Credibility and Fit:
Investigating Brand Endorsements by Brazilian Digital Influencers on Instagram Sarah Camargo Tomaz Cleto | 11008199
Master Thesis – Graduate School of Communication
Master’s Program Communication Science: Corporate Communication Supervised by Theo Araujo
Abstract
Digital influencers endorsing, or promoting a brand on social media has become an increasingly popular method of advertisement. Past research on celebrity endorsements has shown different determinants for effectiveness of endorsements, namely how credible the endorser is perceived (based on perceived trust and expertise), and how the endorser seems to fit with the
brand/product endorsed. The aim of the current study was to answer the question of to what extent credibility (made up of trust and expertise) and brand/product fit could generate positive engagement for brands, through likes and positive comments on Instagram. Since Brazil has the second largest amount of Instagram users, Brazilian digital influencers that endorse brands on Instagram were analyzed. Study 1 in this research, a qualitatively interviewed Brazilian digital influencers based on previous research on credibility and brand/product fit. Findings revealed important strategies that they use to be perceived as credible and fitting with the brand/product. These included, but were not limited to, the transparent disclosure of information about the endorsement, the expertise a digital influencer has, and how well the brand and product type fits with what the digital influencer usually endorses. Study 2, a quantitative study, used content analysis to find the strategies used on the Instagram endorsed posts. Findings show that
trustworthiness does not generate any kind of positive engagement, whereas expertise is able to generate somewhat of a negative kind of engagement in relation to the likes, and fit also
generates some negative engagement both for likes and comments for brands. Implications and further suggestions for research are presented in the discussion.
Credibility and Fit:
Investigating Brand Endorsements by Brazilian Digital Influencers on Instagram
Since its launch, brands are regularly using the social media platform, Instagram, to reach its consumers. Increasingly, companies have been recognizing the many possibilities and
benefits of using Instagram for marketing and communication purposes (Bergström & Bäckman, 2013; Woods, 2016). More often than not, brands are present on the platform through celebrities (i.e. actors) that represent and show off the brands. More recently, brands are also represented through digital influencers who may encourage associations with a brand. These digital
influencers are bloggers who use social media platforms such as blogs, YouTube, Twitter, and Instagram to share their stories and interpretations, as well as to influence attitudes and opinions of others (Uzunoğlu & Kip, 2014). The purpose of the current study, first and foremost, is to fill gaps in research of endorsed posts on Instagram, such as knowing what actually makes the endorsement effective for the brand (Ashley & Tuten, 2015).
The current study will be on endorsements as advertisement strategies in non-US
markets, which is essential for brands by providing deeper understanding of opportunities that lie in international markets. This will be done by focusing specifically on Brazilian digital
influencers that have been gaining fame through their social media accounts focused on lifestyle and fashion (Hughes, 2016). Furthermore, it is relevant for academic and market research to gain knowledge on the platform’s other large markets (Amos, Holmes, & Strutton, 2008; Soares, 2013; Woods, 2016). Therefore, the focus for research is on Brazil – the country with the second largest active and engaged monthly user base on Instagram after the USA. (“Instagram tem 29 milhões de usuários”, 2015).
Due to how little to no research has interviewed digital influencers themselves – especially in the topic of endorsements on Instagram – the current study will take two
approaches (Woods, 2016). In order to understand the strategies and perceptions of the digital influencers themselves on endorsement effectiveness, qualitative semi structured interviews with Brazilian digital influencers will be used. To gain an objective perspective and to analyze the general material and content apparent in the digital influencers’ posts, a content analysis will be conducted.
Most research on effective endorsement to date has focused on celebrity endorsement, and much of the strategies behind the celebrity endorsements are similar to those used for digital influencer endorsements (Kapitan & Silvera, 2015). Effective endorsements, in this case, is characterized as consumer engagement, which represents the responses that followers have to Instagram posts by liking and commenting on them (Ashley & Tuten, 2015; Brodie et al., 2011; Van Doorn et al., 2010; Woods, 2016). Past literature on celebrity endorsements found that the endorsements successfully persuade consumers and generate engagement for the brand when the source is a credible individual, perceived by consumers as trustworthy and an expert (Andersson, 2015; Erdogan, 1990; Uzunoğlu & Kip, 2014). Another determinant strategy for a successful endorsement is the fit, or match, an endorser has to the product and brand endorsed (Amos, Holmes, & Strutton, 2015).
Using these strategies at play for celebrity endorsements, the qualitative and quantitative research will aim at finding similarly effective strategies for successful endorsements on
Instagram for digital influencers – namely, credibility and fit. The findings of the current study will aim at producing valuable implications and advice for both digital influencers and brands that use Brazilian digital influencers for endorsements on Instagram. By determining interaction
effects of the digital influencer with credibility and fit, the current study can search for a possible ideal scenario for digital influencers and brands who want to gain most consumer engagement. Therefore, the research question is as follows:
To what extent do the credibility and brand/product fit of Brazilian digital influencers that endorse a brand on Instagram influence consumer engagement through likes and positive comments?
Theoretical Background Instagram
Instagram, a social media platform used for photo editing and sharing, was created in 2010, by Kevin Systrom and Mike Krieger (Kolawole, 2012). Since the creation of the app, research has focused on many different aspects of how brands use this platform to communicate with its stakeholders. It has “a great impact on the strategic communication objectives such as to create awareness, build relations and strengthen the brand” (Andersson, 2015, p.32), and can be used to increase brand recognition as well as the engagement between consumers and the brand (Kartona & Savary, 2014). Brands can also use social media, such as Instagram, to build and strengthen relationships with consumers (Fred, 2015). Instagram is unlike other kinds of social media in that it does not have an easy way to share information (nor track it), such as the “share” button on Facebook or the “retweet” button on Twitter. It is a rather consumer/user oriented platform, and brands rely more on use generated content branding. One strategy used to
strengthen the brand and reach consumers, is through digital influencers who have the ability to influence others on social media, increasing recognition and value for brands (Chang, 2014; Soares, 2013). Unlike the celebrity endorsements, digital influencers are bloggers who generate
content for the brand (i.e. through reviews) and spread positive word of mouth through endorsements of products and brands.
Digital influencers
The conceptualization of the term digital influencers used for this research follows the definition by Uzunoğlu and Kip (2014). The authors used the Katz and Lazarsfeld (1995) two-step flow theory to explain that, in society, opinion leaders are people who “interpret media information they receive and pass it on to others” (Uzunoğlu & Kip, 2014, p. 592), thus mediating information diffusion (Valente & Davis, 1999). In this case, digital influencers will serve as the individuals who influence the decisions of others as opinion leaders (Goldsmith, Lafferty, & Newell, 2000; Song, Chi, Hino, & Tseng, 2007). Recent research on advertising with influencers, found that “Instagram as a social media channel is one of the first to come to mind when the topic is mentioned [...] due to the visually engaging nature of the platform and the 400 million user base it has accumulated” (Woods, 2016, p.11).
