0
The role of brand equity and product category knowledge in the innovation adoption process
extending the UTAUT2 model on technological innovations
Eva Berends
MSc. Business Development S1681796
1
Preface
This master thesis is not only the end of the master program Business Development; it is also the end of six years of studying in Groningen. At the time I found an interesting thesis theme that captivated me, I also decided that I wanted to do an internship. An internship that had nothing to do with this thesis, actually. Many people warned me about the possible consequences of writing a thesis next to working full-time. Despite all this, -stubborn as I am- I decided to do it. I moved from Groningen to Rotterdam in March, with six intensive and intense months at Unilever’s Consumer & Market Insight department ahead of me. I learned more than I hoped for and got the freedom to initiate and carry out several research projects with the aim of ‘bringing the marketers closer to the consumer’. I kept working steadily on this thesis at night and during the weekends. In the summer I was told that Unilever would pay for my data collection via a professional panel agency. During one of my projects at Unilever, I worked closely together with CG Selecties, a panel agency with over 100.000 respondents in their Dutch panel. I am very thankful for the fact that Unilever paid for this, as this enabled me to gather the data (N=600) within just one day via CG Selecties.
After that, the real fun part was about to begin: the analysis. However, completely contrary to what I initially expected and planned, I found a job while still working for Unilever. In September I started as a qualitative research consultant at InSites Consulting; a market research agency, which makes me carrying out research projects for large FMCG clients again. I can again talk you through the reactions I got regarding this thesis and graduation in general. Of course I wanted to finish it. I would be lying if I said it was an easy process. The more I got involved in my job, the harder it was to catch up with writing. As a consequence, the final part of this master thesis is the result of late night office hours, weekend sessions and a complete lack of a social life. I am pretty sure that if I wouldn’t have made the decision to move to Rotterdam and become an intern at Unilever, my thesis would have been finished months ago. However, I am even more sure of the fact that in that case, I wouldn’t have grown so much on a professional as well as on a personal level, and that I wouldn’t be having a job that I love, actually.
I would like to thank Unilever and CG Selecties for enabling me to gather the data needed for this thesis within such a short time frame, InSites Consulting for hiring me (and thus believing in me) months before finishing my thesis, the people around me (especially Thijs and my parents) for their continuous support and motivation and of course, professor Jo van Engelen for guiding me through this process and enabling me to finish it at my own pace. I would like to thank him for his valuable feedback, flexibility, his pleasant personal approach and the freedom he gave me in writing this thesis. I won’t forget what he said about publishing and will certainly keep in touch about this. Thanks all!
Eva Berends
2
Abstract
This research aims to extend a well-known innovation adoption model with the additional variable brand equity, as prior research stressed the importance of brand equity in innovation adoption. It is somewhat odd that despite the important impact brand equity seems to have on the evaluation and adoption of either incremental or radical innovations, none of the existing technological innovation adoption models developed so far take it into consideration as a variable. The most recent extension of the UTAUT2 model does take three variables into account that have not so much to do with the technology or product itself, which are habit, price value and hedonic motivation. These relate more to the person adopting the innovation rather than to the innovation itself. However, the brand the radically or incrementally innovative product carries is not taken into account. This is where this research comes in, as it aims to extend the UTAUT2 model with the variable brand equity. Also, two moderators were tested as prior research found a moderating effect of both product category knowledge and the degree of innovation (radical versus incremental) on the relationship between brand equity and adoption.
In order to test the role of brand equity, product category knowledge and the degree of innovation in the UTAUT2 model, a survey was developed consisting of validated scales from prior research. In order to allow for more diverse responses and thus an adjusted model with more predictability, four scenarios were developed. The four scenarios were based on two axes: a high versus a low equity brand, and a radical versus an incremental innovation. The scenarios were justified after a pre-test among 10 respondents. Afterwards, 600 nationally representative respondents were selected out of a professional panel agency’s panel and filled out the survey, leading to 150 respondents per scenario.
After carrying out several statistical tests and analyses such as t-tests, regression analyses and a factor analysis, conclusions were drawn. First of all, after adding brand equity, the predictability of the UTAUT2 model significantly increased. It turns out among the other seven variables in the existing and initial UTAUT2 model, brand equity has the largest share towards behavioral buying intention. An additional factor analysis revealed that the initial UTAUT2 model might needs more adjustments than merely adding brand equity. This was already somewhat clear after looking at the correlations between brand equity and the initial variables. Redesigning and validating the model was beyond the scope of this research, as this research merely sheds a first light on the necessity of adding brand equity to the UTAUT2 model. The model is more management friendly, as companies are better able to influence and thus steer brand equity than the initial variables in the model. This opens doors for further research concerning the restructuring of the UTAUT2 model to eventually make it even more predictable and in the same time more convenient and relevant for companies dealing with either radical or incremental innovations.
3
Table of content
1. Introduction 4
2. Research question 5
3. Literature review 6
3.1 Innovation diffusion and adoption 6
3.2 A review of existing adoption models 6
3.3 Research on innovation and brands 8
3.4 Brand equity 8
3.5 The degree of innovativeness 9
3.6 Product category knowledge 10
4. Methodology 11
4.1 Research design and stimuli 11
4.2 Measurement 12
5. Results 14
5.1 The initial variables in the UTAUT2 model 14
5.2 The added variables in this research 14
5.3 Measuring the effect of adding brand equity 16
5.4 Changing the UTAUT2 model 19
5.5 Measuring the moderating effects of the degree of innovativeness 19 5.6 Measuring the moderating effects of product category knowledge 19
6. Discussion 21
6.1 Conclusion 21
6.2 Practical implications 22
6.3 Theoretical implications, limitations and future research 22
7. References 23
8. Appendices 26
8.1 Scenarios 26
8.2 Questionnaire 27
4
1. Introduction
Imagine yourself walking around in an electronics store. Your television recently broke down and you consequently need to buy a new one. The technical knowledge you possess is limited to the basics. You pass by models that are obviously newer and fancier than the one you used to have. You notice familiar brands and brands that you have never heard of before. Which television would catch your attention and which one would you probably end up buying, assuming all specifications seem equal to you? Would it be the one of the rather familiar brand or one of the televisions from an unknown brand? What if your knowledge of televisions in general would be extremely high and the slightly different specifications thus make sense to you? Would your decision differ? In other words, to what extent would the more familiar or renowned brand be more favorable than the unknown brand in both cases? Or would you not mind at all, as it does not involve the purchase of an either exciting or revolutionary product?
