The impact of a social influencer as a sports brand endorser
The influence of perceived expertise, person- and brand familiarity on financial, non-financial andbehavioral change outcomes.
Master’s Thesis
Student Anna Sofia Claus
Student number 10082247
University University of Amsterdam
Faculty Economics and Business
Study program Business Administration (MSc)
Specialization Digital Business
Supervisor M. Lee
Date of submission 23-06-2017
Preface
“Done is better than perfect.”
-Mark Zuckerberg
Statement of originality
This document is written by student Anna Sofia Claus, who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Table of contents
1. Introduction 6 1.1 Problem statement 8 1.2 Theoretical contribution 9 1.3 Practical contribution 10 1.4 Outline 10 2. Literature review 12 2.1 Social influencer 12 2.2 Non-financial impact 15 2.3 Financial impact 172.4 Behavioral change impact 19
2.5 Three models of endorsement effectiveness 22
2.5.1 The expertise of the endorser 23
2.5.2 The familiarity of the endorser 25
2.5.3 The familiarity of the brand 27
3. Hypotheses 29 4. Method 32 4.1 Stimuli development 32 4.2 Pre-test 33 4.3 Study design 35 4.4 Procedure 36 4.5 Measures 38 4.5.1 Dependent variables 38 4.5.2 Manipulation check 39 4.5.3 Control variables 39 4.6 Sample size 40 4.7 Data analysis 40 5. Results 41 5.1 Respondent analysis 41 5.2 Data preparation 41 5.3 Manipulation check 42 5.4 Hypothesis 1 46 5.5 Hypothesis 2 48 5.6 Hypothesis 3 49
5.7 Interrelation dependent variables 50
6. Discussion 52
6.1 Expertise 52
6.2 Person familiarity 53
6.3 Brand familiarity 56
6.4 Interrelation dependent variables 56
6.5 Theoretical implications 57
6.6 Practical implications 58
7. Conclusion 60
7.1 Limitations 61
7.2 Recommendations for further research 63
8. References 64
Abstract
The popularity of social influencers is growing, so does the impact that the social personas generate (Langner, Hennigs & Wiedmann, 2013, Bughin, 2015). Still, the established literature on endorsements primarily focuses on celebrity and athlete endorsers. Besides this narrow scope, the impact of endorsements is only measured on a limited number of outcomes, in a limited number of ways. The aim of the present study is to broaden the scope of endorsement literature by examining, especially the impact social influencers have as endorsers and by providing alternative outcomes, while also deepen the endorsement literature by providing alternative ways of measuring them. The outcomes central to this study are of financial and non-financial nature, but do also consider outcomes related to behavioral change. The customer’s willingness-to-pay was introduced as alternative to the common used financial return measures. The customer’s engagement was central to measuring the non-financial outcome, and the customer’s increased interest in exercising was translating the customer’s behavioral change. The present study focused on three characteristics of social influencer endorsements as independent variables: perceived expertise, perceived person familiarity and perceived brand familiarity. A 2 (expertise: expert/non-expert) x 2 (person familiarity: familiar/unfamiliar) x 2 (brand familiarity: familiar/unfamiliar) between-subjects experimental design was conducted among 284 respondents. Findings show no significant impact of brand familiarity on financial, non-financial and behavioral change outcomes. However, a perceived expert shows to significantly improve the customer’s health behaviors. Additionally, person familiarity leads to a significant increase of the customer’s willingness-to-pay in comparison with unfamiliar influencers.
Keywords: social influencer, endorsements, sports brands, perceived expertise, brand familiarity, person familiarity.
1. Introduction
In today’s advertising environment, it is extremely common for companies to use endorsers as representatives for a brand (Erdogan, Baker & Tagg, 2001). Since social influencers can contribute to the construction of stronger brands, due to their ability to transfer person related associations to the connected brand (Keller, 2005). In the last decades, companies primarily focused on acquiring celebrity endorsers. However, since the increase in popularity of social media platforms, a shift is taking place towards employing social influencers as brand endorsers. Social influencers are endorsing through blogs, tweets and other forms of social media (Freberg, Graham, McGaughey & Freberg, 2011). They are known for their interesting and influential position as role models within social groups (Langner, Hennigs & Wiedmann, 2013). Moreover, famous social influencers such as Chiara Ferragni, tend to have one to several million followers on social media. Therefore, an increasing number of organizations acknowledge the worth of the social influencer as endorser (Bughin, 2015). A good example of a brand that makes use of social influencers as endorsers, is Scandinavian watch brand Daniel Wellington. The brand employs hundreds of different social influencers simultaneously. The various social influencers, appraise the watch in a social media post and offer their followers a unique discount code. With this social influencer endorsement technique as business model, the brand saw its revenues grow with more than 200 per cent in one year (Parker, 2015).
Keller and Lehmann (2003), introduced the value chain model. The model serves to map and monitor how a marketing investment, such as the employment of an endorser, leads to certain returns. The first step in the model is the impact the endorsement might have non-financially, which is on the customer’s mindset, which is also known as customer-based brand equity. Customer-based brand equity contains brand awareness, -associations and -attitudes. Keller (2005) recognized the added value of using endorsers to leverage secondary associations, which
enhances brand awareness and brand associations and thus the customer-based brand equity. The second step, which has been subject to many studies, is the impact endorsements have on financial returns (Agrawal & Kamakura, 1995; Fizel, McNeil & Smaby, 2008; Ding, Molchanov & Stork, 2010; Farrell, Karels, Montfort & McClatchey, 2000; Gerritsen & Rheenen, 2017; Knittel & Stango, 2012; Alberse & Verleun, 2012). Measuring the endorsement’s financial returns varies from analyzing firm’s stock value, profit, price premiums and sales. An element that might be affected by endorsements, but is not covered by the brand value chain, is the impact of the marketing investment on behavioral change. An increasing number of non-endorsement related studies, examined and found that a role model, such as a social influencer, could have extensive impact on changing customer’s behavior by acting as a source of extrinsic motivation (Stulberg, 2014; Turchaninova, Khatri, Uyanik & Pavlidis, 2015). Thus, companies and their endorsements as marketing investment, could potentially lead to financial and non-financial returns, but could also lead to more extensive outcomes such as behavioral change.
