Characteristics of social media influencers that affect engagement with fashion brands on Instagram
---
A content analysis on social media influencers’ brand engagement rates and brand engagement valence, influenced by the frequency of posts, the frequency of sponsored posts
and whether they are a micro or macro influencer
Kiyara Makatita (11417641) Master's Thesis
Graduate School of Communication Master track Corporate Communication Suzanne de Bakker
Abstract
Even though social media afford brands to build up loyalty with consumers, brands are still having difficulties engaging with their consumers online. To solve this problem, both academics and professionals have been recognizing the potential of social media influencers (SMI) as effective endorsers in brand-related communication. Due to their influential reach and their ability to connect with a brand’s target audience, they can serve as effective intermediaries between brands and their consumers. However, professionals do not know what characteristics make a social media influencer influential. The current study attempts to answer to what extent brand engagement can be positively affected by social media
influencers on Instagram? Seventeen SMI’s in fashion and their 64 brand-related Instagram posts have been studied. Three SMI characteristics and their relationship with brand
engagement have been investigated through a content analysis. The three characteristics are frequency of posts, frequency of sponsored posts and whether they are a micro or macro influencer. Brand engagement was measured by brand engagement rate, brand engagement valence and who the engagement was referred to. The results indicate that frequency of posts it not a relevant characteristic to judge influence on. The frequency of sponsored posts has contradictory results: the more posts are sponsored the lower brand engagement rates are, but the higher the chance brand engagement is positive. And last, micro influencers generate higher brand engagement rates and more engagement towards the brand than macro
influencers. This implicates brands should invest in collaborations with micro influencers who have worked with other brands as well to engage their consumers.
Introduction Social media
Social media have changed the way we live our lives. On a daily basis, the amount of time we spend on social media ranks second, just after watching tv and before time spend on eating and drinking (Eror, 2017). Not surprisingly, to reach us, organisations are increasingly recognizing the importance of social media as a central tool use (Killian & McManus, 2015). It is expected that 36 million dollars has been spend on social media marketing in 2017 globally (Eror, 2017). The reason why so much money is worth investing is because social media give brands better communication platforms to build up loyalty with consumers (Erdoğmuş & Cicek, 2012). According to Vernuccio (2014), organisations can build up a strong brand by capitalizing on the strengths of social media in three ways. First, the brand’s image can be established, maintained and reinforced by interactive communications on the platforms. Second, social media can be used to listen to stakeholders and monitor
conversations. Third, brands can use social media platforms in public relations to reach new opinion leaders, such as social media influencers. The present thesis focuses on the third way.
Social media influencers
In the academic world, scholars are increasingly recognizing the possibilities for organisations to collaborate with social media influencers (SMIs) (Booth & Matic, 2011). While influencers used to be described as “everyday consumers who are substantially more likely than average to seek out information and to share ideas, information and recommendations with other people” (Keller, Fay & Berry, p.2, 2007), SMIs now represent “a new type of independent third party endorsers who can shape audience attitudes through the use of various social media” (Freberg, Graham, McGaughey & Freberg, p.1, 2011). The latter definition matches the persuasive power that social media influencers nowadays have and has led to
organisations seeing the potential of influencers as endorsers in their corporate
communication (Freberg et al., 2011; Gräve, 2017). Because of their ability to connect with a brand’s target audience, SMIs can serve as effective intermediaries between brands and their consumers (Erdoğmuş & Cicek, 2012; Uzunoğlu & Kip, 2014).
However, social media have made it more difficult for brands to engage with their public online (Enginkaya & Yilmaz, 2014). It has become harder for organisations to directly reach their consumers because they do not have the needed voice in social media (Booth & Matic, 2011,). Next to that, consumers do not turn to organisations anymore for brand information, but to others online (Enginkaya and Yilmaz, 2014). Collaborating with SMIs could solve this engagement-issue for brands. According to a marketing platform, 92% of the consumers trust messages from influencers more than traditional advertisements or even celebrity
endorsements (Weinswig, 2016). SMIs are trusted more than organisations because they are seen as “one of them” by consumers and their messages are perceived as own experiences (Enginkaya and Yilmaz, 2014; Wu & Wang, 2011; Zietek, 2016). Collaborating with social media influencers has therefore led to new ways for brands to engage with their consumers more directly, more organically and at a larger scale (Adweek, 2015).
Brand engagement
According to Gambetti, Graffigna and Biraghi (2012), brand practices that are carried out with the goal to engage consumers belong to the concept of (consumer) brand engagement. From an organisational point of view, the aim of brand engagement is to establish a strong and lasting relationship with its consumers (Gambetti et al., 2012; Hollebeek, 2011 & Schultz, 2007). As a consequence, consumers become loyal to the brand (Dwivedi, 2015). From a consumer point of view, consumers express their engagement with a brand through certain
behaviours, such as interacting with other social media users about the brand, sharing and even creating brand content. Muntinga and his colleagues (2017) refer to these behaviours as COBRAs, which is short for consumers’ online brand related activities. When consumers feel more engaged with a brand or its product, they are more likely to express that engaged
behaviour. The other way around, engagement behaviour leads to consumers feeling more engaged with the brand (Mangold & Faulds, 2009). Thus, once consumers are engaged with a brand, it is likely they keep on engaging. To conclude, brand engagement is a dynamic and interdependent concept where organisations try to manage the engagement behaviours of consumers and the engagement of consumers drives organisational practices.
Now, to develop relationships with a brand’s consumers, online influencers are becoming part of PR strategies in organisations because of their broad and influential reach (Jarvis, 2006). In other words, through social media influencers, organisations want to engage consumers, so that strong and lasting relationships can be formed and consumers become loyal to the brand. Thus, organisations can boost the impact of their brand messages by working together with influencers (Adweek, 2015). However, as De Veirman, Cauberghe and Hudders (2017, p. 799) say “to increase the message’s impact one should search for the most likeable, credible influencer who has a high value as an opinion leader”. But this is exactly where the practical problem lies. Professionals do recognize the significance of social media influencers, but do not know what the effective metrics are to decide who the most influential people are (Enginkaya & Yilmaz, 2014).
Probst, Grosswiele and Pfleger (2013) argue that practical approaches in identifying online influencers are still in their early stages. Their overview of scientific articles shows that many studies attempt to find out to what extent people could be influential, but only based on the
strategic location within a social network. Therefore, a person’s network size is often seen as an important factor in identifying influencers (De Veirman et al., 2017). The authors add that it is important to ask how these influencers can be characterized? Keller, Fay and Berry (2007) investigated consumers’ daily conversations through a survey in order to identify influencers. The identification was based on three characteristics: the network size, the frequency of communication, and past recommending behaviour. Booth and Matic (2011) combined a person’s strategic location and characteristics in their study. They presented an index valuation algorithm organisations could use to identify influential bloggers. Factors they considered to be influential are views per month, post frequency, popularity of blog post links, how often the blogger is cited, topic-related posts, reader response and amount of comments, and the rank of the influencer. Through an experiment, De Veirman and
colleagues (2017) studied what effect the number of followers has on attitudes towards the influencer and brand, and also looked at the moderating effect of the product type that is promoted. Furthermore, Freberg et al. (2011) studied consumers’ perceptions and found that SMIs are thought of as smart, ambitious, productive, and poised. However, these studies by De Veirman et al. (2017) and Freberg et al. (2011) used prototypes of SMIs for their research instead of real-life influencers. Zietek (2016) interviewed professionals from the field about working with influencers. The participants mentioned several factors that contribute to a successful collaboration with influencers, such as how well an influencer fits with the brand, her/his visual language, how passionate and trustworthy he/she is, the price of collaborating, the amount of creative freedom for an influencer and an influencer’s frequency of
communication. Nevertheless, these findings are based on experiences and not on empirical research.
