Categories and co-branding : one product, many views

(1)

Master’s Thesis

Categories and Co-Branding: One

Product, Many Views

Author: Rizzo, Marco 11592656 MSc BA

Supervisor: dhr. drs. R.E.W. Pruppers

(2)

STATEMENT OF ORIGINALITY

This document is written by student Marco Rizzo who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(5)

Abstract

The way everyone perceives everyday life is different. We shape our decisions and evaluations on what we see and perceive differently, by using clear and reliable references to make sense of the world and what constitutes our decisions as consumers.

Deciding whether we want to buy a new phone rather than a new mp3 player makes a huge difference. What if we would find a product that combines functionalities of both categories? And what if we see that brands that put their logo on that product are just partially related to that category? What would our decision be?

This study delves deeper into what constitutes a categorization decision when contextualized in the setting of a cobranded alliance when discordant information is presented on the product.

Results show that not all categories elicit the same reactions in evaluations and that, in some cases, the discordancy of brand and product related information produces different categorization assessments. The study also delves deeper in the analysis of the behaviour of willingness to pay when related with categorization decisions, in order to understand if (and to what extent) a different categorization of a product produces different assessments of its value.

While for one of the sets of stimuli used for the research impacted the categorization of the items based on their physical appearance, the second set of stimuli used showed greater responsiveness to brand related information.

The diversity of the two categories of stimuli sets casts the doubt that such differences in evaluation might be driven by familiarity with the product presented and by the complexity of the evaluation based on previous experiences.

Different levels of complexity and ambiguity however do not seem responsible for major shifts in willingness to pay in either case, whether the categorization decision is without doubt or when it is more ambiguous.

(6)

1

1. Introduction

1.1 Brand Alliances

1.1.1 Brand Alliances and Categorization

In the last 20 years, research on Brand Alliances has been receiving increasing interest due to its relevance and diffusion. This particular strategy is defined as the short (or long) term cooperation or combination of two or more brands, products, and/or other distinctive proprietary assets (Rao and Ruekert, 1994; Aaker,Keller 1990).

Using a well respected and consistent brand to reinforce the image of your own brand or to get a clearer positioning of an existing/new branded product is a popular tactic nowadays: it can create a strategic advantage for the market success or gain consumer wide acceptance (or rejection), depending whether strategic decisions do take some important aspects into account. Key global players - like LG, Philips, Kenwood, Asus, Huawei, Bose, etc...- have already launched co-branded and/or licensed products, supported by consistent brands with a strong heritage, gained in different but related worlds (such as Ferrari, Prada, Leica, Volvo, Ducati, Caterpillar...). The fact that this many strong players have been consistently using brand alliances (even if for different strategic purposes) highlights how its relevance is still vivid even for consolidated brands. In this study we will focus on a particular kind of brand alliance: co-branding, intended as the joint effort of two companies or brands in creating a separate product, service or experience that contains elements of both brands simultaneously (e.g. Heineken x Lowdi x Ed Banger speaker box, which is a portable speaker that houses a specially designed Heineken bottle) (Aaker, Keller, 1990; Cordeiro et al, 2016; Bouten et al 2011).

Co-branding represents a unique and particular branding strategy through which companies can reinforce an existing association by stimulating the transfer of associations (positive or negative) from brand to product in the mind of the consumer, hopefully bolstering brand image and brand awareness (Cordeiro, Ponchio,Mazzon, 2016; Riley, Charlton and Wason, 2015; Rao, Ruekert, 1994).

According to the core literature on co-branding, the degree of perceived fit between product and brands in the alliance has been identified to be the main guiding variable in the process of co-branded evaluation. Perceived fit can be defined as the consumer’s perception of congruence in brand associations (brand image fit) and product categories (category fit) that the co-branded product is recalling. (Rao and Ruekert, 1994; Aaker,Keller 1990). In the case of co-branding, less obvious categorizations ( e.g. Dr.

(7)

2

Pepper-flavored lipstick by Bonne Belle, which is clearly a lipstick) it could be sometimes difficult for the consumer to relate the two brands when the category to which it is perceived to belong is not clear at first glance. A fictional example might be a co-branded high definition camera, branded by Apple or HP, but produced and labeled by Canon. Will consumers perceive it as a webcam or as a professional photo camera? The stimuli that marketers will place in the context in which the product is presented will drive the evaluation of the product based on plausible relationships between fit, the product’s attributes (contextualized in the perceived category) and the framing of the involved brands. Previous studies have indeed shown that the context in which the consumer is called to action also plays an important role in this process: Swaminathan et al (2011) explored consumers’ reactions to co-branded products featuring brands with complementary or similar attributes. The author also argues that different contexts, either based on product attribute complementarity or similarity when related to the involved brands, may elicit different association transfers in how the consumer evaluates the product.

1.1.2 Research Gap: Brand and Product information might drive Categorization

Previous researches are identified how context can play all important role in the evaluation of branded products, and results testify that consumers can be “pushed” towards different evaluations through some external stimuli. Past literature has approached the topic of co-branded product evaluation through the lens of perceived fit between brand (which is nonetheless an important variable), but has always taken as granted that the product the consumers were evaluating was clearly identifiable in its category. According to Moreau et al (2001) and Gregan-Paxton (1999) marketers can influence association transfer by implementing communication cues that can guide consumers by suggesting the new product’s category. This level of insight is however partially unexplored, and the focus of research has shifted towards the interplay product attributes so to partially drive consumers to categorize co-branded products that have unclear category belonging (Swaminathan et al, 2011). Previous research has however generally loosely focused on such co -branded products (hereby “ambiguous products”), and always on an attribute evaluation level. What is yet to be known is how the framing of brands can influence the categorization of co-branded products when they can be interpreted as being “between two categories” and how much we can stretch the mapping process through the framing of the involved brands based on their meaning to the consumer.

