Preselling a Different Remake: a study of category spanning and similarity of product redesigns

(1)

Preselling a Different Remake: a study of category

spanning and similarity of product redesigns

Department of Economics and Business

Jannick Engel 11850078

August 17th 2018 – Final version MSc. Business Administration

Entrepreneurship and Management in the Creative Industries track Amsterdam Business School – University of Amsterdam

(2)

2 Abstract

This research investigates if category spanning is advantageous for producers or not by testing whether category spanning has any positive effect on market performance. Category spanning is a tactic used by producers to stretch product category boundaries in such a way that a particular product does not belong to a single product category. A theoretical argument is presented which combines the outcome of existing literature on category spanning and on brand extension, thereby explaining the positive effect of category spanning. It is argued that a higher degree of category spanning in a sequentially released product has a positive effect on its market performance. Furthermore, it is argued that a degree of similarity is a positive moderator on this proposed relationship. The theoretical argument is tested by studying a database of original and remake films – consisting of films produced between 1980 and 2016, a total number of 492 sets – and by conducting an OLS regression analysis. Statistical

analysis of the results showed no significant correlation and, therefore, this research lacks statistical support for its theoretical argument. This research concludes that category spanning has no positive effect on market performance. Therefore, this research implies that category spanning might not be advantageous. However, future research is stimulated to provide revisions or look into different research opportunities, as several leading experts suggest that category spanning can indeed be advantageous.

Statement of originality

This document is written by Jannick Engel, 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 responsibly solely for the supervision of completion of the work, not for the contents.

(3)

3 Acknowledgements

This research could not have been completed if it was not for the support and guidance of many, for which I am immensely grateful.

Firstly, I would like to thank my supervisor Dr. Angelo Tomaselli for his guidance and perpetual efforts and insights. Never had I expected a supervisor to be so involved and considerate. I count myself lucky to have had the opportunity to interact with my supervisor so freely and effortlessly.

Secondly, I would like to thank Mr. René Rector for his additional guidance and his versatile views on the basis of conducting research. Our conversations were very inspirational and provided a fresh look on things. It was very useful and sometimes relaxing – especially, our conversations on failed sci-fi remakes –.

Last but, certainly, not least, I would like to express my love to my family and friends for their unconditional support and believe in me. Although, at times your inputs could be exceptionally confrontational, I am certain it has made me a better man in the end.

(4)

4

1. Introduction

Product categories are an essential characteristic of products, as product categories help the producers understand what a certain product consists of and help the consumer to identify a certain product and its boundaries (Cohen & Basu, 1987). Since product categories are important to both producers and consumers, handling and altering product categories is a rather delicate matter. Conventional literature teaches us that producers are to stick to proven and established product categories for them to be successful. However, more recent literature shows that sometimes producers are inclined not to adhere to given product categories even though they are aware of its counterproductivity. In particular, those producers try to stretch product category boundaries by spanning between multiple product categories (Hsu, 2006; Negro & Leung, 2012; Durand & Paolella, 2013, 2016). The reasons why those producers span product categories – as well as the consequences of the spanning of the product categories – differ.

This research tries to provide new insights into the relationship between category spanning and the producers’ ability to perform. Also, this research hopes to clarify why producers span categories, while conventional literature tells those producers to stay within the boundaries of the category. More specifically, this research argues that there is a positive correlation between the degree of category spanning and the market performance of a product. Moreover, this research stresses that a degree of similarity between the categories spanned has a positively moderating effect on the relation between the degree of category spanning and the market performance of a product. Thus, similarity invigorates the effect of category spanning. An explanation for this positive effect of spanning categories on market performance is possibly found in brand extension literature. While a brand extension – or a product redesign – elaborates on the success of prior products within a brand, it is still necessary for a

subsequent product to differentiate from the initial product. Spanning categories proves to qualify as such a meaningful differentiation which allows subsequent products to exploit parent brand success and still achieve significant financial success on their own (Sood & Drèze, 2006; Hennig-Thurau, Houston & Heitjans, 2009). Therefore, it is argued that a brand extension – or a product redesign – offers the right context to explain in what situations category spanning is a beneficial activity. Hence, an extension makes category spanning an activity to be considered by producers in order to achieve an outstanding market performance.

The empirical setting for this research is the motion picture industry. In particular this research focuses on remake films and its original. This empirical setting is chosen because it

(6)

6 is most likely to fit the theoretical argument; the remake films can be considered part of a brand (Bohnenkamp, Knapp, Hennig-Thurau & Schauerte, 2015) and, to make sure the remake does not become an identical copy of the original, remake productions need in some way to be differentiated from the original (Sood & Drèze, 2006). Furthermore, another recent development or trend makes this empirical setting particularly interesting. This recent trend is the motion picture industry’s growing dependence on presold franchises, especially in

Hollywood (Elberse, 2013). According to Elberse (2013) the major film production companies allocate most of their budget to films in franchises with a strong brand as the producers can be more certain that those films will become a blockbuster. This focus can be considered a brand extension strategy. Good examples in the last few decades are the

numerous superhero movies, the Star Wars Saga revival and the Jurassic Park/World reboot. This research consists of a quantitative analysis that includes a database research and a regression analysis. A database with data on remake films between 1980 and 2016 is used to collect the data for this research. With the data in the database, a construct for category spanning is created by using niche width as a measure for category spanning. Furthermore, another construct is created using a Jaccard index to account for similarity between spanned categories in original and remake film. As with Hsu, Negro and Perretti (2012) genre

spanning is acknowledged as the manner in which films can span categories. Other important data from the database is the box office result for the remake film, which is used as a deflated figure. An ordinary least squares regression analysis is executed to test whether an increase in category spanning for the remake – compared to the original – has a positive effect on the box office result of that remake. Also, the OLS regression analysis tests whether the amount of similarity in genres between the original and remake has a moderating effect on the box office result of the remake.

Results show that it might be possible that a certain degree of category spanning can be regarded as having impacted market performance, although it is not a considerable effect or even remotely a determining influence. Also, a degree of similarity among product

categories is possibly not a defining moderator in any way. Unfortunately, results proved not to be significant. Therefore, this research is unable to provide any statements on the effect of category spanning on market performance. Ultimately, this research concludes that it is unable to prove what it was set out to do and that there is no effect between the degree of category spanning and market performance. Among other directions, future research is encouraged to revise in order to give the effects of category spanning on market performance another look.