This past research on social media influencers has helped professionals to guide their
marketing strategies and PR campaigns (Kelly, Fay & Berry, 2007). But what really happens when an influencer endorses a brand on social media? Even though social media are core in the building of brand engagement (Gambetti et al., 2012), empirical research on brand engagement through social media influencers is scarce. In addition to that, influencers are known for using a variety of social media, but the most important platform to follow influencers on is Instagram (Bloglovin’, 2016). Instagram shows great potential as the right medium to be used by organisations for their social media marketing (Çukul, 2015). The platform started as a simple app among friends, but has grown into a global community of consumers and brands where users are able to connect with brands online (Çukul, 2015; Killian & McManus, 2015). Therefore, Instagram could be more valuable than other social media to investigate social media influencers and brands on. Despite all this, social media influencers, specifically on Instagram, in combination with brand engagement has not been investigated yet. Altogether, this leads to the following research question:
RQ: To what extent can brand engagement be positively affected by social media influencers on Instagram?
Theoretical background Instagram
Even though many social media platforms provide their users with the same functions, such as being able to follow other individuals, post media and place comments (Lim, Lu, Chen & Kan, 2015), Instagram distinguishes itself from other social media in several ways. Since its launch in October 2010, Instagram has now become the most popular photo-sharing platform with over 800 million monthly active users (Instagram, 2017; Mejova, Haddadi, Noulas &
Weber, 2015). Vital to photo sharing communities such as Instagram, is engagement. These kind of platforms have grown at an impressive speed and therefore the challenge for
organisations is to get users to communicate around and engage with a brand’s content (Bakhshi, Shamma & Gilbert, 2014). According to Elliot (2014), Instagram is the king in generating engagement for organisations. His research on user interactions and brand-related content revealed that on Instagram, brands get 58 times higher engagement rates than on Facebook and 120 times higher engagement rates than on Twitter.
Both Facebook and Instagram are used to maintain social relationships (Sheldon, 2008). What distinguishes Instagram from Facebook is that the platform is also used for self-expression and to escape from reality (Lee, Lee, Moon & Sung, 2015). One of the most appealing features of Instagram is that users can promote themselves by the content they create (Erkan, 2015). Instagram enables its users to transform mobile pictures into visually appealing images right away (Instagram, 2017). Comparing Instagram to Twitter, Instagram is an image-based platform and Twitter a textual-based one. This creates differences in how users experience the platform. Pittman and Reich (2016) found that image-based networks enhance feelings of intimacy with its users. This can for a large part be explained by Sundar’s (2008) MAIN model; people have the general belief that pictures cannot lie and therefore have more trust in pictures than textual descriptions. The author further explains that images can facilitate social presence, therefore users of Instagram in comparison to Twitter, feel they are communicating with an actual person instead of with an object for example and this leads to more intimacy. And last, because Instagram users share, amongst other things, photos of themselves and their homes, it is perceived as a more personal social medium and therefore more intimate than Twitter (Highfield, 2015).
Brand engagement
According to Muntinga and colleagues (2011), online brand-related activities can be
expressed by consumers at three levels of engagement: consuming, contributing and creating. These levels are not all separate but represent a continuum of activeness. The consuming COBRA type includes the least active online brand-related activities such as viewing, reading, listening, watching, downloading and sending content that is related to the brand. More active activities belong to the contributing COBRA level. The authors consider activities such as rating products and/or brands, joining a brand profile, engaging in branded conversations, and commenting on brand-related content to belong to this contributing level. Behaviours that belong to the creating COBRA level are the most active on this continuum and include publishing, uploading and writing brand-related content such as blogs, pictures, reviews, and articles.
For an organisation, consumers’ engagement with the brand is needed to establish strong relationships. Therefore the higher the engagement, the better. What social media influencers in general do is create content on their own social platform and operate in different forms of advertising, ranging from unpaid to paid (Voorn, 2016). When collaborating with a brand, which is a form of paid advertising, social media influencers create brand-related content, spread this content among their own social network, and receive a compensation for this (Forer, 2017). Thus, social media influencers engage with brands on a creating level. The highest engagement organisations can get from consumers is then at a contributing COBRA level. These activities are applicable to the Instagram platform as well. Consumers can rate a brand-related post from a social media influencer by liking it, join a profile by following a SMI who collaborates with brands, converse with other users about a brand through comments, or just comment on a SMI’s brand-related post.
The number of likes on a post indicates to what extent the content of the post is interesting to users and the number of comments quantifies how much discussion is going on and measures explicit action on the post (Bakhshi, Shamma & Gilbert, 2014). Based on the number of likes and comments a post receives, together with the amount of followers a social media
influencer has, the engagement rate can be determined (Geurin-Eagleman & Burch, 2016). Somya Mehta (2017), Email Marketing Executive of a marketing platform, says the
engagement rate on Instagram is a good way of indicating the impact of the message and how engaging the content is to consumers. In addition to the brand engagement rate, what is commented determines the valence of brand engagement. Through commenting on a brand-related post, users can share their opinions about and experiences with that brand (Hennig-Thurau, Gwinner, Walsh & Gremler, 2004). This is also known as (electronic) word-of-mouth (eWOM), which is characterized as non-commercial and interpersonal communication
between a receiver and communicator about an organisation, brand or product (Arndt, 1967; Goyette, Ricard, Bergeron & Marticotte, 2010). It is important to notice that word-of-mouth is not only positive, but can be negative as well (Hennig-Thurau, Gwinner, Walsh & Gremler, 2004). Whether the content is positive or negative determines the valence of brand
engagement (Van Doorn, Lemon, Mittal, Nass, Pick, Pirner & Verhoef, 2010). Thus, this thesis investigates two dimensions of brand engagement: brand engagement rate and brand engagement valence.
However, consumers do not just engage on social media, there are certain needs that need to be met. First of all, consumers have the desire to not only establish but also maintain social relationships in their own online networks. They want to socially interact with others and make connections (Chu & Kim, 2011; Tsai & Men, 2013). Second, consumers seek for information about products on social media platforms, preferably from others who have
specific knowledge about the product in question (Chu and Kim, 2011). Furthermore, several researchers have concluded that a key factor for consumers to engage with a brand is trust. Pansari and Kumar (2017) developed a consumer engagement framework by reviewing marketing literature. Based on previous academic articles, they state that consumer
engagement is built on relationship management. Their framework suggests that when trust is established in a relationship between consumers and organisations, consumers progress to the stage of engagement. Sashi (2012) linked practitioner views of consumers engagement with several marketing concepts and developed a consumer engagement matrix based on the opportunities of social media. This author also links consumer engagement to relationship marketing and states that consumer engagement “requires the establishment of trust and commitment in buyer-seller relationships” (p.259).
