Author: Ton Jansen Date: 18/09/2012 Sp
2012
[
SPONSORSHIPS AND FIT
]
2 Department of Marketing Master Thesis 18/09/2012 Ton Jansen S1686305 Nassauplein 22, Groningen 06-55580674 tonjansen89@hotmail.com Supervisors: J. Berger D.A. Naydenova
[
SPONSORSHIPS AND FIT
]
3
Management Summary
Which is more effective in sponsoring? Is it the functional match between the company, or the image match, or a good-cause-sponsorship, even if there is no match at all? Amongst other, these questions will be answered in this thesis.
The author has researched the effectiveness of four sponsorship deals, all with a different ‘kind of fit’: Complete Match, Mismatch, Functional Match and Image Match. The companies and event representing these categories were chosen in a pretest, where respondents were asked about the functional and imagery match between the two. For each of the categories, an example sponsorship was found.
The test phase was completed by asking over 150 Dutch students about their opinion on the company (brand) personality. Half (75) of the respondents obtained additional (fictional) information on a sponsorship deal between the company and an event. The brand personality in this thesis was measured by a scale developed by Geuens, Weijters and de Wulf (2009). This scale was chosen for its small amount of items (12), while still keeping explanation power.
The differences in brand personality were measured and the expectation that the Complete Match sponsorship would be most effective was not met. In fact, this sponsorship type was the least effective, even less effective than the Mismatch (which was a sponsorship with a good cause). The Functional Match was the most effective type, but not all brand personality items changed positively, while they did for the Image Match.
4
Preface
I would like to start by thanking all people that have contributed to this thesis, in various ways. This includes the respondents of the pretest, the regular test, but most definitely my supervisor, J. Berger. Without his critiques and help, I would probably still be stuck somewhere between forming a research question and researching it. I would also like to thank D. Naydenova for her (secondary) supervision and helpful suggestions.
5
Inhoudsopgave
Management Summary ... 3 Preface ... 4 I. Introduction ... 8 a. Managerial implications ... 9 b. Scientifical implications ... 10 c. Research Question ... 10 d. Conceptualization ... 10 Sponsoring ... 10 Brand Personality ... 11 e. Structure ... 12II. Theoretical Background... 13
a. Sponsoring and fit ... 13
b. Image transfer ... 14
c. Hypotheses ... 15
Complete match ... 15
Mismatch ... 16
F-Match & I-Match ... 17
Conceptual framework ... 19
III. Methodology ... 20
A. Pretest ... 20
B. Results of the pretest ... 21
6 Bold (Brutaal) ... 28 Ordinary (Gewoon) ... 28 Simple (Simpel) ... 28 Romantic (Romantisch) ... 29 Sentimental (Sentimenteel) ... 29 E. Personality Factors ... 29 Responsibility ... 29 Activity ... 30 Agressiveness ... 30 Simplicity ... 30 Emotionality ... 30 F. Questionnaire ... 31 f. Results ... 32 a. Complete match ... 33 b. Mismatch ... 34 c. I-Match ... 35 d. F-Match ... 36
g. Discussion & Recommendations ... 39
A. Discussion ... 39 Research Question ... 41 a. Recommendations... 42 Limitations ... 43 References ... 45 h. Appendices ... 49 Appendix 1: Introductions ... 49
Blok A: EA Sports (Functionality Match) ... 49
Blok B: Adidas (Image Match) ... 49
Blok C: Apple (Mismatch) ... 50
Blok D: Rabobank (Complete Match) ... 50
Appendix 2a: Complete match ... 51
Appendix 2b: Mismatch ... 51
Appendix 2c: F-match ... 52
8
I.
Introduction
Strong and differentiated brands have better firm performance and (thus) higher brand equity (Madden, Fehle & Fournier, 2006; van Rekom, Jacobs & Verlegh, 2006; Colucci, Montaguti & Lago, 2008). One way to differentiate from the competition is to develop a brand personality that (target) consumers can relate to. Sonnier and Ainslie (2011) have verified this statement by ‘putting a price’ on brand personality by ‘creating a choice-factor model that measures consumers’ willingness to pay for brand-image associations’. Because of their conclusions, it is proven that consumers actually value the brand personality of the company, the same as they value for example quality (though less important).
The personality of companies/brands is effected by different factors, both in their actions and their feel (how do consumers see the company, logo, colours etc). Consumers assign different human characteristics to companies, which allows them to connect with the company on a deeper level. This connection leads to a positive evaluation of the brand (before buying) and thus to a larger chance of purchase.
In this thesis, the focus is on the influence of sponsorship deals on the company’s (brand’s) personality. Gwinner and Eaton (1999), Smith (2004), Donahay and Rosenberger (2007), Dees, Bennett & Ferreira (2010), Carillat, Harris and Lafferty (2010) and Woisetschläger and Michaelis (2012) are a few researchers that already addressed this issue, but the field is not yet fully covered. Especially the factors of success are, although featured in most of the named researches, highly unknown and unpredictable.
When we review past researches, three factors that influence the effectivity of sponsorship come forward: fit, relatedness or congruence (in the remainder of this paper, fit will be used), event identification and the sponsor’s credibility (e.g. Becker-Olsen and Hill 2006; Cornwell, Weeks, and Roy 2005; Simmons and Becker-Olsen 2006; Speed and Thompson 2000; Smith, 2004; Goldsmith, Lafferty & Newell, 2000; Gwinner & Swanson 2003). Since the sponsor’s credibility is influenced by fit (Rifon, Choi, Trimble & Li, 2004) and event identification is basically a choice of the company – if the event is not known by target consumers, it will not be chosen as subject of the sponsorship deal – this thesis focuses merely on the effects of the first factor: fit between the event and the company/brand.
9 either an image level, at a functional level, none of those, or both. Table 1 shows these four possibilities, with the names that will be used in the remainder of this thesis.
Possibilities for fit/non-fit between brands and events
Functional fit
High Low
Image fit High Complete match Image-only (I-match)
Low Functionality-only (F-match
Mismatch
Table 1: Possibilities of congruence (based on Gwinner/Eaton)
Even though several researchers (Ahn & Sung, 2009; McDonald & Gandz, 1991; Gwinner, 1997) agreed with the research of Gwinner and Eaton (1999) that fit consists of multiple components (image and functional fit), none of them has researched the difference in effects between all four ‘kinds of fit’ (to the knowledge of the author) in one study. The most recent research (Ahn & Sung, 2009) pointed this out by stating that a lot is still unknown about (the difference in effects between) functional and symbolic (image) fit.
