of profit and therefore every brand is in search of ways to attract more
customers (to increase sales) or to cut the amount of money being spend on marketing and branding. Whether this is through listening to the customers and let them be king (Kotler & Armstrong, 2009) or through promoting the brand with differentiating brand values (Sharp, 2010) does not matter. Or does it?
This is where this research comes in. It aims to test whether brand consistency has a significant effect on customer-based brand equity, including the mentioned brand image and brand attitude. Brand consistency is divided in three levels, ranging from inconsistent to consistent brand advertising. This could provide an insight in the possible success of the two dominant perspectives in brand–related research: inconsistent brand communications (i.e, personalized advertisements, Kotler & Armstrong, 2009) and consistent brand communications (i.e., integrated marketing communications, Sharp, 2010; Keller, 1993). Furthermore, the present reseach is extended through the inclusion of brand familiarity, as an expected moderating influence.(i.e., Delgado-Ballester, Navarro & Sicilia, 2012). It allows for a distinction of the effects of brand consistency for different kinds of brands, varying in familiarity.
In the present research, the ‘success’ means a higher customer-based brand equity. This is the added value of a brand for a customer as well as for the brand itself (Keller, 1993; Aaker, 1991). Why would a customer pick a certain brand over another brand or think better of one are questions that are related to customer-based brand equity. Very important brand concepts, that are closely linked to customer-based
brand equity according to research, are brand attitude and brand image. These brand concepts will also be tested to enhance the overview of influences of brand
consistency on the most important brand assets. It is also the intention to signify what level of consistency is preferred for all brand concepts. Therefore, the following research question has been drawn:
RQ = What is the influence of different levels of brand consistency on customer-based brand equity, brand attitude and brand image? And how is this influence moderated by brand familiarity?
This research is scientifically relevant because research to date has not been conclusive about what factors play a role in building customer-based brand equity (Christodoulides & De Chernatony, 2010; Raggio & Leone, 2007) and whether or not it is influenced by brand familiarity. Adding to the current amount of research is particularly important because customer-based brand equity is a strong predictor of turn-over, sales and market share. These concepts are very central to brands’ core business nowadays. This research tries to clarify the influence on customer-based brand equity of brand management generally and brand consistency specifically. Both have limited empirical research to build upon and the level of consistency has not even been considered an influence on customer-based brand equity at all.
Furthermore, there is not much literature that stresses the characteristics and conditions of brands that might influence effects of brand consistency on customer-based brand equity. The inclusion of brand familiarity in this research is a little step in filling that gap. It functions as an example of possible mitigating and influencing brand conditions.
For marketing and business professionals and brand enthusiasts, this research might show how and when to use either inconsistency or consistency to have an impact on the brand’s customer-based brand equity and to deploy marketing and branding efforts more effectively. Situations might be different for strong, well-established, very familiar brands and relatively new, unfamiliar ones or even start-ups. To date insights in how these situations might have a different impact on
customer-based brand equity have not yet been shown much (for the limited research, see for example Aaker, 1996; Keller, 1993; Aaker, 1991) and have not been
empirically proven. This research tries to give valuable insights on which brand-related business practice can build. For the general public, lastly, this research might show ‘how brands work’ and what their ‘added value’ actually means.
Theoretical framework
Two perspectives of (in)consistency in branding and marketing
In brand-related research practice nowadays, two perspectives are on the forefront, propagating either inconsistency (Kotler & Armstrong, 2009; Kotler, 1972) or consistency (Sharp, 2010; Keller 1993) in brand communications. Next, both perspectives will be discussed and compared.
Inconsistency
On the one hand, there is the strong, proven conviction that brands should increase the inconsistency in their brand communications and should more directly address members of their target groups. Increasing inconsistency can happen via many brand communications strategies. Examples include personalization (De Keyzer, Dens & De Pelsmacker, 2015; Demmers, Van Dolen & Weltevreden, 2013; Antheunis & Van Noort, 2011), direct mailing and addressing (Piersma & Jonker,
2004; Reinartz & Kumar, 2002), different brand messages based on different brand personalities (Eisend & Stokburger-Sauer, 2013; Malär, Nyffenegger, Krohmer & Hoyer, 2012; Aaker, 1997), location-based marketing (Banerjee & Dholakia, 2008; Thode & Maskulka, 1998) social media marketing and branding (De Vries, Gensler & Leeflang, 2012; Kaplan & Haenlein, 2012; Yan, 2011) and brand-related
user-generated content (Ertimur & Gilly, 2012; Constantinides & Fountain, 2012; Daugherty, Eastin & Bright, 2008). These strategies increase inconsistency because they all require tailoring, personalization and/or customization of brand messages. Those processes, in turn, lead to different mesasges from the same brand, for different target groups or stakeholders, hence inconsistency.
Common brand consistency theory holds, that as long as the level of
inconsistency of brand advertising is not too incongruent with existing schema, it can have either positive or negative effects on advertisement processing and the
motivation (Törn & Dahlén 2009; Dahlén, Rosengren, Törn & Öhman, 2008; Mandler, 1982; Hunt, 1963). This, in turn, might change brand attitude and brand image (Petty & Cacioppo, 1986) and in the end customer-based brand equity. Inconsistency in brand communications and putting the individual customers center-stage might increase customer-based brand equity (Kotler & Armstrong, 2009; Andreasen & Kotler, 2008, Kotler, 1972), through creating new brand associations in customers’ minds, but it can also be too schema-incongruent (Mandler, 1982).
The foundations of the perspective that propagates inconsistency lie within the research by Kotler (1972) and it is therefore colloquially known as the Kotlerian view on marketing and branding (Sharp, 2010). Addressing the demand for
heterogeneity, that is satisfying the needs of segmented target groups, is what brands should do in order to compete for a larger market share and an increase in sales
(Kotler & Armstrong, 2009; Kotler, 1972). Brands would have to become more differentiated and would have to sell to different kinds of customers in the same market, with a possible strategie being customer relationship marketing (N’Guyen & Mutum (2012; Kotler & Armstrong, 2009). The researches of Puntoni, Vanhamme and Visscher (2011), Puntoni, Schröder and Ritson (2010) and Warlaumont (1995) show that there are advantages only of ‘purposeful polysemy’, which emerge through a brand’s efforts to create ambiguity and inconsistency in their communications. Consistency
On the other hand, there is the perspective that brand communications should not be changed for every single target group, but should shed a light on the ‘raison d’être’ of a brand: the reason the brand exists (Sharp, 2010; Keller, 1993). Contrary to the perspective promoting inconsistency, for which the customers are the most
important, this perspective promotes that the brand is the most important element in the brand’s marketing and branding and should be put in the center of the attention.