Now imagine yourself walking around in the same electronics store, still trying to find the perfect new television. You suddenly pass by a product range you have never seen before and therefore it catches your attention. This is something completely new and revolutionary and would save you hours a week cleaning your house! You notice three brands that all offer the particular product; one of your favorite electronic brands that is renowned for producing televisions instead of cleaning appliances, another electronics brand you heard of and one completely unfamiliar brand. You are interested in the product and ask for explanation after which you consider buying the product, as it seems extremely useful to you with regard to cleaning your house. But, which one of the three brands would you probably buy? You might associate your favorite brand with their core business, which has not much to do with appliances. Or do you think that because of your favorable associations, the product should be good and you should thus buy it? What about the unknown brand? Would your decision be different if you possessed more knowledge of the product category? These are all interesting questions and especially relevant for companies that are willing to innovate, which seems to be inevitable in today’s fast changing consumption based society.
There is a general consensus among researchers about the importance of innovation for all companies (e.g.Prahalad and Ramaswamy, 2003; Gourville, 2006). What determines whether an either radical or incremental innovation eventually turns out to be successful is the adoption by new or existing consumers. Many technological innovation adoption models have been developed and extended, first of all merely focusing on technological innovations in the workplace but later on also to study the adoption process of technological consumer products (e.g. Rogers, 1962; Fishbein and Ajzen, 1975; Ajzen; 1985, Davis et al., 1989; Venkatesh et al., 2003). However, all variables in the TRA, TPB, TAM and UTAUT models are related to the technical aspects of the specific technology, product or system to be adopted. Only the UTAUT2 model takes into account three additional variables in a more consumer related context. The aim of this research is to find out whether variables relating more to the particular brand play a role in the adoption process as well, as brand equity was found to have an impact on the way people perceive and evaluate certain products (Erdem and Swait, 1998). Furthermore, Gammoh et al. (2011) and Corkindale and Belder (2009) found evidence concerning the importance of brand equity and reputation in the process of innovation evaluation and adoption. But is it really better to have a strong brand in order to have your either radical or incremental innovation adopted easily? Maybe this merely holds for specific types of innovations or when a certain level of product category knowledge exists. This is something that Gammoh et al. (2011) to a certain extent investigated in their research as well, but not in relation to technological innovation adoption and the UTAUT2 model. These questions will therefore be investigated in this research.
5
2. Research question
The objective of this research is thus to find out whether and to what extent brand equity as an additional independent variable plays a role towards behavioral buying intention, next to the already existing variables in the UTAUT2 model, which will be taken as a starting point. This because brand equity was found to play a role in the process of innovation evaluation and adoption (Erdem and Swait, 1998; Gammoh et al., 2011; Corkindale and Belder, 2009).
In the UTAUT2 model, behavioral intention is referred to as the step before technology use, thus the intention to do so (Venkatesh et al., 2012). In this research, behavioral intention is defined as the eventual buying intention of the innovative product. Furthermore, it will be investigated whether product category knowledge and the type of innovation moderate the relationship between brand equity and behavioral intention as it was found in literature that these variables do influence the relationship between brand equity and consumer’s evaluations (Gammoh et al., 2011). The main research question therefore is:
To what extent are brand equity and its underlying constructs a significant predictor of behavioral intention (buying intention) in the initial UTAUT2 model and how do the degree of innovation and product category knowledge in turn influence this relationship?
As mentioned earlier, brand equity was found to play an important role towards innovation evaluation and adoption. It is thus expected that in the UTAUT2 model, brand equity will play a significant positive role towards behavioral buying intention as well. As will be elaborated upon later, the degree of innovation is expected to negatively moderate the relationship between brand equity and behavioral buying intention, as this was to a certain extent investigated by Gammoh et al. (2011) as well. The same holds for product category knowledge, which is also expected to weaken the relationship between brand equity and behavioral buying intention according to Gammoh et al. (2011).
This thesis has been written in a rather concise manner, as it is the basis and framework for an article in a scientific journal.
6
3. Literature review
Product innovation and service innovation have been a hot topic over the last decade. A lot has been written about the importance and necessity of product innovation and on a more recent basis, service innovation. Prahalad and Ramaswamy (2003) argue that the need to innovate is greater than ever and that value creation through profitable growth can come only from innovation. In today’s hypercompetitive marketplace, companies that successfully introduce new products are more likely to flourish than those that do not (Gourville, 2006). Jeffrey Immelt, chairmain and CEO of General Electric Co., has talked about the ‘Innovation Imperative’, a belief that innovation is central to the success of a company and the only reason to invest in its future. Companies in the 21st century are faced with two incommensurable challenges. On the one hand, they need to be constantly innovative and ready for change. On the other hand, they are expected to create an enduring and recognizable identity that attracts attention in a world saturated with communication (Beckmann and Jensen, 2007).
3.1 Innovation diffusion and adoption
What determines the success of any innovation attempt is the degree to which the new product or service is accepted and adopted by either new or existing customers. It is thus of vital importance to understand the process of innovation acceptance and adoption on an individual level and on the market level, referred to as respectively innovation adoption and innovation diffusion (Rogers, 1962). The question why certain innovations spread more quickly than others and why some innovations do fail is one of the major concerns in the field of innovation adoption and diffusion research today (Gottschalk and Kalmbach, 2005). A lot of research has yet been undertaken in the area of innovation acceptance, which started in the 1940s with studies on the spread of new agricultural techniques. Since then this field of academic interested has developed rapidly incorporating concepts of other social sciences as psychology and marketing (Planing and Britzelmaier, 2011). Rogers (1962) refers to innovation diffusion as the process by which an innovation is communicated through certain channels over time among the members of a social system. He developed the innovation adoption lifecycle consisting of different type of adopters: innovators, early adopters, early majority, late majority and eventually the laggards. Rogers (1983) further states that the diffusion of any innovation through channels over time depends on the innovation itself, the communication channels, time and the social system. Diffusion thus refers to innovation adoption on a larger scale. What is of importance to companies undertaking innovation efforts is how individuals adopt innovations. The five stages of the adoption process according to Rogers (1983) are respectively knowledge (awareness), persuasion (information gathering), decision (whether to try the innovation or not), implementation (actually trying the innovation) and confirmation (whether or not to continue using the innovation). The persuasion and decision stage are usually the main interest of innovation acceptance studies, although recently the consequences of innovation gained increased attention (Rogers, 2003). He furthermore states that five intrinsic characteristics of the specific innovation influence the decision of individuals whether or not to adopt an innovation: relative advantage, compatibility, simplicity versus complexity, triability and observability. These above mentioned diffusion and adoption processes however are rather general which led to the birth of many individual adoption models and extensions. Some of these models are more psychological in nature whereas others were specifically developed for technological innovations or the adoption of information systems.