An established collection of literature, emphasizes the conditions to which endorsements are most effective (Kamins, 1990; Homer & Kahle, 1990; Ohanian, 1991; Till & Busler, 1998; McGuire, 1985; Erdogan; 1999). The essential characteristics leading to optimal effectiveness, are central to three key endorsement models: the match-up hypothesis, the source credibility model and the source attractiveness model. The match-up hypothesis from Kamins (1990), stresses the importance of a fit between the endorser and the brand. Fit is a broad concept and entails, among others, the perceived level of expertise of the endorser (Ohanian, 1991; Till & Busler, 1998; Koernig & Boyd, 2009). The source credibility model builds upon the perceived level of expertise introduced in the match-up hypothesis. Being argued is that, the perceived level of expertise is influencing the purchase intention, brand attitudes and the level of
compliance of consumers (Ohanian, 1990; Woodside & Davenport, 1974; Crisco & Kassinove, 1973). The third model, the source attractiveness model, emphasizes the importance of the familiarity of the endorser. More familiar endorsers tend to be related to higher brand evaluations and purchase intent (Erdogan et al., 2001; Kamins and Gupta, 1994). However, between the three models reigns controversy on which characteristic is most decisive in the endorsement’s success. Influential characteristics introduced by the established literature, are the endorser’s familiarity and the perceived level of expertise. Moreover, the familiarity of the brand central to the endorsement, tends to play a role by influencing purchase intention and brand attitude (Axelrod, 1968; Haley & Case, 1979; Laroche, Kim & Zhou, 1996).
1.1 Problem statement
An extensive number of studies have examined the impact of celebrity endorsements and the conditions under which the endorsement turns out to be most effective. Central to these studies are the endorser’s perceived level of expertise and person familiarity as main drivers of endorsement effectiveness. The existing literature, measures the endorsement’s effectiveness in financial returns and in non-financial returns. A common way to measure the financial impact of the endorsements is to examine changes in stock prices. Whereas, brand awareness, brand associations and brand attitudes are research units for non-financial returns.
Hitherto, little research has broadened the scope of endorsement literature from solely focusing on celebrities and athletes, towards examining the growing number of social influencers and their role and impact. Remarkably, since the number of companies using an influencer is rising (Bughin, 2015). The existing literature repeatedly focuses on just one or two characteristics of the endorser in relation to endorsement effectiveness, therewith lacks to map the probable interrelation between a combination of more than two different features. Moreover,
endorsement literature looks at either financial or non-financial returns, but fails to examine both outcomes in one study. Additionally, the literature leaves the possible impact on customer’s behavior completely out of consideration. The possibility of alternative measurements of financial returns, is another element that stays undiscovered in the majority of literature. The main goal of the present study is to broaden and deepen the existing academic knowledge on endorsements by answering the following question: ‘How does the expertise and
familiarity of social influencer endorsers and familiarity of the endorsed brand, influence a brand financially and non-financially and how does it influence customer behavior?’
1.2 Theoretical contribution
To fill the previously discussed gap in knowledge, the present study aims to identify the relationships between perceived levels of expertise, person familiarity of the endorser and brand familiarity, on various outcomes. The different outcomes are non-financial, financial and behavioral change. The non-financial outcome will comprise the customer’s engagement with the brand. The financial outcome will be translated as the customer’s willingness-to-pay and the behavioral change will include the customer’s interest in exercising. By adding general theoretical insights on the social influencer as endorser, the present study contributes to the existing literature on endorsements. Social influencers are a noteworthy research subject, because of the interesting position social influencers acquired within social groups and the deep impact they tend to make on their followers (Salup, 2014; Bughin, 2015). Even though the existing literature acknowledges that a variety of endorsement characteristics individually leads to higher degrees of effectivity, little research has been done on the impact the different characteristics could have when considered as a group. By emphasizing the interrelation, the present research will extend existing literature on the conditions driving endorsement effectiveness. Moreover, the present study objects to propose an alternative to measure financial
returns, by examining customer’s willingness-to-pay. Lastly, growing attention exists for social media posts influencing customer’s behavior. Nevertheless, the social influencer’s role, which in many cases exists of being responsible for creating the social content, stays out of consideration in endorsement literature. Examining whether the social influencer has an impact on customer’s behavior, will deepen the current knowledge on the impact endorsement campaigns could have.
1.3 Practical contribution
The findings of the research could provide managers with general insights on the impact of social influencers. Moreover, by knowing which combinations of perceived expertise, brand familiarity and person familiarity leads to the highest financial, non-financial and behavioral change returns, will provide managers with guidelines on which type of social influencer to obtain. With the insights provided by the present research, managers will be able to carefully select the endorser best suited for the situation and the desired outcome(s). Employing social influencers is an expensive financial investment, with the given insights, losses related to choosing a less suitable social influencer unable to create effective endorsements, could be prevented.
1.4 Outline
Subsequently to this introduction chapter, a literature review (chapter 2) will follow. The chapter will contain an overview of the existing literature on social influencers, the possible impact on financial, non-financial and behavioral change returns and the three different models emphasizing endorsement effectiveness. The literature review, will result in the hypotheses and the related conceptual model central to this study (chapter 3). Chapter 4, will be dedicated to the experiment and questionnaire used for the present research, followed by presenting the
results in chapter 5. Lastly, the present research will end with chapter 6 and 7 containing the discussion and conclusion.
2. Literature review
Since the introduction of endorsements to the advertisement world, there has been research on the effectiveness of it. Albeit the academic consensus about certain factors influencing the effectiveness, whether this is a consequence of specific characteristics of the endorser, the brand, or maybe both, remains debatable. This literature review will first examine a new and modern type endorser, the social influencer. The review will then touch upon the different outcomes an endorsement could have, divided into: a financial impact, a non-financial impact and the impact on behavioral change. Finally, the three main models on endorsement effectiveness will be discussed.
2.1 Social influencers
Decennia ago, companies started to make use of endorsements, which are a form of advertisement, and are a part of marketing activities. Endorsements are used to manage and enhance brand equity. Managing brand equity means trying to control and influence the meanings that are associated with a brand in the mind of the consumers (Keller, 1993). Aaker (1991) defines brand equity as “the tangible and intangible value that a brand provides positively or negatively to an organization” (p. 1). By using endorsements as marketing activity, companies try to leverage associations connected to known individuals, to that of the company, to control and strengthen brand equity (Keller, 2005).
With the popularity explosion of blogging and the rapid rise of social media platforms like Twitter and Facebook around 2005, and later on Instagram and Snapchat, a whole new endorsement opportunity emerged to the world of advertisement. The opportunity consists of the employment of social influencers to leverage associations and build equity. Freberg, Graham, McGaughey and Freberg (2011), define a social media influencer as “a new type of
independent third party endorser who shape audience attitudes through blogs, tweets, and the use of other social media” (p. 1). Social influencers are being defined as “role models - and even as leaders - inside social groups that affect group (and buying) behavior” (Langner, Hennigs & Wiedmann, 2013). The social influencer makes recommendations on buying or using a certain product. Along with the rise of internet and social media, and consequently the exposure and fame the social influencer tends to get, the line between being a celebrity and being a social influencer is starting to blur. There still are, however, some differences between the two. The deployment of celebrity endorsements, is a popular marketing strategy according to Forbes (2015), mainly because “we [marketers] view celebrities as influential, aspirational and most importantly, marketable”. The same applies to social influencers, but since the social influencer is not a celebrity, but someone inside a social group, it becomes easier for followers to identify with the endorser. Often, social influencers have a smaller audience, in the sense of less followers, but the messages they spread tend to be of a deeper impact than that of celebrities (Salup, 2014). Additionally, consumers tend to trust influencers more than celebrities (Salup, 2014).