Consistency refers to the implementation of common brand identity, brand personality and brand meaning, with according content, in the brand’s
communications (Aaker & Biel; 2013; Delgado-Ballester, Navarro & Sicilia, 2012; Keller, 2008; Keller, 1993). The core values of the brand and its strong belief in satisfying a need of customers should be enough for a brand to attract customers and therefore ensure a stable turn-over and enough sales to have the brand survive and possibly even grow. Furthermore, the communication of a consistent and coherent brand could rely on a vast amount of, in most cases, loyal customers. This is due to the clear brand message the brand is trying to communicate, which is easy to interpret for customers and therefore they know what the brand has to offer and whether the brand is relevant for them. In an environment with so many competing brands and
even more brand advertising, this might be the very key to come out on top (Sharp, 2010; Keller, 1998). Identifying the brand’s own identity, personality and core values might even be used in building a stronger and more successful brand in the way that these concepts nurture new marketing efforts, when the traditional (mass) media are no longer effective in transferring the brand image, because customers are
increasingly avoiding advertising via these channels (Joachimsthaler & Aaker, 1997). Consequently, to maintain brand consistency might end up being not so different as a brand (Kotler & Armstrong, 2009) but to change and to innovate for the sake of targeting potential customers and looking for new markets might end up being not so relevant as a brand (Beverland, Wilner & Micheli, 2015). Relevance is as important as being different. Customers like to have different options to choose from, but they will only consider personally relevant options. Therefore, brands need to and try to be relevant all the time, which allows for inconsistency, but they also have to be different and have to stand out from other brands, which calls for consistency. This shows how important the level of consistency in brand communications can be, but it also shows that more elements in brand communications should be reckoned with. Inconsistency or consistency: a comparison
Conventional brand management wisdom holds that singularity is to be preferred over variety, as is simplicity over complexity and specificity over polysemy (Brown, 2014). This means, that consistency is preferred over inconsistency, but many brands are very inconsistent (or too consistent). They either try to be complex, flexible and inconsistent (for example: Nike or KLM) or single-minded, stable and consistent (for example: Nespresso or Ben & Jerry’s), when they are reaching out to customers. Both strategies do not guarantee that customers will keep on buying the brand, since brand associations can change over time (Keller, 1993; Aaker & Keller,
1990), because these are the result of an adaptive learning process (Van Osselaer & Janiszewski, 2001; Janiszewski & Van Osselaer, 2000; Shanks, Darby & Dickinson, 1996; Pearce, 1994) and schema (in)congruity (Delgado-Ballester, Navarro & Sicilia, 2012; Sjödin & Törn, 2006).
Adaptive learning and schema (in)congruity
Adaptive learning means that through experience with a brand and brand cues, brand associations can change, because brand experiences and cues either meet or fail to meet expectations and brand associations related to those expectations (Van Osselaer & Janiszewski, 2001). Regardless of previous brand experience, brand associations exist as mere predictors of the brand. Through brand experience they either become stronger or weaker or eventually disappear.
Adaptive learning requires repetition, meaning that brand associations have to be met or have failed to be met in a number of occasions, before brand associations start to change (Van Osselaer & Janiszewski, 2001; Pearce, 1994). This brand
association change can, in turn, function as a more precise base for new predictions or expectations about the brand and then the whole process starts anew (Van Osselaer & Janiszewski, 2001). This might imply that consistent brand advertising benefits the adaptive learning process.
Applied to brand advertising, the schema incongruity theory (Mandler, 1982) holds that advertising messages that do not fit the existing brand schema in the
customer’s mind are unlikely to be processed, let alone able to change brand
associations, brand attitude or brand image. Usually, brand schema consist of strong associations that are not easily changed, for example because most customers rather see their brand associations confirmed than challenged and therefore avoid
confirmation (Mandler, 1982). The schema incongruity theory (Mandler, 1982) might provide an explanation for the effects of consistent brand advertising in that it shows that messages are more likely to be processed when they are congruent and consistent.
However, a certain level of incongruity is acceptable (Dahlén, Rosengren, Törn & Öhman, 2008). In that case a message might be processed attentively and be able to change brand associations in the long run. Nonetheless, it depends on the brands, regarding which level of congruity is acceptable (Dahlén, Rosengren, Törn & Öhman, 2008; Dahlén, 2005; Dahlén & Lange, 2004; Lange & Dahlén, 2003).
Brand associations
Brand associations, either desired (whatever the brand wants customers to relate them to) or realized (the actual brand associations in customers’ memory) are very important to brands. They play a critical role for customers in choosing brands, buying brands and evaluating brands. Brands’ marketing and branding activities create and influence these desired brand associations, but these differ for the different target groups. A sports brand, for example, would communicate sports-related
associations to athletes, fun- and price-related associations to customers and yet other associations to sports organisations. This might mean that a lot of desired different brand associations are being communicated (inconsistency), because of the segmented target groups. It can also be that the same sports brand keeps on communicating the same kind of brand associations over and over (consistency), in order to ensure strong links between the brand and the communicated brand associations. From this point of view, brands are the one that build brand associations, but this is not true in reality: customers do. Customers do not even need matching brand associations to consider a brand when buying products (Koll & Von Wallpach, 2013). Brands themselves have only a small influence on brand association building.
Customer-based brand equity: the brand’s added value
Brand equity is a construct based on several concepts and it can be defined in different dimensions and that is the reason there is no existing agreement on what brand equity really is (Christodoulides & De Chernatony, 2010; Raggio & Leone, 2007; Washburn & Plank (2002). The dimensions are financial brand equity
(Fetscherin, 2010; Shimp, 2010) and customer-based brand equity (Christodoulides & De Chernatony, 2010; Leone et al., 2006). The former relates to the incremental added value of a brand in terms of market share, sales and return on investment (ROI) (Aaker, 1996). It provides economic advantages over non-brands (Buil, De
Chernatony and Martínez, 2008). The latter relates to all concepts that have to do with customers and, more importantly, the perceptions of these customers (Zeugner-Roth, Diamantopoulos & Ángeles-Montesinos, 2008).
Brand awareness (Keller & Lehmann, 2006; Percy & Rossiter, 1992), brand image (Martínez & De Chernatony, 2004), brand attitude (Zarantonello & Schmitt, 2010; Matthes, Schemer & Wirth, 2007; Yoo & Donthu, 2001), brand loyalty (Keller & Lehmann, 2006; Pappu, Quester & Cooksey, 2005) are just some examples of concepts that customer-based brand equity can include. In this sense, customer-based brand equity is the incremental added value for a customer with brand knowledge, which is used in brand evaluations and (consequent) purchase situations. In this research, there is a focus on the customer-based brand equity dimension, because it is known to be a major premise for brand-related concepts.