(7)

7 The remainder of this research is structured as follows. After providing an overview of

category spanning and brand extension literature on the effects on and causes of sufficient market performance, this research argues that a brand extension strategy serves as a complementing feature in explaining why category spanning can be beneficial to market performance. Next, by making use of a set of 492 remakes and originals and conducting several regression analyses, this research investigates whether the degree of category

spanning is of influence on market performance and whether similarity in categories spanned invigorates this influence. Finally, results are discussed and implications and future directions are introduced.

2. Literature review

Frequently, products get assigned to a certain product category. Producers categorize products as to make consumers aware of how to recognize and evaluate products and the product characteristics. Even if product categories evolve over time (Loken and Ward, 1990), they help consumers to identify a product and ascertain its value (Schoormans & Robben, 1997). Consumers can use categories to evaluate key product information, which are classified and differentiated in their memory (Alba & Hutchinson, 1987; Cohen & Basu, 1987). Logically, consumers have a strong preference for products with predefined, strong and clear product categories (Schoormans & Robben, 1997).

Given that consumers prefer products with precise categories, one would assume that producers define clear product categories and let their products stick to their assigned product category. However, producers are often inclined to undertake an action which suggests otherwise. This action is called category spanning. When producers choose to span categories it means that their product is not categorized in one category, but rather that the product belongs to multiple product categories (Hsu, 2006). Category spanning is generally

understood to be counterproductive for the market performance of a product. In conventional literature there exists a consensus on how combining product categories and neglecting category boundaries have a negative effect on the product’s ability to perform. It has been shown by a growing number of studies how products that span multiple product categories enjoy less attention and legitimacy and have inferior chances of survival and success (Hsu, Hannan & Kocak, 2009; Negro & Leung, 2012). With such an extensive body of knowledge

(8)

8 one would expect that producers would be wise enough to stick to the product categories that are assigned to and identified with certain products. Therefore, it is surprising that there are still organizations that continue to span product categories given the risks of inferior

performance involved. In recent research several scholars (Hsu, Negro & Perretti, 2012; Durand & Paolella, 2013, 2016) have searched and argued for several reasons that might explain why category spanning continues to be pursued by producers. On the one hand, some of those scholars are advocating a change of perspective, so the concept of category spanning will be assessed differently (Durand & Paolella, 2013, 2016). On the other hand, other

scholars (Hsu, et al., 2012 and Van Venrooij & Schmutz, 2018) – less drastically – argue that boundary fuzziness and the creation of hybrids – i.e. multiple category products – allow producers to span categories productively. Boundary fuzziness refers to the unclarity on category limits. It is suggested that when the fuzziness is high, producers find it easier to span across boundaries (Hsu, et al., 2012). In their visions (Hsu, et al., 2012 and Van Venrooij & Schmutz, 2018), hybridization and hybrid products seem to be more consistently framed as creative ways for producers to introduce new products and enter new markets.

At a first glance, Hsu, Hannan and Kocak (2009) are very conclusive that there are market penalties involved in category spanning, which is in line with conventional literature. These scholars investigate why producers who span multiple categories are likely to suffer from the social and economic consequences. According to them, previous research has only focused on either the audience-side or the producer-side of multiple category membership and its implications. In their research both views are integrated. This integration leads to a more systematic and harmonized explanation of how the market penalties emerge from the audience-side and producer-side processes (Hsu, et al., 2009). The audience-side view explains that the negative effects from category spanning emerge because audience members use traditional product categories to evaluate products. Products that fuse elements of multiple categories are hard to evaluate or classify by audiences and are thus less alluring. The

producer-side view describes that spanning categories diminishes a producer’s capability to target the certain audiences adequately and completely. Conclusively, Hsu, et al. (2009) state that their analysis provides support for the notion that both audiences and producers

contribute to the negative consequences of category spanning. Although, Hsu, et al.’s (2009) results rule in favour of the conventional literature on category spanning, they do leave remarks that might suggest otherwise. Namely, Hsu, et al. (2009) indicate that more research which looks into the compositions of audience members and the relations that will follow is necessary. Especially, the influence it will have on product categories. By pointing out that

(9)

9 significant overlap in elements of different product categories probably has an effect on how audience members observe and understand category spanning products, Hsu, et al. (2009) imply that ambiguity might lead to different results.

A different result is presented in later research by Hsu, Negro and Perretti (2012). In this research Hsu, et al. (2012) find that category spanning allows producers to combine widely recognized and accepted characteristics in unradical, yet different ways that provide various atypical product categories. Hsu, et al. (2012) propose the argument that even though there is a significant risk involved when producers do not comply to category boundaries, particular contextual conditions may improve the realized attractiveness of category spanning. The aim of their study is to develop awareness of how structuration processes of product categories shape the tendency of producers to develop hybrid products instead of

pure-category commodities (Hsu, et al., 2012). In this case hybrid products are defined as products subjected to category spanning – i.e. products with multiple product categories –. In this research Hsu, et al. (2012) test whether a film is more likely to span categories – i.e. genres – when its primary category has higher boundary fuzziness. Specifically, they test whether a film is more likely to span genres if the primary genre is associated with a higher diversity in secondary genres. Furthermore, they test whether producers are more likely to exit the market when they produce a higher amount of hybrid products and they test whether a producer is more likely to achieve an exceptional success when that organization produces a higher amount of hybrid products. Hsu, et al. (2012) conclude that their results point to the existence of several forces stimulating hybridization among producers. More specifically, fuzzier category boundaries are connected with a less well-defined approach for arranging products and therefore it motivates producers to span category boundaries in their products. What is more, high degrees of similarity cause an increase in competitive overlap between producers. As a consequence, the perks of differentiation increase in payoff and thus producers are stimulated to actively seek ways to successfully combine product categories (Hsu, et al., 2012). Overall, the results suggest that producers’ positioning decisions, categorical freedom and constraint in multiple category markets develop gradually and in an iterative manner. In terms of hybridization effects on performance and exit rates, results point out that hybrid products consistently influence the stability of the producers’ position. In other words, producers who rely heavily on higher amounts of hybrid products have significantly less chance of survival. Although Hsu, et al. (2012) efforts show that ambiguous categories stimulate hybridization and that innovating through hybridization creates potential for exceptional performance advantages – thus providing a plausible explanation on why

(10)

10 categories are spanned by producers – there is still much debate on the subject and how it should be approached.