3.2 A review of existing adoption models
The Theory of Reasoned Action referred to as TRA (Fishbein and Ajzen, 1975) was one of the first in its kind. In general, the model aims at predicting individual behavior by postulating that human behavior is based on the systematic use of available information through the formation of beliefs
7 (Planing and Britzelmaier, 2011). The model in its most simple form states that behavior is influenced by behavioral intention and that behavioral intention is in turn influenced by the attitude towards the behavior (an individual's positive or negative evaluation of self-performance of the particular behavior) and the subjective norm (a person's perception that most people who are important to him think he should or should not perform the behavior in question). The model has been criticized by many (e.g. Sheppard et al., 1988; Hale et al., 2003) and later on revised by Ajzen (1985) into the Theory of Planned Behavior (TPB). The main difference between the TPB and the TRA is the additional variable perceived behavioral control (an individual's perceived ease or difficulty of performing the particular behavior) next to attitude towards the behavior and the subjective norm. Perceived behavioral control influences both the behavioral intention and actual behavior and represents the beliefs of a subject that he or she is able to perform the behavior in question or that he or she has actual control over performing the behavior (Planing and Britzelmaier, 2011). Whereas the TRA and the TPB are more general in nature and concern human behavior in a broader perspective, the Technology Acceptance Model (TAM) is an information systems theory that deals with the adoption and usage of a certain (information) technology (Davis et al., 1989). The TAM is also an extension of the TRA (Fishbein and Ajzen, 1975) and states that perceived usefulness and perceived ease of use are the two factors influencing attitude towards using, which in turn influences the intention to use the technology, serving as a mediator towards the actual system usage. Also, perceive ease of use directly influences perceived usefulness. Although the TAM was developed in order to explain the determinants of computer acceptance for job-related purposes, the model is increasingly applied to behavior across a broad range of end-user computing technologies (Planing and Britzelmaier, 2011). Venkatesh and Davis (2000) further extended the TAM into the TAM2 in which they found several antecedents of perceived usefulness, being respectively the subjective norm, the image, job relevance, output quality and result demonstrability. Furthermore, experience and voluntariness were found to be moderators of the relationship between the subjective norm and perceived usefulness (experience) and between the subjective norm and intention to use (both experience and voluntariness). A TAM3 has also been developed by Venkatesh and Bala (2008) which deals with the antecedents of perceived ease of use, being respectively computer self-efficacy, perceptions of external control, computer anxiety, computer playfulness and eventually perceived enjoyment and objective usability. In an attempt to recognize the strengths and weaknesses of virtually all technology acceptance models developed so far, Venkatesh et al. (2003) incorporated Rogers Innovation Diffusion Theory, the TRA and TPB as well as the TAM and several other specialized innovation acceptance models into one unified model, which was consequently referred to as the United Theory of Acceptance and Use of Technology or UTAUT (Planing and Britzelmaier, 2011). As a result of this effort, four key constructs (performance expectancy, effort expectancy, social influence and facilitating conditions) were found to either influence behavioral intention (performance expectancy, effort expectancy and social influence) and use behavior (facilitating conditions and indirectly via behavioral intention). The relationships are moderated by gender, age, experience and voluntariness of use.
As has been mentioned above, the several adoption models described here have been criticized, revised and extended a lot. On a more recent basis for example, Chtourou and Souiden (2010) extended the TAM with the variable fun as an additional antecedent of attitude towards using the innovation and Venkatesh et al. (2012) developed the UTAUT2 model which was developed in order to study acceptance and use of technology in a consumer context. They proposed three additional variables as antecedents of behavioral intention, namely hedonic motivation, price value and habit. They defined hedonic motivation as the fun or pleasure derived from using a technology. Price value was defined as consumers’ cognitive tradeoff between the perceived benefits of the applications and
8 the monetary cost for using them (Dodds et al., 1991). Habit was defined as the extent to which people tend to perform behaviors automatically because of learning (Limayem et al., 2007). In addition to the effects on behavioral intention, habit was found to also influence use behavior directly. The relationship between hedonic motivation and behavioral intention is however moderated by age, gender and experience. The relationship between price value and behavioral intention is moderated by age and gender, the relationship between habit and behavioral intention is moderated by age, gender and experience and finally, the relationship between habit and use behavior is as well moderated by age, gender and experience. With this most recent extension of the UTAUT model, Venkatesh et al. (2012) show that variables relating more to the product and consumer mindset play an important role in the formation of behavioral intention as well. It is thus somewhat odd that there has not been much emphasis in prior research on brand- or product related variables in all of the above-mentioned adoption models and extensions. Only the technical aspects of the product are reflected in the models in terms of perceived usefulness, ease of use, effort expectancy and performance expectancy, some of them including either technical, social and personal antecedents.
3.3 Research on innovation and brands
In a broader perspective, prior research that aimed to link innovation and brand related variables mainly concerned the role innovation plays in building customer loyalty and brand equity and thus refers to innovation as being a driver of loyalty and brand equity. Brands are formed by innovation, and driven by new products, as these provide a temporary competitive advantage (Kapferer, 2001). Beverland et al. (2010) furthermore state that product innovation is vital to ongoing brand equity. Brand equity is a key marketing asset which nurtures long-term buying behavior (Davis, 2000). Thus, there seems to be a consensus about the importance of innovation efforts in the creation of strong brands. But, what if we turn this around? To what extent plays brand equity a role in the acceptance and adoption process of innovations? As has been mentioned above, Venkatesh et al. (2012) showed in their most recent extension of the UTAUT model that consumer or brand related variables do play a role in the adoption of new technological products.
3.4 Brand equity
According to Keller (2008), brand equity is the differential impact of brand knowledge on consumer response to the marketing of the brand. Brand knowledge is the source of such equity and represents the set of associations the consumers hold in memory regarding the brand’s attributes, benefits, perceived quality and attitudes. Such brand knowledge provides value to consumers by enhancing their processing of information and confidence in making purchase decisions (Keller, 2003). In addition, Erdem and Swait (1998) have shown that a strong brand name reduces perceived risk and enhances consumers’ understanding of the product’s benefits. It is thus reasonable to believe that brand equity plays a role in the way people adopt innovative products.