According to the most common customer-based brand equity definition, it is the set of brand associations, brand assets, liabilities and behavior on the part of a brand’s customers, channel members and parent corporation, based on their subjective perception (Park & Srinivasan, 1994), that permits the brand to earn greater volume or
greater margin than it could without the brand name (Kapferer, 2012; Aaker, 1991; Leuthesser, 1988). In other words, customer-based brand equity is the incremental added value for a brand and its customers of a brand to a product, compared to a non-branded product. Aaker (1996; 1991) adds that it can also have backfiring effects and can subtract value instead of adding value as well.
Customer-based brand equity is very important for a brand in the sense that it could function as a base for predictions about future brand performance, of which the financial part is significantly influenced by the social part. This is, because in order for a brand to have financial value, it needs to have customer value, or in other words: the power of the brand lies within the minds and experiences of customers (Kotler & Keller, 2006; Cobb-Walgren, Ruble & Donthu, 1995). It expresses the relationship between the brand and its customers. Customer-based brand equity is also able to create and build ‘stronger’ brands (Hakala, Svensson & Vincze, 2012; Buil, De Chernatony and Martínez, 2008; Arvidsson, 2006).
It is assumed in most research (though limited), that brand consistency has an significant influence on customer-based brand equity (Delgado-Ballester, Navarro & Sicilia, 2012; Netemeyer et al., 2004) and indirectly on sales (Vogel, Evanschitzky & Ramaseshan, 2008; Oliveira-Castro et al., 2008; Cobb-Walgren, Ruble & Donthu, 1995). The influence of brand consistency on customer-based brand equity roots in brand associations, which are a . Both inconsistency and consistency have a way of influencing these brand associations. Inconsistency in brand advertising means trying to attract attention through advertising in an unusual way and trying to create new associations. Consistency in brand advertising means repeating the important brand associations over and over again, so they become strong links to the brand in the customer’s mind.
The elaboration likelihood model (ELM) (Petty & Cacioppo, 1986) and the schema incongruity theory (Mandler, 1982) provide explanations on the influences of brand (in)consistency on customer-based brand equity. Inconsistency enhances the probabilitiy that an brand advertisement will be processed because the inconsistent elements attract attention and that is a prerequisite for a customer’s motivation to process the advertisement attentively (Petty & Cacioppo, 1986). However, when the inconsistency is too incongruent with the existing brand schema (Mandler, 1982), the advertisement is not likely to be processed. Moreover, created or changed brand associations through inconsistent brand communications are unlikely to remain, because of their inconsistent nature. Consistency allures to the already existing brand schema, which facilitates easier and faster advertisement processing (Mandler, 1982), better decoding/interpretations (Malär, Nyffenegger, Krohmer & Hoyer, 2011) and strengthens the already existing brand associations (Petty & Cacioppo, 1986, Mandler, 1982). Because consistency is not as noticeable as inconsistency is and therefore not as much attracting attention and risking to be uninteresting, it has got a higher probability than inconsistency to be processed and to influence brand
associations. (Delgado-Ballester, Navarro & Sicilia, 2012; Sharp, 2010; Petty & Cacioppo, 1986; Mandler, 1982). Therefore, it might have a bigger effect on customer-based brand equity than inconsistency (Delgado-Ballester, Navarro & Sicilia, 2012). This is stated, together with the expected overall effect of brand consistency in the following hypotheses:
H1: Brand consistency has a significant effect on customer-based brand equity. H1a: This effect is significantly different between the inconsistent and consistent levels, but not between the moderately consistent level and the other two levels.
How brand consistency affects brand attitude
Brand attitude is a customer response to obtained brand-related information, which contains feelings, thoughts and associations (Kozary & Baxter, 2010; Cowley & Barron, 2008; Yang & Roskos-Ewoldsen, 2007; Van Reijmersdal, Neijens & Smit, 2007). In other words, it is a summary evaluation of a brand (Spears & Singh, 2004), usually in terms of likability and favorability and it differs in strength. Brand attitude has a significant influence on purchase intentions (Hartmann & Apaolaza-Ibáñez, 2012; Spears & Singh, 2004), brand evaluations (Zarantonello & Schmitt, 2010) and customer-based brand equity (Zarantonello & Schmitt, 2013; Park et al., 2010; Faircloth, Capella & Alford, 2001). Therefore, most brands put a lot of effort into maintaining them (when positive) or changing them (when negative) (De
Pelsmacker, Geuens & Van den Bergh, 2007).
These circumstances and behaviors change the way in which customers process brand advertising (if they do), which makes it difficult for this advertising to change anything. It does not matter whether a brand advertisement is inconsistent or consistent, but, opposite to brand associations, brand attitudes are increased through personal relevance of brand communications. Inconsistency, therefore, is more likely to influence these brand evaluations than consistency.
Inconsistent brand advertising might be fresh and different, which could make it attract attention and enhance the customer’s (Petty & Cacioppo, 1984), but it is also possible that it is only disruptive and causes cognitive dissonance, which in turn leads to a negative attitude towards the advertisement and, consequently, the brand. Consistent brand advertising might influence brand associations because it is easier to process and to decode/interpret (Dahlén, Rosengren, Törn & Öhman, 2008; Petty & Cacioppo, 1984; Mandler, 1982), but the its downside is that it can become
uninteresting or not noticeable and that affects the attitude towards the advertisement and the brand negatively.
It has not originally been considered a part of customer-based brand equity, but like brand image it is very closely linked and increasingly taken into account in customer-based brand equity conceptualizations. Since brand attitude is very closely linked to both customer-based brand equity and brand image (Zarantonello & Schmitt, 2013; Park et al., 2010), it is likely that the influence of the levels of consistency on brand attitude are similar. The only exception is that in this case inconsistent brand advertising is expected to have a bigger effect than consistent brand advertising, because of the evaluative nature of brand attitudes. Based on this, the following hypotheses are stated:
H2: Brand consistency has a significant effect on brand attitude.
H2a: This effect is significantly different between the inconsistent and consistent levels, but not between the moderately consistent level and the other two levels.
How brand consistency affects brand image
Brand image emphasizes what the brand stands for in the perceptions of customers. It shows what the brand means and whatever is connected to the brand within customers’ minds (Kaur & Agarwal, 2014; Martínez & De Chernatony, 2004). Brand image is the set of brand associations from the associative network (Collins & Loftus, 1975), which together hold some meaning and consist of strong(er) and weak(er) ties (Aaker, 1996; Keller, 1993; Aaker, 1991). Those associations could be related to product attributes, brand personality, brand identity, core values,
brand associations usually emerge through experience with or even expectations of a brand (Persson, 2010; Kuhn, Alpert & Pope, 2008). Those expectations are created through the brand’s advertising.