Durand and Paolella (2013, 2016) are contributing to this debate by arguing for a change of perspective on category spanning and its implications. With a change of

perspective, the scholars mean moving away from the conventional prototype view on product categories. The prototype view defines a product by a set of features that are considered part of a certain category. In this consideration each feature is evaluated and scored on how well it fits a category. In this view prototypicality is operationalized by how well given products correspond with intended meaning by a category label and the perceptions that people base their judgement on (Cohen & Basu, 1987). Durand and Paolella’s (2013) argument opposes this prototype view. First of all, they do recognize that research based on the prototype view has provided an extensive foundation of knowledge that is able to explain how product categories impact, limit and shape organizational success (e.g. market performance). However, Durand and Paolella (2013) find that two lines of thought might be able to more clearly explain why producers resort to category spanning activities and productions. This plausible explanation is therefore in support of their argument to change perspectives on category spanning. According to Durand and Paolella (2013) these different views are respectively the causal-model approach and the goal-based approach. Compared to the

prototype view the causal-model approach is a more audience-based theory, which means that inventive associations by audiences of different elements define a category. The basis for the goal-based approach is that audiences’ or consumers’ perceptions of the world is continually changing . Over time, consumers do regularly pursue different, specific goals. When this happens consumers put together different elements of different domains of knowledge more easily – even traversing across time periods – and create new product categories (Durand & Paolella, 2013). Compared to the prototype view category membership in the causal-model or goal-based approach is not merely a checklist of elements to determine whether a product is regarded categorically pure or whether that product belongs to spanned categories. Therefore, category spanning is evaluated differently and the benefits of adhering to product categories must be reconsidered. The causal-model approach implies that categorical purity may be of less value than having fewer elements. Judging from insights of the goal-based approach, producers – as well as consumers – occasionally might be able to make more sense of a product that is assigned to multiple product categories, than when such a specific product is affiliated with only a single category. Based on these new insights on product categories and category spanning provided by the causal-model and goal-based approach, Durand and

(11)

11 Paolella (2013) contend that audiences might be more favourable and accepting towards hybrid organizations who dare to span categories. Thus, by arguing for a change of perspective Durand and Paolella (2013, 2016) prove that category spanning does not

necessarily have solely negative social and economic effects. Accordingly, this can provide a plausible explanation as to why producers keep spanning product categories, something which conventional literature based on the prototype view was unable to do.

Pontikes (2010) also tries to provide a plausible explanation. Although this explanation is less radical as her research is not arguing for a change of perspective. She highlights the role of different audience structures and its relation with category ambiguity. Category ambiguity can be a result of category spanning as continued spanning efforts might lead to discussion about the extent of a category. In this research category ambiguity shows the degree to which a product has an explicit identity or not. An ambiguous category may be universally acknowledged and used, but it is not specifically clear where the boundaries of the category lie, nor has it universally accepted social significance. Furthermore, such a category lacks the qualities to assure the consumers of what to expect and what not. However,

ambiguous categories can be perceived as less constraining and thus offer producers and consumers the opportunity to be more flexible (Pontikes, 2010). In a way category ambiguity and boundary fuzziness are comparable – or related – as they both indicate a product

categories’ degree of unclarity. Also, an ambiguous product category usually has less well-defined category boundaries than an unambiguous product category. Pontikes (2010) concludes that whether a product with an ambiguous category is appealing or unappealing depends on the audience that is assessing the product. To illustrate, if a certain audience uses product categories to search for particular products, category ambiguity might make those products unclear and thus harder to find. On the other hand, if a certain audience values novelty and is looking for innovation, that same category ambiguity can become an asset as it will provide more flexible products. By introducing category ambiguity Pontikes (2010) builds on the effects of audience perceptions on evaluations – i.e. social or market

performance – of a product introduced by Hsu, et al. (2009). However, in opposition of Hsu, et al.’s (2009) view, Pontikes (2010) emphasizes that whether the effect of category spanning and category ambiguity on the social or market performance of a product depends on the composition of the audience.

Category ambiguity – or fuzziness – is also a topic of discussion in the work of Van Venrooij and Schmutz (2018). In their research Van Venrooij and Schmutz (2018) test how and why the effects of fuzziness are conditionally dependent on institutional and professional

(12)

12 differences. They do this by comparing the outcomes of category ambiguity. This research leads to the conclusion that category ambiguity has negative implications for the commercial subfield, but not so much for the artistic subfield. This conclusion is built on the central argument that fields of cultural production can differ in how intensive their classification systems have been institutionalized and hence react differently to category fuzziness. Subsequently, large-scale commercial subfields are likely to have more strongly

institutionalized categories – in that those categories are broadly accepted, understood and more rigid – where category fuzziness can become an inconvenience. Otherwise, in more restricted, artistic subfields which enjoy more dynamic and emergent categories, category fuzziness does not lead to devaluation and has potential performance advantages (Van

Venrooij & Schmutz, 2018). Because of the connection with the artistic subfield audiences are more likely to appreciate category spanning, which leads to more critical acclaim and higher artistic valuations. This research indicates that category spanning could be beneficial in the artistic field and has a similar effect on critical acclaim, as these positive evaluations are a solid indicator of innovation through hybridization of product categories. For producers in artistic subfields fuzziness signals a creative incentive to innovate, which is valued by the market (Van Venrooij & Schmutz, 2018).

The above discussed researches and insights are illustrative for the profound academic debate on category spanning. Contemporary literature reveals that if we fully want to

understand why category spanning and categorical fuzziness is an asset instead of a liability to some producers, we must either understand the context of the market in which those

producers are operating, or we need to leave the prototype view and see category spanning in a different light. In light of the arguments for contextual understanding and the call for a change of perspective presented, it is obvious that the debate on category spanning and its subsequent implications is far from being a resolved issue. Furthermore, category spanning has been researched in a vast variety of markets, ranging from wine (Negro, Hannan & Rao, 2011) to as far as venture capitalism (Pontikes, 2010). However, so far there has been no research on category spanning and hybridization of products that had already proved to be successful, but where – at a later stage – producers still choose to span its product categories. This happens for products which are part of a brand or products that are derived from a parent brand. For instance, the Marvel Cinematic Universe in recent years. Therefore, a plausible explanation as to why category spanning in sequential products can be beneficial might lie in brand extension literature. This is due to the fact that brand extension – or product redesign – is able to generate spill over effects, which will enable extension products to profit from its

(13)

13 strong, successful parent product (Hennig-Thurau, Houston & Heitjans, 2009; Knapp,

Hennig-Thurau & Mathys, 2014; Bohnenkamp, Knapp, Hennig-Thurau & Schauerte, 2015). This research will try to shed a light on the implications of category spanning in the case of sequential products by incorporating brand extension insights. The focus on sequential products is chosen for it is argued that in such a case the results will have a more valuable significance, because of the ability to benchmark with the previous market performance of the initial product. In the end, if the market performance of the sequential product deviates from the market performance of the initial product and there is a significant effect of category spanning, future research can be more conclusive on the effects of category spanning on market performance.