Next to the fact that only one research (to the authors knowledge) compared the effects of image and functional fit-sponsorships on brand personality (Gwinner & Eaton, 1999), this research did not provide a full line of the effects. Only the influence on personality factors was tested, while the several items underlying those factors were ignored. The influence of four kinds of fit on each of these items (and factors) is therefore a new concept, that will give an insight in which factors/items are influenced by which kind of fit.
a. Managerial implications
10
b. Scientifical implications
As said, this research is unique and new because of two things: the choice to research all four kinds of fit in one test, and the fact that all items and factors of brand personality will be included. The definition and description of brand personality will be given in the next chapter of this thesis.
Most researches in this topic focus on either one ‘kind of fit’ or on fit in general (Dees, Bennett & Ferreira, 2010; Woisetschläger & MIchaelis, 2010), not on the different kinds of fit. By making this the topic of the thesis, a new insight in the actual differences between the kinds of fit will be presented. This will hopefully put former researches, that focused on one kind of fit, in the right perspective to each other and give new grounds to compare them on.
c. Research Question
Since the difference between complete match, mismatch, I-match and F-match has yet to be researched in a comparison, it seems like an interesting starting point to find out whether there are differences between the two types of fit when it comes to sponsoring. A comparison between the four types of fit would be a new concept, while the conclusions may help to clarify why researchers have found contradictory results (Becker-Olsen & Simmons, 2006) towards fit.
Except from the new approach of including all four kinds of fit in the research, this thesis also makes a distinction between all separate items and factors in the brand personality scale (in this case of Geuens et al., 2009). By adding all of the items, an option to compare the different ‘fits’ emerges. Keeping in mind these possibilities, the research question for this paper is as follows:
‘What is the influence (if any) of fit between companies and events in sponsorships on the brand personality traits of a brand/company?’
d. Conceptualization
In the following paragraphs, the concepts in the research question will be conceptualized for clarity in the latter part of the research. Both the definition of sponsoring and brand personality will be given as they will be used in this research. These definitions are based on past research on both subjects separately, as well as literature on image transfer.
Sponsoring
11 implies that companies can use sponsorship deals to strengthen their image, but it is hard to turn the image around.
In this research, sponsorship means a cooperation between a profit company and a sports team/event or non-profit organization/event, wherein the profit company utilizes the deal to improve their image or awareness. Sponsorships will be introduced by means of a (mock) press release (much like Gwinner & Eaton, 1999; Roy, 2010), thereby influencing the test group.
Brand Personality
‘Brand personality is the set of human personality traits that are both applicable to and relevant for brands’ (Azoulay & Kapferer, 2003). Fournier (1998) adds that consumers ‘use strong brand personality brands to build relationships with’ and consumers typically seek brands with their own aspirational or ideal personalities (Batra, Lehmann, Singh, 1993). In this light, a brand personality scale can be an important tool for firms to identify target consumers, increase the consumer base and thus brand equity and firm performance.
Jennifer Aaker (1997) was among the first to develop a scale to measure brand personality, basing her ideas on the fact that people often see objects/companies/animals as humans, with an own personality. This led her to make a five-factor scale resembling the famous Big Five human personality scale. This scale became the most used reference for brand personality researches for years to come (Aaker, 1999; Aaker, Benet-Martinez, & Garolera, 2001; Kim, Han & Park, 2001), although the scale was not that generalizable: researches in Germany (Bosnjak, Bochman & Hufschmidt, 2007), Croatia (Milas and Mlacic, 2007) and the Netherlands (Smit, van den Berge & Franzen, 2002) all found (slightly) different scales. Next to the generalizability, Aaker’s research receives criticism for the loose definition of brand personality; characteristics as gender and age were included, as opposed to merely personality traits. Azoulay & Kapferer (2003) and Bosnjak et al. (2007) argue that this loose definition is causing confusion of what is measured: either brand personality or user characteristics.
12 outcome of their research. The describing power relative to the size of the factors is better than Aaker’s scale, which needed around 40 variables to achieve a slightly higher describing power. The size of the factors also allows for a smaller test set-up, instead of a large test set-up which would lead to either less respondents or huge amounts of data.
Figure 1: Brand Personality: 12-item 5-factor scale (Geuens, Weijters & de Wulf)
e. Structure
13
II.
Theoretical Background
This chapter will elaborate on both the problem and the constructs of the problem. This will be done by comparing and critically assessing past research on the subject. First, the construct of sponsoring will be described, with a deeper focus on articles which address ‘fit’ (in relation to sponsoring). Second, the connections past researchers have found with respect to brand personality and sponsoring will be introduced and compared. Last, researches which address a problem similar to the presented problem will be compared and evaluated, to make sure there will be no duplication of results. Also, the different styles and methods of research will be evaluated to enhance this thesis.
a. Sponsoring and fit
Figure 2, based on Simmons & Becker-Olsen (2006), shows that the effect of the fit construct on firm equity (including brand personality) is influenced in two ways; the clarity of positioning and the attitude toward the sponsorship. The former deals with the fact that consumers feel the need to understand the reason for the sponsorship, which in this case means an connection on either functional or image level. The attitude toward the sponsorship evaluates how consumers feel about the sponsorship. Most simply put this means whether consumers feel positive or negative about the sponsorship deal. The authors found that the effect of clarity of positioning (the subject of this thesis) on brand equity is actually stronger than the effect of the actual attitude toward the sponsorship.
Figure 2: Influence of sponsorship fit (Simmons & Becker-Olsen, 2006)
14 (e.g. Johar & Pham, 1999; Cornwell et al., 2006) experience a positive relation between fit and brand/company recall, although Becker-Olsen & Simmons (2002) found that a non-match can also lead to higher recall and Olsen & Thjømøe (2009) found a negative relation between fit and brand/company recall . When looking at the attitude of respondents towards the brand/company, it is clear that a higher fit leads to a better evaluation: Speed & Thompson (2000), Becker-Olsen & Simmons (2002), Simmons & Becker-Olsen (2006) and Olsen (2010) all found a positive relationship. As said (Chapter 1: Introduction), several researches (e.g. Gwinner & Eaton, 1999; McDonald, 1991) have differentiated between ‘kinds of congruence’. Of course, the most obvious confrontation would be between match or no match, but since there are different components of fit, the researchers have tried to find underlying constructs which might explain the concept of fit. Examples of studies that embrace the idea of a multi-construct fit are McDonald (1991), Gwinner (1997) and Speed & Thompson (2000). All of these studies argue that there should be two underlying constructs: functional and image fit. Combined, they explain total fit between the company and the event. Table 1 (Chapter 1) illustrates the ideas of the named researchers.