Related to customer-based brand equity, because it is also based on the associative network, it is that part that is decisive in determining if a brand is
personally relevant and should be considered in the next purchase situation, together with brand attitude and perceived brand quality. It enhances the differentiation of the brand and it might eventually lead to attractive economic and communicative
advantages (Persson, 2010). Therefore, it is a very important brand element and brands (should) do everything they can to influence this.
Brand image could be a either a distinct part of customer-based brand equity, like many researchers assume it to be (Vomberg, Homburg & Bornemann, 2014; Severi & Ling, 2013; Aaker & Biel, 2013; Kapferer, 2012, Franzen, 2009) or it is a very closely linked, but separate concept (Buil, De Chernatony & Martínez, 2013; Guzmán & Samih-Baalbaki, 2012; Keller, 2008; Pappu, Quester & Cooksey, 2005; Yoo & Donthu, 2001; Faircloth, Capella & Alford, 2001; Keller, 1993). Brand associations are a major part of customer-based brand equity, but lack meaning, so brand image could be a valid replacement. It adds meaning through the network of strong(er) and weak(er) ties.
In the present research, brand image is not considered to be part of customer-based brand equity, but to be very closely linked (Buil, De Chernatony & Martínez, 2013; Guzmán & Samih-Baalbaki, 2012; Keller, 2008; Aaker, 1996; Keller, 1993). Because brand image consists of brand associations, it is assumed that consistent brand advertising has a bigger effect on brand image than inconsistent brand advertising. This is based on the aforementioned ways of advertising processing
(Petty & Cacioppo, 1986; Mandler, 1982). Considering all the above, the following hypotheses are stated:
H3: Brand consistency has a significant effect on brand image.
H3a: This effect is significantly different between the inconsistent and consistent levels, but not between the moderately consistent level and the other two levels.
It is all about ‘knowing the brand’
When customers hold favourable, unique and strong brand associations, they are usually very familiar with the brand (Delgado-Ballester, Navarro & Sicilia, 2012; Keller, 1993). Very unfamiliar brands first have to create and build those brand associations and increase brand knowledge of customers, because they do not exist from scratch (Campbell & Keller, 2003). Brand familiarity, therefore, is one of a few major distinctive brand characteristics or features and it is often used to measure marketing and branding effectiveness. Also, it might influence a relationship between brand consistency and customer-based brand equity. It refers to the actual brand knowledge of customers, which is obtained through experience, either direct (i.e., purchase) or indirect (i.e., word-of-mouth) (Campbell & Keller, 2003), and the ability to recall and/or recognize a brand (Torres, Olavarrieta & Barra, 2016).
In this research, it is proposed that very familiar brands have more communicative advantages than very unfamiliar brands (Lange & Dahlén, 2003). These communicative advantages can include easier advertising processing (Delgado-Ballester, Navarro & Sicilia, 2012), lower risk perceptions (Persson, 2010), more favorable attitudes and better liking (Dahlén & Lange, 2004; Kent & Allen, 1994). Like aforementioned, very familiar brands also have a stronger associative network
(Collins & Loftus, 1975), which tend to be stronger and more unique, which in turn increases advertising and brand liking and improves the possibility of brand loyalty (Dahlén & Lange, 2004; Taylor, Geluch & Goodwin, 2004; Kent & Allen, 1994). This associative network might lead to a more favorable brand image, since the brand image is derived from this associative network (Aaker & Biel, 2013; Aaker, 1996; Keller, 1993; Aaker, 1991; Collins & Loftus, 1975).
Furthermore, higher brand familiarity can lead to a higher customer-based brand equity, because better liked, ‘closer’ brands, with stronger associative networks are usually considered more in purchase situations, purchased repeatedly and
considered more relevant for the individual customer (Torres, Olavarrieta & Barra 2016), which even might lead to a high level of brand loyalty. All of this together, it leads to the following hypotheses:
H4: For very familiar brands, the effect of brand consistency on customer-based brand equity will be higher, compared to very unfamiliar brands.
H4a: This moderated effect is significantly different between the inconsistent and consistent levels, but not between the moderately consistent level and the other levels.
H5: For very familiar brands, brand attitude will be higher for all three levels of brand advertising consistency, compared to very unfamiliar brands.
H5a: This moderated effect is significantly different between the inconsistent and consistent levels, but not between the moderately consistent level and the other levels.
H6: For very familiar brands, brand image will be higher for all three levels of brand advertising consistency, compared to very unfamiliar brands.
H6a: This moderated effect is significantly different between the inconsistent and consistent levels, but not between the moderately consistent level and the other levels.
Figure 1
The conceptual model.
Methodology Design & Procedure
To find out whether a causal relationship exists between the level of
consistency in brand advertisements and customer-based brand equity, brand attitude and brand image and whether this relationship is moderated by brand familiarity, a 3 (level of consistency: consistent vs. moderately consistent vs. inconsistent) x 2 (brand familiarity: high vs. low) between-subjects online experiment was conducted. The participants were randomly assigned to one of six conditions.
Participants were recruited via a convenience sampling technique. This happened via Facebook, with a link to the Qualtrics questionnaire. Furthermore, the link has also been distributed to some classes on two foreign universities. There were
H1 + H1a Customer-based brand
equity (CBBE) Brand attitude Brand image Brand consistency Brand familiarity H4 + H4a H6 + H6a H5 + H5a
no conditions which the participants had to meet. Before the start of the experiment, participants had to agree to an informed consent text. The participants were also provided with contact details of the supervising research organization ASCoR (Amsterdam School of Communication Research) in case of any complaints. The experiment did not include a coverstory nor any information regarding the real reasons behind it. At the end, a short debriefing with the actual aim of the research was added, to inform the participants.
They were then asked to fill in a questionnaire which started with questions about their familiarity with either Nike or Under Armour. The questionnaire started with these questions before the randomization took place, so viewing the stimulus material did not influence the brand familiarity of the participants. The participants were then asked to answer some questions about any brand associations they had with either Nike or Under Armour. Again, this was asked upfront, to prevent the stimulus material from having any influence on the participants’ answers. After those questions the participants were randomly assigned to one of six conditions. The participants in each condition saw two brand advertisements related to either Nike or Under Armour (see manipulations). Subsequently, questions about the brand that was shown in the brand messages were asked. One question about the participant’s attitude toward the previous two brand advertisements was asked, in order to be able to control for this variable in the analyses. The total questionnaire took about five to ten minutes to complete.