3. Hypothesis development

The previous section deals with either going for a change of perspective or understanding the context of the market in order to comprehend and pursue category spanning and benefit from its outcomes. In this section it is argued that due to contextual characteristics of the market, category spanning is beneficial for the market performance of products and, thus, profitable for producers. Those contextual market characteristics are to be considered consequences of a brand extension strategy. In this case, brand extension strategy leads to a product redesign. This product redesign represents a situation where a sequential product is introduced to the market that is related to the initial product, but is differentiated in some way (Sood & Drèze, 2006).

However, simply making use of the parent brand’s image in a product redesign is usually not a guarantee for instant success. Product redesign needs contemplation and it is exceptionally elusive which elements will determine the degree of success. Overall, Völckner and Sattler (2006) find that the fit between an extension and the parent brand is the most important determinant of the success of the product redesign. After a close fit comes

marketing support, parent-brand conviction, retailer acceptance and parent brand experience. In addition, Albrecht, Backhaus, Gurzki and Woisetschläger (2013) also conclude that a close fit between the parent brand and its sequential extension is a determinant of product redesign success. Likewise, the resulting success of the extension has an amplifying effect on the parent brand value. Judging from the insights presented, there seems to be a complex, yet

(14)

14 valuable relationship between a parent brand and its extensions through product redesign. Furthermore, historically it has been considered crucial that a producer who engages in product redesign does so in different product categories. According to Boush and Loken (1991) customer evaluations are negatively influenced by the extension’s overlap in product category with that of the initial product of the parent brand. However, Boush and Loken (1991) also conclude that significant variation among the present products on offer has an influence on product redesign evaluations too. These findings stress the importance to differentiate from the parent brand just enough to make the extension product successful.

Also, in brand extension literature there is evidence that previous triumphs in a series of products – or sequentially released products – do have a positive effect on the performance of the succeeding products. Based on this assumed relationship of parent product and

extension there has been a substantial body of research further exploring experience goods in the context of a brand extension strategy (Hennig-Thurau, Houston & Heitjans, 2009; Knapp, Hennig-Thurau & Mathys, 2014; Bohnenkamp, Knapp, Hennig-Thurau & Schauerte, 2015). Every one of those scholars show – by investigating different product redesign (e.g. books, films) – that extension products can benefit from the success of the parent product. In other words, a successful parent product influences the market performance of a product redesign. What is more, Sood and Drèze (2006) argue that in some cases – i.e. experience goods – changing product features for the extension compared to the parent product is valued by the customer.

Since it has been shown that certain contextual market characteristics can lead to positive effects of category spanning (Hsu, et al., 2012), it is argued that a product redesign in combination with category spanning has a positive effect on market performance. In other words, while brand extension literature indicates that a sequentially released product will likely benefit from the success of the initial product if the subsequent product is differentiated in some way, it is argued that these are the specific contextual characteristics that will foster the positive effect of category spanning. In this case, category spanning is the technique that producers use to differentiate the product redesign from the parent brand. Considering these insights, the first hypothesis is:

Hypothesis 1:When a sequential product spans more categories than the original product it derives from, the spanning has a positive effect on the market performance of the sequential product.

(15)

15 Since further extensions to an initial product are the objects of discussion here, it is possible to assume that at least some of the product categories between the first and second – or third – product overlap. According to Slavich and Castellucci (2016) there is an inverted U-shaped relation between similarity between product categories of sequentially released products and the performance of those products. However, Slavich and Castellucci (2016) do stress that – even though products should display some similarity with previously successful and earlier released products – the inclusion of elements of novelty is necessary to earn higher

performance or evaluations. Consumers are likely to expect some degree of similarity

between sequentially released products and will therefore be more positive towards or will be more willing to pay for a product that is comparable to an earlier successful product. It is therefore argued that consumers will evaluate those products more positively if there is a degree of similarity, because consumers will associate the new product with the old one. Nevertheless, although similarity is valued, novelty is still expected. If novelty is not present at all, consumers will penalize those products – or producers – as no novelty is regarded as a sign of not developing an own product or a lack of innovation. On the other hand, if similarity is completely missing, consumers will also penalize those products – or producers – because consumers are unable to relate it to the new product (Slavich & Castellucci, 2016). Hence, the aforementioned inverted U-shaped relation.

Kovács and Hannan (2010, 2015) argue that the negative consequences – or market penalties – of not respecting category boundaries are becoming more severe when categories spanned are more distant and contrasting to one another. Correspondingly, the combination of high-contrast categories leads to increased confusion, because high-contrast categories are met with greater expectations and norms (Kovács & Hannan, 2015). In their research on the effect of niche fitness and audience confusion on devaluation or ignorance of category spanning products Keuschnigg and Wimmer (2017) elaborate on the argument made by Kovács and Hannan (2010, 2015). Interestingly, Keuschnigg and Wimmer (2017) conclude that the confusion mechanism has a relatively more significant role than the niche fitness mechanism in clarifying audience responses towards spanned genres. Therefore, ignoring category boundaries is not necessarily unfavourable for producers. Neither for critical appraisal nor for the market performance (Keuschnigg & Wimmer, 2017). Nevertheless, the scholars do stress that their results indicate that market penalties of category spanning only occur if the spanned categories in question are significantly, culturally distant.

As discussed before, the perceived fit between an extension product and the parent brand is a key determinant for the success of a product redesign. Park, Milberg and Lawson

(16)

16 (1991) emphasize the importance of this fit by providing a more profound understanding of what contributes towards the fitness between initial and extension product. Namely, Park, et al. (1991) conclude that the market performance of a product redesign does indeed depend on the perceived fit with the parent product. More specifically, this perceived fit is a function of product feature similarity. In other words, if a potential, successful new product redesign is evaluated by the market, it appears that consumers individually will include similarity with the parent product into consideration (Park, et al., 1991).