The first time a social influencer was employed, it was in exchange for a goodie bag full of free products. Not much later, in 2015, Nike spent 3,2 billion on what they call ‘demand creation expenses’, including advertisement and endorsement contract costs (Nike INC, 2015). At that same time, CBS reported that the blog The Blonde Salad of Chiara Ferragni’s, one of Nike’s social influencers, became worth 8 million U.S. dollars (CBS, 2015). In the same year of Nike’s and CBS’s announcements, Bughin (2015) reported a rise in the use of social influencer marketing and not without good reason. He found that the recommendations from social influencers have a significant impact on the buying decisions of customers, especially on first-time buyers. The recommendations account for around 26 per cent of the purchases made and
this percentage is increasing. Heretofore, this impact has been underestimated around just 10 per cent (Bughin, 2015). However, Bughin argues that only a small share of social influencers is actually able to generate this astonishing impact. In other words, just a small part of the total amount of social influencers, is responsible for a disproportional share of the generated returns from corresponding endorsements. Additionally, Bughin (2015) found that every product category has a specific, limited amount of successful social influencers. The most influential endorsers are operating in the fashion industry. Even though offline recommendations are, for now, still more influential than social influencer marketing (Bughin, 2015), this could quickly change due to the decreasing consumption of traditional media caused by the increase of the internet consumption (Stempel, Hargrove & Bernt, 2000; Pew Research Center for the People and the Press, 2000; Katz, 1999, Dutta-Bergman, 2004). Therefore, it is of importance to monitor the rapidly growing impact of social media marketing.
Although, measuring the exact non-financial and financial impact of the activities generated by influencers, remains a challenge for most companies, Keller and Lehmann (2003) introduced the brand value chain. The chain seeks to monitor and translate marketing investments into outcomes. This chain shows how a marketing program investment, such as a social influencer endorsement campaign, influences a firm’s brand equity. Keller and Lehmann argue that a marketing program influences the customer’s mindset, has an effect on a company’s market performance and finally has an impact on shareholder value. The chain shows that the impact of a marketing program on brand equity consists of two components: a financial, the company’s market performance, and non-financial part, the customer’s mindset. Besides the financial and non-financial outcomes, an increasing number of research focusses on the impact endorsements have on the health behaviors of the customers. The brand value chain of Keller and Lehmann (2003) does not discuss marketing investments having an impact on the health behaviors of its
consumers. However, Kotler and Zaltman (1971) do. For companies employing social influencers to affect customer’s physical behavior, Kotler and Zaltman (1971) introduced the term social marketing. Social marketing are marketing principles seeking to influence social behaviors to benefit the welfare of the target audience, instead of only that of the organization. Social marketing focuses on changing ideas, attitudes and behaviors that are related to complex social-, health- and/or environmental topics. Sports brands are engaging in social marketing by encouraging its target group to get more active to decrease the risk of deceases linked to obesity. The growing academic interest in the impact of marketing programs on health behavior, is part of a more general trend of corporate social responsibility. Corporate social responsibility entails that companies take a more active stand in being responsible for their direct and indirect social surroundings. Within this trend, companies pay more interest in the health and welfare of their employees and customers (BSR, 2013). Companies pursuing behavioral change is probably the most profound impact of an endorsement campaign that companies generate, at the same time this is also the most challenging.
2.2 Non-financial impact
Following the brand value chain of Keller and Lehmann (2003), this theoretical framework will first consider the non-financial outcomes as a result of endorsements. Keller and Lehmann argue that endorsements can influence the customer’s mindset and build customer-based brand equity. The customer’s mindset includes the brand awareness, the associations customers have with the brand and the attitudes towards it.
Prominent literature on customer-based brand equity is written by Aaker (1991) and Keller (1993). A widely-accepted definition of brand equity is: “a set of assets (or liabilities) consisting of brand awareness, brand associations, perceived quality, brand loyalty and other proprietary
assets” (Aaker, 1991, p. 16). Keller complements this definition by defining brand equity as: “the differential effect of brand knowledge on consumer response to the marketing of the brand” (Keller, 1993, p. 2). Aaker’s definition on brand equity encloses four dimensions of brand equity: brand awareness, brand associations, perceived quality and brand loyalty, which are merged into the customer-based brand equity (CBBE) pyramid of Keller (2001). A social influencer could be employed to reach the highest level of the CBBE pyramid, loyalty. Looking at the first step of the pyramid, the social influencer could increase brand awareness. Keller (2001), describes brand awareness as how easily the brand is being recalled or recognized (depth) and the range of different situations in which the brand comes to mind (breadth). Since the social influencer adds situations to the range of moments the brand could come to mind, they contribute to the breadth of the brand. Additionally, according to Keller (2013), endorsers could be of value in the second stage of the pyramid due to their ability to transfer associations linked to themselves, to a brand. The process of transferring association between people, places or events is called leveraging secondary associations. By the careful recruitment of a social influencer as brand endorser, brands could strengthen the valuable associations that customers have had before the employment of the endorser, but could also leverage associations that are new to the brand, but could of value in the creation of brand loyalty. Therefore, social influencers as brand endorsers, could help in the creation of brand meaning by adding meaningful associations to the brand. For instance, a certain sports brand is known for their high quality, but is also somewhat dull and outdated. By employing a young and popular social influencer, who is interested in sports, the associations of him or her being young and trendy are leveraged to the brand, causing the currently existing associations of being outdated and dull to be diminished. Eventually, the new associations related to the brand, could result in more customers wanting to be actively engaged with the brand (brand loyalty). Thus, as acknowledged by Keller (2003), social influencers can serve as a beneficial technique in
enhancing brand awareness and brand associations. Therewith, the social influencer contributes to the customer-based brand equity building of the company.
2.3 Financial impact
The next elements discussed in the brand value chain of Keller and Lehmann (2003), are that of the company’s market performance and the impact on shareholder value. The two steps resemble the financial outcome of a marketing investment and are translated into stock prices and market capitalization. Despite the brand value chain emphasizing the financial outcomes of endorsements, many of the studies on the effectiveness of endorsements, focus solely on evaluating the non-financial customer’s responses (Farrell, Karels, Montfort & McClatchey, 2000; Agrawal & Kamakura, 1995; Chung, Derdenger & Srinivisan, 2013). Besides a minority of academic literature even considering the financial aspect, there reigns controversy on whether endorsements actually lead to higher economic value for companies or not. A possible explanation for the dissimilar findings on endorsements’ economic value, is presented by Agrawal and Kamakura (1995) stating that measuring the direct financial link between the endorser and the firm’s returns, is almost impossible to measure.