Participants
313 Participants took part in the experiment, 42 of those were excluded in the analysis based on missing data and an outlier analysis. The average age of the participants was 28.48 years old (SD = 10.79). More females have taken part, with
63.8%. Dutch was the most common nationality (74.5%) and the most participants indicated university to be their level of education (65.7%).
Stimulus material
The six conditions in this experiment stem from a manipulation in the level of consistency of the brand advertisements that were shown. The participants either saw two brand advertisements which were highly consistent, moderately consistent or inconsistent. Per condition, two brand advertisements were shown, because
(in)consistency can only exist between two or more messages. Besides that, inconsistency is only possible in the presence of consistency. Moreover, it is
important that the participants already have associations and evaluations of the brands used in this experiment, because they are a significant part of the dependent variables. The level of consistency was based on brand features used (brand logo, brand slogan and person featuring in the advertisements) and the brand associations to which the advertisements adhere. The two inconsistent brand advertisements did not relate to each other and communicated very different brand associations. The two moderately consistent brand advertisements were more similiar, but differed in the amount of brand features used. One of the two had nearly no brand signs. The two consistent brand advertisements had a high similarity and were taken from the same campaign. The communicated brand associations were based on the mission (raison d’être) and vision (where does the brand want to be in the foreseeable future) of both brands.
For the purpose of this experiment it was decided to use existing brands as opposed to earlier research (Delgado-Ballester, Navarro & Sicilia, 2012; Sjödin & Törn, 2006), for the sake of comparability of the two brands and the necessity of the existence of an associative network for the two brands. The associative network is one of the key elements of the dependent variable customer-based brand equity (Aaker,
1996; Keller, 1993). The two brands that were picked after a pretesting of brand familiarity in the sports brands industry were Nike and Under Armour. Nike was meant to be the very familiar sports brand, which is indicated by the percentage of Americans that have personal experiences with this brand: 88% (Yoo & Donthu, 2001). Under Armour was meant to be the relatively unfamiliar sports brand, because the brand declares itself to be relatively new to the European market (although it is the number two sports brand in the United States).
For example, Nike wants to be the sports brand for the day-to-day sports activities for all people. This is resembled in the use of unknown ‘ordinary’ (groups of) people in their campaigns, their slogan (‘just do it’ and ‘find your greatness’) and/or the use of inspiring texts, which might make it easier for people to identify themselves with the person(s) in the brand’s advertisements and the brand itself. This means that one of the inconsistent two advertisements of Nike did not contain any of this (no people, no inspiring texts and no slogan). In fact, it was a user-generated advertisement, relating to different brand associations. The two moderately consistent brand and the two consistent brand advertisement contained these examples, but varied in the amount.
It is different for Under Armour. It is a sports brand for professional athletes, providing them with state-of-the-art sportswear and accessoires. This is still their core business and it is reflected in the use of professional athletes as celebrity endorsers in their brand communications. Under Armour itself identifies innovativeness, passion, endurance as its core brand associations, among others. All of these are usually featured in their campaigns, for example through the brand’s slogan (“I will (what I want)”). Participants in the two consistent brand advertising conditions saw two advertisements that were highly similar regarding the abovementioned brand features
and brand associations. Participants in the two moderately consistent brand
advertising conditions saw two advertisements that differed in the use of these brand features and brand associations. Generally, one of the two advertisements contained less brand features (different or no slogan) and used different brand associations (other than the main associations identified by the brands themselves). They were taken from different campaigns targeted at different groups of customers. Participants in the remaining two inconsistent brand advertising conditions saw one brand
advertisement created by the brand and an user generated advertisement, which was obviously not created by the brand itself. The user generated advertisement did not relate to any of the brand associations identified by the brands themselves nor did it include any brand features other than the brand logo. This was meant to show the difference between desired brand associations (visible in the advertisements created by the brands) and realized brand associations (reflected in the user generated advertisement).
Furthermore, each level of consistency is measured in two conditions: one condition for very familiar brands (Nike) and one condition for very unfamiliar brands (Under Armour). This means that the brand familiarity is manipulated as well. Participants saw either the very familiar brand (Nike) or the relatively unfamiliar brand (Under Armour) in the two brand advertisements. These advertisements were shown at the same time and at the same pace, in order to give the participants the opportunity to look at both ads closely. It might have been interfering to show the advertisements one by one, because then superficial elements like coloring or lay-out might have the participants think that the advertisements are consistent when they are not. Facilitating a closer look in showing both advertisements at the same time was meant to prevent this from happening.
Pretesting stimulus material
In order to find out whether the stimulus material used in this research is interpreted in the way it was meant to be and to be certain that the manipulations in the experiment work successfully, a small pretest was carried out. This pretest was completed by 30 participants, who were randomly assigned to one of six conditions. The average age of the participants was 27.53 years old (SD = 8.74) and 53,3% was male. Almost all participants had the Dutch nationality (83,3%) and most participants went to or are still enrolled in university (60%). It contained a question about the perceived level of consistency of the two advertisements: “what do you think about the consistency of both ads?“ (Törn, 2009; Speed & Thompson, 2000; Gwinner & Eaton, 1999; Kent & Allen, 1994). It also contained three questions for each brand (Nike and Under Armour) about the familiarity with the brand, based on a measure of brand familiarity (Kent & Allen, 1994) (see operationalization of brand familiarity). This measure turned out to be sufficiently reliable (Cronbach’s α = 0.83; EV = 2.27, R² = .76). For Nike, the average score was 5.51 (SD = 0.94) and for Under Armour, the average score was 2.51 (SD = 1.21), on a seven-point Likert scale. Lastly, four questions were asked about the control variables: the participant’s age, gender, nationality and level of education. All questions, which will be explained in more detail in the operationalization section, were based on a seven-point Likert scale, except for the items relating to the control variables. It took about one minute to complete the pretest.
For the level of consistency, the result of an analysis of variance (ANOVA) showed that the three conditions differed significantly in terms of perceived
= 3.60, SD = 1.17; M consistent = 5.40, SD = 2.22). Consequently, level of consistency
was successfully manipulated.
It also turned out that brand familiarity differed significantly for the two brands: t (29) = 9.71, p < .001, with a 95% CI (2.37, 3.63). Nike turned out to be very familiar (M = 5.51, SD = 0.94) and Under Armour turned out to be relatively
unfamiliar (M = 2.21, SD = 1.21). Therefore those two brands were used as the very familiar and the very unfamiliar brand respectively. It means that brand familiarity was successfully manipulated as well.