Most recently, Prendergast, Poon & West (2010) revised the existing research, arguing against researchers that view fit as a one-dimensional construct (Rifon, Choi, Trimble & Lee (2005). Prendergast et al. agree with the view of a multidimensional congruence construct. Up to now, most researchers (McDonald (1991), Gwinner & Eaton (1999), Cornwell (1995)) agree with Prendergast et al. that functional congruence and image congruence are not mutually exclusive.
b. Image transfer
Image transfer occurs when the personality of a brand/company changes because of a certain action; in this case a sponsorship deal. The concept of image transfer has only recently become widely researched (Gwinner & Eaton, 1999; Smith, 2004; Donahay & Rosenberger, 2007; Dees, Bennett & Ferreira, 2010; Carillat, Harris, Lafferty, 2010). The focus in all of these researches was different, as can be seen in table 2.
Researchers Year Focus
Gwinner & Eaton 1999 Pioneering the field
Smith 2004 Building on past research and other fields
Donahay & Rosenberger 2007 Formula 1: effect of team performance and functionality of products on image transfer
Dees, Bennett & Ferreira 2010 NASCAR: fan identification, product involvement and fit between sponsor/driver as dimensions
15 same event
Woisetschläger & Michaelis 2012 Pre/post analysis: the effect of sponsorship congruence on image transfer
Table 2: Past research on image transfer
The general effect of sponsorship congruence on image transfer has been researched (Dees et al., 2010), as has the effect of functional fit (Donahay & Rosenberger, 2007). However, the difference in effect in all four ‘kinds of fit’ has yet to be researched. None of the past research mentions a comparison between the effects of F-match, I-match, Mismatch and Complete match, at least not to the author’s knowledge. The reason no one researched the difference in effect is unknown, because past research hinted at a different influence of the ‘kinds of congruence’ (e.g. Donahay & Rosenberger, 2007).
The researches featured in table 3 do offer a kind of guideline and warn for possible pitfalls along the way. Therefore, even though the subject has not yet been researched, a lot can be learned from the past researches. Combined, they even form a good basis for the separate hypotheses (to be named in the next paragraph: Hypotheses). However, as said, this paper still features a new avenue of research: the influence of the ‘kinds of congruence’ on the different brand personality items/factors.
c. Hypotheses
Keeping the former paragraphs in mind, a few hypotheses (all referring back to table 2) about the influence of sponsorship on brand personality can be drawn up.
Complete match
Firstly, the ‘complete match’, in which both functionality and image of the brand are highly congruent with the event personality. The general belief is that a better congruence should lead to better results in image transfer, which is based on past research (e.g. Becker-Olsen & Simmons, 2002; Johar & Pham, 1999) on different objects: brand recall, brand awareness and brand personality.
16 On the other hand: (image) fit implies a similarity between a brand and an event, and since the image fit is in fact included in the complete match the effect of image change (or transfer) might be hard to see in this case. The effect could be more in the strength of the associations then in the rating. If the test provides contradictory results, this could be a possible explanation.
H1: Brand personality items/factors of a sponsor change positively when the fit between the event and the sponsor is both functional and imagery (complete match).
Mismatch
In their 2007 paper, Martensen, Grønholt, Bendtsen and Juul explore the effects of both positive (complete match) and negative (mismatch) fit. They found that the effect of a mismatch has a negative influence on negative brand traits (personality items) like aggressive and bold. On the other hand, Rifon, Choi, Trimble & Li (2004) found that while a mismatch sponsorship might lack clarity (figure 2), it does make consumers think about the sponsorship. This way, the consumers make new associations and connections with the company; it allows them to reevaluate the brand image and personality. In this light, the mismatch could prove to be effective for company who would like to change their image. However, the effects of the mismatch sponsorship are highly unpredictable, making it a risky operation.
H2a: Mismatch sponsorships changes the brand image of the sponsor; ‘bad’* personality items change up, while ‘good’* personality items change down.
* in this case, ‘bad’ items are personality items that are rated poorly by respondents, while ‘good’ items are the contrary.
17 H2b: Mismatch sponsorships will negatively influence negative-loaded items of brand personality. H2c: The moderating effect of a good-cause event will improve the brand image of the sponsor (brand/company); positive-loaded items will change up.
F-Match & I-Match
In their 2007 paper, Donahay and Rosenberger researched the effects of functional congruency (fit) on the transfer of brand image (personality) in Formula 1. The differences between sponsor and racing team personality turned out to be small and significant for the functional fit, while for the non-functional fit, they remained large (over 1,5 point difference). Regarding this research setting, this would mean that the effect of functional fit on brand personality would be significantly larger than the effect of mismatch. Donahay and Rosenberger (2007) did only include a mismatch sponsorship as a comparison.
Gwinner & Eaton (1999) studied the effects of F-match, I-match and mismatch sponsorships. They found that both F-match and I-match have a greater influence on image transfer than mismatch does. Regarding the difference between F-match and I-match however, they found contradictory results: the regular test – which is comparable to this research - pointed to higher effectivity of the I-match, while the holistic measure – which was tested by a single-item measure of congruence - pointed to higher effectivity of the F-match. A possible explanation is the connection a consumer makes. With a functional congruency, there might be a more direct connection, as the consumer immediately sees why a sponsor is connected to the event. The image connection might be stronger indirectly, because there is no obvious link between the sponsor and the event at first glance, but the feelings about the sponsor and the event are similar. In this thesis, the regular test will be used, because it is impossible to ask respondents about complete match and mismatch in a few questions, and a comparison between the two different measures might lead to different results (as happened in the research of Gwinner & Eaton, 1999).
18 H3a: An F-match sponsorship causes brand personality items/factors to change positively.
H3b: The effect of the F-match sponsorship is expected to be the greatest in the factors Responsibility, Activity and Aggressiveness.
It has yet to be researched which items and factors are influenced by image congruency. Logical reasoning would say that if the image is matched between the company/brand and the event, their brand personality would also be quite similar. In this case, the obvious expectation would be that the effect on the brand personality items is low – meaning not so much items changing significantly – since the items are already comparable.
19
Conceptual framework
To clarify the hypotheses and the relationships that will be researched in this thesis, a conceptual framework has been designed. This conceptual framework shows the types of fit, moderator and hypotheses in the research design.
Figure 3: Conceptual Framework
Legend: Good-cause event H2a/b H2c Image match
Complete match Mismatch
Functional Match H1
H4
H3b
H3
20
III.
Methodology
This chapter will describe the method of testing for this research. Also, the choices for particular methods will be explained. The testing phase of the research consists of 2 parts: a pretest, in which the brand(s) and event(s) will be chosen, and a regular test, in which the brand’s personality will be tested.