As mentioned earlier, brand equity is regarded as being a key marketing asset that nurtures long term buying behavior (Davis, 2000). Understanding the dimensions of brand equity and then investing to grow this intangible asset raises competitive barriers and drives brand wealth (Yoo and Donthu, 2001). Thus, for firms, growing brand equity is a key objective achieved through gaining more favorable associations and feelings among target consumers (Falkenberg, 1996). However, the literature on brand equity, although substantial, is largely fragmented and inconclusive (Christodoulides and Chernatony, 2009) and perhaps the only thing that has not been reached with regard to this concept is a conclusion (Berthon et al., 2001). The most clear distinction that has been made so far in research is the one between consumer-based (e.g. Yoo and Donthu, 2001) and firm-based brand equity (e.g. Simon and Sullivan, 1993). The latter involves the financial value brand equity creates to the business
9 whereas consumer-based brand equity is considered to be the driving force of increased market share and profitability of the brand, and is based on the market’s perceptions (Christodoulides and Chernatony, 2009). The conceptualisations of consumer-based brand equity have mainly derived from cognitive psychology and information economics with dimensions for example being brand awareness, brand associations, perceived quality and brand loyalty (Aaker, 1991; Yoo and Donthu, 2001) and brand benefit clarity, perceived brand quality, brand benefit uniqueness, brand sympathy and brand trust (Burmann et al., 2009). It is reasonable to assume that these constructs do play a role in the adoption process of innovations, as consumer’s responses in general were found to be positively affected by them (Buil et al., 2013). Furthermore, Corkindale and Belder (2009) state that there is evidence for the influence of corporate reputation and brand equity on the success of new products but this has mostly been in the context of well-established firms and brand extensions. Gammoh et al. (2011) investigated the role of incremental versus radical innovation with relation to brand equity and innovation adoption, thus not merely focusing on brand extensions. They did research on how brand equity levels influence the evaluation of continuous versus discontinuous innovations of new products and the moderating effects of consumer’s product category knowledge on this relationship. They first of all found that consumer’s evaluations of the new product (i.e. brand attitude, perceived quality and purchase intentions) are significantly higher for high equity brands compared to low equity brands. In the case of high equity brands, consumer’s evaluations do not differ between continuous versus discontinuous innovations. For low equity brands however, consumer’s evaluations are higher for discontinuous innovations compared to continuous incremental innovations (Gammoh et al., 2011). Furthermore, when consumers do not possess a lot product category knowledge, their evaluations of both continuous and discontinuous innovations do not differ for high as well as low equity brands. When consumers do possess product category knowledge though, their evaluations of discontinuous innovations are higher compared to continuous innovations in case of a low equity brand. In case of a high equity brand and product category knowledge, there is no difference in evaluations for both discontinuous and continuous innovations. Furthermore, Corkindale and Belder (2009) found that the strength of the corporate brand has a significant impact on the likelihood of adoption of new services. But, to what extent are these findings extendable to technological products in order to fit into the UTAUT2 model? The research by Gammoh et al. (2011) focused on consumer’s evaluations, but to what extent are favorable associations the same as eventually adopting the new technological product? In order to find this out, this research aims to extend the UTAUT2 model (Venkatesh et al., 2012) with the variable brand equity in addition to performance expectancy, effort expectancy, social influence, facilitating conditions, hedonic motivation, price value and habit as a predictor of behavioral intention. This model was chosen as it is applicable to a consumer context as well due to the most recent extension. The existing moderators age, experience and gender will not be taken into account as the aim of this research is merely to find out whether brand equity influences behavioral intention in the UTAU2 model. This leads to the first hypothesis:
H1 = In the existing UTAUT2 model, the additional variable brand equity will be significantly positively related to behavioral intention
3.5 The degree of innovativeness
The moderators that will be tested are drawn from Gammoh et al. (2011) and are respectively product category knowledge and the type of innovation (radical versus incremental) as these were found to impact the relationship between brand equity and consumer responses (Gammoh et al., 2011). An incremental innovation is regarded as being a product that provides new features, benefits or improvements to the existing technology in the existing market (Garcia and Calantone, 2002). A radical innovation is an innovation that embodies a new technology that results in a new market
10 infrastructure and thus makes it radically different than current products in the market. It is assumed that the level of innovativeness of new product introductions does play a moderating role in the effect brand equity has on innovation adoption, as previous research has widely documented that consumers are faced with uncertainty about the costs and the benefits of adopting and evaluating new products (Hoeffler, 2003). Discontinuous or radical innovations present consumers both with uncertainty and opportunity for benefits, whereas in the case of a continuous or incremental innovation, uncertainty and opportunity for benefits are lower (Hoeffler, 2003). It is thus interesting to find out whether the type of innovation also moderates the relationship between brand equity and innovation adoption in the UTAUT2 model. The following hypothesis could thus be stated:
H2= The degree of innovation (radical versus incremental) will moderate the relationship between brand equity and behavioral buying intention in such a way that the more radical an innovation is, the weaker the relationship between brand equity and behavioral intention
3.6 Product category knowledge
This also holds for product category knowledge, as Gammoh et al. (2009) found that in the case of low product category knowledge, the evaluations of incremental and radical innovations do not differ between high and low equity brands. However, in case of existing product category knowledge, the evaluations for radical innovations are higher than incremental innovations for low equity brands. For high equity brands, there is no difference in evaluations. It is thus important to take product category knowledge into account as well in this research on the extension of the UTAUT2 model. This then leads to the following hypothesis:
H3= Product category knowledge will moderate this in such a way that the higher the product category knowledge, the weaker the relationship between brand equity and behavioral intention
11
4. Methodology
In order to find out whether to reject or accept the hypotheses, data were collected using a survey. Four scenarios were developed in order to allow for more diverse responses. These were pre-tested first. The survey used consisted of validated scales from prior research concerning brand equity (Yoo and Donthu, 2001), product category knowledge (Gammoh et al., 2011) and the UTAUT2 model (Venkatesh et al., 2012). 600 respondents filled out the survey, leading to 150 respondents per scenario.