Operationalization
Manipulation checks. The manipulation check for level of consistency consists of only one question: “what do you think about the consistency of both ads?” (Törn, 2009; Kent & Allen, 1994). The participants could indicate the perceived level of consistency on a seven-point Likert scale, with answer possibilities ranging from “1 (very inconsistent)” to “7 (very consistent)”. The manipulation check for brand familiarity is based on a measure of brand familiarity (Kent & Allen, 1994).
Customer-based brand equity. In the present research, the scale of customer-based brand equity of Yoo and Donthu (2001) is used, since this is one of the most widely adopted customer-based brand equity measurements and it measures all elements identified by Keller (1993) and Aaker (1996). Customer-based brand equity consists of four key elements: brand awareness, brand loyalty, brand
associations and perceived brand quality (Pappu, Quester & Cooksey, 2005; Yoo & Donthu, 2001; Aaker, 1996; Keller, 1993; Aaker, 1991).
The scale of customer-based brand equity (Yoo & Donthu, 2001; Guzmán & Samih-Baalbaki, 2012) contained ten items, relating to the four key elements. Brand awareness was made of one item, with the following question: “How well do you
know the brand?”, with answer possibilities ranging from “1 (Totally not
knowledgeable)” to “7 (Totally knowledgeable)”, based on a seven-point Likert scale. Brand loyalty consisted of three items, which were written down as statements on which the participants were asked to rate themselves. The three statements were “I consider myself to be loyal to this brand”, “This brand would be my first choice” and “I will not buy other brands if this brand is available”. Again, the answer possibilities were based on a semantic differential seven-point Likert scale, ranging from “1 (totally disagree)” to “7 (totally agree)”. Brand associations were measured with three items, again formulated as statements on which the participants were asked to rate themselves. “Some characteristics of the brand come to my mind quickly”, “I can quickly recall the symbol or logo of the brand” and “I have difficulty in imagining the brand in my mind” were the used items. Again, the answer possibilities were based on a semantic differential seven-point Likert scale, ranging from “1 (totally disagree)” to “7 (totally agree)”. Like the measure for brand familiarity, these three items were asked for each brand and before the randomization and the possibility to view the stimulus material. For perceived brand quality also three questions were included. This time the. participants were asked to indicate their personal opinion about three different evaluations of the brand they saw in the two brand advertisements: “very low (very high) quality”, “very low (very high) functionality” and “very bad (very good) performance”. The answer possibilities were again based on a seven-point Likert scale, with lower scores indicating a negative opinion and higher scores indicating a positive opinion about the brand’s perceived quality.
The scale of customer-based brand equity of Yoo and Donthu (2001) turned out to be sufficiently reliable (Cronbach’s α = .74), with an average score of 4.37 (SD = .99). A principal component analysis (PCA) was carried out (EV = 2.02, R² = .67).
For this scale, brand awareness was excluded, because that improved the reliability of the scale significantly. When the closely linked concepts brand attitude and brand image would be added to the scale of customer-based brand equity (Yoo & Donthu, 2001), the reliability of the scale improves (Cronbach’s α = 0.82). A principal
component analysis (PCA) was carried out (EV = 2.63, R² = .66) for this scale as well. Also the average score improves (M = 4.56, SD = .96). Again, brand awareness was excluded, in order to improve the reliability of the scale. Since there is not (yet) enough empirical backing to start using the scale containing brand attitude and brand image, the original scale will be the one used in this research.
Brand attitude. Since this can be seen as a concept separate from customer-based brand equity as well (Faircloth, Capella & Alford, 2001) a scale of brand attitude is derived from the research of Matthes, Schemer and Wirth (2007). This scale consists of four items and the participants could answer on a seven-point Likert scale for each of these items. The seven-point Likert scale indicated a more negative personal brand attitude when the scores were lower and a more positive brand attitude when the scores were higher, with 4 being “neutral” or “I don’t know”. The scale’s items were “very unattractive vs. very attractive”, “very bad vs. very good”, “very unpleasant vs. very pleasant” and “very negative vs. very positive”. A principal component analysis (PCA) was carried out (EV = 3.55, R² = .89). The scale of brand attitude (Matthes, Schemer & Wirth, 2007) turned out to be very reliable (Cronbach’s α = .96). The average score on this scale was 5.12 (SD = 1.16).
Brand image. Brand image is measured with six items, which added up to the scale of brand image, developed by Martínez and De Chernatony (2004). A separate scale of brand image has been used, because it might not turn out as being a part of customer-based brand equity (Faircloth, Capella & Alford, 2001). The items
for the scale of brand image (Martínez & De Chernatony, 2004) had answer possibilities on a seven-point Likert scale as well. The answers ranged from “1 (totally disagree)” to “7 (totally agree)”. The items were written down as statements and the participants had to indicate their personal opinion about these statements, which all related to the brand shown in the stimulus material. The statements were as follows: “This brand provides good value for money”, “There’s a reason to buy the brand instead of others”, “The brand has personality”, “The brand is interesting”, “I have a clear impression of the type of people that buy this brand” and “The brand is different from competing brands”. A principal component analysis (PCA) was carried out for this scale (EV = 3.49, R² = .58) (Martínez & De Chernatony, 2004), which had an average score of 4.66 (SD = 1.03).. The scale turned out to be very reliable
(Cronbach’s α = .85).
Brand familiarity. Brand familiarity’s scale is derived from the research of Kent and Allen (1994). This scale has been widely adopted and has been used in many research articles concerning brand familiarity (i.e., Delgado-Ballester, Navarro & Sicilia, 2012; Simonin & Ruth, 1998). It consists of three items, which all together measure brand familiarity. Sometimes the same scale is used to measure brand awareness, because that concept is very close to brand familiarity and can be
measured through asking almost the same set of questions. Of course, this scale had also answer possibilities, based on a point Likert scale. This time the seven-point Likert scale indicated a lower brand familiarity when the scores were lower and a higher brand familiarity when the scores were higher, with 4 being “neutral” or “I don’t know”. This interpretation is comparable to the one for the scale of brand attitude (Matthes, Schemer & Wirth, 2007). The scale’s items were “very unfamiliar vs. very familiar”, “very inexperienced vs. very experienced” and “totally not
knowledgeable vs. totally knowledgeable”. The scale of brand familiarity (Kent & Allen, 1994) was very reliable for both Nike (Cronbach’s α = .84) and Under Armour (Cronbach’s α = .96). A principal component analysis (PCA) was carried out: the average score on the brand familiarity scale for Nike (EV = 2.32, R² = .77) was 5.50 (SD = 1.13) and the average score on the brand familiarity scale for Under Armour (EV = 2.80, R² = .93) was 1.58 (SD = 1.31).