Even if we know that the perceived fit in a brand alliance is influenced by the context in which it is presented, literature is silent regarding antecedents in the evaluation that are triggered by the interaction

(8)

3

of two individually meaningful and known brands in a product that does not immediately communicate its identity through its features.

If one of the ways it is possible for marketers to act directly on the categorization transfer is through a correct framing of attributes, it is also interesting to assume that is possible to influence the perception of category (which implies influence on the evaluation of the alliance) acting directly on the framing of the involved brands.

1.2 Categories, Brands and Connections

1.2.1 Problem Statement

Product categories constitute the mental labels that consumers use in order to make sense of products, and they use these labels to recognize familiar patterns in new products they see, based on what the actual item looks like and on the brand(s) advertised on it. Products that are born as a result of a brand alliance are sometimes difficult to categorize, since they might report conflicting information or simply because they are not perceived to be something that can be labelled immediately and reliably.

When such ambiguity arises, consumers rely on the information they have available: what they see about the product and what they know and feel about the brands on it. What cue will consumers value more in terms of their decision and how strongly they identify the product with a known category might be different from one context to the other and from one alliance to another. While some alliances might be easy to understand for consumers, literature and every day life show that Brand Alliances come in many forms, and it is logical to assume that different forms of brand alliances will trigger different categorization cues. Having identified this gap in growing on this literature this research will be committed to answer the question:

“

In the case of a co-branded product, how and how much do consumers rely on the product’s physical appearance and the framing of the involved brands in order to categorize it?”

1.2.2 Subquestions

In order to make the logical connection between theoretical concepts clearer, we can outline a set of sub questions related to the research proposition:

(9)

4

1. What is co-branding and in which conditions is a preferred strategy?

2. What are the components of perceived fit and according to which variables are they ranked in terms of importance?

3. Why is category based processing so crucial for product evaluation?

4. When can we consider a product to receive added value after a category evaluation?

1.3 Contributions

1.3.1 Theoretical Contribution

My work will be built on the two main blocks of literature, namely literature on categorization theory and on brand extensions and brand alliances (even if it will be mainly focusing on co-branding examples). Research on these complementary topics might benefit from my work since it could outline dynamics that have not been considered when analyzing the categorization process at a consumer response level. Studying category-based processing’ s antecedents in this context and with these premises is also an opportunity to focus on variegated brand pairings in order to identify interesting stimulus - response relationships in terms of co-brand evaluation. This contribution looks appealing at a theoretical level, since it could offer interesting insights to be better studied to co-branding literature. Moreover, multiple recent sources make a strong call for broader data collection, in order to test existing findings in a across samples with different brand knowledge, product category, familiarity and consumer involvement conditions (Gregan Paxton, Roedder John, 1997) (Xiao,Lee 2014; Singh, Kalafatis, Ledden, 2014).

The objective of this research is to focus on possible categorization cues that might prove to be more effective than others in guiding the categorization process. Testing research hypotheses by incorporating different forms of co-branding in our scope (e.g. ingredient branding) could further be valuable in terms of contribution to existing literature.

As a final contribution, given the nature of this study, it would be valuable and beneficial for the existing literature to investigate the how consumers react to asymmetries in a co-branding relation. Previous studies have also been exploring how consumers evaluated low fit brand alliances: this work will be also looking at brand asymmetries on a categorisation perspective, in order to explore whether the asymmetrical condition triggers different responses to categorisation cues based on category and brand. For instance, if the product is not evaluated positively, asymmetry in the co-branding relation may drive attribution of failure to only one of the two brands in the partnership.

(10)

5

1.3 Contributions

1.3.2 Managerial Contributions

The results of this study will hopefully better clarify whether in which cases and under which circumstances marketers will find advantage in leveraging on stressing the perceived fit measure rather than exploiting the effect of category based processing on attitude formation. As a consequence for this main contribution I hope to provide, I would also like to mention that both the theoretical and practical contributions might be further broadened by maximizing the heterogeneity of analysed categories: following this design it could be possible for us to find some product categories where the effects of categorisation and fit prove to be stronger or weaker than average.

This knowledge could guide managers in assessing with more precision the brand with which they are planning to start an alliance with, as well as the market in which it would be more profitable to introduce the product in order to fully exploit this (expected) synergistic effect.

Further contributions of the study at the managerial level can be seen when we consider the context analysed: having to work with multiple combinations of categories, the stimuli that will be designed for this study could bring to light key associative interactions across brands, categories, product and context. These conditions and the potential scope of the study might give management a hint about how to better communicate and suggest the category of a new co-branded product in order to maximize the effect of brand association transfer on the evaluation.

Concretely, I expect this research to broaden the possibilities of current positioning strategies in terms of:

● Brand Building

More insights on the analysed scenarios can help marketers to understand new ways to identify and position their brand through a product fitting its category, by co-branding with a second brand coherent with the values of the host, in order to reinforce its current position.

● Brand Extensions

The variety of categories and associations this study can cover is surely wide enough to explore unconventional associations in terms of brand meaning and image, an aspect that could provide interesting brand extension opportunities.

(11)

6

Knowing more about the topic I want to tackle and about brands in general and how they play their role in consumers Minds. Co-branded products may sometimes be difficult to categorize, because of the way they look at the functions they have a but more importantly the brands that represent them.

1.4 Research Outline

The first following chapters will be committed to a more detailed exploration of the theoretical foundations and findings of previous authors regarding co-branding strategies for success (both at a communication level and at a strategic positioning level) and category based processing, with special attention to the phenomenon of analogical transfer of knowledge.

From the core literature, this study will explore more in detail the process of analogical transfer of knowledge and the analysed moderating effect of perceived fit on evaluations of co-branded products, in order to outline a basis from which the research hypotheses will be built.