In the case of positive word-of-mouth behaviour, where brand engagement valence is based on, trust is also proven to be a key factor. Chu and Kim (2011) explain that to fulfil the need to establish and maintain relationships in social media, consumers share product information and brand experiences. In other words, consumers engage in word-of-mouth, because they want to socially interact with others on social media. Again, the underlying mechanism here lies in social relationships. Trust is one of the focal dimensions of social relationships that positively influences word-of-mouth behaviour (Nisbet, 2006). In online environments too, trust is essential to users to exchange information with others (Ridings, Gefen & Arinze, 2002). Through results from a survey, Chu and Kim (2011) proved that trust in social relationships with others in a social network has a positive effect on word-of-mouth behaviour. Thus, social media create a need for consumers that they fulfil by engaging in (positive) word-of-mouth. But this behaviour is explained by the fact that consumers engage when they feel they can trust the relationship with the other person.
Based on this, the current study perceives trust in relationships as a key antecedent for consumers to engage with the brand through both liking and (positively) commenting. Therefore, organisations should develop a personal connection with consumers (Fan &
Gordon, 2014). Brands have been trying to accomplish this by collaborating with social media influencers, who can establish this trust.
Opinion leaders and social media influencers
The transmission of information through influencers is not a new concept. Lazersfeld, Berelson and Gaudet (1944) developed a two-step flow of communication model and
highlighted the special role of opinion leaders in the network of personal relationships. They suggested that ideas often flow from mass media to opinion leaders and from opinion leaders to the less active sections of the population. Later one, Katz (1957) identified three qualities opinion leadership is based on. These qualities can all be linked to social media influencers as well. The first quality refers to a leader’s competence or expertise. Opinion leaders have the ability to gate keep interventions, change social norms and to accelerate changes in behaviour (Valente & Pumpuang, 2007). They are able to bring in new information, ideas, and opinions, disseminate them down to the masses, and influence the opinions and decisions of others (Song, Chi, Hino & Tseng, 2007). The definition of a social media influencer already captures one of the biggest competence, namely shaping attitudes of audiences (Freberg et al., 2011). By sharing content through their own social media channels, SMIs do not only have the ability to shape corporate brand perceptions (Booth & Matic, 2011) but also to strengthen a brand’s image (Brorrson & Plotnikova, 2017). The second quality of opinion leaders is their social position; this refers to who they know, who knows them and how accessible they are. Social media influencers are known to have a big network through which they can reach a
large audience (Keller, Fay & Berry, 2007). This indicates SMIs are known by a lot of others in the network.
And third, the leader’s values and traits. In general, opinion leaders possess a number of traits, including authority, likeability, scarcity and consistency (Schaefer, 2012). These
characteristics can be linked to the results of Wu and Wang’s (2011) and Zietek’s (2016) studies on online influencers. SMIs are authoritative because consumers trust brand-related opinions from influencers more than corporate messages. Then, because internet users consider SMIs as “one of them", SMIs are likeable as well. According to social identity theory, members of a group tend to be favoured over others who are not members of that group (Turner, Brown & Tajfel, 1979). Next, according to experts from the field, good social media influencers are exclusive and do not work with too many brands (Zietek, 2016). In other words, they should possess some kind of scarcity. And last, online influencers are consistent because of their frequent communication with their followers (Keller, Fay & Berry, 2007; Uzunoğlu & Kip, 2014; Woods, 2016).
It is believed that social media influencers can be as influential as opinion leaders, because they share the same qualities. Furthermore, brands want to build on the trust influencers have established, hoping the brand-related content that is distributed by an influencer is perceived as trustworthy as well (De Bakker, 2017). The next section will elaborate on which SMI characteristics facilitate this trust in order for consumers to engage with a brand.
Hypotheses
Vollenbroek, de Vries, Constantinides & Kommers (2014) investigated influence in social media communities and concluded that an actor has influence in their network if that actor
actively communicates with others. This can be explained by a consumer’s expectation of constant delivery of information when they, in any form, subscribe to any of a brand’s services (Pulizzi, 2013), when this expectation is fulfilled, influence is likely to occur. But how is the concept of frequent communication related to brand engagement? Frequent communication seems to play an important role in developing trust (Holmes, 1991), which is needed for consumers to engage with brands. The underlying mechanism of this lies in that through communication, expectations and perceptions can be aligned and consequently trust will be established (Etgar, 1978). Morgan and Hunt’s (1994) commitment-trust theory states that when past communication from the other has been frequent, this leads to greater trust in the relationship. This theory is applicable to social media influencers as well, as they are known for frequently posting on their social media accounts (Zietek, 2016). Through that, the relationships with their followers will gain trust. Based on this, it is believed that posting more frequently establishes trust between a social media influencer and consumers, and therefore consumers are more likely to like and positively comment on brand-related posts. This leads to the following hypotheses:
H1A: the more frequent a social media influencer posts the higher the brand engagement rates are
H1B: the more frequent a social media influencer posts the more positive brand engagement valence is
Next to frequent communication, the content influencers produce could have an effect on brand engagement as well. In the case of unpaid advertising, an influencer’s personal style is expressed without sponsorships or free branded goods (Marwick, 2013). Marwick (2013)
further explains that this non-sponsored content leads to consumers perceiving the influencer as authentic. This is advantageous for the influencer’s perceived trustworthiness by
consumers. In the case of paid advertising, influencers get compensated by money or products (Forer, 2017). Charlotte Meindersma (2017) is the owner of a law advisory office and a lawyer herself. In one of her blogs she writes that last year, the Reclamecode Social Media (Code for Advertisement on Social Media) clarified the rules regarding paid advertising on social media. Influencers should use the well-known hashtags #ad and #sponsored to indicate the content in sponsored, but less-known hashtags such as #spon, #paid, #sample, #adv, and #prom are accepted as well (RSM). Furthermore, a self-made hashtag and regular text are approved as well, as long as it is clear the collaboration between a brand and influencer is sponsored. Examples are: ‘gifted by organisation/brand x’ and ‘this is in collaboration with organisation/brand x’. In another blogpost Meindersma (2017) mentions the recently launched “in paid partnership” feature of Instagram. Influencers can tag the brand they collaborated with on a sponsored post in this new type of disclosure. Using this new feature, both brands and influencers meet the regulation of advertising on social media she says. Thus, influencers must now clearly mention when a post is sponsored, however this can have negative
consequences.