Attitude towards the advertisements. In scientific research, the same scales are used for brand attitude and attitude towards the advertisements. Therefore the scale of brand attitude of Matthes, Schemer and Wirth (2007) was used, with the same four items and the same answer possibilities, again on a seven-point Likert scale. The only difference is that the items do not relate to the brand, but to the overall attitude towards the two brand advertisements the participants got to see in each of the six conditions. This concept is not included in the actual research, but it is added to the questionnaire in order to be able to control for this concept when the average score of the participants is too high. Liking the advertisements (too) much might influence the results. The scale of attitude towards the advertisement (Matthes, Schemer & Wirth, 2007) was very reliable (Cronbach’s α = .95). A principal component analysis (PCA) was carried out (EV = 3.46, R² = .87). The average score was 5.04 (SD = 1.27).
Control variables. In the present research, the participants were asked to fill in four questions relating to the control variables age, gender, nationality and level of education. Through collecting these data, it is possible to control for any of these variables if necessary and it also enables a more precise randomization, which in turn creates six more or less equal research populations in each of the six conditions. Participants were asked to write down their age (in years) in numbers and to name their gender, which is created as a dichotomous variable. Furthermore, the
participants had to write down their nationality and pick their level of education from a list of options (either Dutch or English/International), which could be either a completed study or the one they are currently enrolled in.
Results Manipulation checks
In order to be able to check whether the manipulations that were used in this research have been successful, a question has been asked concerning the perceived level of consistency. The question was: “what do you think about the consistency of both ads?”. An analysis of variance (ANOVA) showed that the perceived level of consistency differed significantly between the three conditions: F (2, 268) = 53,41, p < .001 (M inconsistent = 2.89, SD = 1.32; M moderately consistent = 4.16, SD = 1.34; M consistent =
5.05, SD = 1.57). Because this was as intended, the manipulation for level of consistency was successful.
Randomization checks
In order to be able to exclude any other influences on the expected effects than the influence of the independent variable, some control variables and possible influences were included in the questionnaire. This way it is possible to check if these variables influence the effects or whether it is necessary to control for them in the analyses. Furthermore, it allows for a check of the randomization process, which should have created six similar conditions. The three levels of consistency did not significantly differ on the control variables, the moderator variable brand familiarity and attitude towards the advertisements. Regarding age, there was no significant difference: F (2, 268) = .22, p = .805 (M inconsistent = 28.52, SD = 11.04; M moderately consistent
significant difference as well: F (2, 268) = .81, p = .448 (M inconsistent = 1.40, SD = .49;
M moderately consistent = 1.31, SD = .47; M consistent = 1.37, SD = .49). The level of education
did not yield a significant difference as well: F (2, 268) = 4.58, p = .011 (M inconsistent =
5.54, SD = .74; M moderately consistent = 5.87, SD = 1.05; M consistent = 5.51, SD = .85). It
means that there are no influences on the expected effects of the control variables and that here is no need to control for them in the analyses. Furthermore, it shows that the randomization process has worked successfully and that the six experimental
conditions are equal and therefore comparable.
For the moderator variable brand familiarity, the mean scores on each level of consistency were checked in order to make sure that brand familiarity was equally divided. This was the case, because there was no significant difference in brand familiarity: F (2, 268) = .22, p = .800 (M inconsistent = 3.39, SD = 2.25; M moderately consistent =
3.36, SD = 2.50; M consistent = 3.57, SD = 2.27). The attitude towards the advertisements
was measured, because high (low) advertisement liking might result in higher (lower) brand attitudes, which might influence the results. It was relatively high (M = 5.04, SD = 1.27), but it did not differ significantly for each of the levels of consistency: F (2, 268) = 15.63, p = .105 (M inconsistent = 4.81, SD = 1.15; M moderately consistent = 5.10, SD =
1.26; M consistent = 5.50, SD = 1.28). It means that it was equally divided over all
conditions as well. Both brand familiarity and attitude toward the advertisements were equally divided and had no influence on the expected effects. Therefore, again, it was a successful randomization and it was not necessary to control for these variables. Testing hypotheses
For this research, six hypotheses were drawn, of which three had to do with main effects of brand consistency on customer-based brand equity, brand attitude and brand image. The other three hypotheses proposed a moderating influence of brand
familiarity on the relationship between brand consistency and customer-based brand equity, brand attitude and brand image. An overview of the (dis)confirmation of the hypotheses are presented below the discussion of the results, in table 1 and 2. For more information on the results, see appendix D, tables 1, 2, 3 and 4 and figure 2. The effect of brand consistency on customer-based brand equity
The first hypothesis (H1) stated that a significant effect of brand consistency on customer-based brand equity exists. Furthermore, an additional hypothesis (H1a) stated that there would be a significant difference between the inconsistent and consistent levels and no significant difference for both levels with the moderately consistent level. The effect of brand consistency on customer-based brand equity did turn out to be significant (F (2, 268) = 13.46, p < .001) in an analysis of variance (ANOVA). A post-hoc Scheffé test showed that there is a significant difference in effect on customer-based brand equity between the inconsistent (M = 3.97, SD = .86) and consistent conditions (M = 4.68, SD = .92) and between the inconsistent and moderately consistent conditions (M = 4.46, SD = 1.04). There is no significant difference in effect on customer-based brand equity between the moderately consistent and consistent conditions. This means that the effect on customer-based brand equity is higher for the moderately consistent and consistent conditions than it is for the inconsistent conditions. Therefore the first hypothesis (H1) is confirmed, as well as the additional hypothesis (H1a).
The effect of brand consistency on brand attitude
The second hypothesis (H2) stated that brand consistency has a significant effect on brand attitude. Subsequently, an additional hypothesis (H2a) stated that there would be a significant difference between the inconsistent and consistent levels and no significant difference for both levels with the moderately consistent level.
The effect of brand consistency on brand attitude was measured with an analysis of variance (ANOVA), which showed a significant effect (F (2, 268) = 7.13, p = .001). A post-hoc Scheffé test showed that there is a significant difference in effect on brand attitude between the inconsistent (M = 4.77, SD = .93) and consistent conditions (M = 5.39, SD = 1.24). There is no significant difference in effect between the inconsistent and moderately consistent conditions (M = 5.20, SD = 1.20) nor between the
moderately consistent and consistent conditions. This means that the effect on brand attitude is higher for the consistent conditions than it is for the inconsistent conditions and that the effect of the moderately consistent conditions is not significantly different from either one of those. Therefore the second hypothesis (H2) is confirmed. The additional hypothesis (H2a) is partially confirmed.