4.1 Research design and stimuli
A 2x2 approach (incremental versus radical innovation and high versus low equity brand) was chosen (Gammoh et al., 2011) in order to allow for more diverse responses and to correctly measure the variable degree of innovation (i.e. the degree of radicalness of the innovation). After pre-tests justified the scenario setting being valid, the respondents were randomly assigned to four experimental conditions: a high equity brand with an incremental innovation, a high equity brand with a radical innovation, a low equity brand with an incremental innovation and a low equity brand with a radical innovation. The respondents were given a scenario including a brand and an innovation and were told to imagine themselves in this situation. Only a short product introduction was given including prices, which were about the same for both brands. There was only a price difference between both types of innovations, although this was controlled by stating in all scenarios that the respondents had to pretend they had the monetary resources available to buy the products. Both the brand name and the type of innovation were manipulated in the four scenarios, as will be elaborated upon next.
The high equity brand was chosen based on a ranking using The Best Global Brands of Interbrand and is Samsung. Most people recognize Samsung and it is mainly renowned for its televisions and cellphones. The low equity brand did not appear in the ranking mentioned above and is the Finnish brand Salora. Salora produces amongst others televisions, tablets and mp3 devices. In order to find out whether Samsung was regarded as being the higher equity brand and Salora as being the lower equity brand, pre-tests were carried out amongst N=10 respondents. After explaining the brand equity construct of Yoo and Donthu (2001), the respondents were asked to rank and discuss both Samsung and Salora on its underlying constructs. It turned out that in all cases and for all constructs, Samsung scored much higher than Salora. This is mainly due to the fact that Salora is rather unknown in The Netherlands.
In order to allow for more diverse responses concerning the innovativeness, a radical and in incremental innovation were given. A television was chosen in this study due to its relevancy to the participants. The incremental innovation involves the newest 4k resolution technology, implying an ultra HD screen. Despite being innovative, it is not revolutionary and thus regarded in this study as being incremental. The radical innovation in this study is something completely new: a television that allows for the ultimate 3D experience in your own home including smell features. This does not exist yet and is not something like simply putting on 3D glasses. In order to find out whether these are indeed regarded as being respectively incremental and radical, N=10 respondents were again asked to rank both innovations on a 8 point Likert scale. It turned out that all 10 respondents mentioned the 3D experience as being much more radical than the 4k technology (e.g. all respondents rated the 3D experience higher than 5, whereas the 4k resolution innovation was only rated higher than 5 by half of the respondents).
12 In total, N=600 respondents were randomly selected out of a professional panel agency’s national representative database. This number is large, especially when using the formula of Tabachnick and Fidell (1996) on the minimum number of respondents needed to carry out a multivariate analysis (50+8m), where m is the number of variables in the model being tested. In this research, this would mean that 122 respondents would be enough to carry out a regression analysis. However, as it involves four different scenarios, around 150 respondents per scenario are needed in order to come to solid conclusions.
As it involves 600 randomly selected participants via a professional panel agency’s national representative panel, the sample is nationally representative as well in terms of gender, age, income level and educational background. The respondents were randomly assigned to one of the four scenarios, eventually leading to N=150 per scenario. Each scenario consisted of the same survey questions based on the operationalizations which will be elaborated upon in the next section. All respondents filled out the survey on the Internet. The amount of men and women was equally spread and thus 50/50. All age categories were represented with M=41.9. The age category spread can be found below.
Age category Percentage
16-20 3 % 21-25 14 % 26-30 17 % 31-35 6 %6 6 %6 % 6 % 36-40 8 % 41-45 6 % 46-50 8 % 51-55 12 % 56-60 16 % 61-65 10 %
Table 1 Age distribution Table 2 Education distribution
50.3 % of the respondents were highly educated (i.e. HBO or WO), as can also be seen in the education level table The response-rate was 100 %, as a professional panel agency took care of the recruitment of these 600 respondents and the respondents thus got compensated for participating. Therefore, no over-recruitment was needed to meet the threshold of N=600.
4.2 Measurement
The research approach taken in this research is theory testing, as it was tested whether an additional variable should also be taken into account as a significant predictor of behavioral intention. A survey was developed in order to be able to test the above mentioned hypotheses. All variables and underlying constructs in the survey are based on existing and validated questionnaires using an 8 point Likert scale. An 8 point scale was used for two reasons. It first of all forces respondents to choose as there is no clear neutral middle. Secondly, respondents seem to be reluctant to choose extreme descriptors for their response (Friedman and Amoo, 1999), which still leaves us with six descriptors to choose from.
Each variable consists either of a certain amount of questions (in case of merely one construct) or a certain number of constructs, all with their own questions, in case of brand equity and its four
Education level Percentage
Elementary school 1 %
High school degree 27 %
MBO 22 %
HBO 16 %
13 underlying constructs. As the entire survey consisted of validated scales from prior research, the survey as such was not tested for validity again.
For the existing variables in the UTAUT2 model (performance expectancy, effort expectancy, social influence, facilitating conditions, hedonic motivation, price value and habit), the original scale of Venkatesh et al. (2012) was used. However, when critically looking at this scale, it can be observed that the operationalization of performance expectancy does not necessarily involve expectancy but rather perceived performance and thus to a large extent perception. That being said, it was decided to stick to the initial operationalization of the UTAUT2 model for all existing variables in order to ensure consistency. However, the operationalization of performance expectancy was borne in mind during the remainder of the analysis, as this could potentially bias the outcome, since brand equity deals with constructs related to perception as well (Yoo and Donthu, 2001).
In order to measure brand equity, the scale of Yoo and Donthu (2001) was used which consists of the constructs brand awareness, brand associations, perceived quality and brand loyalty, as these constructs appeared consistently in many prior research articles on brand equity (Christodoulides and de Chernatony, 2009). Each construct was measured using an existing set of items developed by Yoo and Donthu (2001).
Furthermore, in order to measure product category knowledge, the scale of Lichtenstein et al. (1990) was used as this scale was also used in the research of Gammoh et al. (2011), on which the hypothesis regarding the moderating effect of product category knowledge was based. Finally, the type of innovation was measured giving people the ability to rank the innovation in terms of newness on an 8 point Likert scale, with 8 being extremely radical and 0 being extremely incremental. All variables can be found in the picture below. As can be seen, only brand equity consists of four underlying constructs. The other variables are constructs on its own, only measured by a certain amount of questions.
14
5. Results
In order to start analyzing the data gathered in this study, 11 new interval variables were computed consisting of the means of the underlying constructs in case the Cronbach’s Alpha exceeded the minimal threshold. For all tests, an alpha level of p=.05 was used with N=600. The partial least squares loadings and cross-loadings for the variables in the initial UTAUT2 model (Venkatesh et al., 2012) were all above .7, which led to the decision to compute the new interval variables in this research by just taking the means of the underlying constructs, as high loadings signify that the separate constructs are all close to 1 in the variable formation. Therefore, calculating the mean of the constructs is sufficient enough and additional factor analyses were therefore not considered necessary and thus not carried out.