In context of the two-step flow model, Carr and Hayes (2014) point out that opinion leaders in social media receiving compensation could have negative consequences for the information transmission and even his/her ability to influence other people in his/her social network. A great part of the literature about sponsored content is in regard to blogposts by bloggers. Overall, the negative consequences due to sponsorship disclosures are: less favourable attitudes towards the brand (Campbell, Mohr & Verlegh, 2013), less favourable attitudes towards the message, and lower source credibility perceptions (Hwang & Jeong, 2016). Uribe,
Buzeta and Vélasquez (2016) compared implicit and explicit sponsorship disclosures. Their results indicate that explicit mentioning of a sponsorship negatively influences the SMI’s trustworthiness and the message’s effectiveness. Thus, sponsored posts can have negative consequences for both the influencer and the brand. This can be explained by the fact that in general, advertisements lead to more negative attitudes (Wojdynski, 2015). Wojdynski (2015) found that sponsorship disclosures lead to recognizing the message as an advertisement, and this negatively influences attitudes toward the organisation, quality of the message, and credibility of the message. If source credibility is negatively influenced, so is interpersonal trust between two parties negatively influenced (Giffin, 1967). The majority of consumers is only willing to engage with a brand if they feel they can trust the communicator (Heller Baird & Parasnis, 2011), which in this case is the social media influencer.
Furthermore, sponsored content can have negative consequences on word-of-mouth behaviour as well. Zengin and Zengin (2017) investigated advertisement from brands through
influencers specifically on Instagram and how users react to this. Participants pointed out that the amount of ads shared by users on their own feed is excessive and they also felt irritated about this amount. They added that the presence of too many advertisements disrupts their photo browsing on the platform, which contributes to their annoyance. To overcome these negative feelings, consumers might engage in negative word-of-mouth. Sundaram, Mitra and Webster (1998) explored the underlying motivations of this behaviour. Their results indicated that consumers engage in negative word-of-mouth communication to reduce feelings of anger. Participants explained that they shared negative experiences to ease their anger, anxiety and frustration. Based on this, it is expected that when social media influencers post too much sponsored content, followers will engage less with a brand because their trust in the influencer
diminishes, plus consumers will get annoyed and therefore engage in negative word-of-mouth in order to reduce those negative feelings. This leads to the following hypotheses:
H2A: the more frequent a social media influencer posts sponsored content the lower brand engagement rates are
H2B: the more frequent a social media influencer posts sponsored content the more negative brand valence is
Types of influencers
In identifying social media influencers to collaborate with, the size of one’s network is often mentioned and found to be one of the key factors (Brorrson & Plotnikova, 2017; Keller, Fay & Berry, 2007; Liu & Tarigan, 2016; Uzunoğlu & Kip, 2014). Macro influencers are the ones with a huge following base and have “turned their social media platforms into fortunes and empires” (O’Connor, 2017). For brands, a bigger virtual presence increases the chance to achieve a competitive advantage (Enginkaya & Yilmaz, 2014). Thus, the more people they can possibly reach the better. Generally speaking, an actor has influence in a network if that actor has many contacts (Vollenbroek et al., 2014). Nevertheless, this relationship is more nuanced as it appears. De Veirman and colleagues (2017) looked with more detail into the size of an influencer’s network and the effect it has on a SMI’s likeability and consumers’ attitudes. Based on two experiments, they confirmed that a higher number of followers leads to higher perceptions of popularity and therefore also higher likeability of a SMI. However, they point out this does not mean that the influencer is also perceived as an opinion leader. In other words, a higher number of followers does not automatically mean the SMI also has more influence on their following.
According to Zietek’s (2016) interviews with experts in fashion influencer marketing, working with micro influencers is key for fashion brands. Shane Barker from Forbes (2017) describes micro influencers as “everyday consumers who have a significant social media following of anywhere between 1,000 and 100,000”. He adds that the reason why micro influencers have gained interest of brands too is because their following seems to be relatively more engaged compared to influencers with a huge amount of followers. Again, this higher engagement can be explained by the trust followers have in micro influencers. Zietek (2016) explains that because micro influencers are perceived as having specific knowledge in a specific field, their opinions are trusted more than those of macro influencers. In addition to that, micro influencers are perceived as authentic because they do not make a living out of Instagram and collaborating with a brand is therefore not for monetary reasons. The more authentic a person is perceived, the more trustworthy he/she is thought of as well (Liu & Perrewe, 2006). This leads to micro influencers being trusted more than macro influencers.
Then, micro influencers could be more effective in generating engagement towards a brand than macro influencers. As mentioned before, consumers have the need to get information about products on social media platforms. Preferably from others who have specific knowledge about the product in question (Chu and Kim, 2011). Micro influencers have a smaller amount of followers because they are opinion leaders in smaller markets (Zietek, 2016), are known for being very passionate about their niche (Kuiper, 2017), and are trusted sources for product recommendations (Pierucci, 2017) For these reasons, the chances are higher engagement is more brand-related for micro influencers because consumers perceive them respectively as having more expertise, being more authentic, and being more
trustworthy. Therefore, consumers might turn to micro influencers to engage in brand-related word-of-mouth instead of to macro influencers. When they need brand-related information,
they could feel micro-influencers are more suitable for this. In addition to that, big social media influencers are becoming just as famous as film stars, singers, and high-end fashion models (Saul, 2016). In other words, they are becoming celebrities. Khamis, Ang and Welling (2017) compared social media influencers to celebrities in their study and say that “the SMI works to generate a form of “celebrity” capital” (p. 202). Because social media influencers are also being perceived as celebrities, their followers might want to engage with them instead of with a brand, as they would do with a celebrity. Based on all this, the following hypotheses have been formulated:
H3A: Micro influencers generate higher brand engagement rates than macro influencers
H3B: Micro influencers generate more brand engagement towards the brand than macro influencers
Method
The current study conducted a content analysis with a directed approach. This means the analysis was based on theory or relevant research findings to guide the coding (Hsieh & Shannon, 2005). Even though it is common to perform an experiment to answer a research question that expects a causal relationship (‘t Hart, Boeije & Hox, 2009), a content analysis was chosen. The reason for this is that through a content analysis, phenomena are studied through an unobtrusive and nonreactive way (Babbie, 1992). In the case of an experiment, the researcher creates a situation which has a more or less artificial character. The researcher determines who the participants are, what is going to happen during the experiment and what the circumstances are (‘t Hart, Boeije & Hox, 2009). In this thesis, social media influencers and their characteristics have been studied as they are in real life. Also, the engagement
behaviours by Instagram users and/or followers have been carried out by themselves, without any interference of the researcher. Thus, the advantage of a content analysis over an
experiment in this case is that SMI characteristics and brand engagement behaviours have been investigated such as they happen in real life.
Sample
“For fashion brands, it’s the responsive, visual channels like Instagram that have been proving themselves most profitable” says Jo Middleton (2016) from Digital Marketing World Forum (DMWF). Linking this to social media influencers, in the fashion industry, referrals by others has been regarded as the most powerful force to influence. Therefore, in fashion too, the question rises why some are more successful in transmitting new fashion and styles into their network than others? (Wiedmann, Hennigs and Langner, 2010). For these reasons, social media influencers on Instagram that work with fashion brands form the population of this investigation.