This theoretical foundation will then be contextualized with both fictional and real examples of brand alliances (the set will include successful products as well as products that have failed) which will be considered as the starting point where we can apply existing knowledge. These examples serve also as a benchmark for comparing different stimuli and reactions with regard to primary data collection and analysis.

(12)

7

2. Literature Review

The increased complexity that co-branding implies (even though this is highly dependent on the involved brands, the category and the product itself) represents an interesting phenomenon when contextualized in a not so trivial context like hybrid products: will consumers be driven in their categorization by the brands’ fit with the product and themselves or will they just focus on what the product looks like?

2.1 Literature on Brand Alliances and Co-Branding

Co-branding is a strategy that stems from the concept of brand alliance, with the difference that we refer specifically to co-branding when we consider the pairing of two or more brands to form a separate and unique product.

A combination of two well known brands can influence consumers by increasing their perception of quality when comparing the co-branded product with a mono-branded one from the same category. Research on co-branding has been exploring this particular strategy in a myriad of scenarios, stretching and analyzing how brands can influence expectations about a product’s features and experience, being a meaningful mean to convey the involved brands’ identities and values.

This stream of literature has mainly focused on two central topics, the first being how customers’ perceptions of a co-branding alliance might be guided by their perceptions of the two parent brands and vice versa and secondly on the relative merits of co-branding with respect to other new brand development strategies. The most important finding in these fields is that a meaningful association of brands with respect to the product can transfer to the co-branded product positive prior evaluations about the brands. (Mitchell and Olson, 1981; Selart, Kuvaas, 2003).

Following this rationale, marketers are expected to choose the partner for their brand alliances as carefully as they can, since the link that stands between the evaluation of the product, the fit between the brands and the product itself needs to be evident and strong for consumers that have the first contact with this new product. This link is strong when we consider a mono branded product, or a product whose brands are closely related to the product’s category and/or that fit well together (like Nike and Apple, which find common ground in sport-related technology for runners, but also Coca-Cola and McDonald’s).

Co-branding examples, however, imply the presence of two brands that simultaneously influence the consumer’s experience that now becomes more difficult to rationalize automatically , especially when the involved brands belong to different categories (Gammoh, Voss, Chakraborty, 2006).

(13)

8

This finding relates also to the model of associative network memory model (Anderson & Bower, 1973), which identifies elements relevant to the elaboration of information as nodes connected by links whose strength may vary according to different contexts in which the evaluation takes place.

Stronger links in the minds of consumers activate more frequently and have the benefit of linking existing information with newly acquired one creating a new association. Aaker (1990) and Keller (1993) enrich this theory by contextualizing brands as associative networks as well, triggering nodes through link that connect values the brand represents as well as possible recollection of similar brands in the category. Although these associations are mostly unconsciously recalled, links and nodes can also be triggered by contextual external stimulation (e.g., any kind of advertising) or indirectly through the activation of contingent nodes (i.e. ideas that are logically closely linked). (Selart and Kuvaas, 2003) Just like any other piece of information, a brand provides meaningful cues on what to expect from a product at many levels of the evaluation, such as its perceived quality, its relationship with brands that operate in the same category and (in general) acts as a meaningful cue to make sense of the unknown. Combining two “similar” brands (e.g. two brands that operate in the same product category) on the same product produces more meaningful relationships between the product and the idea the consumer forms in his mind about it (Yong, Buda, 2000; Theodore et al, 2012). It is however to be noted that the presence of two brands on the same product is likely to trigger, on average, twice as much the associations that a single brand would, increasing the difficulty of understanding which brand is to be considered “more important” (or dominant) in the alliance.

Consequently, we can infer that co-branding partnerships are complex entities that are not always easy to categorize for consumers despite the effort produced by marketers in the search and framing of “fitting” brands due to the multiple information sources consumers can base their evaluation on.

In this case, research has shown that brand fit was more influential than product fit in some categories mainly depending on the degree of overall familiarity of the consumer, and its importance even increased when comparison between high and low familiarity brands was taking place.

Research widely agrees on the finding that a level of high brand ﬁt is more likely to result into a better evaluation of the co-branded alliance and to an overall better perception of value by the consumer (Simonin & Ruth,1998; Baumgarth ,2004; Desai &Keller, 2002).

Most studies indeed revolve around fit, a multidimensional variable that has been tested and analysed across categories and at various levels of influence. Research agrees that there currently is not an universal definition of fit, as it is a rather stratified concept that can be perceived differently when the product or advertising context are altered. The most agreed definition defines it as the perception of consumer of the degree of logical consistency of the two brands with respect to the product.

(14)

9

In other words, perceptual fit can be outlined as the degree to which the co-branded extension “fits” the categories of both parent-brand and host brand.

The stream of research that has explored the influence of a high, low or average level of perceptual fit on co-brand evaluations widely agrees on brand image fit having relevant and interesting impact on the evaluation of products that featured a cooperation of two similar brand partners, indicating that the positive effects of co-branding manifest more in these conditions. (Thompson & Strutton, 2012) In most studies, fit was conceptualized as the perceived fit between the partner brands with respect to the category that was meant to host it, meaning that the interaction between brands and categories might be perceived with a graded approach, where the individual can rationalise brands as representatives of categories and can represent more or less accurately the category.

This demonstrates that brands might indeed have varying levels of category breadth, consequently with different levels of success when complementary or similar combinations are employed.

Categorization theory, however, holds fast that consumers are more likely to prefer products that can be clearly associated with one of the two brands in the partnership, since an easier (and thus preferred) path to categorization is available: in the case of ambiguous products, this might not be the case.