Despite the level of difficulty, Agrawal and Kamakura (1995) seek to measure the economic effect by assessing the impact the announcement of a new endorsement contract has on the firm’s stock value. Agrawal and Kamakura find by using an event study methodology, that in general, the employment of an endorser has a positive effect on a firm’s stock returns. Which means, that endorsements are a worthwhile investment in advertising (Agrawal & Kamakura, 1995). Both, Fizel, McNeil and Smaby (2008), and Ding, Molchanov and Stork (2010), redo the research from 1995 by Agrawal and Kamakura, using the same methodology to evaluate a firm’s returns after a contract announcement. Nonetheless, the study of Fizel et al., differs by
examining the use of conventional celebrity athletes instead of A-rated megastar athletes. Both studies, Ding et al. and Fizel et al., find no significant returns around the announcement dates, and therefore contradict the findings of Agrawal and Kamakura (1995). While all three studies (Agrawal & Kamakura, 1995; Fizel, McNeil & Smaby, 2008; Ding, Molchanov & Stork, 2010), measure the stock value at the time of the announcement of the endorsement contract, they do not capture the impact of value changes subsequent to the announcement. To overcome this shortcoming, Farrell, Karels, Montfort and McClatchey (2000) extend the focus from one specific moment in time, to a longitudinal approach on firm value. The study looks at the relationship between Tiger Wood’s performance during several tournaments and the returns on the companies for which he endorses. Farrell et al. (2000) conclude, that the market value of endorsements continues well beyond only the impact associated with the announcement day. Probably because a higher performance of the endorser leads to higher exposure, which triggers sales and firm value (Gerritsen & Rheenen, 2017; Farrell, Karels, Montfort & McClatchey, 2000). Chung, Derdenger and Srinivasan (2013), also examined the long-term economic value generated by professional golf star Tiger Woods. They found that there is a significant so called
Tiger Woods endorsement effect on profit, but also on prices. Not only did contracting Tiger
Woods result in an additional profit of $103 million from 2000 to 2010, the endorsements also resulted in Nike being able to command a price premium of more than 2.5 per cent compared to its competitors.
Knittel and Stango (2012), examined the effect of a celebrity endorser involved in scandals have on a firm’s value. Whereas most research looked at the stock value as an isolated entity, Knittel and Stango compared the changes in stock value to data provided by Google Insights. Google Insights was used to see whether changes in stock value are related to endorsement-related news. They found that stock value losses are greater on days where the intensity of
endorsement-related news was higher. Knittel and Stango classified this result as: “statistically significant and economically substantive” (p. 1). Alberse and Verleun (2012) follow the example of Knittel and Stango (2012), and try with their paper The economic value of celebrity
endorsements to complement earlier research on the topic. Aside from taking the effects on just
stock returns into consideration, they try to consolidate their findings by considering the endorsement effects on corporate sales too. The findings of this paper are in line with the conclusion stated by Agrawal and Kamakura (1995): generally, contracting a celebrity for endorsement purposes has a positive effect on sales, increases the stock returns and therefore is considered as an investment that pays off. Despite the majority of the studies examining changes in stock prices and sales, Grange (2001) took another approach and looked at market share. He found that the market share of Nike golf grow from below 1 per cent to 3.9 per cent in only eight months time, in the time Tiger Woods changed from Titleist to Nike as ball sponsor.
2.4 Behavioral change impact
Even though, little is known about the effect of specifically a social media endorser as influencer of physical activity behavior, more and more general research on physical activity behaviors, has shown the strong impact of social surroundings on the level of activity of an individual (Stulberg, 2014; Turchaninova, Khatri, Uyanik & Pavlidis, 2015). Since social influencers are part of one’s social group (Lagner, Hennigs & Wiedmann, 2013), they could possibly have such a profound influence. For a social influencer to reach this form of impact, the endorser should be prominent as a socialization agent, also known as role model. Moschis and Churchill (1978), define socialization agents, as entities who transfer norms, attitudes and behavior to the consumer of the message. The role model does not, necessarily, need to be in direct contact with the learner. Bandura (1986), calls this a vicarious role model and argues the
role model’s influence on the consumer via electronic or print media. Studies examining the impact of vicarious role models on adolescents, found that the socialization agents can have a significant impact on the career aspirations, educational choices and the self-view of the consumer (Bush, Martin & Bush, 2004). The role models are accountable for the behavioral intentions of the learner, these intentions can be both positive and negative. The favorable intentions could lead to increased sales of the brand, spreading positive worth of mouth and increased brand loyalty. Moreover, according to Durantini, Albarracín, Earl & Mitchell (2006), a role model perceived as expert has a greater positive effect on behavioral change, than a non-expert.
Once the social influencer has positioned itself as vicarious role model, it needs to motivate the individual. Motivation consists of two parts, intrinsic and extrinsic, which can be divided into four categories according to the self-determination theory of motivation (Deci & Ryan, 1985): a-motivation, other-determined extrinsic motivation, self-determined extrinsic motivation and intrinsic motivation. After passing the a-motivation phase, which is characterized by a complete lack of motivation, people move into the other-determined extrinsic motivation phase, which contains the impact of social influencers. In this phase, external factors such as rewards and punishments, provide motivation. The third level of motivation is that of self-determined extrinsic motivation. The third phase is characterized by the enjoyment the individual gets from the activity because of the social aspect or by understanding the health benefits. The most desired level of motivation is that of intrinsic motivation. In this phase, the motivation is based on the appreciation of the activity itself and is therefore independent of any external factors.
Because social influencers are extrinsic sources of motivation, the effectiveness of them being drivers of behavioral change, is being questioned. Since research shows, that intrinsically
motivated people are more easily actively engaged than extrinsically motivated people (Ryan, Kuhl & Deci, 1997; Kilpatrick, Hebert & Jacobsen, 2002). In line with this notion, according to different findings, especially women who already regularly exercise, pay more attention to sports and exercise related posts, than their non-exercising counterparts (Berry, Spence & Stolp, 2011). However, even though the groups of already regularly exercising women are more likely to pay attention to social influencers endorsing sports products, this group is less sensitive to incentives coming from an extrinsic motivation source, since they already reached higher levels of behavioral change (Petri, 1991; Deci & Ryan, 1985). Whereas, especially the group of non-exercising individuals, would benefit from the social influencer extrinsically motivating them (Berry et al., 2011), this is often not the group which the social influencer tends to reach. However, in the beginning, the behavioral change of an individual is relying on extrinsic forms of motivation, according to Petri (1991) and Deci and Ryan (1985). Only later in the process, the motivation leading the change should transform into intrinsic motivation to be of long-term impact. Thus, the extrinsic motivation provided by the social influencer, could be of value at the starting point of changing the physical behavior of the consumer.