A.
Pretest
Before the actual test, a pretest will be conducted among (about) 20 students. The goal of the pretest will be to find brand/event combinations that are suitable to represent the four ‘kinds of fit’. A sample size of 20 is comparable to other researches with the same focus: Gwinner and Eaton (1999) also used 20, while Olsen and Thjomoe (2011) only had 13 respondents in the pretest,.
This test set-up means there has to be at least one suitable match for each of four different ‘matches’ (complete, functionality, image and mismatch). If not, another pretest would have to be conducted with different brands to finalize the subject brands and events for the testing phase. The brands and events were chosen in a random manner, but will definitely be known by students, the core group in the research. This is of course a prerequisite for a positive or any outcome at all. Also, the chosen brand/event combinations are expected to have a certain match between them. The combinations only show the F-match and I-match, as these matches are the hardest to find. Since there will be stronger and weaker matches, the ‘complete match’ and ‘mismatch’ categories will be featured by the best match (highest average) and the worst match (lowest average). By choosing this setting, the two matches in the middle (F-match and I-match) are found quite easily, and the other matches can be derived from there. The F-match will be represented by the highest difference in F- and I-match in the advantage of functionality, while the I-match-combination will be chosen by the opposite. Table 3 shows the possible brand/event combinations.
Brand Event Expected Match
EA Sports (gaming) Dutch Soccer Cup Image-only
EA Sports (gaming) World Cyber Games Functionality-only
Adidas World Online Soccer Championships Image-only
Adidas Dutch Hockey League Functionality-only
iPad (Apple) Dance Valley Festival Image-only
iPad (Apple) World Day for Audiovisual Heritage Functionality-only
Rabobank Amstel Gold Race Image-only
21 To determine the functional and image match between brand and event, the scale developed by Gwinner & Eaton (1999) will be used. This 7-point Likert scale is useful because it is the only one to differentiate clearly between functional match and image match, while other scales (e.g. Fleck & Quester, 2007) do not differentiate the two items, but ask questions about fit in general.
Gwinner & Eaton (1999) proposed six questions during their research, half describing functional match and the other half describing image match.
Functional match questions:
(1) It is likely that (participants) in the (event name) use (brand name) during the (event name). (2) When I watch the (event name), I often see (brand name) being (used).
(3) (Brand name) is not a product that (participants) in the (event name) would consider (using).
Image match questions:
(1) The (event name) and (brand name) have a similar image.
(2) The ideas I associate with (brand name) are related to the ideas I associate with the (event name).
(3) My image of the (event name) is very different from the image I have of (brand name).
The answers of the respondents will be added together, before dividing them by 3, getting a mean score for both functionality and image (table 5).
B.
Results of the pretest
The results of the pretest (table 4) were in some cases surprising, as some image scores were lower than expected beforehand, sometimes even lower than the functional score.
Brand Event Functional score Image score Difference
EA Sports Dutch Soccer Cup 5,5 3,8 1,8*
World Cyber Games 5,2 4,4 0,8
Adidas World Online Soccer Championships 2,6 4,7 2,1**
Dutch Hockey League 4,7 3,7 1,0
iPad (Apple) Dance Valley Festival 3,3 3,7 0,4
World Day for Audiovisual Heritage 2,1 2,7 0,5*** Rabobank Amstel Gold Race 4,3 3,6 0,7
KvK Startersday 6,1 4,2 1,9****
22 Even though the results were not as expected upfront, they still offer possibilities for further research. Some of the results offer a good starting point to research the main goal of this paper: comparing the different ‘kinds of congruence’. The bold printed combinations provide the most suitable results; the stars behind them indicate to which ‘kind of congruence’ they will be related. The single star corresponds with the F-match. EA Sports has the second highest functional score, while scoring low on image. Although not the highest difference, the combination of a high functional score (5,5) and low(er) image score (3,8) does provide a suitable brand/event sponsorship venture to research the influence of a functional match.
The second, two-star sign corresponds with the I-match. Adidas does match the image of the World Online Soccer Championships quite well (4,7 – highest image match), but is not deemed to be functional in any way (2,6). Therefore, this combination will be used to describe an I-match.
The three stars are appointed to the combination of iPad (Apple) and the World Day for Audiovisual Heritage, since this combination clearly is the worst in both functional congruence and image congruence. Therefore, it is the most suitable choice to represent the ‘mismatch’ category.
The last, four-star sign represents the complete match. Even though a high difference between functional congruence and image congruence is present, both are still among the highest in their respective categories (functionality is actually the highest scoring). Therefore, Rabobank sponsoring the Kamer van Koophandel Startersday will represent the ‘complete match’ category. These four combinations will form the basis for the next part in this research: the actual testing phase.
C.
Testing: Phase 2
To test the change in perception consumers experience, an experiment that tests the effects of specific elements seems the right method. The experiment will be structured as follows: 1 control group will fill out a questionnaire about well-known brands with a clear personality. A second group will be shown the same test, but only after reading about a (fictitious) sponsorship deal – Gwinner & Eaton (1999) and Roy (2010) have proven this to be an effective, reliable method. The results will be compared and tested on significant differences.
23 Figure 2: Brand Personality: 12-item 5-factor scale (Geuens, Weijters & de Wulf)
The items represented in this scale will individually be discussed in the next section of this chapter (Personality Items). The four (to-be-)tested companies will be compared for every item (control group), and every factor separately to make individual item-conclusions more interesting. Because the test was in Dutch, the used Dutch terms will also be introduced in the Results section.
All respondents will rate the four brands from 1 to 7 on the 12 different Likert-scales, just as Geuens et al. (2009) did in their research. This means the test consists of four blocks of 12 questions, making it a total of 48 questions. While Geuens et al. (2009) just needed to test their new scale, this research requires a comparison between two different set-ups (one with and one without a press release about the sponsorship deal). The two different test set-ups are shown underneath (figure 3).
Figure 3: Testing Phase: Graphical Representation
Method of testing
To test the hypotheses stated in chapter IIc, different methods will be used. First of all, a One-Way-Anova will explore the variance of means between the test group and the control group (no press release shown). This way, the actual influence of the press release can be explained for each of the
EA Sports regular
Adidas regular
iPad with press release
Rabobank with press release
Test
1
EA Sports with press release
Adidas with press release
iPad regular
Rabobank regular
Test
24 12 variables. Because the amount of respondents will exceed the critical point necessary to conduct a Anova-test without the variables being normally distributed, no problems are expected with the validity of these results. Still, it is good to have a control test; the One-Way-Anova test results will be verified by conducting an Independent-Samples t-test, which should provide the same results. Following the ANOVA, some variables will vary significantly, while others will not. The results will reflect a clear distinction between the four different tested ‘congruences’. When it is clear which variables are significant, a comparison will be made to find out which match has the greatest influence with regards to a change in brand personality. An example of this could be that an I-match has more influence on a change in the ‘Active’ factor then an F-match has.