The impact of brands on categorization as a process is further to be explored and constitutes a crucial piece of knowledge to shine more light upon, because brands can communicate their presence in a plethora of touchpoints nowadays. (Swaminathan et al,2015)

2.2 Literature on Categorization

The ability to categorize efficiently and reliably is important, because is the way consumers can compare product features and attributes with the prior knowledge and / or experience within that same category. Research has proven that consumers tend to categorise products and novelties to just one category: even when a product may not be easily attributable to just one category, consumers elaborate this evaluation based on the previous experience and the context to make sense of the ambiguity. To simplify a complicated environment and to make understanding more efﬁcient, customers tend to categorize by brand association or image given their knowledge and previous experience on perceived similarity and resemblances (Sheng & Pan, 2009).

The ultimate goal of categorizing objects is to create a logical mental structure that identifies the maximum similarity of members within each category, while minimizing the similarity of objects belonging to different categories.

(15)

10

This is especially important for consumer behaviour literature, since the manner in which consumers categorize products will eventually influence their inferences about the product at multiple levels of evaluation.

Two fundamental assumptions that underlie categorization theory are that a central function of categories is to support inductive inferences and that categories are psychologically built and rationalised such that through them, is possible to optimize the predictability of features. In other words, categories serve the purpose of providing reliable information about uncertain experiences and also help to “know what to expect”.

However, not every product can be categorized reliably and without effort: for example, when the physical appearance of a product differs significantly from known category members (or presents characteristics that might be expected outside from that category), assessing the product’s category may be difficult or uncertain.Innovative or unusual appearance of a new product can hinder categorization, as the consumer will have to make sense of it, relying on all available category cues and contextualizing the experience in his associative network. (Rajagopal, Burnkrant, 2009)

It has been just partially explored how multiple categories can be recalled by a single object, as well as the cues that “suggest” to which categories the object could potentially belong.

Both psychology and marketing literature have observed that when products exhibit properties that belong to more than one category, consumers will try to categorize it in a single, pre-existing category. This consideration of an ambiguous product as an instance of one unique category intuitively leads to the generation of different expectations with regard to its evaluation than it would have if the product was considered as an instance of a second distinct category.

This insight implies that individuals faced with this uncertainty might think about an ambiguous product as belonging to a category that was previously “stored” in their minds, even if the product itself was not originally meant to belong to that category.

Two main factors may be considered relevant enough be the basic variables responsible for a single category over a multiple category approach: category familiarity and the nature of the category cue. (Gregan Paxton ,2005).

Research has defined category familiarity as the extent to which a consumer can reliably infer attributes and evaluations on the category: the more familiar a consumer is with a category, the more likely this category will be recalled to his/her mind when attempting to categorize a novelty.

Familiarity has also been identified as a dynamic entity in the process of a multiple-category inference process: the generation of inferences from one category to the other requires consumers to change (or “update”) the characteristics of the dominant category they are analyzing. (Malt,1995; Gregan Paxton ,2005)

(16)

11

In other words, if any product presents features that were not considered in the category before, an observer might consider that feature to be now part of that category, if the link is plausible enough. Research has also shown that the ability to adapt pre existing categories decreases as the number of previous experiences within that category increases.

This finding suggests that consumers will be less likely to question high-familiarity category features than low-familiarity ones.(Gregan Paxton, 2005; Lajos et al, 2009)

The second variable explored is the nature of the categorization cue, which can be defined as the aspect that consumers consider more reliable when assessing a category. Stemming from previous research, we know that as long as the inferred category represents a reasonable and viable answer, no additional categories will be taken into account.

Even if this insight is undoubtedly valuable, it is still not easy to predict which category will be used by consumers to make sense of the unknown.

What has been explored by previous research is that category membership can be expressed perceptually (usually by visual depiction) or conceptually through the identification of a mental category label (e.g “snack” or “fruit” when considering an apple). (Gregan Paxton ,2005; Thompson, 1989, Mazanec & Schweiger, 1981)

It is also valuable to mention that artifact categories (defined as every category referring to strictly artificial elements, like a car) are generally identified according to the function they are expected to perform, which is crucial to the case we want to analyse since we can logically infer that every co-branded product can be somehow referenced to an artificial category. (Malt,1995)

In other words, we might ideally say that anything that is able to make a phone call has a marginal probability to be categorized as a phone.However, physical appearance has also proven to be the critical element for the inference of what an item can do, considering the mental category label as a storage of backup information that can add detail to this evaluation. (Rajagopal et al., 2009)

Even if conceptual inferences play an important role in the categorization process, not all inputs will be equally weighted when this attribution is ambiguous: current research has argued that consumers will generally treat a perceptual input as a more reliable tool for ambiguous product inferences than they consider a conceptual input.

However, physical appearance of products and their perceived scope are not the only elements consumers evaluate in the process. Extensive literature has investigated the effect that brands exert on product preference and expectations, a consideration that is magnified when we considered a co-branding scenario, where two brands have to be rationally conceptualized in one single product.

(17)

12

2.3 Categorization of Co-branded ambiguous products

The identification and placement of a product in a precise category is nowadays a topic that proves to be central when considering the prior and post purchase expectations and satisfaction of consumers in any given purchase.

Benchmarking different features a consumer expects among the category itself represents a crucial point in the evaluation of both novelties and brand new entrants in the category.

For example, in the smartphone market consumers evaluate common features that they will take for granted to all devices (or the features that they expect to contextualize in that category) such as the phone’s storage space or its ability to take more or less professional-looking pictures.

Categorization cues play an especially important role in this process, since they allow consumers to change the context of their evaluation and to assess a standard that the product might or might not exceed. Even if this inference is almost taken as granted for products that clearly belong to a single category, the process becomes more complicated and stratified when we take into account products that are not immediately recognized in just one category: ambiguous products. We can define an ambiguous product as a product whose physical appearance or conceptualization embraces more than one product category, nowadays often by integrating two or more technologies in the same product.

A few years ago smartphones themselves could have been considered as ambiguous products, as the idea of a mobile phone that could also surf the web, take pictures, send emails and even serve a portable gaming console was ambiguous and disorienting for consumers.