A social influencer operating as a role model can also have a negative impact on the receiver of the message. Research shows that trends such as fitspiration posts, in which the social influencer tries to encourage and extrinsically motivate the receiver by using quotes and training photographs to attain a fitness goal, could negatively affect the self-esteem and self-image of women (Hausenblas, Campbell, Menzel, Doughty, Levine & Thompson, 2013; Tiggerman & Miller, 2010; Teixera, Carraça, Markland, Silva & Ryan, 2012; Schreiber & Hausenblas, 2015; Adams, 2015). Finding the right balance is difficult, since using a text or quote combined with a photo where the endorser is using the product, is also a very common way to endorse sports gear.
As discussed by Keller (2003), social influencers serve as secondary sources of customer-based brand equity, by leveraging associations from the social influencers, to the brand. By doing so, social influencers are able to enhance brand awareness and brand image. Besides social influencers boosting the customer-based brand equity, research shows that contracting endorsers and leveraging their associations, could also lead to a higher quantifiable economic value of the company (Agrawal & Kamakura, 1995; Farrell et al, 2000; Chung et al, 2013; Knittel & Stango, 2012; Alberse & Verleun, 2012). Finally, there is a growing attention for the influence of the social surroundings of individuals on physical activity. Social influencers being part of this social surrounding as vicarious role models, could influence consumers by motivating them extrinsically into exercising more often (Stulberg, 2014; Turchaninova et al, 2015; Bandura, 1986; Deci & Ryan, 1985; Petri, 1991). Thus, deploying social influencers has non-financial outcomes in the form of increased customer-based brand equity, financial outcomes such as higher profits and share prices, but also could lead to behavioral change related to customer’s health.
2.5 Three models on endorsement effectiveness
Whether the social influencer’s endorsements are effective and leading to the desired returns, is primarily based on three leading models: the match-up hypothesis, the source credibility model and the source attractiveness model. The models consider the various dynamics that have impact on the success of the endorsement. The models show that certain characteristics of the endorser, positively affect the receptiveness and thus the effectiveness of the endorsements (Erdogan, 1999).
2.5.1 The expertise of the endorser
The first theory that is controlling endorsement literature, is the match-up hypothesis. Kamins (1990) introduced the Match-Up Hypothesis, which entails the importance of a sense of fit between the celebrity endorser and the endorsed brand. The fit between endorser and product is not only based on attractiveness, as has been suggested by Erdogan, Baker and Tagg (2001), nor should it only consider perceived expertise (Ohanian, 1991; Till & Busler, 1998). For the endorsement to be effective, there should be congruence between the complete perceived image of the endorser and the image of the endorsed product (Erdogan, 1999). Especially athlete endorsers are being used as research topic, because they are perceived as an obvious match on expertise (leading to high credibility), when endorsing sports products (Koernig & Boyd, 2009). Till and Busler (2013) examined the match-up factor expertise. They looked at the fit between the endorser and the product, being athletes endorsing energy bars, versus actors endorsing energy bars or candy bars. The results show that expertise is an important match-up factor when it comes to brand attitude, purchase intent and the fit between product and endorser. Nevertheless, Fizel, McNeil and Smaby (2008) contradict the Match-up Hypothesis of Kamins (1990), by testing the notions of trustworthiness and expertise. The findings suggest that there is no significant greater positive economic effect when an athlete is endorsing sports products, than when an athlete is endorsing non-sports products.
The second model that emphasizes the expertise of the endorser, is the source credibility model. The source credibility model contends that the credibility of the endorser consists of the perceived level of expertise and its trustworthiness. The level of credibility determines the effectiveness of the message. More specifically, the higher the level of perceived credibility, the more persuasive the source. Erdogan defines expertise as “the extent to which a communicator is perceived to be a source of valid assertions” (p. 298) and comprises
knowledge, experience and the skills of the communicator. To which extent the communicator truly is an expert matters little, the extent to which the consumer perceives him or her as an expert is what matters most (Hovland et al., 1953; McCracken, 1989; Ohanian, 1990; Erdogan, 1990). Trustworthiness refers to “the honesty, integrity and believability of an endorser” (Erdogan, 1990, p. 297). According to Friedman and Friedman (1978), trustworthiness is the most influential determinant of the model. Ohanian (1991) disagrees after finding that trustworthiness has no significant effect on the purchase intention of the customer. Instead, Homer and Kahle (1990) argue that expertise is more important.
A broad range of literature examined the impact of expertise. First, Speck, Schumann and Thompson (1988) argue that expert endorsers enlarge the awareness of the brand. According to Ohanian (1990) “the source’s perceived expertise has a positive impact on attitude change” too (p. 42). For instance, that the level of expertise tends to have a positive influence on the perception of the product’s qualityand the purchase intention. In comparison, Ohanian (1991) found that among other endorser characteristics, such as attractiveness and trustworthiness, expertise was the only factor significantly associated with customers’ purchase intention. Findings of Till and Busler’s study (1998) support those of Ohanian (1991), and claim that perceived experts endorsing a product matching this level of expertise will be more effective. This is probably, because a perceived expert is shown to be more persuasive. For example, in terms of selling, customers’ purchase intention increased significantly when an expert salesperson was involved, versus a non-expert (Woodside & Davenport, 1974). Experts being more persuasive is in line with findings of Crisci and Kassinove (1973), suggesting that the perceived level of expertise influences the level of compliance. Respondents have a higher compliance with the recommendations of a communicator that is perceived an expert, than with a non-expert.
Thus, research shows that expertise is an important characteristic affecting the customer’s beliefs and behavior and could predict the endorser’s effectiveness. However, the findings of Amos, Holmes and Strutton (2008) shed a slightly other light on the importance of expertise. Based on their research, expertise finishes only in 3rd place, behind the impact of ‘negative information’ on the endorser and its trustworthiness. Nevertheless, Amos et al. (2008) do report that the influence of expertise still is significant.
2.5.2 The familiarity of the endorser
The source attractiveness model is an extension of the source credibility model and consists of the familiarity, likeability and similarity of the endorser (McGuire, 1985). The idea behind the source attractiveness model is that of identification. Due to the desire to identify with an attractive endorser, the message of the source is being accepted. Attractiveness is not only referring to physical attractiveness, but also to athletic prowess and/or personality traits for instance (Erdogan, 1999). With likeability is meant the “affection for the source as a result of the source’s physical appearance and behavior” (Erdogan, 1999, p. 299). Research have found that likeability is closely related to the trustworthiness and level of expertise of the endorser. Thus, likeable endorsers are being perceived as more trustworthy and are possessing higher levels of perceived expertise. Similarity is defined as “a supposed resemblance between the source and the receiver of the message” (Erdogan, 1999, p. 299).