Clearly, in the case of sequentially released products and product redesigns it is important that the subsequent product is affiliated with the initial product by the consumers. Therefore, similarity is an important feature between the products, as a degree of similarity will contribute positively to the fitness between the extension and the parent product. As a clear fit between sequentially released products is a determinant of the market performance of subsequent products, any form of similarity becomes an essential necessity to establish market success for an extension. Thus, for a series of products and redesigns, a degree of similarity is key in brand extension literature. Interestingly, in category spanning literature a degree of similarity in categories seems to generate understanding among consumers. Moreover, consumers seem to value similarity in multiple category products. Still, consumers value some degree of novelty and thus appreciate innovative behaviour. Furthermore, if – culturally – distant categories are spanned it will lead to confusion among consumers. Those consumers will be less able to recognize a products’ value and thus be more likely to reject a product. Therefore, the second hypothesis is:

Hypothesis 2: When a sequential product spans more categories than the original product it derives from, the similarity between the categories of a sequential product and an original product has a positive moderation effect between the spanning and the market performance of the sequential product.

4. Data and method

4.1 Research design

This research is considered to be a quasi-experimental research. Accordingly, this research will be conducted with the use of an ordinary least squares (OLS) regression analysis. The dependent variable in this analysis will be the market performance of a remake film, as to be

(17)

17 determined by the box office performance collected from the Internet Movie Database

(IMDb). The independent variable will be a variable that correctly signifies whether or not a remake is subject to category spanning. More specifically, the independent variable puts a score on whether the remake differs from the original in genre in total number of genres used or use of other genres. This will indicate if category spanning has been applied to the remake or not. As previously discussed, once category spanning for a particular film has been

determined, it can be derived that hybridization is present because multiple category products and hybrid product are essentially the same (Hsu et al., 2012). Altogether, this regression analysis will determine whether category spanning has an effect on the perceived box office success of a sequel product, in this case being the remake film.

4.2 Empirical setting

The empirical setting of this research is the motion picture industry. In particular, this research concerns films that have been remade into another motion picture for whatever reason. Therefore, the units of analyses in this research are the sets of original film and remake film. This empirical setting is chosen, because it possesses some unique

characteristics that are valuable to this research.

First of all, remakes are a specific and unique kind of product. A remake implies that it is an identical reproduction of the contents of the original. Therefore, it tells a story that has been told before. According to Bohnenkamp, et al. (2015) remakes differ fundamentally from other extensions, because remakes offer familiarity and limited sensation value due to a lack of novelty. This lack of novelty arises, because remakes do not elaborate on or expand a known story, such as with sequels. Neither do remakes offer a particular new kind of sensory experience, as with book adaptions. However, completely identical remakes – as for instance the shot-for-shot remake of Psycho – do not seem to be attractive to consumers (Bohnenkamp, et al., 2015). Rather, remakes still need a feature of novelty in order to be a successful

extension product – or product redesign –. Product category spanning might offer this novelty, as has been illustrated by Hsu, et al. (2012) with the introduction of hybrid products.

Secondly, according to Elberse (2013) the motion picture industry is increasingly reliant on the extent of particularly strong parent brands – appropriately being dubbed as presold franchises by Elberse (2013) – to produce blockbusters. By exploiting the successes of the presold franchise, the major film production companies are trying to deal with the high uncertainty that is an inherent part of the motion picture industry. It might well be that a remake is regarded as a more simple form of exploiting a presold franchise compared to a

(18)

18 sequel or an adaption, but it is still a contemporary form of production that is widely used and it is still able to perform at the box office. Therefore, it enables producers to deal with the uncertainty in their industry. With the growing reliance on presold franchises and the endless search for blockbuster productions, the remake as an extension has become an interesting unit of analysis.

4.3 Methodology

4.3.1 Sample and data collection

The underlying population of this research is all remake films – of minimum 70 minutes of length – produced since 1980 until 2016 in the USA (production country) or in the USA together with other countries. The total number of remakes in the population is 492 films. The final sample comprises 483 films and was drawn on a convenience sample – i.e. all remakes were selected for which the data allowed to test the relation between category spanning and market performance –. The data on the sets of original and remake films are collected from a database on film data which has been compiled by and belongs to Dr. Angelo Tomaselli, who is an Assistant Professor at the Amsterdam Business School. The data has been stored in an OpenOffice file.

4.4 Variables and measures

4.4.1 Dependent variable

Market performance. To determine the effect on success, this research considers the total box office results as a valid measure for financial success. In order to operationalize market performance, data from the database is used. However, in the database there are several measures that report different box office figures. For instance, there is a variable that depicts total box office performance, a variable that represents opening week box office performance and a variable that illustrates weekend box office performance. In this research the variable boxoffice_def_tot will be used as the variable that serves as a justifiable depiction of the market performance of a remake film. The variable boxoffice_def_tot is the total box office result of a film, both domestic – i.e. first country of production – and foreign – i.e. worldwide –. Furthermore, all values of boxoffice_def_tot for every remake film are present and all values are deflated. Therefore, the market performance of all remake films between 1980 and 2016 are available in the database and represent correct measurements for the analysis.

(19)

19

4.4.2 Independent variable

Niche width. Central in this research is whether or not deviating from original product categories has an effect on the performance of a sequel product. To operationalize this

deviation, this research will look at whether or not the sequel product is subjected to category spanning. This means that the remakes will be scored on distinction – i.e. if the total genres of the remake compared to the original deviate –. According to Tang and Wezel (2014) niche width is a valid measure to score category spanning, because it has already been used by Hsu, et al. (2009) and it is based on Simpson’s (1949) index of diversity. This index of diversity is a standard measure of the variety of a distribution over a collection of discrete categories. Generally, one could argue that the more genres a film claims, the more common its character is (Hsu, 2006). However, Tang and Wezel (2014) note that a specific film might be

categorized different by one agent than by another, because it is susceptible to their interpretations respectively. In this research genre categorization per film is based on data provided by the Internet Movie Database (IMDb), thus only one agent. First of all, the grade of membership (GoM) is calculated for every film. The grade of membership is a score between 0 and 1 and represents the degree to which a film belongs to a certain genre – i.e. full genre membership only exists for one genre films –. To calculate the GoM one is divided by the total number of genres that a particular film has – e.g. if a film has three genres then the GoM is 1/3 –. Next, with the use of the GoM niche width is calculated. Niche width is one minus the sum of the grade of membership in each category elevated at the square term – e.g. if a film has three genres then the niche width is 1 – [(1/3)^2 + (1/3)^2+(1/3)^2] –.