A total of seventeen fashion influencers on Instagram have been investigated. Six of those influencers were macro influencers. Abby Crain (2018, January 10), from the Wall Street Journal, writes that reaching one million followers on Instagram is a coveted mark for many influencers. Therefore, all macro influencers in this sample had above one million followers on Instagram at the time of data gathering. These macro influencers were selected partially based on Forbes’ list of top social media influencers in fashion (Forbes, 2017). Then the Instagram accounts of these Forbes’ listed influencers were searched through to find more macro influencers. The other eleven influencers were micro influencers and had between 10.000 and 100.000 followers on Instagram at the time of data gathering. To select them a
search on Google was done [e.g. “best fashion micro influencers on Instagram”]. This search led to three different lists where the micro influencers were chosen from.
Posts from a two-week period (6th of November 2017 thru 19th of November) from each influencer were investigated as well. A distinction was made between sponsored and non-sponsored posts. From each non-sponsored post the corresponding number of likes, number of comments, valence of comments and to whom the comments referred to were studied as well. A total of sixty-four sponsored Instagram posts and 8096 comments have been coded and analysed. Thirty-two of these sponsored Instagram posts were from macro influencers and the other thirty-two posts were from micro influencers. Only comments that contained text and were in English, Dutch or German were coded.
Data gathering
To be able to gather enough data, all seventeen influencers had to have an open account, be active daily on Instagram and had to have collaborated with (a) fashion brand(s). It was noticed that many posts in December were giveaways because of the upcoming holidays and many more users commented on these kind of posts. Therefore December was not a
representative month to study brand engagement and a random two-week period in November 2017 was chosen. Screenshots were taken from the profile pages of all the influencers
including their number of followers. All posts from those two weeks were separately screenshotted. The screenshots included the image/video, tags in the image/video, and the caption. The codebook (Appendix I) illustrated the differences between a non-sponsored and sponsored post. A sponsored posts can be recognized in three ways. First, it has included one of the following hashtags in the caption: #ad, #sponsored, #spon, #paid, #adv or #prom. Second, it has included the “paid partnership with” feature in the sub-header. Third, a
self-made hashtag or text has been included that clearly mentions the collaboration with the brand is sponsored. From sponsored posts only, the corresponding number of likes and number of comments were also screenshotted. For each sponsored post, a hyperlink was provided to access all the comments on Instagram. All screenshots were taken on December 13th 2017 and saved in a private file. By then, all Instagram posts had already reached their maximum engagement. Branded posts receive 75% of all comments within 48 hours and high-quality content reaches the same amount within even 13 hours (Osman, 2017).
Operationalisation
A list of all the influencers was made including their Instagram account names and whether they were a micro or macro influencer. This was based on the amount of followers on
December 13th 2017. The frequency of communication was measured by counting the amount of posts that were uploaded from 6th of November 2017 thru 19th of November. As said before, these posts were screenshotted and saved into a private file. Thus the researcher only had to count the amount of posts that were in the specific file. Whether a post was sponsored could be recognized by the three previously mentioned ways. The amount of sponsored posts was then measured by counting how many of the posts were sponsored per influencer.
To test the brand engagement rate, only the engagement rates of posts that were in
collaboration with a brand have been analysed. For each branded post, the number of likes, number of comments and the number of followers from that influencer were coded. Then, in the codebook was instructed how the brand engagement rate could be calculated for each post. This is done by adding a post its total number of likes and comments, dividing that number by the number of followers of the social media influencer and multiplying that by one hundred to get the rate in percentages (Laurence, 2017, August 22). Brand engagement valence was
measured by the tone of the comments which could be either positive, negative or neutral. Positive comments express happiness, excitement, interest and/or amusement. Examples are good experiences followers have had with the brand or expression of the love they have for the influencer. Negative comments express anger, hate, disappointment and/or sadness. Examples are complaints about the product or towards the influencer or brand. Neutral comments do not express whether they like or do not like the content/post and often included questions. The codebook also included a list of words that are commonly used in positive and negative comments on Instagram posts of a social media influencer. To measure the brand valence, the neutral comments (n = 484) were recoded into missing values and only positive and negative comments were used for analysis. A total of 7612 comments were eventually used to test hypotheses 1B and 2B. Furthermore, the comments could refer to either only the social media influencer, only the brand and/or its product(s), both the social media influencer and the brand and/or its product(s) or to neither of them/not clear. For analysis of hypothesis 3B, the comments were divided into brand related and non-brand related. In this case, positive, negative, and neutral comments were all used for analysis (n = 8097).
Reliability
This study has performed a intercoder reliability test using Krippendorff’s Alpha on 10% of the total sample by two coders. When using Krippendorff’s Alpha, total sample size, presence of multiple coders and missing data are not problematic (Hayes & Krippendorff, 2007). The sample consisted out of seven Instagram posts. One Instagram post was from one macro influencer and 222 comments were coded from this. The other six Instagram posts were from two micro influencers, and 298 comments were coded from these posts. All comments were numbered beforehand to assure both coders coded the same comment each time. The average
Krippendorff’s Alpha showed a value of 0,82 and represents a good intercoder reliability. A more detailed overview of the reliability test and its values can be found in Appendix II.
Results
To be able to answer the research question, several analyses have been conducted to test the hypotheses. The first hypothesis tests the relationship between a SMI’s frequency of posting and brand engagement rate. Table 1 shows the frequency of posting per influencer and the average brand engagement rate per influencer. The average frequency of posting is 29.82 posts in two weeks, the lowest frequency of posting is 12 posts in two weeks and the highest frequency of posting is 39 posts in two weeks. As seen in Table 1, the lowest average brand engagement rate per influencer is 1.04% and the highest brand engagement rate per influencer is 8.90%. From all posts, thus not per influencer, the average brand engagement rate is 3.53%, the lowest brand engagement rate is 0.64%, and the highest brand engagement rate is 9.62%.
A simple linear regression was calculated to predict the brand engagement rate based on the SMI’s frequency of posting (H1A). The results show that frequency of posting, b* = .53, t = -4.85, p < .001, 95% CI [-.19, -.08], has a significant strong association with brand
engagement rate. A significant regression equation was found (F(1, 61) = 23.555, p < .000), with an R2 of .279 meaning that 27.9% of the variance is explained by frequency of posting. However, when frequency of posting increases one unit, a social media influencer’s brand engagement rate decreases with -.137. Meaning that the more frequent a social media
influencer posts the lower brand engagement rates are. This is contrary to what was predicted. Thus, hypothesis 1A cannot be accepted.
Table 1.
The frequency of posts, frequency of sponsored posts and average engagement rate per influencer (N = 17)
Name social media influencer
Frequency of posts Frequency of sponsored posts Average brand engagement rate Macro influencers Weworewhat 34 5 1.92% Songofstyle Juliahengel Camilacoelho Imjennim Carodaur Micro influencers Hannahliza Savinachaiyj Kateogata Maryorton Candidlychan Thefashionbum Lilymontasser Alyssainthecity Iiindiefoxx Joandkemp Themoptop 39 22 27 21 33 18 18 17 24 20 17 16 19 12 15 14 9 6 2 2 8 6 2 2 3 2 3 2 2 2 6 2 1.04% 1.81% 1.95% 3.53% 3.25% 3.78% 1.57% 2.14% 8.90% 5.18% 4.60% 3.77% 5.87% 4.46% 6.35% 6.33%
Hypothesis 1B predicts that the more frequent a social media influencer posts the higher the chance is brand engagement valence is positive. 7612 comments were analysed to test this
prediction, 76 comments were negative and 7536 comments were positive. The dependent variable is categorical, brand engagement valence can either be positive or negative, which is why was chosen for logistic regression. The analysis shows the logistic model is not
significant (χ2 (3) = 2.537, p = .111) and is not a good fit to the data according to the Hosmer and Lemeshow test (p = .002). Based on the results, it can be said that the effect of frequency of posting on brand engagement to be positive is not significant (Exp(B)=1.022, Wald = 2.576, df = 1, p = .109). Therefore, hypothesis 1B is rejected.