5.1 The initial variables in the UTAUT2 model
The underlying items of the variables in the initial UTAUT2 model were all found to be reliable using Cronbach’s Alpha, indicating that the internal consistency among the statements was high enough (α >.6) to compute new interval variables based on the means of the constructs. The alpha’s, means and standard deviations of the initial UTAUT2 variables can be found in the table below.
Variable Cronbach Alpha (a) Mean (M) Standard deviation (SD)
Performance expectancy .92 3.78 1.88 Effort expectancy .953 5.72 1.86 Social influence .955 3.39 1.99 Facilitating conditions .838 5.53 1.70 Hedonic motivation .953 5.59 1.90 Price value .929 5.15 1.80 Habit .938 3.81 2.08
Table 3 Descriptives initial variables
Performance expectancy represents the degree to which using the technology will provide benefits to the respondents in performing certain activities, whereas effort expectancy refers to the degree of ease associated with the use of the technology (Venkatesh et al., 2012). However, as mentioned earlier, it can be observed that the operationalizations of both performance and effort expectancy rather involve perception and not expectation. This was taken into consideration in the remainder of the analysis. Social influence represents the extent to which respondents perceive that important others believe they should use the particular technology, whereas facilitating conditions refers to respondents’ perceptions of the resources and support available to perform a behavior (Brown and Venkatesh, 2005). Hedonic motivation and price value respectively refer to the fun or pleasure derived from using a technology (Brown and Venkatesh, 2005) and consumers’ cognitive tradeoff between the perceived benefits of the application and the monetary costs for using them (Dodds et al., 1991). Finally, habit refers to the extent to which people tend to perform behaviors automatically because of learning (Limayem et al., 2007).
15
5.2 The added variables in this research
The scales of respectively brand equity, product category knowledge and the degree or type of innovation were all found to be reliable, which can be found in the table below.
Variable Cronbach Alpha (a) Mean (M) Standard deviation (SD)
Brand equity .92 4.47 1.86 Degree of innovativeness .839 5.04 1.87 Product category knowledge .927 4.76 1.70 Behavioral buying intention n.a. 4.22 2.20
Table 4 Descriptives newly added variables in this research
In order to measure brand equity (a=.92, M=4.47, SD=1.86), the scale of Yoo and Donthu (2001) was used, consisting of four constructs being respectively brand loyalty (a=.941), perceived quality (a=.937), brand awareness/associations (a=.867) and overall brand equity (a=.975). Both innovations of Samsung scored significantly higher on brand equity than both innovations of Salora (p=.00), which is in line with the outcome of the pre-test done in order to justify the scenarios. Within the broader variable brand equity, brand loyalty is the largest contributor with β=.294 as a simple regression analysis reveals. An additional question regarding brand preference has been asked in the survey in order to control for possible bias due to a preference for either Samsung or Salora, but this control variable did not strongly correlate with any other variable, thus it was not taken into consideration during the remainder of the analysis.
The scale of the newly added variable product category knowledge (M=4.76, SD=1.7) which was derived from Lichtenstein et al. (1990) was found to be reliable with a=.927, and refers to the knowledge of the broader product category the respondents already possess. The scale used to measure the degree or type of innovation (M=5.04, SD=1.87) was also found to be reliable with a=.839, which is of major importance as this did not involve an already existing and thus validated scale. A simple independent samples T-test reveals that the 3D TV of both Samsung and Salora scored significantly higher than the 4K TV of both brands on type of innovation (p=.00), which is a likely outcome as this innovation already turned out to be indeed perceived more radical than the 4K innovation in the pre-test.
The final variable measured in the survey was behavioral buying intention (M=4.22, SD=2.20), referring to the degree to which the respondents would be willing to buy the product using the price stated and assuming they have the budget available to do so. There is no Cronbach Alpha measure for this variable as it merely consisted of one question regarding buying intention. The intention to buy the innovation was significantly higher for Samsung (p=.00), regardless whether the innovation was either radical or incremental, which is an interesting yet expectable finding when reading the literature on the importance of brand equity (Corkindale and Belder, 2009). The degree of innovation (incremental versus radical) did not contribute to a significant difference in means regarding the degree of behavioral intention (p=.07), implying that although the behavioral intention for the incremental 4K innovation was somewhat higher than for the radical 3D television, this was not significant. Regarding the scenarios, the incremental 4K innovation of Samsung scored the highest on behavioral intention, followed by the radically innovative 3D TV of Samsung, the 4K TV of Salora
16 and eventually the 3D TV of Salora. The ANOVA reveals that the incremental 4K innovation of Samsung scores significantly higher than the other scenarios on behavioral intention (p=.00), implying that in this case, a high equity brand introducing an incremental innovation leads to the highest buying intention. This is due to the fact that both a high equity brand and an incremental innovation separately lead to a higher buying behavioral intention.
5.3 Measuring the effect of adding brand equity
In order to find out whether brand equity is a significant predictor of behavioral intention and whether adding brand equity increases the predictability of the UTAUT2 model, regression analyses were run. In order to do so, several assumptions needed to be met beforehand.
Variable Linearity according to scatterplots Levene’s test of homogeinity VIF Tolerance
Performance expectancy Yes .586 2.891 .346
Effort expectancy Yes .031 2.485 .402
Social influence Yes .066 2.764 .362
Facilitating conditions Yes .021 2.626 .381
Hedonic motivation Yes .006 2.960 .338
Price value Yes .249 2.262 .442
Habit Yes .921 2.774 .361
Brand equity Yes .278 2.034 .492
Type of innovation Yes .751 1.094 .914
Product category knowledge
Yes .107 1.480 .676
Table 5 Regression analysis assumptions
All assumption regarding the type of variables, N, linearity, homoscedasticity and multicollinearity were met, as can be found in the table above. All variables involve interval variables and the minimum number of respondents needed with the given number of variables (50+8m) is by far exceeded with N=600 (Tabachnick and Fidell, 1996). Furthermore, the assumptions regarding linearity, homoscedasticity (except effort expectancy, facilitating conditions and habit) and multicollinearity were met. Although multicollinearity existed for all variables, the VIF and tolerance level were still within the acceptable range, which indicates that multicollinearity is not a problem. According to normality tests, the data was not normality distributed, but as both the skewness and the kurtosis of all variables sticked between -1 and 1, the data was treated as being normally distributed.