Categorization ambiguity takes place when information about a new product makes it difficult for a customer to intuitively attribute novelties to a single, existing category.

The studies of which stimuli might elicit different sets of responses uncovered that evaluations of a set of information happens through a three-stage process: the first step is the identification of relevant attributes for consideration, the second being the subsequent evaluation of the level of each attribute considered, eventually combining this information to form an overall evaluation. (Yong, Buda, 2000: Hoch,Ha 1986)

At each stage, consumers can be faced with ambiguity and may not be able to evaluate attribute levels intuitively, either because the attribute is not straightforwardly suggesting something concrete about the product or is entangled with other attributes that might throw additional confusion in the picture. This insight is especially important when contextualized in the case of a co-branded ambiguous product, as it communicates a set of convergent or divergent inputs (both brand and category wise): understanding

(18)

13

what are the main cues employed by consumers in order to categorize a new item can thus become a critical point in research.

When attributes are uncorrelated or negatively correlated with each other, information processing demands will be much greater: understanding how inference and preference interplay in such a complex environment also shines light on relative comparison, which is nowadays a common phenomenon especially in crowded and/fast moving markets, where competition presses for more and more product differentiation or sophistication with respect to the category. (Swaminathan et al, 2015)

In the case of co-branded ambiguous products however, other factors influence the evaluation and categorization process, such as on the degree of fit between the involved brands and whether one of these brands is dominant over the other. Last but not least, the way the product itself is presented to consumers (I. E. The context in which the evaluation happens) can lead different consumers assigning the same product to two different categories.

In the case of ambiguous products, consumers will be much more likely to analyse both the physical appearance (i.e. “ does this product look like a phone?) and the brand the project is labeled with (“ is this brand suggesting a category? Do I expect this brand to sell a quality product in this category?”) in order to assess to which category the product belongs to and to evaluate it based on the previous experience.

(19)

14

3. Hypotheses Development

The previous chapters described how researchers have explored how consumers react to physical appearances of products and to brands, as well as how these evaluations can subsequently affect the evaluation of a co-branded product. The leveraging of associations from a set of multiple brands in the context of the same product is indeed the main point around which co-branding strategies revolve, with the ultimate goal of obtaining more favorable evaluations towards the co-branded product. Previous research has also investigated how categories also play a relevant role in both the evaluation of the product and the assessment of its value, since the ability to refer to categories allows consumers to reliably benchmark product values in a broader and familiar mental context. (Keller, 2003). This context provides fertile ground for transfers of relevant associations from the individual brands to the product itself, which can (or is hard to) be placed in a category of product of similar features and where the analysed brands are likely to be more dominant. A comprehensive approach for fully understanding categorization of product can be found in the concept of fuzzy set theory (which will not be developed further in this work), whose basic premise is that there are no clear boundaries between members and non members of a category set. (Childers, Viswanathan, 1999)

The categorization process in itself is not straightforward and does not follow strict rules of logic, but is rather driven by impressions and associations that arise from an unfamiliar and sometimes discording mixture of stimuli. Measures of product categorization are considered to be at overall product level and, therefore, hardly provide insight on the attributes that significantly influence the labelling of a product as a category member. (Childers, Viswanathan, 1999)

The physical capabilities and look of a product could convince the observer that the product belongs to a category where one would expect all the products to be able to perform a certain function; however, certain physical features could challenge this view and place the item in a similar (or completely different) category.

However, the categorization of a co-branded product revolves around the evaluation and contextualization of the brands involved in the partnership. Brands are (by definition) associated with a certain category of products in which they might or might not excel: as an example, technological devices of unclear look marketed by GoPro are more likely to be identified as cameras, with respect to the same devices where a smartphone brand would appear.

This perceptual fit of product look and brands is a key factor in the process, as the involvement of more prestigious brands could significantly influence the perceived value of the product analysed.

(20)

15

In the case of cobranded products, these two crucial components for categorization play a role that is not easy to disentangle: two brands might send different or even contrasting signals, a perception that can be weakened or reinforced by the physical features of the product on which the two brands appear. An interesting scenario would indeed be one where the two brands involved in the partnership to not openly conflict with each other, but are rather simply associated with different categories. Considering the physical looks of the product and the brands involved, consumers might react more positively when there is a fit on both product and brand dimensions rather than a fit perception on a single one, as they could perceive the benefit of both brands in the partnership instead of focusing on just one.

What would happen if one of the two brands in the partnership were to be more dominant than the other, suggesting that the product belongs to the same category as the one the dominant brand is associated with? And what if the physical looks of the product would reinforce or discourage this evaluation? Building on these speculations we can further explore what is the real relationship between brand driven category evaluations, product features and categories.

3.1 Perceived Category Similarity

The first input consumers analyse in an unfamiliar context is what their eyes perceive. In the context of this research, what the product looks like is without doubt a crucial determinant in the process of understanding the category to which the product belongs to. When linking category membership with attributes, literature strongly stresses sized feature-based attributes as the main conductors of typicality (Medin and Smith 1984): logically reasoning, different features and different looks would trigger different attributions. The presence of a certain attribute on the physical appearance of a product can drastically change the idea a consumer forms about a product with which he/she is not familiar with; to some people, a product looking like an mp3 could be seen as a phone the very moment a display with a number pad is added to it. When the product looks recalls more than one category, evaluation and assessment become harder challenges, since the observers will be confronted with the task of analyzing and making sense of features that do not (always) belong to just one category. This line of reasoning allows us to conclude that the more the two categories will be apart and not related, the harder these tasks will be, implying that different levels of similarity to category stereotypes will significantly influence the categorization decision. Some features or functionalities might also belong to both categories or neither, depending on how common these features might be in the mind of the consumer: this condition might be more influential for categories of products that are not often object of

(21)

16

evaluation.