Respondents
In this paragraph, the obtaining of responses and the assignment of respondents to a certain test will be explained. It is important that both groups are homogenous, or at least comparable in a way. Therefore, the target group for this research will be students, and students only will be asked to respond to the test. All respondents which are not considered to be part of the target group will be filtered out when the results are screened (the definition of a student is a person between 16 and 30 years old, still in school). There are enough opportunities in the university to find respondents, and comparable researches also used relatively small samples, ranging from 30 (per researched brand – Geuens et al., 2009) to 184 (Gwinner & Eaton, 1999).
The majority of the results have been obtained by an internet-based survey with a program called Qualtrics. This program contains a feature which is especially interesting for the purpose of this research, in that it provides random assignment of respondents to questions. This way, it remains uncertain for each respondent whether he/she will take the regular test, or the one with the press release. Since this split is done for each respondent and each combination separately, there are essentially (2x2x2x2) possible different test combinations. At the end, it is sure to have included each possibility the same amount of times (half).
25
Descriptives
The descriptives of the complete set of respondents are shown underneath (table 5). As you can see, there are 154 respondents, all of which are in higher education in the Netherlands (or starting coming year).
Descriptives
Survey progress Amount of respondents
Complete 154
Incomplete; of which 48
Before 1st set of questions 36
During 1st set of questions 5
2nd set 5
3rd set 2
Gender Amount of respondents
Male 91 (59,1%)
Female 63 (40,9%)
Education Amount of respondents
High school 21 (13,6%)
Community College 41 (26,6%)
University of Applied Sciences 43 (27,9%)
University 49 (31,8%)
Average age 22.411
Table 5: Descriptive statistics sample
As said, it is important for the research that the test and the control group are comparable (in both size, and descriptive). Since the test and control groups are not the same for every test set-up (see Methodology: Respondents), there are four different tables to describe the similarity between the control and the test group: one table for each of the four ‘matches’. These table are shown underneath.
26 F-match Age Education Gender
Control Group 22,70 2,73 1,51 Test Group 22,12 2,83 1,31
Total 22,41 2,78 1,41
I-match Age Education Gender Control Group 22,65 2,78 1,42 Test Group 22,17 2,78 1,4
Total 22,41 2,78 1,41
Mismatch Age Education Gender Control Group 22,60 2,79 1,42 Test Group 22,22 2,77 1,4
Total 22,41 2,78 1,41
Complete match Age Education Gender Control Group 22,38 2,87 1,44 Test Group 22,44 2,69 1,38
Total 22,41 2,78 1,41
Table 6: Similarity between control and test group
The tables show satisfactory results: for each of the matches, the control and test groups are comparable – the only problem could be the gender difference in the F-match groups. The average age differs not more the 0.6 years (on a range of 14), and the education gap is always smaller then .18 (on a range of 4).
D.
Personality Items
In this section, the brand personality scale of Geuens et al. will be analyzed by introducing the items in the scale separately. The used (Dutch) translations of the items will be given, as well as a short description of whether the item is positive or negative.
Down-to-earth (Nuchter)
For most of the items, the translation is the first ‘hit’ in the English/Dutch dictionary Ectaco (an internet-based service). For ‘down-to-earth’ however, this was hard, since there was no translation found – since it is not a single word, but more of a saying (in English, anyway). Therefore, a regular dictionary was used. Here, the Dutch definition of ‘down-to-earth’ was ‘nuchter’ – this term was used in the test phase.
27 Dutch company, Rabobank, puts more emphasis in fitting their image to Dutch culture, and will score higher on this item.
Stable (Stabiel)
The item Stable can easily be translated as a word, but also as a meaning. All English-Dutch translators used (Ectaco, Babylon, Google translate, ‘regular’ dictionary) give the Dutch ‘Stabiel’ as the best translation for Stable. Stability is seen as a positive personality trait: being stable gives a feeling of safety, a good company which remains ‘on top’, or at least in the market.
Responsible (Verantwoordelijk)
The same four dictionaries were used for ‘Responsible’ as for ‘Stable’, and all of them firstly give ‘Verantwoordelijk’ as the best translation. This means ‘Verantwoordelijk’ was used in the questionnaire as well. Being Responsible has again no double meaning; it’s positive to be responsible, be it for the company, for the customers, or for the environment. Therefore, a higher score is more positive than a lower score.
Active (Actief)
Active is a more difficult item to explain, does it deal with the company itself or does it concern their products? Probably, a combination of the two is meant here. Therefore, ‘Werkzaam/ werkend’ is not the best translation, but the more general ‘Actief’ is. It is not hard to assess whether Active is a positive or negative trait: higher activity means more products, more variations, better service etc.
Dynamic (Dynamisch)
Dynamic is another item that can be translated almost literally from English to Dutch: ‘Dynamisch’ essentially means the same, according to the four used dictionaries. Being dynamic is often seen as a positive trait: being adaptive to the environment and being energetic are two examples of this trait.
Innovative (Innovatief)
Innovative is easily translated into the Dutch ‘Innovatief’. Again there is no discussion: all four dictionaries give ‘innovatief’ as the only translation. Innovativeness of course is a positive trait; new products and new features are always appealing to consumers.
Aggressive (Agressief)
28 marketing, take-overs etc.: the same goes for ‘agressief‘ in Dutch. However, ‘strijdlustig’ could also be included in the translation of aggressive, as this word completes the English term. Since one has to be chosen, it is best to choose the one which has the most complete description: Agressief. Then the second part: is aggressive a positive or a negative trait? In essence, being aggressive seems a negative personality trait, but can you translate this negativity to brand personality? Since Geuens et al. compared their item aggressive to the more negatively loaded Unpleasant (d’Astous & Lévesque, 2003), Ruthless (Davies, Chun, Vinhas da Silva & Roper, 2004) and Annoying (Smit, van den Berge & Franzen, 2002) it seems there is a negative sound to this item. This means a lower aggressiveness is favorable for companies.