Kapitan and Silvera (2015), found that many individuals consider digital influencers more trustworthy than celebrities, and that more than 84% of millennials “report that user generated content [such as these by digital influencers] influence their purchasing decisions” (p. 12). The authors explain that this occurs due to Kelman’s 1961 internalization process, a
phenomenon that occurs when the consumers are persuaded by an endorsed message and
internalize it by adopting the endorser’s beliefs. (Kapitan & Silvera, 2015). The process is made up of two steps: first the consumer identifies with the source (in this case the digital influencer), “through mechanisms such as familiarity or attractiveness” (p.2). Then, the consumer
internalizes the content of that message as they “weigh an endorser’s authenticity and adopt the message as if it were their own” (Kapitan & Silvera, 2015, p.2). Furthermore, the content that the
digital influencers share, is found to be more powerful than the content shown in advertisements (Hall, 2010; Woods, 2016). This occurs because message sharing has drastically changed with the introduction of social media, empowering communication among peers and trustworthy electronic word of mouth that is more likely to be free from manipulation compared to advertisements (Uzunoğlu & Kip, 2014).
Credibility
Past research on the credibility of traditional celebrity advertisement found that when the celebrities were viewed as credible sources of information by the consumers, then the
endorsements were effective and consumers were actually influenced (Amos, Holmes, & Strutton, 2008; Schmallegger & Carson, 2008). This can be attributed to the internalization process by Kelman (1961). The process explains that “perceptions of high source credibility, honesty, trustworthiness, and expertise are key inputs into consumers’ tendency to internalize an endorsement message”, which in turn makes the endorsement more effective (Kapitan & Silvera, 2016, p.7).This occurs when the audience perceives endorsed content to be believable and
credible, and therefore, deems it authentic, having a positive feeling towards the endorsement and even the endorser (Kapitan & Silvera, 2016).
When researching celebrity endorsement credibility, either a foundational source model (known as the source-attractiveness model) (Erdogan, 1999), or a source-credibility model is taken into consideration (Ohanian, 1990). The source-attractiveness model suggests that an endorsement message will be effective when an individual is similar to, familiar with, and likes the endorser. The source-credibility model posits that the effectiveness will depend on the trustworthiness and expertise of the endorser (Erdogan, 1999). In the current study, the latter will be used due to the fact that trustworthiness and expertise are both constructs shown effective
in celebrity endorsement research (Goldsmith et al., 2000; Scott, 2010). On the contrary, the source-attractiveness construct needs further research due to the uncertainty of its nature and scope (Amos, Holmes, & Strutton, 2008).
Past research on endorser credibility has found evidence that highly credible sources can persuade more than non-credible sources (Ohanian, 1990), and that “when considered jointly, expertise and trustworthiness are presumed to embody the source credibility construct” (Amos, Holmes, & Strutton, 2008, p.216). In fact, when it comes to digital influencers, blog readers “view the information shared by bloggers as one of the few forms of real, authentic
communication” (p. 599), and this information is perceived by consumers as peer recommendations, rather than paid advertisement (Uzunoğlu & Kip, 2014).
Many studies on celebrity endorsement have analyzed credibility as a bi-dimensional concept (the dimensions being trustworthiness and expertise) and found its positive effect on engagement and brand evaluation (Lafferty & Goldsmith, 1999; Spry, Pappu & Cornwell, 2011). Others have separately shown the positive effect that expertise (Eisend & Langner, 2010;
Rossiter & Smidts, 2012) and trustworthiness (Chao, Wührer, & Werani, 2005) have on engagement and brand evaluations. The current study will look at the two separate dimensions that “underscore the concept of source credibility” (Ohanian, 1990, p.41). It will also look at a third determinant for effectiveness of celebrity endorsements: product fit, meaning how much the endorser and their content match with the brand/product endorsed (Amos, Holmes, & Strutton, 2008; Bergkvist & Zhou, 2016; Scott, 2010).
Trustworthiness. Ohanian (1990) describes trustworthiness as “the listener's degree of confidence in, and level of acceptance of, the speaker and the message” (p.41). Trustworthiness is “perceived when someone is consistently honest or truthful and these characteristics are again
demonstrated over time” and there needs to be “continual consistency” in what the individual is saying and showing (Scott, 2010, p.309). Trust is important for endorsements, in that it results in the sender of the message being accepted by the receiver (Ohanian, 1990).Thus, digital
influencers are only able to influence those who accept and trust in them as influencers. Most past research has found that there is a positive effect between trustworthiness and the effectiveness of endorsements (Amos, Holmes, & Strutton, 2015; Chao, Wührer, & Werani, 2005; Erdogan, Baker, & Tagg, 2001). Yet, there are studies that have found that this is not true, such as Rossiter and Smidts (2012), who argue that the consumers would already be aware that the endorsers were paid, therefore not trusting the endorsement. The mechanism of
trustworthiness’ importance seems to be true for celebrity endorsements since, untrustworthy celebrity sources do not successfully persuade. This might therefore hold true for digital influencers, if they lack acceptance and, therefore, trust, they will not be able to effectively persuade consumers.
In summary, a crucial dimension to endorsement effectiveness is the perceived
trustworthiness of the digital influencer. Yet, trustworthiness is a subjective construct that is measured based on what the receiving audience perceives to be trustworthy and honest (Ohanian, 1990). This research explores the strategies that an influencer uses to be perceived as trustworthy by consumers. In order to do so, an antecedent of trust, or a testable measurement of what
actually leads to the feeling of trust, must be established. Past research has found different
factors that can affect and predict trust, including confidentiality, sincerity, honesty, and integrity (Erdogan, 1999; Moorman, Deshpande, & Zaltman, 1993).
One important form of attaining trust is through being transparent and fully disclosing information, thus being honest (Miller, 2010). Research on trustworthiness of digital influencers
on YouTube found that “trustworthiness and transparency work hand in hand. Honesty, integrity, and believability are core characteristics in a good endorser” (Fred, 2015, p.13). Bergkvist and Zhou (2016) explained that although consumers might now be more aware of paid endorsements, they also believe that celebrities have a choice in what to endorse. Digital influencers often disclose information on the endorsement, for example, by explaining whether or not it is merely paid advertisement. Since the reaction of trust through disclosing information leads to positive results for endorsers, one can predict that this will lead to more engagement to the endorsed posts, in the form of more likes and higher presence of positive comments (Miller, 2010) Therefore, the first hypothesis proposes that:
H1: The presence of disclosure of information in endorsed posts will be associated with higher levels of consumer engagement through a) likes, and b) positive comments towards the brand, on digital influencers’ Instagram posts, than endorsed posts without the presence of disclosure.
Additionally, there are various strategies at play by the digital influencers to disclose information (Woods, 2016). Thus, a sub-research question will be created in order to further understand these strategies through an interview with Brazilian digital influencers, and to be used for the content analysis. This will give a better insight into what makes up disclosure (the act of disclosing information) and how to measure it on Instagram along with consumer engagement.
RQ1: What strategies do digital influencers use on Instagram for presenting disclosure on endorsed posts, and how do these strategies influence the amount of a) likes and b) positive comments directed at the brand on the endorsed Instagram posts?