Familiarity is the “knowledge of the source through exposure” (Erdogan, 1999, p. 299) and is one of the three factors that influences the effectiveness of the communicator’s message debated by McGuire (1985). Familiarity contributes to the overall perceived credibility of the endorser (Ohanian, 1990), and is therefore an impactful construct. Often, familiarity and likeability, are
being treated as entwined, however Amos et al. (2008) perceive familiarity as an unconnected element in their study. Kamins and Gupta (1994) do this as well, and find that the congruency between product and endorser is positively affected by the familiarity of the endorser. In other words, endorser familiarity acts as an important factor in generating effective endorsements. According to Siemens, Smith, Fisher and Jensen (2008), familiarity as a construct is more powerful than the perceived expertise of the endorser in terms of perceived credibility. This is consistent with Erdogan, Baker and Tagg (2001) highlighting the importance of celebrity familiarity as a fundamental criterion in selecting the right celebrity endorser for a product. Familiarity is also an important cause of higher brand evaluations and purchase intent (Erdogan et al., 2001; Kamins and Gupta, 1994).
McGuire states that influencers who are well-known to consumers are more persuasive than when they are not, because of their recognizable persona. However, TapInfluence (2017) contradicts McGuire, after studying the trend on so called micro-influencers and comparing the impact of these low familiarity influencers, to more popular personas. Micro-influencers are being defined as social personas with between 10.000 and 100.000 followers on, besides others, Instagram. The influencers scoring high on familiarity have somewhere between the 1 and 10 million followers. Despite of the relatively smaller reach of the micro-influencer compared to the celebrity influencer, the impact measured in likes is remarkable: micro-influencers tend to have a like-rate of 2.4 per cent, whereas celebrity influencers are stuck around 1.7 percent. Keller Fay Group in cooperation with Experticity and Berger (2016), conducted a study to evaluate the impact of the recommendations made by influencers. He found that around 82% of consumers are ‘highly likely’ to follow the recommendation made by the micro-influencer, probably because customers perceive micro-influencers as credible, believable and knowledgeable. Thus, states Tapinfluence, celebrity endorsers could be helpful in increasing
brand awareness, while micro-influencers could account for more engagement and brand loyalty, but also increase the product sales. Additionally, they found that micro-influencers tend to be more active in niche markets, compared to celebrity influencers operating in the mass categories.
Interesting findings of Pritzen (2012) show that the perceived expertise of the endorser is higher in case of an unfamiliar endorser, than with a familiar endorser. According to Pritzen, a possible cause for this surprising outcome could be the overexposure to celebrity endorsements. The assumption is based on the high number of endorsements performed by celebrities, which negatively impacts the endorsers trustworthiness (Tripp, Jensen & Carlson, 1994; Tripp et al., 1994; Mowen & Brown, 1981). Additionally, Pritzen’s study also contradicts the research of Kamins and Gupta (1994) and Erdogan et al. (2001) on the effects of familiarity on purchase and change in attitudes. The findings of Pritzen show that, while just the opposite was expected, the unfamiliar endorser leads to a higher purchase intention, attitudes and perceived credibility.
2.5.3 The familiarity of the brand
While the previously discussed literature covers endorser familiarity, brand familiarity tends to influence the effectiveness of an endorsement as well. Brand familiarity is part of consumer knowledge, which can be defined as “information relating to the products and to the brands which are stored in the memory of the consumer” (Engel, Blackwell & Miniard, 1995). Axelrod (1968) and Haley and Case (1979) reported about the correlation between brand familiarity and brand choice. The correlation between brand familiarity and eventually, purchase intention has to do with the three stages of consumer decision introduced by Wright and Barbour (1975). The first stage is that of the pool of alternatives, also known as the consideration set (Howard & Sheth, 1969). Out of the large number of brands an individual knows, only a few brands are
being considered to be purchased. If a person is not familiar with a brand, the brand cannot be part of one’s consideration set. When a person is aware of a certain brand, the possibility of the brand being considered for purchase is increasing. Thus, a brand scoring high on familiarity has a greater chance of being part of a customer’s consideration set and eventually a greater chance of being purchased compared to a brand that scores low on familiarity. Moreover, brand familiarity is found to be positively influencing consumer’s confidence towards a brand. In other words, brand familiarity is increasing “the state of the buyer feeling sure of making a judgement on the quality of a particular brand” (Laroche, Kim & Zhou, 1996, p. 115). Finally, according to Laroche, Kim and Zhou (1996), brand familiarity is positively related to brand attitude. Even though research of Zammit (2016) supports the findings of Laroche et al (1996), she argues that the link between brand familiarity and purchase intention, mediated by brand attitude is only a weak positive relationship. Thus, brand familiarity influences the consumer’s attitude towards the brand, and eventually what brands they buy.
3. Hypotheses
Based on the literature review, this chapter will introduce the different hypothesis central to this study and the conceptual model.
As indicated in the theoretical framework, an endorser could serve as secondary source to transfer associations from him or herself to the brand. The effectiveness of this process and its impact on financial, non-financial and behavioral change outcomes is influenced by the social endorser’s level of expertise. It appears to be so, that the level of expertise of the endorser positively influences brand awareness (Speck et al, 1988; Ohanian, 1990; Till & Busler, 2000). Not only does the perceived level of expertise influence the non-financial equity of a brand, it also affects the financial returns of a company. Starting with Crisci and Kassinove (1973), individuals tend to have more compliance with recommendations provided by perceived experts than people have with their non-expert counterparts. The compliance leads to a higher purchase intention (Ohanian, 1991). Eventually, being a perceived expert leads to higher sales, according to Till and Busler (1990) and Woodside and Davenport (1974). Additionally, since people are more in compliance with an expert and their messages (Crisci & Kassinove, 1973), this present research tests a positive relationship between expertise and outcomes on behavioral change. Therefore:
H1: The difference in all three different outcomes is higher when the perceived expertise is high, as opposed to when the perceived expertise is low.
Despite the positive relation between expertise and the different outcomes, person familiarity tends to generate a contradicting effect. Pritzen (2012) argues that the effects on the three different outcomes, is higher with an unfamiliar endorser, than with a familiar one. Remarkably,
the endorser scoring low on person familiarity creates more engagement and more positive brand attitudes (Pritzen, 2012; TapInfluence, 2017). Besides, high person familiarity lowers purchase intentions (Pritzen, 2012) and ultimately the number of sales (TapInfluence, 2017). Lastly, extensive research is done on the effects of sports related endorsements and women’s self image (Hausenblas, Campbell, Menzel, Doughty, Levine & Thompson, 2013; Tiggerman & Miller, 2010; Teixera, Carraça, Markland, Silva & Ryan, 2012; Schreiber & Hausenblas, 2015; Adams, 2015). Person familiarity is leading to reversed effects on behavioral change. Meaning, in some cases, the endorsements of the social influencers is leading to low self-esteem. The above discussion is leading to the following hypothesis:
H2: The difference in all three different outcomes is lower when the perceived person familiarity is high, as opposed to when the perceived person familiarity is low.