Table 1 also shows the frequency of sponsored posts per influencer. The average amount of sponsored posts per influencer is 6.35, the lowest amount of sponsored posts per influencer is 2 and the highest amount of sponsored posts per influencer is 9. A linear regression was used to test if the amount of sponsored Instagram posts significantly predicts a SMI’s brand engagement rates (H2A). The results show that the amount of sponsored posts, b* = .39, t = -3.31, p < .005, 95% CI [-.55, -.14], has a significant, moderately strong association with brand engagement rate. The amount of sponsored posts explains 15.2% of the variance (R2=.15, F(1, 61)=10.934, p<.005). Also, for every unit the amount of sponsored posts increases, brand engagement rates decrease with -.344. This means that the more posts of a social media influencer are sponsored, the lower brand engagement rates are and hypothesis 2A can be accepted.
To test hypothesis 2B "the more posts of a social media influencer are sponsored the more negative brand valence is" a logistic regression was done. Again, because the dependent variable “brand engagement valence” is categorical (negative/positive) a logistic regression was chosen. The results show the logistic model is significant (χ2 (1) = 18.446, p < .001) but only explains 2.3% of the variation in the outcome (Nagelkerke R2 = .023). The Hosmer and
Lemeshow test shows the model poorly fits to the data (p = .014). Nevertheless, the results show there is a significant effect of the amount of sponsored posts on the valence of brand engagement (B = .183, Wald = 18.550, df = 1, p < .001). In more detail, for a social media influencer who posts more sponsored posts, the chance to receive positive comments is 1.201 times higher than for a social media influencer who posts less sponsored posts
(Exp(B)=1.201, p<0.001). Even though this effect is significant, it is contrary to what was expected because it was expected the chances would be lower to receive positive comments. Therefore, hypothesis 2B cannot be accepted.
Hypothesis 3A predicted that micro influencers have higher average brand engagement rates than macro influencers. Table 1 shows the brand engagement rates for both macro and micro influencers. To test the prediction, an independent-samples t-test was conducted. The results showed that micro influencers (M = 4.97, SD = 2.23) report significantly higher brand engagement rates than macro influencers (M = 2.11, SD = 1.20), t (47.83) = 6.35,
p < .000, 95% CI [1.95, 3.75]. Micro influencers have an average brand engagement rate of 4.97% per post and macro influencers have an average brand engagement rate of 2.11% per post. Based on these results, hypothesis 3A can be accepted.
A Chi-square test of independence was calculated comparing the amount of brand-related comments between micro and macro influencers (H3B). For this analysis all comments (positive/negative/neutral) were used (n = 8096). From the 1711 comments on posts of micro influencers, 476 comments were brand-related and from the 6385 comments on posts of macro influencers, 1067 comments were brand-related. A significant interaction was found (X2 (1) = 107.95, p < .000). In more detail, significantly more comments on branded posts of micro influencers were brand-related (27.8%) than comments on branded posts of macro
influencers (16.7%). Thus, micro influencers generate more engagement towards the brand than macro influencers and hypothesis 3B can be accepted.
Conclusion and discussion
This thesis asks to what extent brand engagement can be positively affected by social media influencers on Instagram? The reason this question is asked is because social media
influencers are becoming part of PR strategies to develop relationships with a brand’s audience (Jarvis, 2006). However, professionals are having difficulties determining what the effective metrics are to decide who the most influential influencers are (Enginkaya & Yilmaz, 2014). To be able to answer the research question, three different characteristics of social media influencers have been investigated: the frequency of posts, the frequency of sponsored posts, and whether they are a micro or macro influencer. It was expected that these
characteristics positively influence brand engagement based on previous research that says these characteristics positively influence trust, which is a key factor in brand engagement. The main conclusions will now be discussed.
First of all, there are noteworthy differences between micro influencers and macro
influencers. Based on this thesis, it can be concluded that micro influencers generate higher brand engagement rates than macro influencers and also generate more engagement towards the brand than macro influencers. Micro influencers and their opinions are trusted more than macro influencers because they do not collaborate with brands for monetary reasons (Zietek, 2016), therefore they could be seen as more trustworthy. Also, micro influencers are
perceived as having specific knowledge in a specific field (Zietek, 2016), therefore consumers could prefer turning to micro influencers when they need brand-related information because they believe micro influencers have a certain expertise. Furthermore, macro influencers are
the ones who have turned their social media platforms into fortunes and empires (O’Connor, 2017) and are increasingly reaching the status of celebrities (Saul, 2016). Working with big influencers who are perceived as celebrities could be disadvantageous for brands, because this could be the reason why consumers engage more with the social media influencer than with the brand. It is important for brands to know that a higher number of followers does not always mean a SMI also has more influence (De Veirman et al., 2017). The findings implicate it is more valuable for brands to collaborate with micro influencers than with macro
influencers. This means that smaller brands can work with influencers too, since these
collaborations require a lower budget (Barker, 2017). Above that, this study proves it is worth it for smaller organisations to invest in these collaborations. Larger organisations, who have a bigger budget to spend, should collaborate with several micro influencers. This way, they can still reach a wider public, but the chances are higher this public is more engaged than that of one macro influencer.
Second, collaborations between brands and social media influencers also create challenges. Influencers must include a sponsorship disclosure in their post when they received
compensation for it by a brand. The current investigation found that the more posts of a SMI are sponsored, the lower brand engagement rates are. This can be explained by research that says sponsorship disclosures negatively influence attitudes toward the brand and towards credibility of the message because the content is recognized as an advertisement (Wojdynski, 2015). Then, if the message, which is distributed by the influencer, is not perceived as
credible, the interpersonal trust between consumers and the social media influencer is also negatively influenced (Giffin, 1967). This finding provides evidence for why organisations should look for influencers who are exclusive by not working with too many brands and do not post too many sponsored posts (Zietek, 2016).
Third, it can also be concluded that the more posts of a social media influencer are sponsored the higher the chance is brand engagement is positive. It was expected that this chance would be lower, because positive word-of-mouth is likely to occur when trust is established between a communicator and receiver (Chu & Kim, 2011; Ranaweera & Prabhu, 2003 & Tsai & Men, 2013) but a sponsorship disclosure negatively influences trust in the online influencer and attitudes towards the sponsored product (Carr & Hayes, 2014). However, an explanation for the current finding could be that followers have so much trust in social media influencers and their opinions (Wu & Wang, 2011; Zietek, 2016), the negative effect of a sponsorship
disclosure weakens or even disappears. In addition to that, according to Abidin (2016), influencers are known for not only creating textual and visual narrations of their personal, daily lives but also for advertising products and services. Consumers could therefore be aware of the fact influencers also create sponsored content and therefore not negatively comment about it. Another explanation could be that followers do not even notice when an Instagram post is sponsored. Evans, Phua, Lim and Jun (2017) concluded that not all disclosures are as effective for consumers to recognize the content as advertisement. Then only if the consumer comprehends the post is sponsored and also remembers the sponsorship disclosure, this negatively influences eWOM.