Expertise. Ohanian (1990) described expertise as “the extent to which a communicator is perceived to be a source of valid assertions” (p.41), and defined its synonyms as competence,
qualification, and being well informed. The expertise of an endorser is present when they are specialized in a specific area of knowledge, or have done enough research to give their opinion on the topic (Scott, 2010).
Research shows that an endorser who has more expertise and is perceived as an expert can be more persuasive and increase engagement with the brand (Amos, Holmes, & Strutton, 2015; Erdogan, 1999; Ohanian, 1990). Most research posits that expertise indeed has a positive effect on endorsements (i.e. Rossiter & Smidts, 2012). Much like the internalization process used for trustworthiness, consumers will cognitively internalize a message more easily and effectively if they believe in the appeal that the endorser knows what they are talking about (Kapitan & Silvera, 2016). When individuals perceive a celebrity as an expert, they are more likely to agree with them, which can increase brand attitude and the likelihood of intentions to purchase this brand/product (Amos, Holmes & Strutton, 2008). Celebrity endorsements have been shown to depend on perceived expertise for effectiveness, since individuals prefer to take advice from competent people (Uzunoğlu & Kip, 2014).
The same mechanism is at work for bloggers as digital influencers. Digital influencers are considered as experts due to how they generally speak on topics in which they have previous knowledge and experience on (Droge, Stanko, & Pollitte, 2010; Hsu & Tsou, 2011). In their research about digital influencers, Uzunoğlu and Kip (2014) found that the role that digital influencers might have as experts is essential for endorsements to be effective. The authors argue that this supports brands in increasing the “intimacy of brand communication” (p.595).
Furthermore, they specified that these digital influencers, as experts, “mainly cover more specific subjects, and give advice, useful information, tips or insightful comments, based on professional experience, wisdom and observations” (p. 221). When digital influencers are perceived to be
experts in the brand/product they are endorsing, their reaction seems to be more positive, therefore one can predict that this will lead to more engagement to the endorsed posts, in the form of more likes and positive comments for endorsed brands. Thus, one can hypothesize that:
H2: The presence of expertise in endorsed posts will be associated with higher levels of consumer engagement in a) likes, and b) positive comments towards the brand on digital influencers’ Instagram posts, than endorsed posts without the presence of expertise.
Much like disclosure, there are various strategies that differ in the ways that digital
influencers are presenting their expertise and need to be further understood. By gaining a deeper insight into what makes up expertise according to the digital influencers, one can comprehend what strategies to look for in the content analysis when exploring consumer engagement on Instagram posts. Thus, the second sub-research question is as follows:
RQ2: What strategies do digital influencers have on Instagram to demonstrate expertise on endorsed posts, and how to these influence the amount of a) likes and b) positive comments?
Brand/Product fit. Other than the source credibility model by Ohanian (1990), the next most important factor found to be essential for endorsement effectiveness is the fit between an influencer and the product and brand (also known as the match up hypothesis) (Amos, Holmes, & Strutton, 2008). Synonymous to congruency and similarity, the fit means the “degree of similarity or consistency between the celebrity and the brand (or product category)” (Bergkvist & Zhou, 2016, p.9). It is important to note that fit is dependent on the particular situation that the endorsement occurs – in other words, a celebrity can fit with a brand/product for a specific endorsement but not others (Bergkvist & Zhou,2016).
Research has repeatedly indicated that the effectiveness of endorsements relies on the brand/product fit of a celebrity endorser (Till & Busler 2000; Zwilling & Fruchter, 2013).
Friendman and Friendman (1979) found that when consumers perceive a celebrity as matching with the brand/product the endorsement was more effective and resulted in better results for the brand overall. Endorsed brands are better evaluated by its consumers when a celebrity and the brand/product actually fit (Kirmani & Shit, 1998). Additionally, digital influencers as endorsers are believed to be picking brands that they truly do use and/or enjoy, thus fitting their personality and lifestyle. The power that digital influencers have, lie in the fact that consumers believe that the influencers have tested and use(d) the brand/product (Uzunoğlu & Kip, 2014).
The match that a celebrity has with a brand/product has also been shown to affect the actual credibility of that celebrity (Berghkvist & Zhou, 2016). This happens because when a consumer identified with the endorser, consumers not only felt familiar with the celebrity and believed them, but also wanted to be them (Erdogan, 1999; Lee & Thorson, 2008). In the same way that celebrity endorsers are able to match with different products and brands based on their previous knowledge, digital influencers can also be considered to fit and match with the brand/product that they endorse. Consumer engagement in the form of likes and positive comments will once again be a measure of attaining endorsement effectiveness, since it will reflect the approval of the consumers for the brand/product fit of the digital influencers’ Instagram posts. Thus, the following hypotheses are proposed:
H3: The brand/product fit reflected by digital influencers and their endorsed posts will be associated with higher levels of consumer engagement in a) likes, and b) positive comments directed at brand on digital influencers’ Instagram posts, than endorsed posts without the presence of fit.
Prior to analyzing the content and consumer engagement on Instagram, it is important to know the strategies already at play by the digital influencers. Again, a sub-research question is
used in order to guide the qualitative research, and understand the various strategies that digital influencers use to demonstrate their brand/product fit.
RQ3: What strategies do digital influencers have on Instagram for demonstrating there is a brand/product fit on endorsed posts, and how to these influence the amount of a) likes and b) positive comments?
Study 1 Methods
The current study takes the approach towards studying digital influencers’ credibility and fit qualitatively. For this, Brazilian digital influencers were interviewed via email, with the purpose of finding new insights into the strategies that they use for endorsements. The dimensions used were sensitizing concepts, which suggest a guidance for the semi-structured interview. These concepts were the main determinants found in the theoretical background, namely, trustworthiness and disclosure, expertise, and brand/product fit. Qualitative interviews were used to understand the strategies which can be possibly have an effect on endorsed posts. Due to how this trend of endorsements on social media by digital influencers is a fairly new phenomenon, it is important to understand the strategies at play (Uzunoğlu & Kip, 2014). Thus, interviews were chosen to further understand the processes digital influencers use for strategies and approaches to endorsement.
The participants were sampled through a Google search of the major digital influencers in Brazil on Instagram (i.e. “Brazilian digital influencers with most followers on Instagram”). This search led to two important websites that resulted in the sampling pool. The names of the digital
influencers were collected from two rankings12. From those lists, 24 digital influencers (or their representatives) were contacted based on whether they had more than 15 thousand followers, and were invited to participate in the research. Four participants responded, showing interest in being a part of the study. Each participant then received an email with a letter of informed consent and were told to reply if their consent was given, before receiving the interview guide. It is important to note that, although their Instagram profiles are all open to the public, anonymity was
maintained for all participants, as they were all given an identification number in order to protect privacy. The semi-structured interview guide was made up of eight (8) open ended short
questions (found in Appendix A). The questions corresponded to a specific dimension of
effectiveness as the sensitizing concepts (trustworthiness/disclosure, expertise, and brand/product fit) taken from theoretical background on the topic. An example of a question which corresponds to expertise is “What strategies have you developed to be seen as an expert, as a digital
influencer on Instagram?”. The questions were sent on the 2nd of May, 2016, and responses were
received between one to two weeks later.