In the present study, the impact of brand familiarity on the three dependent variables is tested. Research of Laroche et al (1996), shows that brand familiarity has a positive interaction with the attitude towards a brand. Moreover, high brand familiarity tends to increase the purchase intention (Zammit, 2016; Haley & Case, 1979). Resulting in the following hypothesis:
H3: The difference in all three different outcomes is higher when the perceived brand familiarity is high, as opposed to when the perceived brand familiarity is low.
Expertise Financial outcome H1 + Brand familiarity Person familiarity
Figure 1. Conceptual model displaying the relationships between endorser perceived expertise, person familiarity, brand
familiarity and the three different outcomes being: financial, non-financial and behavioral change.
Non-financial outcome
Behavioral change outcome H2 - H3 + H3 + H3 + H2 - H2 - H1 + H1 +
4. Method
To enable answering the central hypotheses and related research question, an empirical experiment was conducted. Prior to the main experiment, a pre-test was conducted and a focus group were formed to develop suitable scenarios. The actual experiment consisted of a 2 (expertise: expert/non-expert) x2 (person familiarity: familiar/unfamiliar) x2 (brand familiarity: familiar/unfamiliar) between-subjects design. In the experiment, respondents were asked to observe an endorsement by a social influencer on social media. Subsequently, the respondents were asked to fill out a questionnaire.
4.1 Stimuli development
To create stimuli that could successfully represent the various manipulations for the experiment, a set of criteria was determined that would have to be met in order to optimize the various manipulations of the experiment. Firstly, nor the endorser or the brand should evoke extremely positive or negative emotions and/or attitudes, to prevent bias towards the presented scenarios. Furthermore, the various endorsements should be realistic, nonetheless hypothetical. Since, when the endorsement is perceived as unrealistic this could interfere with the attitudes and emotions that the scenario is trying to evoke (i.e. if someone perceives the information as something that could never occur in real life, he or she will possibly have a different attitude towards an endorser trying to stimulate the respondent in exercising more often). However, the presented endorsement should be hypothetical at the same time, to make sure that the attitudes and cognitions that respondents might have with previous real life campaigns, would not interfere with the currently presented endorsement.
Based on the above criteria, eight scenarios were created combining social influencer endorsers and brands (table 1). The endorser selected to represent expertise and familiarity, is Dutch track
running athlete Dafne Schippers. In 2015, she became the first Dutch athlete to win the 200 meters sprint with a European time record. She became immensely popular by this achievement and became highly familiar to the Dutch population due to the extensive coverage she got on news- and talk shows. To embody the unfamiliar expert, Kara Koucher was selected. Since running is a niche sports discipline, chances are low that the general Dutch public would be familiar with an American athlete, who performed quite moderately throughout her career and who’s most recent successes were recorded a decade ago. These characteristics make Kara Koucher an expert in the field of running, but in all probability unknown to the respondents. Chiara Ferragni was chosen to resemble a non-running expert, however well known to the public, due to her enormously popular fashion blog, The Blonde Salad. The fourth endorser was the fictitious person Sofia Wayne. Sofia Wayne was created to reach a high person unfamiliarity and non-expert response rate. Nike and Oiselle were the brands represented in the experiment. Nike, being globally an immensely popular sports brand, served as the familiar brand, whereas Oiselle, an American women track brand, is used to resemble brand unfamiliarity.
Brand familiarity Person familiarity
High Low
Expertise
High
High Dafne Schippers
endorsing Nike
Kara Koucher endorsing Nike Low Dafne Schippers endorsing
Oiselle
Kara Koucher endorsing Oiselle
Low
High Chiara Ferragni
endorsing Nike
Sofia Wayne endorsing Nike Low Chiara Ferragni endorsing
Oiselle
Sofia Wayne endorsing Oiselle
4.2 Pre-test
First, the eight different scenarios were examined in a pre-test (N = 22), to see whether the
Table 1. Table displaying different kinds of social influencers, corresponding brands and criteria on expertise,
stimuli portrayed the manipulation goals adequately and whether the manipulations were being interpreted as intended. The pre-test consisted of both a paper version and an online version of the experiment, provided via Qualtrics, in which four different scenarios per person were randomly shown. Low expertise, low person/brand familiarity scenarios should score significantly lower than the midpoint of the 5-point Likert scale (i.e. below 3). High expertise and high person/brand familiarity scenarios should, vice versa, score significantly above 3. The majority of the combinations of the endorser’s perceived level of expertise and familiarity, met the criteria which were set beforehand. More specifically, the scenarios concerning athlete Dafne Schippers were evaluated high on perceived expertise and person familiarity, whereas lesser known athlete Kara Goucher scored reasonably high on perceived level of expertise, but low on person familiarity. Also, the fictitious person Sofia Wayne met the desired results concerning person familiarity and the perceived level of expertise. The respondents evaluated Sofia low on perceived level of expertise and low on person familiarity. However, the selected person to represent an endorser with a low level of expertise, but high level of person familiarity, Chiara Ferragni, was not meeting the desired outcomes set prior to the pre-test. Namely, the pre-test revealed that all 22 respondents participating in the pre-test were unfamiliar with Chiara Ferragni and did not know who she was. Therefore, the scenarios concerning Chiara as the non-expert, familiar endorser, did not score significantly above 3. Remarkable, since she has over 8.5 million followers and is often mentioned as one of the most influential social personas of her time (Sowray, 2015; Cochrane, 2016; Indvik, 2016). Whilst further analyzing the result of the pre-test, it became apparent that the answers given to the question making up the financial outcome scale, was very low: 9 of the 22 participants answered their willingness-to-pay was €0.
group served to evaluate the general quality of the experiment and the accompanying questionnaire, to find a solution for the possible underperformance of the financial outcome scale and to find adequate replacement for Chiara Ferragni.
Concerning the overall quality, the participants mentioned two possible adjustments concerning the flow of the questionnaire. One suggestion was dividing the welcome message from the message defining what social influencers are. The focus group mentioned that this could increase the number of respondents paying careful attention to the important information on social influencers. The second suggestion that was proposed, was showing the endorsement together with the corresponding questions, instead of putting a page break in between. By doing so, the respondents did not need to switch back and forth between pages, and could therefore improve the answers provided. Another adjustment, concerned the financial outcome scale. The suggestion was coming from male respondents. Since the product being endorsed is a sports bra, a lot of male respondents answered €0 when asked to fill in their willingness-to-pay for the endorsed product. In collaboration with the focus group, the question was changed into: “assume you want this product (for yourself, girlfriend, sister, mother, female friend), what would you pay (€) for the sports bra endorsed by [name endorser]”. Lastly, a more profound change was made finding the right replacement for social influencer Chiara Ferragni. After elaborating with the participants, decided was to replace Chiara Ferragni with Anna Nooshin. Ferragni probably being too much of a niche fashion icon, whereas Anna Nooshin starred in a famous Dutch television show called Expeditie Robinson. Therefore, Anna Nooshin will probably be more familiar to both, men and women.