The effect of brand consistency on brand image
The third hypothesis (H3) stated that brand consistency has a significant effect on brand image. Subsequently, an additional hypothesis (H3a) stated that there would be a significant difference between the inconsistent and consistent levels and no significant difference for both levels with the moderately consistent level. The results of an analysis of variance (ANOVA) turned out to be not significant (F (2, 268) = 3.28, p = .039), which means that there is no significant effect of brand consistency on brand image. Furthermore, there are no significant differences between the levels of consistency (Minconsistent = 4.43, SD = 1.08; Mmoderately consistent =
4.79, SD = .95; Mconsistent = 4.75, SD = 1.04), so all levels of consistency have a
non-significant effect on brand image. Because these results are contrary to what is posited in the third (H3) and the additional hypotheses (H3a), both are disconfirmed.
The fourth hypothesis (H4) stated that the brand familiarity would moderate the effect of brand consistency on customer-based brand equity, in the sense that very familiar brands would have higher customer-based brand equity than very unfamiliar brands. An additional hypothesis (H4a) stated that there would be a significant
difference between the inconsistent and consistent levels and no significant difference for both levels with the moderately consistent level, concerning the moderating influence of brand familiarity. A multiple analysis of variances (MANOVA)
(including customer-based brand equity, brand attitude and brand image as dependent variables) showed that this is not the case, considering the non-significant result (F (2, 265) = 4.31, p = .014). It turned out that there only was a significant main effect of brand consistency on customer-based brand equity (F (2, 265) = 14.28, p < .001). The main effect of brand familiarity on customer-based brand equity was non-significant as well (F (1, 265) = 6.12, p = .014). This means that there is no significant
moderating effect of brand familiarity on the effect of brand consistency on customer-based brand equity. Splitting the file for the different levels of brand consistency did not lead to any significant results as well. For all brands, the effect of any level of brand consistency on customer-based brand equity would not be significantly different. Therefore, the fourth hypothesis (H4) and the additional hypothesis (H4a) are both disconfirmed.
The moderating influence of brand familiarity: brand attitude
The fifth hypothesis (H5) stated that the brand familiarity would moderate the effect of brand consistency on brand attitude, in the sense that very familiar brands would have higher brand attitude scores than very unfamiliar brands. An additional hypothesis (H5a) stated that there would be a significant difference between the inconsistent and consistent levels and no significant difference for both levels with the
moderately consistent level, concerning the moderating influence of brand familiarity. The result of a multiple analysis of variances (MANOVA) (including brand attitude, customer-based brand equity and brand image as dependent variables) turned out to be non-significant (F (2, 265) = 2.30, p = .102). The main effect of brand consistency on brand attitude turned out to be significant (F (2, 265) = 7.31, p = .001).The main effect of brand familiarity on brand attitude turned out to be non-significant (F (1, 265) = 1.32, p = .252). The results do not show a significant moderating effect of brand familiarity on the effect of brand consistency on brand attitude. Splitting the file for the different levels of brand consistency did not lead to any significant results as well. For the effect of any level of brand consistency on brand attitude, it does not matter whether the brand is very familiar or very unfamiliar. The fifth hypothesis (H5) and the additional hypothesis (H5a), therefore, are both disconfirmed. The moderating influence of brand familiarity: brand image
The sixth hypothesis (H6) stated that the brand familiarity would moderate the effect of brand consistency on brand image, in the sense that very familiar brands would have higher brand image scores than very unfamiliar brands. An additional hypothesis (H6a) stated that there would be a significant difference between the inconsistent and consistent levels and no significant difference for both levels with the moderately consistent level, concerning the moderating influence of brand familiarity. The same multiple analysis of variances (MANOVA) (including brand image,
customer-based brand equity and brand attitude as dependent variables) showed that there is no significant result (F (2, 265) = 5.70, p = .004). Once again there is no main effect of brand consistency on brand image (F (2, 265) = 3.58, p = .029). Brand familiarity had no main effect on brand image as well (F (2, 265) = 2.07, p = .151). This means that there is no significant effect of brand consistency on brand image, so
there can not be a significant effect of brand familiarity on that non-existing effect. Splitting the file for the different levels of brand consistency did not lead to any significant results as well. For brand image, neither the level of brand consistency nor the brand familiarity matters. This means that the sixth hypothesis (H6), which stated the opposite of the results, and the additional hypothesis (H6a) are disconfirmed.
Table 1
An overview of the (dis)confirmation of the H1, H1a, H2, H2a, H3 and H3a.
Hypothesis p Confirmed/disconfirmed H1
Brand consistency has a significant effect on customer-based brand equity.
.000* Confirmed
H1a
This effect is significantly different between the inconsistent and consistent levels, but not between the moderately consistent level and the other two levels.
.000* .001* .317
Confirmed
H2
Brand consistency has a significant effect on brand attitude.
.001* Confirmed
H2a This effect is significantly different between the inconsistent and consistent levels, but not between the moderately consistent level and the other two levels.
.039 .001*
.540
Partially confirmed
H3
Brand consistency has a
significant effect on brand image. .039 Disconfirmed
H3a This effect is significantly different between the inconsistent and consistent levels, but not between the moderately consistent level and the other two levels.
.069 .113 .972
Disconfirmed
Note. * = p ≤ .001 The p-values for the additional hypotheses are based on the results of a post-hoc Scheffé test.
An overview of the (dis)confirmation of the H4, H4a, H5, H5a, H6 and H6a.
Hypothesis p Confirmed/disconfirmed H4 For very familiar brands, the effect of brand consistency on
customer-based brand equity will be higher, compared to very unfamiliar brands.
.014 Disconfirmed
H4a
This moderated effect is
significantly different between the inconsistent and consistent levels, but not between the moderately consistent level and the other levels.
.101 Disconfirmed
H5
For very familiar brands, the effect of brand consistency on brand attitude will be higher, compared to very unfamiliar brands.
.102 Disconfirmed
H5a This moderated effect is significantly different between the inconsistent and consistent levels, but not between the moderately consistent level and the other levels.
.031 Disconfirmed
H6
For very familiar brands, the effect of brand consistency on customer-based brand equity will be higher, compared to very unfamiliar brands.
.004 Disconfirmed
H6a This moderated effect is significantly different between the inconsistent and consistent levels, but not between the moderately consistent level and the other levels.
.098 Disconfirmed
Note. * = p ≤ .001. The p-values for the additional hypotheses are based on the results of a post-hoc Scheffé test and split files.
Conclusion
The main focus of this research was the relationship between brand
consistency and customer-based brand equity, brand attitude and brand image. It was expected that there would be a significant effect of brand consistency on all three