Furthermore, the niche width also ranges from 0 to 1. However, in this case a larger value indicates a larger niche width, which would indicate category spanning – thus, not full genre membership –. To determine whether the remake spanned more genres than the original, the niche width of the original is subtracted from the niche width of the remake.

4.4.3 Moderating variable

Similarity. In this research it is argued that when the remake and original film overlap in genres this will enhance the box office performance of the remake. This overlap is operationalized as a degree of similarity. Thus, similarity in genres between the original product and sequel product has a positive effect on the market performance of the sequel product. Similarity is calculated in the same way as has been done by Slavich and Castellucci (2016). This means that to measure similarity a Jaccard index is created. First, the data on the genres for the original film and the remake have been collected. Next, a variable is created

(20)

20 which scores the amount of genres that are present in the description of the original film and of its remake. Subsequently, a variable is created that contains the total number of genres of the original film and the remake combined. To create this variable each genre is only counted once, as to make sure that if both the original and the remake have a same genre it is not double weighted. For instance, if the original has 3 genres, the remake has 4 genres and the overlap is 2 genres then this set gets a score of 2/5 – i.e. 2 overlapping genres plus 1 non overlapping original genre plus 2 non overlapping remake genre equals 5 genres in total –. When both the variables are created the Jaccard index represents the sum of shared genres between original and remake over the total number of genres in the same set of original and remake. This gives a reported ratio that represents a score of similarity for genres in both films.

4.4.4 Control variables

To control for other possibilities and to rule out any alternative explanation several control variables are included in the analyses. Accordingly, by including those control variables in the analyses, it is concluded that the positive effect on market performance cannot be explained in any other way. The control variables included are star power and recency.

Star power. The variable star power accounts for any experience a high ranking crew member – e.g. screenwriter, director, top billing actor – has in previously, financially

successful film productions. Star power is operationalized by scoring the top billing crew on whether they have been involved in one – or multiple – of the top ten grossing films per year since 1980. The total star power for a remake film is to be understood as the sum of the individual star power of the director, actor(s) and writer(s) which has been previously accumulated. According to Tang and Wezel (2014) the blockbuster experience gained by actors and director functions as a proxy for the skills and expertise available and can therefore affect the market performance of a film. Similarly, Slavich and Castellucci (2016) state that in their research on apprentices’ similarity in their products to their former masters’ products and its influence on critics’ evaluations, they need to control for the apprentices’ former

evaluations and previous experience. Evidently, because of these arguments it is imperative to control for this possible relation between star power and market performance, as to make sure that the effect measured is really due to category spanning.

Recency. The variable recency accounts for the timespan between the release of the original and the remake film. Also, it accounts for the differences in timespan between the sets of

(21)

21 original and remake film in the analysis, as a given remake film is not necessarily released after exactly the same passage of time as all the other remakes. Recency is operationalized as the difference between the year the original film is released and the year the remake film is released. In other words, the number of years that have passed between the release of the original and remake film. This analysis is controlled for the time that has passed between the set of films as the difference in timespan might be a plausible explanation as to why a certain remake seems to perform better than another. Likewise, Slavich and Castellucci (2016) control for years elapsed since an apprentice left the master, as value created with the exposure to the master’s skills might decrease over time. The same holds for remake films, although it works a little differently. According to Bohnenkamp, et al. (2015) research on film brand extension has proven that recency is positively linked with the market performance of sequels. Adapting this mechanism to the remake concept means that recency can likely be beneficial as well as counterproductive. For instance, a huge timespan between releases might instigate attracting a new target audience, but the huge timespan might also mean that

familiarity of the original is lost. On the other hand, a small timespan can ensure the familiarity of the original is still of value. However, because of the small timespan novelty and, thus, the sensation value of the remake will be narrowed (Bohnenkamp, et al., 2015).

5. Results

In table 1 the descriptive statistics for the analyses is presented. Of the remake films 36,0% have a GoM of 1/3 – which is the most frequent number of genres in a remake – and only 11,6% of remakes have a full genre membership – i.e. their GoM score is 1 –. For the original films a valid 32,7% have a GoM of 1/2, which means that original films in approximately a third of the cases have 2 genres. Also, full genre membership is a little higher for original films than for remake films. Namely, 18,4% of the cases of original films have a GoM score of 1. The difference in GoM score between the original films and remake films is for 141 cases 0. This is the most frequent difference measured and accounts for 29,2%. In 29,8% of the cases the grade of membership of remake films is higher than the grade of membership of original films. Furthermore, in 41,0% of the cases subtracting the GoM score of remake films from the GoM score of original films equals a higher GoM score for original films.

Subsequently, descriptive statistics for the niche width provide numbers for original and remake films as follows. 177 cases of remake films have a niche width of 0,667 – or 2/3

(22)

22 – and this represents 36,0% of the cases. Obviously, this corresponds with the GoM score of 1/3, seeing that the niche width score is the inverse of the GoM score. For that reason, the numbers also show that 11,6% of remake films have a niche width score of 0. Likewise, in case of the original films the research reports that 158 cases – or 32,7% – have a niche width of 0,500 and that 89 cases – or 18,4% – have a niche width of 0. Furthermore, the difference in niche width – to no surprise – is most frequently 0, thus 29,2%. The niche width is in 41,2% – which is ~ 41% – of the cases higher for remake films than for original films. Moreover, the niche width is in 29,6% – which is ~ 29,8% – of the cases higher for original films than for remake films. Again, this corresponds with the scores that are found for grade of membership, as it has been established that the GoM score is the inverse of the niche width score.

Last but not least, the descriptive statistics for the Jaccard index – which scores the similarity of genres between original and remake – report the following. In 106 cases – which is 21,9% – the Jaccard index scores 0,500 – or 1/2 –. Furthermore, in 4,8% of the cases the Jaccard index scores 0 and in 16,4% of the cases the Jaccard index scores 1. This comes down to 23 and 79 cases respectively. Accordingly, this means that 106 cases overlap in half of their genres, 23 cases of remake films and original films do not overlap in genres and in 79 sets of films 100% of their genres overlap.