Because unfamiliar products like the ones used in this study often contain features or suggest functionalities derived from different categories, consumers will have to rely on multiple category labels to produce a logically sound evaluation.

The focal point of this analysis is that a category label guides the attention on features and details, enhancing the perceived relevance of the features within the most obvious category in their view while decreasing the level of relevance attributed to the features of other categories. Mapping information from one of the perceived categories to the product, observers might value a certain feature as more evident or indicative as a category label, gradually leading them to think of the object as a whole. Calling back to the initial speculation of this paragraph, we can thus formulate this hypothesis.

H1 A brand alliance product will be categorized more strongly towards a certain product category when the product features of that category are presented as dominant than when they are not (i.e. when both categories are dominant or the opposite category is dominant)

3.2 Perceived Brand Dominance

Brands can become such a strong indication of a category that they can even become a label for an entire set of products with similar characteristics (just think of Kleenex as a synonym for paper tissues). The drive toward this level of specificity is significant, simply because the ability to identify objects with ease lowers the consumer's need to evaluate brand-specific information. (Alba and Hutchinson, 1987). Some brands commit to one and only one category, while others become associated with multiple ones, which implies that the combination of two brands belonging to logically linked categories can trigger different reactions based on the context in which the brand is perceived and considered prototypical (i.e their representativeness in relation to a product category). There are differences across brands when prototypicality is concerned. The prototypicality of brands is significantly related to personal preference, which is a great influencer in how much attention and effort are spent towards the memorization and recall of a brand (Alba and Hutchinson,1987).

Research about brand extensions revealed that consumers tend to evaluate more highly those associations that involve a fit between the product and the brand that more evidently defines the category of product (Aaker, Keller,1990). This meaningful and reinforced association can be a crucial determinant in evaluating which category the product belongs to, since it is the easiest and more reliable route for an observer to find a category for the product. Alba and Hutchinson (1987) note that consumers buying a

(22)

17

product from a specific category (but do not feel sufficiently committed to compare brands), tend to go for the brand that seems the most typical in the category, because it is easier to recall. In a setting of co-branding consumers are thus more likely to perceive this fit basing their evaluation on products and brands that “point” in the same direction, rather than when the two concepts are different. We can thus conclude that the hypothesis is:

H2 A brand alliance product will be categorized more strongly towards a certain product category when

brands related to that category are presented as dominant than when they are not (i.e. when both categories are dominant or the opposite category is dominant)

3.3 Moderated Mediation And Interactions

Previous authors conceptualized product categories as sets of similar objects that have degrees of membership based on specific category attributes: the extent to which a product is considered a member is influenced by the extent to which a product possesses an attribute. (Alba and Hutchinson, 1987) The different effects that brands and product may cause in the categorization process of co-branded products never happen separately. The evaluation that consumers take when faced with multiple stimuli is mostly intuitive and is rarely a reasoned and balanced assessment. The product is always considered as a whole, and for different observers these stimuli can have different weights in understanding which product category the analysed item can belong to.

Is it thus logical to assume that these effects might interact with each other, or even that just one of them proves to be more influential in the assessment of category.

Different levels of visual and/or perceived dominance of one brand over the other might suggest a stronger tie of the product to one of the categories the brands recall, which might be (or not be) in line with what the product actually looks like. Vice versa, the look of the product might depict one or even both brands involved as perfectly fitting (or completely inappropriate) with regard to the category the product itself is suggesting.

Which one of the two ways consumers might perceive products that have these kind of characteristics is yet to be confirmed.

For the purpose of this research, we will delve deeper in the latter, in which perceived product similarity directly influences categorization while acting as a moderator in the relationship between perceived brand dominance and categorization.

(23)

18

The rationale behind this research choice is that, logically speaking, the observers will have an easier time in categorizing a product they are unfamiliar with by basing their evaluation on the simplest and less requiring task, i.e. answering the basic question “what does this look like?”. Only then, the effect of brands is expected to be taking place, as the respondents could focus on whether the two brands make sense by having a “solid” perceptual basis on which to base their assessment. The opposite process (having a first evaluation based on what the brands recall in their mind) would be difficult and somewhat forced, as this approach is not expected to provide a sound common ground onto which a complete and reliable evaluation can be developed.

Previous research seems indeed to be pointing in this direction, as the goal of categorization has been defined as the attempt to maximize similarity within categories while reducing similarities across different categories (Medin and Schaeffer 1978; Rosch and Mervis 1975).

Basing our insight on this definition, it would be illogical to assume that such process can be originated by polyvalent entities as brands (which might operate in multiple categories and can be perceived differently in different contexts) and not by the much more reliable input provided by a visual stimulus. Even if little research has been performed on how the effects of multiple category cues and brand related information can be processed together when the categorization decision takes place, past research has been exploring the individual effects of each variable during this process.

In the literature review, evidence that category cues can point in a “direction” when trying to provide contextual information to consumers was found, leading to the conclusion that individual stimuli can elicit different decisions in the categorization process.

This study, however, focuses also on the interaction of said central categorization cues, by hypothesizing that when both variables provide information relating to the same category, different effect sizes can be found in the final evaluation and in the role of each variable on the outcome.

In order to focus our attention on the implications of possible findings and by basing the hypothesis over past research, we can thus conclude that the third hypothesis is to be formulated as the following.

H3 Categorization scores will be more indicative of one category when Perceived Category Similarity and Perceived Brand Dominance are directed towards the same product category than when they conflict.

(24)

19

3.4 Effect on Outcome Variable: Categories and Willingness to pay

Understanding more deeply how respondents analyse concordant and discordant categorization cues can reveal interesting insights about their perception of value of the alliance that is being evaluated.