The participants responded to all questions via email. Participant 1 is a female Brazilian digital influencer with over 600k followers. She is a mother and teacher (with over 15 years of experience) who mostly creates content related to motherhood, such as giving advice on
parenting. Participant 2 is a young Brazilian digital influencer with almost 500k followers. She is a journalist (for over seven years) that posts content directed at a young female audience on a
1Hughes, T. (2016, March). SERMO Digital Influencer Index 2016. Retrieved May 10, 2016, from https://issuu.com/talkpr/docs/sermo_digital_influencer_index_2016_6b5b9e3bfa9b85 2MZ.id Casting. (n.d.). Retrieved May 10, 2016, from http://mzid.com.br/
range of different topics such as cosmetics and young female lifestyle. Participant 3 is a male Brazilian digital influencer with more than 15k followers, who works as a journalist for a Brazilian lifestyle magazine for almost ten years. He shares content related to a young male audience on a range of different topics, such as fashion, and lifestyle. Participant 4, is a female Brazilian digital influencer with over 400k followers. She is a journalist for over 10 years, and editor of a large Brazilian fashion magazine, and shares content primarily about make-up. Results
The responses are further examined below under the sensitizing concepts used
(trustworthiness/disclosure, expertise, and brand/product fit). Open coding was used in order to find key words of the interview responses that related most to the current study and could therefore be used for creating the codebook used in Study 2.
Personal branding, in this case, represents the marketing choices a digital influencer can make in order to advertise themselves as a brand, creating an online presence (Labrecque, Markos, & Milne, 2010). In order to understand the general strategies they have for
endorsements, participants were asked about the strategies they use to create and strengthen their personal brand on Instagram and to be an effective digital influencer (Montoya, 2002).
Participants 2, 3, and 4, focused more on their audience, explaining that they aim at creating a personal branding that interacts with and reflects the audience. Participant 3 went as far as
explaining that by knowing your audience you can give them what they want to see and make the content relevant to them. Participant 1 alternatively explained that by creating a strong personal brand, digital influencers can “strive for credibility” and define what she wants her audience to associate her with. She specified that the way to achieve a positive and strong personal brand, is by being honest and trustworthy as to who she is.
Trustworthiness/Disclosure. The digital influencers were asked about what strategies they have in order to be perceived by their followers as trustworthy, and how they approach transparency of the endorsements. All participants explained the importance of being honest in endorsing products that they actually like. Participant 4 stated that “information is key, and I only post products and brands that I really like / tried / trust / find relevant to share”. Participants 1 and 3, emphasized trustworthiness being built slowly by consistently being honest and
transparent about endorsing what they actually like and continuously use.
Participants were further interviewed about their strategies to be perceived as trustworthy and how they disclose information of endorsements on Instagram. All participants agreed that it is essential to disclose when products were gifted to them, and when an endorsement is directly paid advertisement. Participants 1, 2, and 3, stated that when the post is paid advertisement, it is clearly stated somewhere, such as “brand X invited us to talk about this” or “look at what X sent me today”, or even using the hashtags #ad #publi (the latter being the Portuguese translation for the word ad). Participant 3 also clarified that he doesn’t always post everything he receives from brands – only products he actually likes and wants to keep using. Finally, Participant 4 explained that she uses a hashtag specific to her blog to explain that the content is part of her sponsored collaboration with brands/products.
Expertise. The expertise that digital influencers have on the specific topics and
brands/products that they endorse can highly influence the way their audience perceives them to be credible. Participants were first asked about what they believe to be their fields of expertise; the responses were introduced in the above description of who each participant is. Then, the interview more specifically asked about the strategies they have developed in order to be perceived as an expert through their posts on Instagram. Most participants explained that the
individual posts in themselves might not reflect how they are experts, but their different social media accounts and background information reflects that. Some posts include more information on their expertise than others, such as Participant 4’s posts about being an exclusive guest journalist at the premier of a new make-up product. She explained that “being in the industry for all this time, both as an editor and blogger, is what I believe to be my biggest strategy [to
demonstrate expertise]”. Participant 3 on the other hand, relies on his honesty and credibility to go hand in hand with his expertise and that by being original and trustworthy, his expertise remains relevant. Participants 1 and 2 described the fact that, for many years, they share content based on the field of expertise in which they work (motherhood and journalist on female fashion & lifestyle), they do not use a specific strategy, but simply remain naturally true to their
competences.
Brand/product fit. Most participants agreed that when deciding which brands to publicly endorse on Instagram, it is essential that they actually are users of the brand. Yet, there were variations between the answers as to whether the use is the sole determinant, an additional determinant, or not a determinant at all for picking brands/products to endorse. Participant 1 stated that if the brand/product does not fit with what she usually uses and likes, then there is “no deal”. Whereas Participants 2 and 4 stated that interest in and use of the brand/product are the most important determinant as to whether it fits with them/their blogs and whether it will be shared on their profile. On the other hand, Participant 3 did not connect the fit to his personality and what he uses. He measured fit based on what is “most relevant to [his] audience”.
Respondents varied in their answers about how they demonstrate a consistent fit between themselves and the brands/products. Participant 1 explained that she always aims at adding her own personal opinions and reflect her personal branding when endorsing a product – once again,
connecting to how she first uses a product, and when she thinks it fits with her, she then shares it with others why they should also use it. Participant 2 similarly described that there is a fit within what she is constantly sharing on social media (mostly related to what she is interested in) and that things are usually consistent in that they are similar topics and brands/products that she shares. Participant 3 explained that when communicating his fit, he has no strategy since he believes his followers already know who he is and what fits with him, so he shares things of his personal interest and to his liking only. Last, Participant 4 mentioned her expertise once again, stating that her match and fit with brands/products lie mostly on her experiences, knowledge, and use of certain “favorite” brands/products that she talks about most.
Study 2 Methods
Codebook. Following Study 1’s qualitative interviews, Study 2 takes a quantitative approach of using content analysis to study the three dimensions of credibility and fit. The first step was to take the key findings from study 1 and translate them into possible variables for coding, which can be found in Appendix B.
Data Collection. The sample was made up of ten Brazilian digital influencers on Instagram. Five of them taken from an index3 of most followed Brazilian digital influencers
currently in the world; the other five taken from a digital agency4 focused on digital influencers. The sample units are made up of N = 200 Instagram posts (20 per digital influencer) that featured
3Hughes, T. (2016, March). SERMO Digital Influencer Index 2016. Retrieved May 10, 2016, from https://issuu.com/talkpr/docs/sermo_digital_influencer_index_2016_6b5b9e3bfa9b85 4MZ.id Casting. (n.d.). Retrieved May 10, 2016, from http://mzid.com.br/
an endorsement. The posts were chosen starting with their most recent post featuring an endorsement, then saving all posts with endorsements up to the number of 20 per influencer.