Table 1: Descriptive statistics

Variables* Mean (SD) Min Max

GoM remake 0,428 (0,230) 0,143 1,000

GoM original 0,489 (0,264) 0,125 1,000

GoM difference 0,057 (0,305) -,833 0,800

Niche width remake 0,572 (0,230) 0,000 0,857

Niche width original 0,511 (0,264) 0,000 0,875

Difference niche width -,059 (0,303) -,800 0,667 Jaccard index similarity 0,544 (0,273) 0,000 1,000

*Regressors of analysis of likelihood category spanning positively effecting market performance

In table 2 the descriptive statistics for the control variables are reported. The control variables are star power and recency. Surprisingly, in 78,3% of the cases – or 385 in total – total star

(23)

23 power is 0, meaning that no one working on the remake had previously worked on one of the top 10 blockbusters between 1960 and 2016. Furthermore, in only 0,8% does the total

experience in top 10 blockbusters for the production team of a remake equal 3 – which is the highest score –. As for recency, in 25 cases the years that elapsed between the release of the original and the release of the remake is only 4. This amounts to 5,1%. The least amount of years elapsed between the releases of original and remake is 1, which happened in 2,6% of the cases – or 13 cases –. The largest timespan between original film and remake film is 113 years and this happened in only 1 case – or 0,2% –.

Table 2: Control variables

Variables* Mean (SD) Min Max

Star power 0,26 (0,543) 0 3

Recency 23,74 (17,485) 1 113

*controlled for as possible determinants of market performance in regression analysis

In table 3 this research summarizes the key figures of the first regression analysis. In the first model hypothesis 1 is tested. In this regression analysis niche width difference is used as the independent variable and star power and recency are used as control variables. What

immediately becomes apparent is the relatively low R square score of ,113. This means that model 1 possibly explains for only 11,3% of the change in the dependent variable, which is total box office performance. Therefore, this research can already conclude that there is no correlation between the independent variables and the dependent variable. Subsequently, what is thereafter observed in further analysis, is that niche width difference has an unusually deviating significance level (p=0.542), which is way too high in order for results to be significant. Appropriately, the significance level of star power has a score which would indicate significant results (p<0.05). The same goes for recency, which has the same significance level (p<0.05). These striking results mean that the difference in niche width between the original and remake film does not effect the total box office performance. Also, this means that a higher degree of category spanning in a sequential product – compared to the product it is derived from – does not lead to an increase in market performance. Therefore, the first hypothesis is rejected.

For what it is worth, in model 2 the second hypothesis is tested. In this regression analysis niche width difference is used as the independent variable, the Jaccard index –

(24)

24 scoring similarity between genres – is used as the moderating variable and star power and recency are, still, used as control variables. In the model summary R square is reported to be ,114. This means that model 2 – including the moderator – is presenting nearly the exact same number as model 1 (R2=0,114 over R2=0,113). Therefore, this research can already conclude that model 2 does not differ in explaining the change in the dependent variable. Consequently, it is has become evident that in the second model there is still no correlation between

dependent and independent variables. In further analysis, this research finds that star power and recency are still showing appropriate significance levels (p<0.05). However, the Jaccard index and, thus, the degree of similarity in genres has a reported significance level of ,693 (p=0.693), which means that regression analysis results for similarity are not significant. What is more, the significance level of niche width difference is even higher in model 2 (p=0.627) than in model 1 (p=0.542). This means that results for the degree of category spanning are still not significant. For these reasons, this research suggests that a degree of similarity in genres between two subsequent products is not a moderator of the relation between niche width difference and box office performance. Also, it is established that a higher degree of similarity – in categories of two related products that use category spanning – does not enhance the presupposed positive effect of category spanning on market

(25)

25

Table 3: Regression analysis output niche width difference and Jaccard index

Variables Model 1 Model 2

R square ,113 ,114 R square change ,113 ,000 F 20,408 15,318 Sig. ,000* ,000** B β Sig. B Β Sig. Constant 27500728,64 _× ,000 29846117,44 _× ,000 Niche width difference -5643905,45 -,026 ,542 -4659954,86 -,022 ,627 Star power 34552905,04 ,289 ,000 34618970,97 ,290 ,000 Recency 658264,76 ,178 ,000 657140,70 ,177 ,000 Jaccard index similarity × × × -4185117,88 -,018 ,693

*Predictors: niche width difference, star power, recency

** Predictors: niche width difference, star power, recency, Jaccard index similarity

In table 4 this research offers an attempt to produce alternative results by excluding some sets of original and remake films, which can be regarded as potential outliers. These sets are discovered and, subsequently, isolated by making use of standardized residuals and considering every value that exceeds a threshold value of 3,000 as a potential outlier. By making use of this threshold value a number of potential outliers (n=12) are identified. These outliers are the remake films: Transformers, Dances with Wolves, The Jungle Book, Fatal

Attraction, The Fugitive, 3 Men and a Baby, Tarzan, Airplane!, Sommersby, Meet the Parents, Planet of the Apes and What Women Want. Subsequently, a regression analysis is

conducted excluding the outliers. However, the results show only a minor improvement of the R square score (R2=0,168 over R2=0,114). What is more, the figures presented for niche width difference and Jaccard index similarity – although slightly improved – are still not significant, because both figures show scores that represent significance levels that are too high (p=0.539 over p=0.627 and p=0.504 over p=0.693, respectively). Both control variables report the same significance level as before (p<0.05). Interestingly though, the reported negative effects of niche width difference and Jaccard index similarity seem to have shifted to positive effects.

(26)

26 This might be due to the impact that the potential outliers (n=12) have on the dependent variable, as those remakes performed considerably well at the box office. All in all, in this analysis without potential outliers, the regression analysis is still not capable of producing desired results and, therefore, also in this case both hypotheses are rejected. Even so, it is highly disputable whether excluding these outliers (n=12) from the analysis is entirely fair, because these films can be considered as remakes and, thus, offer valid data. Therefore, this research states that these sets of original and remake films should be included into the analysis.