The setup of the study allows us to separate the sample in groups that identify the product with more certainty with respect to other groups, whose categorization scores lie in the middle points of the scale. It could be argued that products whose score lie on the extreme ends of the scale could be valued very differently: this statement in strengthened by the consideration that respondents would be very sure of the category belonging of the alliance, thus they would base their evaluation of value on the average price the item would have when associated to other members of its category (in other words, the average willingness to pay for a product belonging to one category could always be more or less than the average willingness to pay for another item in another category). It is however to be noticed that the stimuli used in the study always report two brands on the same product, whose features always recall more than one category at once. This uncertainty and the fact that the item can be seen as belonging to more than one category could influence the perception of value of respondents, which is an interesting consideration to analyse.

Moreover, even when categorization scores lie at the end of the scale (thus indicating a more certain category identification) there could still be a significant effect of the other brand over the dominant one which could then be translated in different evaluation of value by the respondent.

It is however important to focus on the extent to which this effect might happen rather than whether it might take place: in order to meaningfully contribute to the study final results, we will focus on the effect size and on the comparison of said effect on the products which are evaluated more certainly.

Basing the following formulation on the above reasoning, hypothesis 4 will be tested:

H4 The influence of Perceived Product Similarity and Perceived brand dominance on Willingness to Pay will be mediated by Co-Branded alliance categorization, in the sense that Willingness to Pay will be higher when categorization is strongly directed towards a product category than when it is more ambiguous.

Based on the formulation of the above hypotheses for this research, we can conceptualize them graphically into a conceptual model:

(25)

20

3.5 Conceptual Model

Perceived Product Similarity

Perceived Brand Dominance

Co Branded alliance

Categorization Willingness to Pay

H1 H4 H2 H3

+

(26)

21

4. Methodology

4.1. Research design

The focus of this study revolves around understanding if and how brands and product features interact significantly with the perception of a category when different degrees of the former variables are embodied in the product itself at the same time. Concrete consequences of this evaluation might then be reflected in the possible alteration of willingness to pay in response to the perception of a different category. This research objective was contextualized in a setting where the product analysed was the result of a co-branded alliance.

In order to fulfil this research purpose, the appropriate research format was determined to be an experimental design.

The main focus of the experiment was to compare evaluations in terms of perceived category and brand dominance across two different categories when the object of the observation was indeed the outcome of a co-branded alliance.

Concretely, respondents were called to fill in survey where they would have to evaluate two pictures of co-branded products belonging to different categories, seeing each picture separately in order not to influence their evaluations on the individual stimulus. Each picture would also include a short verbal description that would give some information with regard to the actual functionality of the product. In sum, the technical setup of the design was determined to be a 3x3 with product category replications. This variety of visual and verbal stimuli were then grouped into nine combinations (per stimuli set), exploring all the possible degrees in both dominance of brand and category similarity as well as category belonging.

In order to test the hypotheses of this study, the first variable to be identified was perceived category similarity. This variable was meant to express how much the physical features of the product where representatives of either category A or B. The first two levels at which this variables was tested were indeed “Dominance of Category A” (where the product would recall more prominently features and looks recognizable in category A) and “Dominance of Category B” . The middle point in this distinction was an intermediate level where both features belonging to category A and B where are presented in the product. Across the three levels of this variable, different looks and descriptions of the products were utilized in order to better accommodate the representation of this variable.

Likewise, the respective dominance of brands involved in the partnership were evaluated at three different levels, reflective of visual dominance of Brand A, brand B or equal dominance of the brands.

(27)

22

To properly represent each of these three levels the pictures were edited by adding logos and primary brand elements on the products so that they could express a stronger belonging to either brand. In both cases, the terminology “dominance” is to be considered as indicative of the condition (i.e. a stimulus at the first level of evaluation of both variables would be referred to as a stimulus where category A and Brand A are dominant). For the purpose of this study and to better understand how the research was structured, from now on we will refer to Category A and to Category B to represent this dichotomy. As far as brands are involved, Brand A is considered to be uniquely or significantly associated with product category A (i.e. a brand like Baskin Robbins is associated with the ice cream category). The same rationale is to be applied to Brand B.

4.2. Stimuli Development

The initial format of the study consequently involved three sets of nine pictures, implying a 3x3x3 experimental design which was inclusive of a total of 27 possible combinations.

Going into detail with regard to categories , the first set of stimuli involved the mix of a sports brand with a technology brand (Adidas and Samsung respectively) to depict the hybrid of sports equipment and technological devices. The second set of stimuli involved food, by depicting these stimuli as either cookies or chocolate and involving Oreo and Milka as representative brands. The third set of stimuli mixed features belonging to a digital camera to the ones belonging to a smartphone, taking GoPro and Nokia as representative brands. This final set of brands and product representations were chosen starting from a much broader group: the initial stage of this design involved 20 pairs of brands that could find a common point in a product category that could be seen as a plausible product the two brands could market together. With regards to this last aspect about the appearance of the product, the basic requirements for a product image to be considered as a possible candidate to be a stimulus were many: it was however imperative that the product could be “understood” without the need of textual information, since the role of text in the study as part of a stimulus was to give additional context, not to be the main source of information. The product should also present some aspects that would link it to a category rather than another one and it was meant to be a product that could be considered realistic and appealing. The brands that were picked in order to be part of this study were brands whose perception and familiarity were meant to be not detrimental to the results of the evaluations. For this reason, brands whose general perception was to be negative or that were not widely known were discarded. Ideally, the brands to be used in the study would have been brands strongly associated with a category, whose perception was not negative and whose name was known.

(28)

23

The initial phase of the experiment involved a Pretest both in qualitative and quantitative terms. The qualitative Pretest was simply meant to ask a small group of respondents (30) to define in one word one of the twenty-seventh stimuli above mentioned. Based on feedback and means in responses, it was possible to establish a standard quality and style for the images, to ensure that was possible for most respondents to not get lost and aimlessly wonder what the product actually was. This provided a rough but on-hand impression on how much clarity the pictures were able to recall and it allowed to avoid too much confusion in the mind of respondents.