First of all, a correlation matrix (see next page and appendix) shows that all independent variables are significantly and highly correlated with the dependent variable behavioral buying intention, with correlation coefficients being at least r=.450. Brand equity shows the highest correlation with r=.742. The correlation numbers indicate that there is relationship between the independent variables towards behavioral intention, although this does specify the direction of the relationships. Therefore, several regression analyses were carried out in order to measure the impact of brand equity and of both moderators, being respectively product category knowledge and the type of innovation.
17 Variable Perfor mance expect ancy Effort exptancy Social influen ce Facilita ting conditi ons Hedonic motivati on Price value Habit Brand equity Buying intentio n Performan ce expectancy 1 .414 .752 .417 .570 .559 .690 .602 .627 Effort expectancy .414 1 .353 .722 .687 .569 .458 .479 .456 Social influence .752 .353 1 .406 .492 .533 .705 .564 .588 Facilitating conditions .417 .722 .406 1 .707 .556 .506 .506 .450 Hedonic motivation .570 .687 .492 .707 1 .648 .598 .573 .600 Price value .559 .569 .533 .556 .648 1 .633 .589 .644 Habit .690 .458 .705 .506 .598 .633 1 .612 .626 Brand equity .602 .479 .564 .506 .573 .583 .612 1 .742 Buyin intention .627 .456 .588 .450 .600 .644 .626 .742 1
Table 6 Correlation matrix. Please refer to the appendix for the raw SPSS output including significances The first regression analysis conducted is on the initial UTAUT2 model without the brand equity addition. The adjusted R² or coefficient of determination is .556, meaning that almost 56 % of the variance of behavioral intention is explained by the variance of the independent variables. The ANOVA confirms that at least one of the regression coefficients and thus also the R² of the population is not zero (p=.00). When taking a look at the independent variables separately it can be observed that most independent variables are significant predictors of behavioral intention, with the exception of effort expectancy (p=.693) and facilitating conditions (p=.229). Of the initial variables in the UTAUT2 model with a significant impact on behavioral intention, price value has the biggest influence on behavioral intention, with β= .345, followed by hedonic motivation (β=.226), performance expectancy (β=.224), habit (β=.142) and social influence (β=.131). This leads in the initial UTAUT2 model to the following formula:
Behavioral intention = -.402 + .224 * performance expectancy + .131 * social influence + .226 * hedonic motivation + .345 * price value + .142 * habit + e.
This implies that the higher the mentioned independent variables, the higher the degree of behavioral intention.
A second regression analysis was run in which the independent variable brand equity was taken into consideration as well, as an addition to the initial variables in the model. Adding brand equity as a variable increases the adjusted R² or coefficient of determination to .654, which implies a difference of almost .1. However, this is in line with the outcomes of the correlation matrix, as brand equity heavily correlated (r=.742) with behavioral buying intention. Also, it does not have to be the case that the .1 change is merely due to adding the constructs underlying brand equity. The question thus remains whether adding brand equity leads to a net effect on the adjusted R² as well.
This was analyzed more thoroughly by running a step-wise regression analysis, as the order in which the independent variables are added to the regression analysis is of utmost importance. The variables
18 were added stepwise, with the first variables being the ones with the highest correlations towards behavioral buying intention and low correlations towards the other independent variables. The order in which the variables have been added together with the change in adjusted R² can be found in the table below.
Model R Adjusted R
Square
Brand equity .742 .550
Brand equity, price value .785 .615
Brand equity, price value, performance expectancy
.800 .638
Brand equity, price value, performance expectancy, hedonic motivation
.803 .643
Brand equity, price value, performance expectancy, hedonic motivation, facilitating conditions
.806 .646
Brand equity, price value, performance expectancy, hedonic motivation, facilitating conditions, social influence
.808 .649
Table 7 Step-wise regression. Please refer to the appendix for all regression statistics
The first variable being added is brand equity with r=.742. This leads to the adjusted R² already being .550. More than half of the variance of behavioral buying intention is thus due to brand equity on its own. The next variable being added is price value with r=.644, increasing the adjusted R² to .615. Next, performance expectancy with r=.627 was added, increasing the adjusted R² to .638. It turned out that both habit and effort expectancy were not listed in the output of the regression analysis, meaning that they both did not contribute to a change in the adjusted R². After performance expectancy, hedonic motivation (r=.600) was added, which increased the adjusted R² to .643. Afterwards, facilitating conditions was added (r=.450), increasing the adjusted R² to .646, and eventually social influence was added (r=.588), leading to the adjusted R² ending up being .649. Compared to the regression analysis on the initial UTAUT2 model, it can be concluded that the adjusted R² has increased from .556 to .649; almost being a change of .1. The formula after adding brand equity in a step-wise method logically becomes:
Behavioral buying intention= -.725 + .536 * brand equity + .263 * price value + .134 * performance expectancy + .183 * hedonic motivation - .124 * facilitating conditions + .103 * social influence + e. This implies that half of the total effect on behavioral intention is due to brand equity. It can therefore be stated that brand equity is indeed a significant (p=.00) predictor of behavioral intention in this extended UTAUT2 model and that H1 is thus confirmed. It can even be concluded that when comparing both R²’s using Fisher’s Z-test, the R² of the model including brand equity is significantly higher (p=.006) than the R² of the initial UTAUT2 model without the brand equity addition.
The fact that only brand equity already accounts for an adjusted R² of .550 and that brand equity correlates with all initial variables implies that the UTAUT2 model might need some adjustments. This is due to the fact that questions belonging to the underlying constructs of brand equity have an overlap with questions underlying the initial variables in the UTAUT2 model. This was already suggested in the methodology chapter as most questions involve perception rather than actual behavior and brand equity thus somehow load on most of the initial variables. A factor analysis was carried out
19 to find out whether the variables in the UTAUT2 model together with brand equity could be remodeled into different variables.
5.4 Changing the UTAUT2 model
As can be derived from the acceptance of H1, adding brand equity to the UTAUT2 model does increase its predictability significantly. However, due to the fact that brand equity heavily correlates with the initial variables in the UTAUT2 model and that the step-wise regression analysis even did not take list habit and effort expectancy, a factor analysis was run in order to find out whether the UTAUT2 model could be redesigned into different factors. When doing a Principal Component Analysis, all factors load on the same component, including the constructs of brand equity. A Varimax rotation was carried out as this would lead to a basic structure of loadings and because of the fact that the former step-wise regression was successful. Five components can be observed afterwards.