Independent Variables. For disclosure, participants of Study 1 emphasized clearly stating an endorsement is an ad– thus the variable Advertisement was created (19% of posts had an indication of advertising). Furthermore, participants also revealed they disclose information on whether an endorsed product was a gift, which corresponds to the variable Gift (17% of the posts specified a gift was given). Instagram posts by digital influencers can often include an explanation that they were invited to participate in an event, which was translated to the variable Invitation (6% of posts indicated an invitation). Participants also suggested the use of a call to action for followers, resulting in the variable Call to Action (CTA) (20% of posts included a call to action). A last form of disclosure is informing whether an endorsement is part of an actual campaign in partnership with a brand (15% of the posts mentioned a campaign). Thus, Campaign was the fifth variable that makes up the dimension Disclosure (representing trustworthiness as part of credibility). These variables were all coded based on the caption of the endorsed Instagram posts by Brazilian digital influencers. All variables were coded dichotomously, (0 = no form of disclosed information present; 1 = disclosure of information was present).
The variables Experience and Environment were used to correspond to Expertise. The first variable reflects the participants’ explanation that their strategy for being perceived as an expert stems from the fact that they have so many years of experience in the business. The data used for this variable was taken either from the participants’ responses to the interview, or from the respective digital influencers’ blogs in their “About Me” pages which included such
information. This variable was coded as either 0 = zero to five years of experience, and 1 = more than five years of experience of working in their respective areas of expertise; 71% of the digital
influencers had more than five years of experience. The latter variable, Environment, was taken based on the justification that professionalism also reflects an expert (36% of the posts indicated expertise in the environment). Thus, this variable was coded 0 when the environment in which a participant was, i.e. simply taking a selfie, did not reflect professionalism; whereas it was coded 1 when professionalism was present in the image, i.e. image taken in a studio.
Brand/product fit, which was measured through four variables, was measured based on information given by the participants in Study 1. Since participants explained that they generally endorse products that they actually use, the variable Use was created. When they mentioned use in the post caption, it was coded as 1, when they did not specify that they use the brand/product endorsed, it was coded as 0. Digital influencers also explained that they aim at matching their personal branding with the brands/products that they endorse, thus the second variable, Brand Type was coded as 1 if the brand endorsed is known for being luxurious/high end, and 0 if it is not (23% of the brands were high end). Additionally, the variable Product Type was coded by using product themes such as 1 = drinks and food (9% of the posts); 2 = fashion (48%); 3 = cosmetics (20%); 4 = make-up (9%), and 5 = others (15%), based on the product endorsed on the post. Last, the digital influencers also emphasized Consistency in endorsing similar
brands/products (82% of posts being consistent with the general content and endorsements). Thus the posts were further coded as to whether these were all consistent in endorsement content (0 = no consistency, 1 =consistent in brand/product type to the other posts analyzed).
Dependent variables. The dependent variables represent the consumer engagement that can be predicted by the strategies used by digital influencers (the independent variables). The consumer engagement variables therefore are the Amount of Likes on each post and Mean Positive Comments about the brands on each endorsed post. For the latter, all comments in the
posts were analyzed, those that were positive in relation to the brand were coded as 1, whereas all other comments about the brand that were neutral and negative were coded as 0. After an initial analysis, the results suggested high multicollinearity between the dependent variable Amount of Likes, and the independent control variable, Amount of Followers ( p = .836).
Alternatively. correlation of some independent variables, i.e. Expertise, and the control variable, Amount of Followers, was quite low ( p = .223).Therefore, a variable was aggregated into one variable measuring the ratio of Amount of Likes divided by Amount of Followers (Likes per Followers). The same was done for the dependent variable Mean Positive Comments and Amount of Followers, producing the variable Comments per Followers.
Inter-coder reliability. In order to ensure that all variables would be accurately and objectively coded, a second coder (a Brazilian student of advertising) coded 10% (N = 20) of the sample. The table in Appendix C shows the figures for the inter-coder reliability tests. A
crosstabs analysis of all coded variables was used in order to find the percentage of agreement; the Kappa and Krippendorff’s Alpha tests were also used to find the reliability of each variable coded. The variables Gift, Invitation, and Consistency showed good reliability scores with Krippendorff’s alpha above .800. For the variables Environment and Use the agreement
percentage was slightly lower and for CTA even the Kappa and Alpha figures were abnormally low. These low figures can be attributed to how often the image and caption can be misleading and imply information, without not making it explicit. An example is how influencers can often imply to be wearing an outfit, but it is not made clear whether they wear it on a daily basis, or for the endorsement post only. The variables still remained in the analyses in order to test their relation to the other variables Findings for the sub-research questions can be summarized in Tables 1 and 2; findings for the hypotheses can be found in Table 3.
Results
Trustworthiness/Disclosure. The first sub-research question (RQ1) aimed at finding what strategies digital influencers use in order to present disclosure on the endorsed posts. It was also used to understand the extent to which these strategies influence consumer engagement in a) likes and b) more positive comments for the brand/product. A multiple linear regression was conducted to test if the independent variables for trustworthiness/Disclosure predicted the Amount of Likes and Mean Positive Comments in a post (Likes per Followers and Comments per Followers, respectively). The model for RQ1a testing the dependent variable, Likes per
Followers, was significant (F(3, 196) = 4.372, p < .01, R2 = .063) explaining 6% of the variance in amount of likes. Gift (b* = .137, t = 1.967, p < .05), CTA (b* = .185, t = .576, p < .01), and Campaign (b* = -.143, t = -1.793, p < .05), all significantly predicted the amount of likes in a post. Disclosure of information on a Gift and CTA increased the amount of likes whereas disclosure of information on a campaign decreased the amount of likes in the posts. A multiple linear regression was also used to assess the ability of the independent variables for Disclosure to predict the Mean Positive Comments (Comments per Followers) for the brand. The model as a whole was significant (F(1, 198) = 26.184 p > .01) and explained 11% of the variance (R2 = .117) in the positive comments that appeared in the posts. The independent variable, Gift, significantly predicted a higher number of positive comments (b* = -.342, t = 5.117, p < .01), making a significant positive unique contribution to the prediction of the Mean Positive Comments.
Hypothesis 1 (H1) proposed that endorsed posts that disclosed of information in the caption, would be receiving a) more likes, and b) more positive comments directed at the brand/product. As mentioned above, five variables (Advertisement, Gift, Campaign, Invitation, and CTA) were aggregated in sum, to form the variable Disclosure. Once the variable was computed, it was also used as an independent variable to test H1a and H1b with the dependent variables, Likes per Followers and Comments per Followers. The overall model was not
significant for the dependent variable Likes per comments (F(1, 198) = 3.432 p > .05), nor for the dependent variable Comments per Followers (F(1, 198) = 2.606 p > .05). This means that there was no association between the amount of information disclosed and the amount of consumer engagement in the endorsed posts, and that Disclosure did not make a significant contribution to the Amount of Likes nor Mean Positive Comments in the endorsed posts. Thus Hypothesis 1a and 1b were rejected.
Table 3: Summary of Linear Regression Analysis for Variables Predicting Consumer Engagement (N = 200)
Likes per Followers Comments per Followers Variable B SE B β B SE B β Disclosure .003 .001 .130 3.37 .000 .114 Expertise -.003 .001 -.189** -5.21 .000 -.025 Fit -.001 .001 -.132 -2.82 .000 -.257** *p < .05. **p < .01.