Hence, the frequency of sponsored posts leads to somewhat contradictory results. On the one hand too many sponsored posts leads to less people engaging with the post, but on the other hand the more posts of an influencer are sponsored, the higher the chance consumers place positive comments and positively engage with the brand. In practice this means that brands can collaborate with influencers who work with (many) other brands too. Even though less consumers will engage, the ones who do, engage positively with the brand. It is a trade-off brands need to consider. However, there are other characteristics brands can look at as well.
Based on this study, if it is a micro influencer who posts many sponsored posts, the
engagement could be higher again because micro influencers have a more engaged following than macro influencers.
Fourth, it was expected that the more frequent a social media influencer posts, the higher brand engagement rates are. SMIs are characterized, amongst other things, by their frequent communication with their followers (Keller, Fay & Berry, 2007; Uzunoğlu & Kip, 2014; Woods, 2016). Findings from Vollenbroek and his colleagues' research (2014) found that in specifically social media communities, a person has influence in their network if that person communicates actively with others. Also, Morgan and Hunt's (1994) commitment-trust theory says that if communication from the other person is frequent, this establishes greater trust in a relationship. However, the current study found the opposite. It seems that the more often a SMI posts, the lower brand engagement rates are. This could be explained by how Instagram and its (new) algorithm work. If a social media influencer posts more frequently the chances are higher some of those posts are uploaded on random times of a day. According to Morales (2017), if a post is upload on less-active times of a day, the content is also less likely to be engaged with. He explains that in the new algorithm, a new post is first shown to a small percentage of the followers. Then, it measures the engagement of that small group of followers and compares it to the other posts of the SMI on similar days and times. If it is getting less engagement, the post is shown to a smaller percentage of the followers and is shown further down somebody's homepage. Consequently, less people see that SMI's post and less people can engage with the post. Thus, once the engagement is “too low” according to Instagram’s new algorithm, it will be more difficult for the post to still get a lot of
engagement. In practice this means that brands should be careful collaborating with influencers who very actively post, because the chances are high many consumers will not
even see the branded content. If a brand does decide to work with an influencer who posts frequently, they should investigate at what times the influencer gets the highest engagement rates. Then they can decide on a specific time the influencer should upload the brand-related post.
Last, the effect of frequency of posting on positive brand engagement was found to be non-significant. This however does not mean there is no relationship between the two. The findings did show a small and positive effect on brand engagement valence by frequency of posting. Perhaps the sample was too small to find a significant effect here. But as mentioned before, especially in the case of Instagram, more frequent communication does not always lead to positive consequences. Other factors that can affect engagement have to be taken into consideration as well, such as the visual content and value of the content (Morales, 2017).
Limitations
Based on previous research, the role of trust is considered as one of the key mediating factors between social media influencers and brand engagement. However, the current study has not investigated and taken this into analysis itself. Therefore, it can be assumed that trust explains some of the effects of social media influencers on brand engagement but this cannot be proven with evidence from this study. For future research it would then be advised to test if trust plays a mediating role between frequency of communication, amount of sponsored posts, being a micro or macro influencer and brand engagement on Instagram. If then this is the case, this implicates that influencers who are perceived as trustworthy are effective opinion leaders to endorse brands.
The current research attempts to investigate relevant characteristics that contribute to a social media influencer's influence, however there are other important factors that are relevant too but could not be taken into consideration, because a content analysis does not make this possible. Vollenbroek and colleagues (2014) for example found that traits such as having authority, expertise and being credible are influential actors, but if followers actually perceive the SMIs as having these traits has not been investigated. Research done specifically on influencers in fashion emphasizes the importance of authenticity (Marwick, 2013). Good influencers are distinguished from bad influencers based on how authentic they are perceived. This study has not included consumer perceptions/attitudes, while as mentioned in the
beginning, organisations and consumers both play a role in brand engagement (Mangold & Faulds, 2009).
Therefore, it is recommended for future research to also study (perceived) traits of SMIs in the context of brand engagement through for example a survey or experiment and combine this with a content analysis. It is advised to use real-life influencers to guarantee a real as possible situation. Then, findings from both the survey/experiment and content analysis can be combined to investigate consumers’ perceptions and attitudes towards a social media influencer/brand and actual behaviours regarding that influencer/brand, to get a more complete picture.
Despite these limitations, the present study also made some important contributions. The method of this thesis is what distinguishes the study from other research. First of all, through a content analysis, social media influencers and their influence are studied in an unobtrusive and nonreactive way in comparison to for example surveys, interviews and experiments (Babbie, 1992). Real-life situations have been studied using data that brands also have access
too. This thesis shows that investigating social media influencers and brand engagement is accessible to both science and practice. Second, the advantage of the directed approach is that some of the current findings can support theories from previous research and give more insight into social media influencers and their influence in general, and on Instagram specifically.
Furthermore, the way brand-related content is created, distributed, and consumed has changed. The likes and posts of influencers on social media is how a brand’s image and reputation are now largely formed (Muntinga, Moorman & Smit, 2011). But 55% of the professionals say they have a limited or even no understanding of the identification of online influencers (Lewis PR, 2011). The most important contribution of this study it that is has shed more light on what makes a social media influencer influential. Interviews have been held with professionals about influencers based on their experience in the work field (Zietek, 2016), but empirical evidence that proves what works for a brand is lacking. The current study can answer the question asked by De Veirman and her colleagues (2017) about who is the most likeable, credible influencer with a high value as an opinion leader. This would be the micro influencers who collaborate with other brands too.
Literature
Adweek (2015, July). 10 Reasons Why Influencer Marketing is the Next Big Thing. Retrieved from
http://www.adweek.com/digital/10-reasons-why-influencer-marketing-is-the-next-big-thing/
Arndt, J. (1967). Role of product-related conversations in the diffusion of a new product. Journal of Marketing Research, 4(3), 291-295. Doi:10.2307/3149462
Babbie, E., & Mouton, J. (1992). The practice of social science. Belmont, CA: Wadsworth. Bakhshi, S., Shamma, D. A., & Gilbert, E. (2014, April). Faces engage us: Photos with faces
attract more likes and comments on instagram. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (pp. 965-974). ACM.