The second phase of the Pretest involved the first draft of the survey, which included three main blocks of 9 pictures each. Each respondent was exposed to one block, and all the stimuli in the block were exploring different levels of perceived category similarity (which was a logical choice, since measuring different levels of perceived brand dominance would have meant that each respondent would have reviewed products that would have had minor changes in terms of logos, but that basically looked the same).

After a preliminary analysis of results, it became clear that the scores given in the first draft of the survey were very discordant with regard to stimulus set number three. It was then concluded that this category and these brands were not reliable enough to provide the sound data to be analysed later in the final stage of the research. This finding led to the decision to drop the third stimuli set in favour of a less extensive but more reliable analysis.

4.2.1 Quantitative Analysis of Pretest

In order to understand what were the preliminary evaluations to be formulated and in order to assess the degree of understanding of the stimuli, as well as the extent to which respondents could comprehend references to categories, the preliminary stage of the survey mainly involved questions about the functionalities of the products shown as well as more straightforward questions about their understanding. As an additional core measure, two separate slider questions directly measuring categorization (In your opinion, to what extent is this product representative of the following category?)

(29)

24

were used in order to assess on a more precise scale how much the products were perceived to be part of each category.

After collecting the relevant data (with a total of 75 respondents) several analysis were ran in order to understand if the measures adopted in the protest where suitable for a later stage analysis.

The first step in the preliminary analysis was to understand if questions related to functionalities and understanding where showing encouraging correlation values as well as satisfactory Cronbach’s Alphas. Positive results would have meant that respondents were evaluating the stimuli coherently and could understand the separation in categories. Correlation values would also give a clearer overview on whether the single items used in the Likert scale could actually measure different constructs or show contradictory behaviours, while Cronbach's Alpha on the other hand constitutes the main measure for reliability to be used in Likert scale questions.

Questions related to functionality were composed of six items each, three of which would refer to functionalities exclusive of category A, while the other three would be reflective of functionalities associated with Category B: respondents had to evaluate how much the product was able to perform the task according to their perception. Theoretically, it was hypothesized that products scoring highly in one function would be identified as members of the category where that function was representative. In the case of stimuli set number two (which was related to the food category), no functionality related questions were formulated since food has no additional functionality but the one to be eaten.

Likert scale items related to the understanding of the product were composed of four items, which were formulated similarly with each other and all tested the degree of ease the respondent had in understanding his/her visual perception.

Likewise, questions related to familiarity with the brands involved in the alliance as well as Likert scale questions whose aim was to understand the perception of brand dominance in the product were employed in the Pretest.

As described above, the first step in the analysis was to look at the values of Cronbach’s Alphas across stimuli set for questions on functionality and understanding of the products.

For functionality questions, Stimuli set 1 reported more than satisfactory values for Cronbach’s Alphas for both groups of functionalities (α =.812 and .871 respectively), while Alphas belonging to Stimuli set 3 scores significantly lower (α =.500 and .173 respectively).

These initial results immediately showed that this last stimulus set was difficult to evaluate on a functionality level. This insight cast doubt on how easy was for respondents to really understand the visual input provided. Such suspect was then reinforced by the values provided in the correlation matrix among these same items.

(30)

25 Stimuli set 3: Correlation matrix for functionalities related to Phones (left) and cameras (right)

This output shows even slightly negative correlation among the items, highlighting a significantly low reliability in the assessment of category through these items.

The core analysis, however, aimed at understanding if and how product and brand condition would prove to be relevant for the assessment of both categories: thus, two separate multifactorial ANOVA referring to both category A & B using slider questions scores as dependent variable were executed. These tests were run across all stimuli set. Results for stimuli set number 3 showed once again results that were not satisfactory.

When analysing the behaviour of category A scores (left graph), Category A+B condition is identified too closely to category A across all brand conditions, showing that little to no difference in evaluation was experienced and that these stimuli set is hard to identify when more than one category is recalled by the product. Moreover, scores for category A are significantly lower than expected, developing in a range between 63 and 65 across all brand conditions. Likewise, scores for category B are too closely place to the middle point of the scale, showing once again that the stimuli are difficult to perceive and analyse as they were intended. This tendency is repeated in the right graph, where the dependent variable is the slider question score for category B. Even if scores relating to the dominance of category B score sufficiently well, this behaviour is not reflected in the other two conditions, which again score too closely

(31)

26

to each other and in sections of the scale that do not match the expected scores (ideally, the lowest value for the condition of category A similarity should have been at least below 50). These results were further analysed in order to understand whether the product or the brand condition drove the behaviour of results. Thus, as a complementary analysis in order to understand extent of the difference in the evaluation of categories originating from a single measure, repeated measures ANOVAs where employed.

In this analysis, a factor constituting the difference in the two scores was included as dependent variable to assess how closely the products were perceived.

From the graphs, is easy to understand how the change in evaluation from one category to the other is too small to reliably confirm the separation in category perception. In either graph, no score ever falls in the lower end of the scale, meaning that the stimuli presented never suggested the belonging of the product to be to category A. Additionally, it can be inferred that respondents never strongly identifying the stimuli with category B, which suggested that the stimuli employed were not giving separate inputs and thus could be identified to be too close to each other in terms of categorization.

It was then concluded that for the purpose of consistency and reliability of the final study, this last set had to be dropped from the final questionnaire.

The insights gathered from these preliminary analysis determined to be crucial for the draft of the final questionnaire: even if the pictorial stimuli were not changed (basing this decision on the rationale that they would constitute the main source of information of respondents when called to evaluate) some slight rework was applied to the verbal description of combinations in order to better define the functionalities of the products shown.

Categories and co-branding : one product, many views

Master’s Thesis