Bold (Brutaal)
In the same factor as Aggressive, the second (and last) item is Bold. Bold might be the hardest item to translate, since there seem to be no direct translations in Dutch. ‘Stoutmoedig’, ‘Brutaal’ and ‘Stout’ all have something in them to describe Bold. Because ‘Stoutmoedig’ is kind of an old word for daring, it does not cover the negative loading of Bold in this case. ‘Brutaal’ en ‘Stout’ both do, with the former being more negative than the latter. Since Geuens et al. found a negative loading for the entire underlying factor (Aggressiveness), ‘Brutaal’ was used in the questionnaire. This negativity already reveals that a lower score on ‘Bold’ is better.
Ordinary (Gewoon)
The literal Dutch translation of Ordinary (‘Ordinair’) does not cover the term at all. Therefore, another word was used in the questionnaire: ‘Gewoon’. All of the four dictionaries used ‘Gewoon’ as the primary translation of Ordinary. The second part is harder; because it is hard to assess Ordinary as either positive or negative. As said before, Dutch culture (White & Boucke, 2006) sometimes rates acting normal as a positive quality, while in other cultures it might be an asset to ‘think outside the box’. Therefore, it is hard to say whether being Ordinary is a quality or not. Since the research is set in the Netherlands, it is assumed higher scores are better in this case.
Simple (Simpel)
29
Romantic (Romantisch)
The next item, Romantic, is easily translated into the Dutch ‘Romantisch’ according to all four dictionaries. Also, the Dutch term covers most of loading of the English Romantic. Therefore, ‘Romantisch’ was used in the questionnaire.
As a company, being romantic is neither positive nor negative; at least not for the four selected companies. There are products, like candles, wine and restaurants for which ‘being romantic’ is a good quality. However, none of the companies produce ‘romantic’ products, so the expectation is a low score for each of the companies. The personality item is not of interest for them.
Sentimental (Sentimenteel)
The translation of Sentimental forms no discussion whatsoever: ‘Sentimenteel’ is not only the ‘best’ translation according to the four dictionaries, but also covers the English Sentimental. The positive/negative discussion is the same for Sentimental as it is for Romantic. For some products, it might be positive to be seen as Sentimental, but most products do not gain from this (or even suffer). Since the score might not really influence a positive view of consumers, it can be expected that the companies have low scores (they do not have to focus on improving this item).
E.
Personality Factors
Together, groups of personality items form personality factors (figure 2). These factors do not require translation, as they are not part of the questionnaire directly – merely indirect, via the items. However, they do require some explanation, since they are used to describe the companies. In this case, it is important to notice that because the results are based on Dutch respondents, cultural differences do not have to be taken into account. If a rise of a particular item takes place, this is either positive or negative, according to the (in this case) Dutch customers. For which company this change takes place is irrelevant, at least for the Dutch market.
Responsibility
30
Activity
The second factor, Activity, also consists of three items: Active, Dynamic and Innovative. This factor is also seen as a positive treat for companies/brands. Geuens, Weijters and de Wulf (2009) argue that both the underlying items and the entire factor are valued better by customers when the scores are higher. All four companies will be rated the same since social difference is very low in this factor, according the named authors.
Agressiveness
The remaining three factors, starting with Aggressive, all consist of two personality items. Bold and Aggressive, the two items underlying Aggressiveness, are both negatively loaded according to the creators of the scale, Geuens, Weijters and de Wulf (2009). This means lower scores on these items, and the factor, are more valued by consumers. This conclusion is valid throughout different kinds of industries, meaning no difference between the test companies has to be taken into account.
Simplicity
For Simplicity, a cultural difference is present. As for the item ‘Down-to-earth’ in the Responsibility factor, both items (Simple and Ordinary) in the Simplicity factor are rated more important in the Netherlands (White & Boucke, 2006) then in the rest of the world. Therefore, Rabobank will probably have higher scores on this item – they are Dutch-based.
Emotionality
31
F.
Questionnaire
The questionnaire that was used for the test is fairly straightforward: respondents had to fill out four ‘descriptive’ questions (to make sure they were inside the target group), and four times a Likert-scale table – each corresponding with one brand/event combination – similar to the one shown in table 7.
Totaal mee oneens Mee oneens Beetje mee oneens
Neutraal Beetje mee eens
Mee eens Totaal mee eens Nuchter/Realistisch Stabiel Verantwoordelijk Actief Dynamisch Innovatief Agressief Brutaal Gewoon Simpel Romantisch Sentimenteel
Table 7: Likert-scale table
Before each Likert-scale table, a short introduction of the brands was shown. For the control group, this was merely an introduction of the company. For the test group, an extra paragraph of text was added: a press release, announcing a sponsorship deal.
Since both the author and the respondents are native Dutch, the questions in the questionnaire did not require translation from English to Dutch, and neither did the introductions to the company. In the appendix, all introductions are shown, as well as their English translations for English readers.
The questions in the pretest, however, were written in English (Chapter IIIA) by Gwinner and Eaton (1999), and were translated the same way as the brand personality items. Online dictionaries (translating machines) were used, and combined with the English knowledge of the author, the most logical versions were chosen. The English questions and their Dutch translations are given underneath.
Functional match questions:
32 (2) When I watch the (event name), I often see (brand name) being (used).
Als ik naar (evenement) kijk, zie ik dat (merk naam) vaak gebruikt wordt.
(3) (Brand name) is not a product that (participants) in the (event name) would consider (using). (Deelnemers) zouden niet overwegen om (merk naam) te gebruiken tijdens (evenement).
Image match questions:
(4) The (event name) and (brand name) have a similar image. (Evenement) en (merk naam) hebben een soortgelijk imago.
(5) The ideas I associate with (brand name) are related to the ideas I associate with the (event name).
De ideeën die ik associeer met (merk naam) zijn gerelateerd aan de ideeën die ik associeer met (evenement).
(6) My image of the (event name) is very different from the image I have of (brand name). Het beeld dat ik heb van (evenement) is heel anders dan mijn beeld van (merk naam).
f. Results
This chapter will lead the reader through the results of the testing phase. Firstly, the comparison of means between the control group and the test group will be given. Since a test has to be conducted for each variable, it means 12 One-Way-Anova (for each brand/event combination) tests to find out which variable(s) change significantly after showing a press release (e.g. after announcing a sponsorship deal). For each of the four brand/event combinations, the same test is conducted. After this test, a clear view on the significant changes can be given, as well a conclusion about which brand personality items change for which combination.
33
a. Complete match
The first comparison of the control group and the test group deals with the complete match. The complete match was tested by a fictional sponsorship deal between the Rabobank and the Kamer van Koophandel Startersdag (Chamber of Commerce Startersday), which was found to be both a functional and an imagery match in the pretest (chapter IIIb).