Component Variables
1 (internal and external importance of product)
Performance expectancy + social influence
2 (internal and external ease of use of the product)
Effort expectancy + facilitating conditions + hedonic motivation
3 (price) Price
4 (habit) Habit
5 (brand equity) Brand equity
Table 8 Variables after Varimax rotation. Please refer to the appendix for all loadings.
This additional analysis also reveals that there is still a lot potential to develop the UTAUT2 model even further by combining variables and renaming them. However, it is beyond the scope of this research to do so.
5.5 Measuring the moderating effects of the degree of innovativeness
In order to find out to what extent the degree of innovation moderates the relationship between brand equity and behavioral intention, a regression analysis was run including a newly created variable consisting of a multiplication of brand equity with the degree of innovation. The interaction variable is not significant (p=.306), implying that the type of innovation does not moderate the relationship between brand equity and behavioral buying intention. The variable type of innovation itself is not a significant independent variable either (p=.382), leading to the conclusion that type of innovation is not a significant predictor of behavioral intention itself either. This leads to a rejection of H2. This is somewhat striking, as Gammoh et al. (2011) did find that the degree of innovation plays a role. They stated that for a low equity brand, adoption is higher when the innovation is radical compared to an incremental innovation. This is true in a way that it can also be observed in the descriptives regarding the scenarios. The 3D TV (radical innovation) of Salora (low equity brand) does score higher on behavioral buying intention compared to the 4K (incremental) technology, but not in a way that the degree of innovation can be regarded as a significant moderator in the extended UTAUT2 model.
5.6 Measuring the moderating effects of product category knowledge
The same type of analysis was used in order to measure the moderating effects of product category knowledge on the relationship between brand equity and behavioral buying intention. The interaction effect (being the newly created multiplication of product category knowledge and brand equity) was found to be significant (p=.011), implying that there is a moderating effect. The variable product category knowledge itself is not a significant predictor of behavioral intention (p=.647), which
20 is somehow logical as the importance of brand equity has been stressed in prior research. If this would have been significant, product category knowledge would have had a negative impact on behavioral intention, which is somewhat in accordance with prior research on the role of product category knowledge (Gammoh et al., 2011). Product category knowledge is merely a moderator affecting the relationship between brand equity and behavioral intention in such a way that the higher the product category knowledge, the stronger the relationship between brand equity and behavioral intention (β=.041). This is somewhat striking as it is in contrast to what was hypothesized and H3 is thus rejected.
After running all analyses, it can be concluded that adding brand equity to the UTAUT2 model significantly increases the adjusted R² or coefficient of determination and thus the predictability of the model. The degree of innovation does not play a moderating role, whereas product category knowledge does, but in a way that it merely strengthens the relationship between brand equity and behavioral intention. Only H1 was therefore confirmed, as can be found below.
Hypothesis Accepted or rejected
H1: In the existing UTAUT2 model, the additional variable brand equity will be significantly positively related to behavioral intention
Accepted
H2: The degree of innovation (radical versus incremental) will moderate the relationship between brand equity and behavioral buying intention in such a way that the more radical an innovation is, the weaker the relationship between brand equity and behavioral intention
Rejected (not significant)
H3: Product category knowledge will moderate this in such a way that the higher the product category knowledge, the weaker the relationship between brand equity and behavioral intention
Rejected (was found to be a positive moderator instead of a negative moderator)
21
6. Discussion
6.1 Conclusion
The aim of this research was to find out whether and to what extent brand equity plays a role in the adoption of technological innovations, and more specifically in the UTAUT2 model, which is the most recent extension (Venkatesh et al., 2012). This because prior research had stressed that brand equity had an impact on the way people perceive and evaluate certain products (Erdem and Swait, 1998). Also, Gammoh et al. (2011) and Corkindale and Belder (2009) found evidence concerning the importance of brand equity and reputation in the process of innovation evaluation and adoption. However, all existing innovation adoption models merely contain variables related to the product or user, rather than to the brand, which is a bit odd. This research tried to find out whether brand equity as an additional independent variable had a positive significant impact on behavioral buying intention, and whether adding brand equity next to initial variables increased the predictability of the UTAUT2 model. The UTAUT2 model was chosen, as this model was already extended with three variables (habit, price value, hedonic motivation) that do not necessarily deal with the technological innovation itself and is thus regarded as being a good starting point for further improvement.
Furthermore, two moderators were tested, being the degree of innovation and product category knowledge, as Gammoh et al. (2011) found that both have an influence on the relationship between brand equity and eventual innovation adoption. 600 nationally representative respondents were asked to fill out a survey. Each of them was randomly assigned to a scenario, being either a radical innovation from a low equity brand, an incremental innovation from a low equity brand, a radical innovation from a high equity band and an incremental innovation from a high equity brand.
It was found that adding brand equity increases the predictability of the UTAUT2 model from .556 to .649, which was a significant increase. Also, brand equity ended up being the variable adding the most variance with β=5.36 to behavioral buying intention according to a step-wise regression analysis. The step-wise regression analysis also revealed that habit and effort expectancy were no longer listed and thus do not account for a net effect on behavioral buying intention in this model.
It turned out that brand equity strongly correlated with the other independent variables. This was already mentioned in the methodology part, as questions underlying for example performance and effort expectancy all involve perception rather than actual behavior. Since the constructs underlying brand equity involve perception as well, the correlation between the various independent variables was high. An additional factor analysis including a Varimax rotation was carried out to find out whether the UTAUT2 model could be adjusted in a way that all variables are still taken into account but in different components. This resulted in five components with either one or more variables.
The complete adjustment of the UTAUT2 model as such was beyond the scope of this research as this research merely stressed the importance of adding brand equity. There is therefore enough room for further development of the model by combining and renaming some of the initial variables after having added brand equity as an additional variable.
Next, the moderating effect of the degree of innovation was measured, as it was hypothesized that the higher the degree of innovation (thus, the more radical the innovation), the weaker the relationship between brand equity and behavioral buying intention. This turned out not to be the case in this research, which is a bit striking taking into account the findings of Gammoh et al. (2011). It could be observed that the radical innovation of the low equity brand Salora indeed scored higher on behavioral buying intention compared to its incremental innovation, but not in such a way that the degree of