Expertise. The same process was used for H2a and H2b and two variables were used which answered the sub-research questions about strategies, namely Expertise and Environment. Multiple linear regression was used to test RQ2a to measure the ability of the independent variables Experience and Environment to predict the dependent variable Likes per Followers.
The model as a whole was significant (F(1, 198) = 7.720 p < .01) predicting 3% (R2 = .038) of the variance in amount of likes. The independent variable Years of Experience was significant (b* = -.194, t = -2.778= p < .01), meaning that when a digital influencer is more experienced, one could possibly predict a smaller amount of likes in the posts. Experience and Environment were tested for RQ2b to assess the ability to predict the Mean Positive Comments (Comments per Followers), the model was significant (F(2, 197) = 7.788 p < .01) and explained 7% of the
variance (R2 = .073). Both independent variables were statistically significant, with Environment
(b* = -.219, t = -3.146 p < .01) recording a higher beta value than Experience (b* = .194, t = 2.795= p < .01). This meant that whenever the post was set in an environment that reflected expertise, it would receive less positive comments for the brand, whereas if the digital influencer had more years of experience, the post received more positive comments directed at the brand.
Once computed, Expertise was tested as an independent variable, in predicting the dependent variables Likes per Followers and Comments per Followers. When testing H2a, the results for the overall model predicting Likes per followers, was significant (F(1, 198) = 7.328, p < .01), explaining 3% of the variance (R2 = .031). Expertise had a statistically significant
prediction to Likes per Followers (b* = -.189, t = -2.707, p < .01), meaning that when there was more indication of expertise, there were less likes on the post. When testing H2b, as a predictor of Comments per Followers, results to the model as a whole were not significant (F(1, 198) = .123, p > .05), demonstrating no unique significant association between the indication of expertise and amount of comments in posts. Thus Hypothesis 2 was partially supported.
Fit. The variables used to test RQ3a and RQ3b were Use, Consistency, Brand Type, and Product. Multiple linear regression was used to assess the ability of the independent variables to predict the Likes per Followers in posts. The variance explained by the model as a whole was
11% (F(2, 197) = 12.247, p < .01, R2 .111). The variables Type of Brand (b* = -.268, t = -3.995 p < .01) and Consistency (b* = -.200, t = -2.980, p < .01) significantly predicted the amount of likes, in that posts had a smaller amount of likes when a fit was demonstrated. The results for RQ3b were similar, when testing the ability of the independent variables to predict the Comments per Followers. The model as a whole was significant (F(3, 196) = 6.050, p < .01), explaining 8% (R2 = .085) of the variance in predicting positive comments for the brand. The independent variables Use (b* = -.150, t = -2.175, p < .05), Product Type (b* = .202, t = -2.880, p < .01), and, Brand Type (b* = -.168, t = -2.383, p < .05) predicted the decrease in positive comments in posts.
Hypothesis 3a proposed that the independent variable Fit (as an aggregate of the four above variables) would predict the dependent variable Likes per Followers. The results show that Fit does not predict Likes per Followers, as the model was not significant (F(1, 198) = 3.496, p > .05). Last, a multiple linear regression was conducted to test hypothesis 3b for the Comments per Followers as the dependent variable and the aggregate, Fit, as the independent variable. The model was significant (F(1, 198) = 5.019, p < .01) and explained 6% of the variance (R2 = .066) in positive comments for the brands that can be predicted by the fit. The variable Fit (b* = -.257, t = -3.748, p < .01) was statistically significant, meaning that there is a unique contribution from the fit of the digital influencer resulting in less positive comments for the brand. Thus,
Hypothesis 3 was partially supported.
Discussion
The current study tied together, not only what is found in celebrity endorsement research, but also the different ways in which endorsements can be studied. Brands can spread a positive image of themselves and their products when collaborating with digital influencers, often gaining
a trustworthy and authentic presence on social media (Chang, 2014). The overarching aim was to offer a better understanding of the effects that credibility and fit have on Instagram endorsements by Brazilian digital influencers.
Trustworthiness/Disclosure. The first key finding reveals that disclosure had no effect on the amount of likes on posts nor positive comments towards a brand itself. This means that independent of whether or not the digital influencer was disclosing any sort of information about the endorsement, the amount of likes and positive comments directed at the brand did not
change. When taking a closer look at the different strategies used by the digital influencers, the disclosure that an endorsed product was a gift generated more likes but less positive comments about the brand. Additionally, disclosure of whether the endorsement was part of a partnership or campaign, resulted in less likes on the post. Yet, no strategy overall had an effect strong enough to predict consumer engagement.
The current study rejects research on the importance of trustworthiness of the endorser, and that the disclosure of information is crucial to effectiveness (Amos, Holmes, & Strutton, 2015; Chao, Wührer, & Werani, 2005; Erdogan, Baker, & Tagg, 2001). These findings are in line with research by Rosetti and Smidts (2012) who found that trustworthiness did not have such a strong effect after all on ensuring an endorsement is effective. This can be explained by how, increasingly, this line between what audiences see as advertisement, or not, becomes blurred. As a result, brands are able to place products strategically without affecting the
audience’s attitude towards the brand. In a study about the use of influencers in advertising, the author found that “this blurring of the lines between what is a genuine endorsement and what is a paid one through content-rich platforms is what makes influencer marketing so powerful”
endorsements might not be the same for digital influencers, since the latter are seen as relatable individuals that have a bigger influence on their audience than celebrities (Hall, 2010; Woods, 2016). Consumers are aware now of these blurred lines and that digital influencers are likely to have been paid for the endorsement. This validates Ohanian’s (1990) finding that credible sources are not always the most effective sources.
It is important though, that influencers are transparent and aim at being perceived as trustworthy, since the backlash that can occur if they are not honest, can be detrimental both to the influencer and the brand (Woods, 2016). Thus, an implication for digital influencers is to remain consistent in that they disclose information and try to remain as a credible source of information, mostly in order to avoid negative consequences that can arise from dishonesty. For brands, these findings show that these blurred lines can be taken to their advantage, since the digital influencer seems to be seen as an authentic source of information, independent of whether or not they disclose endorsement information (Woods, 2016).
A limitation to the present research is that it did not study the effects that the endorsement could have had on the actual sales and profits of the brand. Future research could take a step further and explore the benefits that these Instagram endorsements, if any, might have on actual purchase and revenue for the brands. The current study was limited to the strategies found through the short interviews, whereas if the size of the sample and interview increased, more strategies for perceived trustworthiness could be revealed. Furthermore, an addition can be done to this research by increasing the sample, as well as analyzing both posts with and without endorsements to find any variation in the amount of likes and comments the digital influence might be receiving.
Expertise. Key findings for expertise reveal that, by using strategies to be perceived as an expert, the Brazilian digital influencers generally receive less likes, but has no effects on the positive comments towards the brand. More specifically, the years of experience that the digital influencers have, had an association to consumer engagement: negatively by decreasing likes, but positively by increasing positive comments for the brand. Findings also show that when the Instagram post and the digital influencer reflected more expertise through the environment in the post, then the positive comments towards the brands decreased. This overall finding rejects the conclusions by Amos, Holmes, and Strutton (2008), who found that when the influencer is perceived as an expert, consumer engagement can increase, through an increase in brand attitude and the intention of purchasing the brand/product. It also rejects literature stating that expertise was proven to be extremely effective and endorsements by experts could be more persuasive (Rossiter & Smidts, 2012).