Before the actual test results, the first column shows the base personality scores of Rabobank, showing the brand personality scores without the complete match sponsorship. The second column shows the scores from the test group – with the complete match sponsorship, while the third shows whether the differences between the scores were significant.
Rabobank scores well on the important items in the Netherlands: down-to-earth, stable, responsible, ordinary and simple are the highest scores of the four companies. Aggressiveness and bold are also better than other companies’ scores, because in this case lower is better. On dynamic, innovative, romantic and sentimental, Rabobank has problems to keep up with the other companies. If hypothesis 1 proves to be true, all items should differ for the better: higher scores for all items but aggressive and bold.
Although the expectancy beforehand was that a lot of significant changes would occur because of the press release, the contrary is true. Only two out of twelve items show significant variation between the control group and the test group. The One-Way-Anova and the Independent Samples t-test (appendix 2a) provide the same results, meaning the used test is not the cause of this surprising outcome.
Personality Item Without PR With PR Significance Down-to-earth 5,70 5,70 1 Stable 5,71 5,64 0,671 Responsible 5,60 5,62 0,897 Active 5,36 5,57 0,163 Dynamic 5,16 4,90 0,108 Innovative 5,10 4,09 ,000 Aggressive 3,09 2,44 ,001 Bold 3,00 2,87 0,523 Ordinary 5,01 5,13 0,575 Simple 4,66 4,42 0,249 Romantic 2,78 2,73 0,829 Sentimental 2,60 2,64 0,867
34 Now, the actual differences between the control group and the test group may not be as significant as expected, there are only two personality items that did change significantly: Innovative and Aggressive. Respondents that were shown the test with the press release rate Rabobank 4.09 on the item Innovative, while respondents that did not see the press release gave the company a score of 5.10. This means the complete match did not produce expected outcomes here either; a favorable item changed down, while it was expected to changed up.
For Aggressiveness, this is different: the test group does rate Rabobank as being less aggressive (2.44) than the control group (3.09). This is a favorable outcome for the company, as being aggressive is contradictory with being ‘Simple’, ‘Down-to-earth’ and ‘Stable’; items that are rated high for Rabobank.
b. Mismatch
The second ‘matching’ possibility is the mismatch, in which a company (product) has neither imagery nor functional congruence with the sponsored event. In this research, the mismatch is represented by Apple and the World Day for Audiovisual Heritage (UNESCO). With a functional congruence of 2.1 and an image congruence of 2.7, this was the most suitable combination to represent this category.
Underneath, the base scores for Apple are given, which will be compared with the results of the test group (second column). Apple scores high in the factors Rabobank scored relatively low: Active, Dynamic and Innovative are the best scores for Apple, while the ‘Dutch’ related down-to-earth, ordinary and simple are fairly low. Romantic and sentimental are, though not of much use to the company, also quite high.
Combined with the hypothesis, the expectation is that Active, Dynamic and Innovative will not change by much, while the ‘bad’ personality items will get a boost (down-to-earth, stable, responsible, ordinary, simple).
Table 9 underneath shows the differences between the control group and the test group on these significant items. Though the results are striking, they are also irregular: there is only one factor with no significant differences, while out of the four other factors at least one item changed significantly.
Personality Items Without PR With PR Significance Down-to-earth 4,31 4,62 0,184
Stable 5,31 5,60 0,104
35 Active 6,18 5,94 0,164 Dynamic 5,94 6,22 0,027 Innovative 6,18 6,48 0,031 Aggressive 3,64 4,27 0,015 Bold 4,47 4,09 0,079 Ordinary 3,09 2,43 0,003 Simple 3,12 2,68 0,076 Romantic 3,05 3,30 0,242 Sentimental 3,13 3,44 0,179 Table 9: Mismatch
The first difference is the item Responsible, which changes from 4.9 to 5.64. Certainly in times where being responsible is of rising importance, this can be seen as a favorable change for companies. Secondly, Activity is the only factor that has two differing items: Dynamic and Innovative. Is the factor is seen as a positive personality trait, a higher score is ‘better’ for a company. In the case of Apple, the sponsorship with the World Day of Audiovisual Heritage caused the already high scores on Dynamic and Innovative to become even higher: 5.94 to 6.22 (Dynamic) and 6.18 to 6.48 (Innovative).
An item that did not change favorably by the mismatch is Aggressiveness. As mentioned before, the negative loading of this item means a lower score is actually for favorable. Apple’s score on this item actually rose in comparison to the control group (3.64 to 4.27).
c. I-Match
An I-match can be described as a sponsorship between a brand/company and an event with roughly the same image. Here, image has to be seen in the widest sense of the word: it can relate to the area both events are from (a used example was Grolsch/FC Twente), the customers the event and the company attract, or try to appeal to etc. The pretest showed good results for an I-match, Adidas and the World Online Soccer Championships were put together (both soccer, both mainly male-oriented) – an image congruence score of 4.7 and a functional congruence score of 2.6 are nice results to base conclusions upon.
36 The I-match press release causes 7 out of 12 items to differ significantly: down-to-earth, active, aggressive, bold, ordinary, romantic and sentimental. Out of the five factors, two factors change the most (Aggressiveness and Emotionality), with all underlying items differing significantly (in red – table 10).
Note that both Aggressive and Bold, which together form the factor Aggressiveness, differ significantly between the two group. The direction in which the items change is favorable in this case; down. Table 15 provides the means for both the control group and the test group for the I-Match. The second factor of which all items change is the Emotionality factor. This factor is not of much use to the companies, but it is still worth noticing the connection between the I-match and the factor Emotionality. Apparently, the I-match has a positive influence here. The remaining significant changes (Down-to-Earth, Active and Ordinary) are all part of different factors.
d. F-Match
The F-match is the last ‘kind of congruence’ between a brand/company and an event. To refresh the readers’ memory: a functional match between a brand/company and an event is mainly related to whether the products of the company can be used before, in, or after the event. In this research, the category is represented by EA Sports (game manufacturer) with the Dutch Soccer Cup. Although the functionality of a football game in an actual match is hard to see, respondents rated the sponsorship with 5.5 on the functional scale and 3.8 on the image scale – the biggest difference (after Rabobank/KvK Startersday) and the highest functionality score (again after Rabobank/KvK Personality item Without PR With PR Significance
37 Startersday). Therefore, this combination seems to be a good example of a functional match. The base test results of the brand (EA Sports) are shown in the first column of table 11, and will be used in the comparison – items written in ‘Bold letter type’ are significant.