4.3 Study design
independent variables: expertise, with the levels expert and non-expert, brand familiarity, with the levels familiar and unfamiliar, and person familiarity, with the levels familiar and unfamiliar. The independent variables are being manipulated in eight different scenarios. A between-subjects design was the most suitable for this research, since it eliminates any bias caused by spill-over effects resulting from, for instance, respondents comparing various scenarios with each other while filling out the questionnaire (Kantowitz, Roediger & Elmes, 2009). The final eight different scenarios are being displayed in table 2 (see Exhibit 1 for a more extensive overview with the examples of the different scenarios). In this experiment, the participants were asked to review social influencer endorsements. Subsequently, they were asked to fill out a questionnaire.
Scenario Expertise Person Familiarity Brand familiarity Endorser Brand
1 High High High Dafne Schippers Nike
2 High Low High Kara Goucher Nike
3 High High Low Dafne Schippers Oiselle
4 High Low Low Kara Goucher Oiselle
5 Low High High Anna Nooshin Nike
6 Low Low High Sofia Wayne Nike
7 Low High Low Anna Nooshin Oiselle
8 Low Low Low Sofia Wayne Oiselle
4.4 Procedure
The experiment was created using the online survey platform Qualtrics. First, the respondents are being informed about their participation in a research, followed by the introduction of basic definitions of the terms social influencer and endorsement. The respondents are being told that they will view an endorsement. The experiment starts with three general questions, asking the
Table 2. Table displaying the final eight different scenarios with the corresponding endorser, brand and
respondents about their gender, age and country of origin. Then, follows a set of fifteen manipulation check questions asking the respondents about their awareness on both, the brand and the endorser that are central to that specific scenario and their social media usage. All fifteen questions are rated on a 5-point Likert scale, running from 1 ‘strongly disagree’, to 5 ‘strongly agree’.
After the first set of questions, the endorsement is shown. The endorsement is presented within the Instagram template: showing the name of the various endorsers, the Instagram logo, the elapsed time after uploading and several buttons for navigation. The manipulation of either expertise and person familiarity is done by changing the social influencer who is making the endorsements. The perceived expert endorser with high person familiarity is being performed by Dutch Olympic track running athlete Dafne Schippers, the perceived expert with low person familiarity is the American athlete Kara Goucher. The non-expert with high person familiarity is famous blogger Anna Nooshin, whereas the non-expert with low person familiarity is the fictitious person Sofia Wayne. The various endorsers are photographed in similar ways. All four different endorsers are wearing a similar sports bra, which clearly displays the logo of a brand. Manipulating the brand familiarity is done by changing the brand that is being displayed on the sports bra. The presented brand is in case of the unfamiliar brand scenarios Oiselle and in case of brand familiarity Nike. Besides the display of the brand’s logo on the sports bra, the logo is also incorporated in the corner of the picture itself. See exhibit 1 for a clear overview of the different scenarios.
After the endorsement is shown, the respondent is asked to answer another nine questions on a 5-point Likert scale. The nine questions cover questions about the engagement towards the brand, which gives insights on the non-financial impact of the social influencer. But also covers
a question emphasizing the respondent’s willingness-to-pay (financial return) and several questions on whether the respondent feels like participating more in physical activities after seeing the endorsement (behavioral change).
4.5 Measures
4.5.1 Dependent variables
To measure the financial returns generated by the social influencer, one open question was used, asking the respondent about what amount (€) they are willing to pay for the endorsed product.
The non-financial impact is measured by asking the respondents about their engagement with the endorsed brand. Three self-constructed questions on a 5-point Likert scale asks the respondent about whether they would seek more online brand content after the endorsement, whether they would recommend the brand to their friends after the endorsement, and whether they would feel proud to be a customer of the brand after seeing the endorsement. The Likert scale ranged from 1 ‘strongly agree’, to 5 ‘strongly disagree’. The three items combined resemble the non-financial outcome created by the endorsement.
The last dependent variable that has been measured by the experiment, is that of effects on behavioral change. Two questions on a 5-point Likert scale are used. The first question emphasizes whether the respondent feels more interested in running after being exposed to the endorsement. The second question asks the respondents whether they feel like exercising more often after reviewing the endorsement. Both questions are ranged from 1 ‘strongly agree’, to 5 ‘strongly disagree’.
4.5.2 Manipulation check
The independent variables central to this research are the social influencer’s perceived level of expertise, brand- and person familiarity. The perceived expertise of the endorser is measured by two items: to which degree the respondent considers the endorser to be a running expert and to which degree the respondent considers the endorser to be a credible source in recommending sports products. Both, brand- and person familiarity, are measured by asking the respondent if they ever heard about the brand/endorser and to which degree they are familiar with the brand/endorser. All questions ranged from 1 ‘strongly agree’, to 5 ‘strongly disagree’.
4.5.3 Control variables
The degree to which the respondents use social media, serves as a control variable. The respondents are asked to answer to which degree they make use of social media, they follow influencers and brands on social media, and whether they ever considered purchasing or purchased items/services based on social media recommendation posts.
Other control variables are the participant’s country of origin and their gender. Both variables are measured by means of one open question. One’s country of origin was recorded by asking: ‘What is your country of origin (0 = Other, 1 = Dutch)? It could be possible that differences in world regions lead to a different attitude towards endorsements. One’s gender is recorded by askin: ‘What is your gender?’ Where 0 = Female and 1 = Male. Since, female respondents might react differently to endorsements, than their male counterparts.
4.6 Sample size
between 18 and 35, since this group tends to make the most use of social media platforms, and therefore will be familiar with social influencers and their endorsements (Duggan, 2015). The gender distribution of the sample will be a representation of the Dutch population and thus will be around a 50/50 distribution (CBS, 2016). The sample is selected by non-probability sampling and is a combination of a convenience and a snowball sample. The experiment has been distributed by means of the internet (e.g. e-mail, Facebook, etc.) and in print. The sampling method used is not the most appropriate for the research, but due to time restrictions this form of sampling is the most effective. However, this will have consequences for the generalizability of the findings (Bryman, 2012).
4.7 Data analyses
The data gathered from the experiment will be analyzed with IBM SPSS Statistics 24. But previously to the SPSS analysis, the data should be checked and cleaned (i.e. remove missing data) and the items should be recoded. Subsequently, general descriptives of the variables will be examined, to get insights on the male/female ratio, country of origin, average age and total number of respondents. A reliability analysis will follow to evaluate the consistency of the items. Followed by analyzing by means of one-sample t-tests, if the various manipulations have worked as how was initially intended. Finally, a multivariate analysis of covariance (MANCOVA) will be performed to test the three hypotheses using expertise, brand- and person familiarity as independent variables, financial, non-financial and behavioral change outcome as dependent variables and respondent’s country of origin, age and social media usage as covariates.