This might be occurring since the digital influencers are seen as friendly relatable
individuals, not celebrities. As they grow in popularity, professionalism and years of experience, they might be perceived as less of a familiar face, and more as an actual celebrity. These
findings are quite important for digital influencers, since their expertise not only seems to have little effect on consumer engagement, but it can actually have a negative effect. These findings also have managerial implications for brands, in that they can always aim at working with digital influencers that are newly popular, in order to remain relevant.
A strong limitation of this research lies in the small amount of measures used to analyze the digital influencers’ expertise. Future research should extend the measurements of expertise, similar to the extension of the trust dimension, by increasing the sample and interviews used. Additionally, it can be quite difficult to objectively measure expertise through the content shown
on Instagram only. Future research could use the various social media accounts that Brazilian digital influencers endorse brands/products on, in order to find the different strategies used across social media platforms.
Brand/product fit. The last key findings of the current study relate to the fit that the digital influencer has with the brand/product endorsed, which had no association with the amount of likes, and had a negative association with the positive comments about the brand. Evidence was found that the consistency in the type of brand that is endorsed, and the type of brand endorsed, actually had a negative effect on amount of likes. Additionally, consistency, brand type, and use of the brand/product also decreased the positive comments. This rejects the findings by Uzunoğlu and Kip (2014) who explain that use of the brand/product is an essential determinant to the fit of the influencer and the endorsement, increasing consumer engagement. This also contradicts findings by Woods (2016), who interviewed agencies specialized in digital influencers, and found that brands stress the fit between the endorser and the brand when
choosing endorsers, which can highly influence the reactions and engagement of consumers. A tentative explanation for this contradiction, is that the audiences that follow digital influencers on Instagram are not reacting much to the endorsements on Instagram, and are therefore not highly reactive towards the fit that the influencer has. The comments on the Instagram posts for example, were mostly compliments for the digital influencers, or conversations that do not revolve around the actual post. The endorsed posts of digital influencers do not generate much engagement for brands themselves. Brands and digital
influencers should be aware of the little value given to fit by the audience. While it is important, for credibility, that the digital influencers endorse a brand/product that the influencer is likely to actually be using, there seems to be negative effects on the consumer engagement based on the
fit that the influencer reflects. Future research would need to further analyze this phenomenon in order to check for any underlying determinants to the fit between a digital influencer and the brand/product. This has a managerial implications for new brands that need recognition. Success of endorsement can mean at least gaining recognition and discovery, by easily reaching
consumers. Even if there is no fit between a digital influencer and the brand, the brand will at least be gaining recognition, validating one of the main reasons for brands to use digital influencers on Instagram (Andersson, 2015; Kartona & Savary, 2014).
A crucial limitation found, was that it was sometimes misleading to understand whether or not the digital influencer was actually using the product, or if they were merely pretending to use the product. Future research could aim at new ways to discovering if the endorser is using the product or different ways that they clearly show they use it. It could also ask audiences how much they perceive the endorser to actually be using the brand/product, as part of an extension to this study where it actually interviews consumers to find what they perceive from these digital influencers. It would also be important to analyze the different types of consumer engagement and effects that an endorsement on Instagram can create.
The current study aimed at understanding the extent to which the credibility of a digital influencer and the fit a digital influencer has with what they are endorsing can be associated with consumer engagement. Study 1 faced certain limitations such as how the interviews had to be done via email due to the digital influencers living in a different country as the researcher, and the constraints in their schedules. Study 2 faced issues in how the strategies used
(trustworthiness, expertise, brand/product fit) had been previously tested as perceived
information, but this research aimed at finding an objective way to measuring these strategies. To improve this and add to the existing research on digital influencers on Instagram, future research
could conduct a third study which would be analyzing the audience. A survey could be used in order to ask the Brazilian followers of these digital influencers how they perceived
trustworthiness, expertise, and brand/product fit.
The current research has important managerial insights for brands such as the increased possibilities to use digital influencers on Instagram to the best of their advantage, mostly in gaining recognition. It also has insights for digital influencers, in understanding how they value strategies, such as disclosing information, might not be as valuable after all since their posts are not affected by it. As Instagram continues to grow and its influence on the market increases, it will be necessary to continue studying the sources of endorsements (digital influencers) and their audiences. It is important that this phenomenon of connecting social media and brands together continues to be empirically explored to benefit consumers and markets alike.
References
Amos, C., Holmes, G., & Strutton, D. (2008). Exploring the relationship between celebrity endorser effects and advertising effectiveness: A quantitative synthesis of effect size. International Journal of Advertising,27(2), 209-234.
Andersson, L. (2015). To be or not to be engaged.
Ashley, C., & Tuten, T. (2015). Creative strategies in social media marketing: An exploratory study of branded social content and consumer engagement.Psychology &
Marketing, 32(1), 15-27.
Bergkvist, L., & Zhou, K. Q. (2016). Celebrity endorsements: a literature review and research agenda. International Journal of Advertising, 1-22.
Bergström, T., & Bäckman, L. (2013). Marketing and PR in Social Media: How the utilization of Instagram builds and maintains customer relationships.
Brodie, R. J., Ilic, A., Juric, B., & Hollebeek, L. (2013). Consumer engagement in a virtual brand community: An exploratory analysis. Journal of Business Research, 66(1), 105-114. Chang, C. Y. (2014). Visualizing brand personality and personal branding: case analysis on
Starbucks and Nike's brand value co-creation on Instagram.
Chao, P., Wührer, G., & Werani, T. (2005). Celebrity and foreign brand name as moderators of country-of-origin effects. International Journal of Advertising, 24(2), 173-192.
Droge, C., Stanko, M. A., & Pollitte, W. A. (2010). Lead Users and Early Adopters on the Web: The Role of New Technology Product Blogs. Journal of product innovation
management, 27(1), 66-82.
endorsers’ attractiveness and expertise. International Journal of Advertising, 29(4), 527-546.
Erdogan, B. Z. (1999). Celebrity endorsement: A literature review. Journal of marketing management, 15(4), 291-314.
Erdogan, B. Z., Baker, M. J., & Tagg, S. (2001). Selecting celebrity endorsers: The practitioner's perspective. Journal of advertising research, 41(3), 39-48.
Fred, S. (2015). Examining Endorsement and Viewership Effects on the Source Credibility of YouTubers.
Goldsmith, R. E., & De Witt, T. S. (2003). The predictive validity of an opinion leadership scale. Journal of Marketing Theory and Practice, 11(1), 28-35.
Goldsmith, R., Lafferty, B. & Newell, S. (2000) The impact of corporate credibility and celebrity credibility on consumer reaction to advertisements and brands. Journal of Advertising, 29(3), pp. 43–54.