With the test taken, the second column of the table underneath shows the results of the One-Way-Anova test of the data (which was verified by an Independent Samples t-test). The amount of significant changes is remarkable for a functional match; nine out of twelve items have shown significant change, in all factors. The three items in the factor Responsibility and the two items in Aggressivity all differ – making these the most influenced factors by the F-match. Activity (two out of three items), Simplicity (one of two) and Emotionality (one of two) all slightly change, but not all items underlying them differ significantly.
In the table underneath, the actual differences are shown (columns 1 & 2). The most important fact to notice in the table is that all significant means of the second column are higher than the ones in the first column, even the negatively-loaded Aggressive and Bold. This points to some ‘collateral damage’ of the sponsorship: most of the results are positive, but they come at the price of being seen as more Aggressive and Bold.
Personality Item Without PR With PR Significance
39
g. Discussion & Recommendations
The final section of this paper consists of two parts; the discussion, in which the results are interpreted and a final answer on the research question(s) will be given; and a recommendation part, which goes one step further: what should the next step in this area of research be?
A.
Discussion
In the introduction, one research question was formulated. The answer to this question will be given in the following subchapter, as well as the conclusions about the hypotheses. Firstly, the hypotheses will be discussed, before giving an overall conclusion of this paper in the form of an answer to the research question.
Complete Match
In the results section, the influence of a complete match did not provide expected results. Only two personality items differed significantly, with only one of these two changing positively (Aggressive). Therefore, H1 must be rejected: Brand personality items/factors of a sponsor do not change positively when the fit between the event and the sponsor is both functional and imagery (complete match).
A possible explanation for the results is that Rabobank already had a quite favourable image before testing (a real ‘Dutch’ image with high scores on Down-to-earth and Ordinary), and one sponsorship deal cannot change this image. It is quite possible that the sponsorship did have effect in other areas, like brand recall or a stronger image (instead of a different image), but these were not tested.
Mismatch
The mismatch category provided useful insights: the brand personality of Apple changed on five out of twelve items, significantly changing their image in the eyes of the consumers. Because the event which Apple sponsored is essentially a ‘good cause’, it is not strange to see the change in the item ‘responsible’ because of the sponsorship (as argued in the introduction). And because the sponsorship might be somewhat far-fetched, but still related to Apple, the changes in ‘innovative’ and ‘ordinary’ can be explained.
40 Apple and a ‘good cause’ is not that ‘ordinary’, to say the least. For the charity, it might even be an extraordinary deal to increase awareness.
Beforehand, a positive change of positively loaded items was expected. The results shows that some positive items do change positive, but not all of them change, and ‘ordinary’ even changes negative. Therefore, H2a has to be rejected, as none of the ‘good personality items’ changes down. If anything, the opposite is true; while ‘good’ items like innovation change up, ‘bad’ items like ordinary change down. H2b also provides negative results: one negatively loaded item does change negatively (Aggressive), but the other (Bold) shows no change. H2c could be accepted, a lot of positive items change positively.
As past research pointed out, it is hard to predict what the outcomes in the mismatch category will be. In this case, Apple would profit from the sponsorship, but the next (mismatch) sponsorship might not be as successful, which makes the mismatch sponsorship a risky operation.
F-Match
With nine out of twelve items differing, including two complete factors (Responsibility and Aggressiveness), the F-match is the most effective in this research. Also, all nine items are higher when consumers were shown the press release (with the sponsorship). In the case of the Aggressiveness factor this could be seen as negative, but in general the items change favourably for the company.
H3b stated that an F-match sponsorship should change the brand personality factors Responsibility, Activity and Aggressiveness. This expectation is (almost completely) supported by the results; three out of three Responsibility items differ, two out of three Activity items and two out of two Aggressiveness items. This means H3 can be accepted.
I-Match
41
Research Question
Lastly, a comparison will be made between the four matches, to see which items change for which ‘kind of matches’. Hopefully, this will give an insight in the possibilities of the different kinds of sponsorships, and companies will benefit from the area of research by making better informed decisions about sponsorship deals.
Past research showed the complete match to be most favourable, but in this particular research the F-Match and the I-Match both provided better results. Although this is no reason to immediately change beliefs, it will be interesting to see whether these results lead to a different view on sponsorship congruency.
Table 12 underneath shows the differing items for each kind of match, and in which direction this change took place.
Complete match Mismatch F-match I-Match
Rabobank/Startersdays Apple/IDOAH* EA/Dutch Soccer Cup Adidas/WOSC
Down-to-earth x x ↑ ↑ Stable x x ↑ x Responsible x ↑ ↑ x Active x x ↑ ↑ Dynamic x ↑ x x Innovative ↓ ↑ ↑ x Aggressive ↓ ↑ ↑ ↓ Bold x x ↑ ↓ Ordinary x ↓ x ↑ Simple x x ↑ x Romantic x x x ↑ Sentimental x x ↑ ↑
Table 12: Item changes per match *= good cause sponsorship
Please keep in mind that these results are not generalizable: as this study was merely exploratory, the research design was simplified to gain a first comparison between differences between four ‘kinds of fit’. Generalizability was compromised in order to find these results and possibly start a discussion in this field of research.
42 (less) Activity positively, I-Match influence the Aggressivity factor in a positive way, and the not really useful factor Emotionality. And although they are in the same factor (Simplicity), it is notable to see that Ordinary differs only through the I-Match, while Simple differs only when a F-Match sponsorship is shown.
In the light of all the foregoing results and conclusions, the research question: ‘What is the influence (if any) of (different kinds of) fit on the brand personality traits of a brand/company?’, is almost answered.
All four test combinations experience a change in their brand personality. The change is the weakest in the case of a complete match sponsorship, and the strongest in the case of an F-Match. With the F-Match however, not all items differ favourably (Aggressiveness). For the I-Match, all items do change favourable, even the Aggressiveness factor. The (charitable) mismatch is regarding the results the 3rd most effective, but the results are not comprehensive and therefore quite hard to predict. This result means that even if a mismatch sponsorship is with a charity or good cause, it is still less effective than F-match or I-match sponsorships.
Comparing the results and conclusions of this research to past research, some major differences can be identified which allow and possibly even require further investigation. First of all, the complete match did not provide the comprehensive and total change that was expected beforehand. Although this came as a surprise, some past researches did find similar (contradictory) results regarding the complete match. A possible explanation for the ‘disappointing’ results regarding the Complete Match could be that strong image Rabobank has build up already. In hindsight, the choice for Rabobank might not have been the greatest choice, since their image (in Dutch society) is seen as very favorable, and hard to improve.
Still, this result could be seen as another wake-up call that shows not yet everything is known about the issue. More research is needed to improve knowledge, and to find out what the missing link between ‘success’ and ‘failure’ of a sponsorship is.