Table 4: Regression analysis output without potential outliers*

Variables Model 1 Model 2

R square ,167 ,168 R square change ,167 ,001 F 31,286 23,549 Sig. ,000** ,000*** B β Sig. B Β Sig. Constant 21745164,41 _× ,000 18560238,34 _× ,003 Niche width difference 6114153,55 ,035 ,413 4754812,56 ,027 ,539 Star power 32875695,15 ,337 ,000 32780461,90 ,336 ,000 Recency 696671,05 ,232 ,000 698166,87 ,232 ,000 Jaccard index similarity × × × 5652596,68 ,029 ,504

*Twelve sets of original and remake films are excluded from analysis due to exceptionally high scores **Predictors: niche width difference, star power, recency

*** Predictors: niche width difference, star power, recency, Jaccard index similarity

(27)

27

6. Discussion and conclusion

This research has set out to prove that category spanning can be beneficial for the market performance of a product in particular markets. The objective was to show that for brand extension products – or product redesigns – the market performance is positively influenced by the fact that the redesign has a higher degree of categories spanned than the product it is derived from. Also, it has been argued that similarity between the product categories of the sequential product and the product categories of the initial product has a moderating effect on the relationship between the degree of spanned categories and market performance. The higher the degree of similarity, the higher the moderating effect on the established

relationship. In the same way as Hsu, et al. (2012) have proven that hybrids in Hollywood can be financially successful products, this research has been eager to conclude that brand

extensions and presold franchises do in fact enjoy higher market performance due to using the spanning of categories as its feature of novelty. This research is unable to provide any

conclusive statistical evidence that backs the theoretical argument of this research as the results of the analyses have shown to be not significant. Therefore, this research withholds to make any hard statements on the presupposed effects or expected relationships. Overall, this research finds that more contemplation on the topic – theoretically and methodologically – is needed to create a situation where a fruitful research with the necessary significant results can be conducted.

The first of regression analyses on difference in niche width between original and remake film and box office performance of the remake was to show a combination of a larger niche width and a higher box office performance. However, the analysis shows a negative effect of difference in niche width on box office performance instead of a positive one. That being said, with an standardized error of three times the size of the coefficient and a

significance level that is far beyond anything that is remotely acceptable, these results are not significant and, therefore, not conclusive. Removing the outliers – that have become apparent in the first regression analysis (n=12) – changes the negative effect of difference in niche width on box office performance to a positive one in a subsequent regression analysis. However, the significance level still points to non-significant results and it is highly

questionable or disputable whether it is fair to leave the outliers out of the analysis as they are justly remake films and, thus, offer valid data.

Secondly, further regression analyses should have been able to expose the positively moderating role of similarity between product categories of the original film and the remake

(28)

28 film on box office performance. In other words, similarity should have been capable of

enhancing the effect of category spanning on market performance. However, also for this argument it is not the case. The analysis shows a possible negative effect of similarity on box office performance instead of a positive effect. Also, this negative effect is turned into a positive one in the analysis without outliers. Moreover, the negative effect of similarity displayed has a significance level that indicates that results are far from significant. Actually, the model including similarity as a moderator has a higher significance level than the model which only includes niche width as an independent variable.

Because of the growing body of research on the possibility of positive effects of category spanning on market performance (Hsu, et al., 2012; Negro & Leung, 2012;

Keuschnigg & Wimmer, 2017) and the overwhelming evidence of successful brand extension strategy efforts and product redesigns (Albrecht, et al., 2013; Knapp, et al., 2014;

Bohnenkamp, et al., 2015), this research still has some prudent believe in the theoretical argument presented. As of now, this research is hesitant the right methodology was chosen or whether the chosen methods have been executed correctly. Confidence is little to none and therefore revisions are kindly requested. However, it must be admitted that it could be equally possible that the theoretical argument needs revisions as well. For the time being, due to the fact that results are not significant, the hypotheses presented are both rejected. Subsequently, this research reaches the conclusion that there is no effect of category spanning on market performance or any positive relationship between them. Just as well, a degree of similarity in product categories – or product category overlap – is by no means a moderator of the

presupposed relationship of category spanning and market performance.

This research could have had several implications for category spanning and brand extensions literature, were it not for the fact that results were not significant. For instance, this research could have contributed to category spanning literature by showing a possible

explanation as to why producers keep spanning product categories. In that way, this possible explanation serves as a critique or counterargument for the conviction of conventional

literature that category spanning is counterproductive for performance in every situation (Hsu, Hannan & Kocak, 2009; Negro & Leung, 2012). In its own manner, significant proof of a positive effect between category spanning and market performance could have invigorated and elaborated on the argument of Hsu, et al. (2012) that producing hybrid products – i.e. multiple category products – can in certain markets and certain conditions be beneficial. Equally, results could have contributed to brand extension literature by showing that innovation and novelty is a necessary key element in establishing a successful extension

(29)

29 product – or product redesign –. Therefore, this significance of novelty could be a

confirmation of Sood and Drèze’s (2006) conviction that an element of newness is needed in an extension for a brand extension strategy to work. Furthermore, if similarity had been a positive moderator between category spanning and market performance, it could serve as a vivid illustration that consumers do in fact value brand extensions because of their relation and resemblance with the parent product. This could further emphasize that the conclusion that product redesigns can benefit from the success of the parent product is likely to be correct (Hennig-Thurau, et al., 2009; Knapp, et al., 2014; Bohnenkamp, et al., 2015).

Obviously, the biggest limitation of this research is the fact that results are not significant. Therefore, the hypotheses are rejected and any effect proposed or argument presented cannot be confirmed. Whether this is due to incorrect theoretical argumentation or faulty methodology is not clear. Therefore, future research is encouraged to provide

conclusiveness and – if possible – to revise the theoretical section or methodology. Another limitation of this research is the lack of sufficient product redesign related variables in the analysis. For instance, according to Albrecht, et al. (2013) overall extension fit, consumer’s involvement in the extension category and a reciprocal spillover effect are also to be

considered determinants of brand extension success. Furthermore, according to Keuschnigg and Wimmer (2017) a variable that should be included in research on the effects of category spanning is a variable that accounts for the uncertainty problem or confusion mechanism. This is because product category spanning producers are ambiguous in such a way that consumers experience difficulty forming expectations and therefore are likely to stay away (Keuschnigg & Wimmer, 2017). This research has no such variable in the analysis on the effect of category spanning. Last but not least, this research does not control for any demographics, public availability and technological developments. These might all be variables that can potentially influence market performance.

Directions for future research are, firstly, potentially revising the theory and the methodology introduced in this research. As has been mentioned before, it is unclear whether there are any mistakes in the theoretical reasoning or in the way the research was executed. Although, it is unclear whether a mistake is made, this research holds that it might well be that with some alterations there still could be a positive effect between category spanning and market performance observable. Especially, because Hsu, et al. (2012) have already

concluded in their earlier research that hybrid products are capable of generating significant financial success. Furthermore, in terms of broadening the knowledge on brand extension, future research could look into the exact dynamics between parent brand and a specific

Preselling a Different Remake: a study of category spanning and similarity of product redesigns