Hall, T. (2010, March 22). 10 Essential Rules for Brands in Social Media. Retrieved April 25, 2016, from
http://adage.com/article/digitalnext/10-essential-rules-brands-social-media/142907/
Hsu, Y. H., & Tsou, H.-T. (2011). Understanding customer experiences in onlineblog environments. International Journal of Information Management, 31,510–523. Hughes, T. (2016, March). SERMO Digital Influencer Index 2016. Retrieved April 8, 2016,
from
https://issuu.com/talkpr/docs/sermo_digital_influencer_index_2016_6b5b9e3bfa9b85 Instagram tem 29 milhões de usuários ativos por mês no Brasil. (2015, November 09). Retrieved
April 8, 2016, from http://g1.globo.com/tecnologia/noticia/2015/11/instagram-tem-29-milhoes-de-usuarios-ativos-por-mes-no-brasil.html
Kamins, M. A., & Gupta, K. (1994). Congruence between spokesperson and product type: A matchup hypothesis perspective. Psychology & Marketing,11(6), 569-586.
Kapitan, S., & Silvera, D. H. (2015). From digital media influencers to celebrity endorsers: attributions drive endorser effectiveness. Marketing Letters, 1-15.
Katona, Z., & Sarvary, M. (2014). Maersk Line. California management review, 56(3), 142-156. Kolawole, E. (2012, April 11). Instagram, Facebook and a prediction about Brazil. Retrieved
April 8, 2016, from https://www.washingtonpost.com/blogs/innovations/post/instagram-facebook-and-a-prediction-about-brazil/2012/04/11/gIQAJJYqAT_blog.html
Labrecque, L. I., Markos, E., & Milne, G. R. (2011). Online personal branding: processes, challenges, and implications. Journal of Interactive Marketing, 25(1), 37-50.
Lafferty, B. A., & Goldsmith, R. E. (1999). Corporate credibility’s role in consumers’ attitudes and purchase intentions when a high versus a low credibility endorser is used in the ad. Journal of business research, 44(2), 109-116.
Lee, J. G., & Thorson, E. (2008). The impact of celebrity–product incongruence on the
effectiveness of product endorsement. Journal of advertising research, 48(3), 433-449. McGinnies, E., & Ward, C. D. (1980). Better liked than right trustworthiness and expertise as
factors in credibility. Personality and Social Psychology Bulletin, 6(3), 467-472. Miller, R. (2010, June). Developing the Culture of Trust in which Large-Scale Gains Become
Possible. In Conference on Corporate Communication 2010(p. 245).
Montoya, P., Vandehey, T., & Viti, P. (2002). The personal branding phenomenon. Personal Branding Press 2002.
Moorman, C., Deshpande, R., & Zaltman, G. (1993). Factors affecting trust in market research relationships. the Journal of Marketing, 81-101.
Ohanian, R. (1990). Construction and validation of a scale to measure celebrity endorsers' perceived expertise, trustworthiness, and attractiveness.Journal of advertising, 19(3), 39-52.
Rossiter, J. R., & Smidts, A. (2012). Print advertising: Celebrity presenters.Journal of Business Research, 65(6), 874-879.
Schmallegger, D., & Carson, D. (2008). Blogs in tourism: Changing approaches toinformation exchange. Journal of Vacation Marketing, 14(2), 99–110.
Scott, P. B. (2010, June). Getting a squirrel to focus: How to communicate with and persuade today’s listeners. In Conference on Corporate Communication 2010 (p. 307-320).
Soares, K. V. (2013). As blogueiras como formadoras de opinião: novas visões de moda, beleza e consumo na sociedade contemporânea.
Song, X., Chi, Y., Hino, K., & Tseng, B. (2007, November). Identifying opinion leaders in the blogosphere. In Proceedings of the sixteenth ACM conference on Conference on
information and knowledge management (pp. 971-974). ACM.
Spry, A., Pappu, R., & Bettina Cornwell, T. (2011). Celebrity endorsement, brand credibility and brand equity. European Journal of Marketing, 45(6), 882-909.
Till, B. D., & Busler, M. (2000). The match-up hypothesis: Physical attractiveness, expertise, and the role of fit on brand attitude, purchase intent and brand beliefs. Journal of advertising, 29(3), 1-13.
Leveraging blogger engagement. International Journal of Information Management, 34(5), 592-602.
Valente, T. W., & Davis, R. L. (1999). Accelerating the diffusion of innovations using opinion leaders. The Annals of the American Academy of Political and Social Science, 566(1), 55-67.
Valente, T. W., & Pumpuang, P. (2007). Identifying opinion leaders to promote behavior change. Health Education & Behavior.
Van Doorn, J., Lemon, K. N., Mittal, V., Nass, S., Pick, D., Pirner, P., & Verhoef, P. C. (2010). Customer engagement behavior: Theoretical foundations and research directions. Journal of Service Research, 13(3), 253-266.
Woods, S. (2016). # Sponsored: The Emergence of Influencer Marketing.
Zwilling, M., & Fruchter, G. E. (2013). Matching Product Attributes To Celebrities Who Reinforce the Brand. Journal of Advertising Research, 53(4), 391-410.
Appendix A – Interview Guide
Investigating Endorsements by Brazilian Digital Influencers on Instagram Interview Guide
1) What do you consider as the most important aspects when choosing to endorse a brand on your Instagram profile?
2) What strategies do you have for creating and strengthening your personal branding on Instagram?
3) Whether a brand initiates contact with you or you initiate contact with the brand, how do you decide which brands to publicly endorse on your Instagram profile?
4) What are the ways that you demonstrate and communicate that there is a similar and consistent fit between you and any products/brands?
5) What strategies do you have on Instagram to be transparent and ensure you are perceived as trustworthy by your followers?
6) In what ways do you reveal endorsement information, and how do you disclose when a product is a gift or a purchase?
7) What topics and/or products/brands would you consider to be enough of an expert in, in order to share your opinions and ideas on Instagram about it?
8) What strategies have you developed to be seen as an expert, as a digital influencer on Instagram?
Appendix B – Codebook
This research aims at understanding effective strategies that Brazilian digital influencers use for endorsements on Instagram. Credibility (made up of the variables trustworthiness/disclosure and expertise) and Fit are the dimensions at focus for
understanding what makes an endorsed post effective. Endorsed posts include any Instagram post in which an individual wears or uses a brand and mentions it or recommends it (Ha, 2015; Kapitan & Silvera, 2015). Examples are an image in which a brand is tagged in the picture (i.e. an individual wearing a shoe by Dior, tagging the brand), or a caption
mentioning/recommending a brand (i.e. “Thank you @Dior for sending me these shoes. I love wearing them all day, so comfy!” “#Dior”). All posts used in this analysis ar e already chosen as endorsed pictures. The following is the Codebook to be used for coding the Instagram posts of 10 Brazilian digital influencers for a content analysis. A total of 200 posts were taken, and a link to all posts is included in order to faci litate the access to each post more than once. The Codebook includes descriptions of each variable to be coded, as well as the directions on how to code them. Note: all posts are to be coded from the version of the application, Instagram.com, thus facilitating data collection; as such, descriptions correlate to what is shown on the website (i.e.
https://www.instagram.com/p/BFrN_qFw9Ws/?taken-by=camilacoutinho) Part 1 of coding (Sheet 1): General
1) Coder
a. Write down the name of the person coding this post 2) Post