Barker, S. (2017). Using Micro-Influencers To Successfully Promote Your Brand. Retrieved from https://www.forbes.com/sites/forbescoachescouncil/2017/09/29/using-micro-influencers-to-successfully-promote-your-brand/#7cd0b06e1763
Barry, J. (2014). 4 Archetypes of Top Social Media Influencers. Retrieved from http://blog.socialcontentmarketing.com/top-social-media-influencers/
Block, E. (2018, January 27). 35 Top Fashion Influencers for When You’re in Need of Fresh Inspo [Web log post]. Retrieved from http://www.whowhatwear.co.uk/best-fashion-influencers
Bloglovin’ (2016, November). Bloglovin’ Consumer Survey November 2016 – We Asked, They Answered: The Global Power of Influencers. Retrieved from
https://influence.bloglovin.com/the-global-power-of-influencers-introducing-thenovember-2016-bloglovin-consumer-survey-ac0d62657296#.taudr2rop
Bakhshi, S., Shamma, D. A., & Gilbert, E. (2014, April). Faces engage us: Photos with faces attract more likes and comments on instagram. In Proceedings of the 32nd annual ACM conference on Human factors in computing systems (pp. 965-974). ACM. Birtwistle, G., & Freathy, P. (1998). More than just a name above the shop: a comparison of
the branding strategies of two UK fashion retailers. International Journal of Retail & Distribution Management, 26(8), 318-323.
Booth, N., & Matic, J. A. (2011). Mapping and leveraging influencers in social media to shape corporate brand perceptions. Corporate Communications: An International Journal, 16(3), 184-191. doi:10.1108/13563281111156853
Brorsson, A., & Plotnikova, V. (2017). Choosing the right social media influencer.: A quasi-experiment to explore the impact of influencers’ different characteristics.
(Dissertation). Retrieved from http://urn.kb.se/resolve?urn=urn:nbn:se:lnu:diva-65092 Campbell, M. C., Mohr, G. S., & Verlegh, P. W. (2013). Can disclosures lead consumers to
resist covert persuasion? The important roles of disclosure timing and type of response. Journal of Consumer Psychology, 23(4), 483-495.
doi:10.1016/j.jcps.2012.10.012
Carr, C. T., & Hayes, R. A. (2014). The effect of disclosure of third-party influence on an opinion leader's credibility and electronic word of mouth in two-step flow. Journal of Interactive Advertising, 14(1), 38-50. doi: 10.1080/15252019.2014.909296
Chu, S. C., & Kim, Y. (2011). Determinants of consumer engagement in electronic word-of-mouth (eWOM) in social networking sites. International journal of Advertising, 30(1), 47-75. doi:10.2501/IJA-30-1-047-075
Crain, A. (2018, January 10). What Happens When You Reach a Million Instagram Followers. Retrieved from: https://www.wsj.com/articles/now-you-too-can-get-1-million-instagram-followers-1515599740
Çukul, D. (2015). Fashion Marketing in Social Media: Using Instagram For Fashion Branding (No. 2304324). International Institute of Social and Economic Sciences. De Bakker, S. (2017, May 11). Dit zijn 3 belangrijke onderzoeken over influencer marketing.
Retrieved from https://www.marketingonline.nl/achtergrond/dit-zijn-3-belangrijke-onderzoeken-over-influencer-marketing
De Veirman, M., Cauberghe, V., & Hudders, L. (2017). Marketing through Instagram influencers: the impact of number of followers and product divergence on brand attitude. International Journal of Advertising, 36(5), 798-828.
Dessart, L., Veloutsou, C., & Morgan-Thomas, A. (2015). Consumer engagement in online brand communities: a social media perspective. Journal of Product & Brand
Management, 24(1), 28-42.
Drewry, E. (2017). Why Powerful Brands and Influencers Choose Instagram to Connect with Audiences. Retrieved from:
https://www.forbes.com/sites/emilydrewry/2017/09/28/why-powerful-brands-and-influencers-choose-instagram-to-connect-with-audiences/#3505e35f4f3c
loyalty intentions. Journal of Retailing and Consumer Services, 24, 100-109. doi:10.1016/j.jretconser.2015.02.007
Elliot, N. (2014, April 29). Instagram Is The King Of Social Engagement [Web log post]. Retrieved from
https://go.forrester.com/blogs/14-04-29-instagram_is_the_king_of_social_engagement/
Enginkaya, E., & Yılmaz, H. (2014). What drives consumers to interact with brands through social media? A motivation scale development study. Procedia-Social and Behavioral Sciences, 148, 219-226. doi:10.1016/j.sbspro.2014.07.037
Erdoğmuş, İ. E., & Cicek, M. (2012). The impact of social media marketing on brand loyalty. Procedia-Social and Behavioral Sciences, 58, 1353-1360.
doi:10.1016/j.sbspro.2012.09.1119
Erkan, I. (2015). Electronic word of mouth on Instagram: customers’ engagements with brands in different sectors. International Journal of Management, Accounting and Economics, 2(12), 1435-1444.
Eror, A. (2017, April 13). Are Instagram Influencers Bad For Fashion? [Web log post]. Retrieved from https://www.highsnobiety.com/2017/04/13/fashion-influencers-instagram-social-media/
Etgar, M. (1978). Intrachannel Conflict and Use of Power. Journal of Marketing Research (JMR), 15(2), 273-274.
Evans, N. J., Phua, J., Lim, J., & Jun, H. (2017). Disclosing Instagram influencer advertising: The effects of disclosure language on advertising recognition, attitudes, and behavioral intent. Journal of Interactive Advertising, 1-12.
Fan, W., & Gordon, M. D. (2014). The power of social media analytics. Communications of the ACM, 57(6), 74-81. doi:10.1145/2602574
Federal Trade Commission. (2017). The FTC's Endorsement Guides: What People Are Asking. Retrieved from https://www.ftc.gov/tips-advice/business-center/guidance/ftcs-endorsement-guides-what-people-are-asking
Flynn, L. R., Goldsmith, R. E., & Eastman, J. K. (1996). Opinion leaders and opinion seekers: Two new measurement scales. Journal of the Academy of Marketing Science, 24(2), 137-147.
Forbes (2017). Top Influencers of 2017. Retrieved from https://www.forbes.com/top-influencers/#15624fa772dd
Forer, L. (2017, February 22). The Lowdown on Social Media Influencers [Infographic] [Web log post]. Retrieved from https://www.marketingprofs.com/chirp/2017/31547/the-lowdown-on-social-media-influencers-infographic
Freberg, K., Graham, K., McGaughey, K., & Freberg, L. A. (2011). Who are the social media influencers? A study of public perceptions of personality. Public Relations
Review, 37(1), 90-92. doi:10.1016/j.pubrev.2010.11.001
Gambetti, R. C., Graffigna, G., & Biraghi, S. (2012). The grounded theory approach to consumer-brand engagement. International Journal of Market Research, 54(5), 659-687. doi:10.2501/IJMR-54-5-659-687
Geurin-Eagleman, A. N., & Burch, L. M. (2016). Communicating via photographs: A gendered analysis of Olympic athletes’ visual self-presentation on Instagram. Sport Management Review, 19(2), 133-145. doi:10.1016/j.smr.2015.03.002
Giffin, K. (1967). The contribution of studies of source credibility to a theory of interpersonal trust in the communication process. Psychological bulletin, 68(2), 104.
doi:10.1037/h0024833
Goyette, I., Ricard, L., Bergeron, J., & Marticotte, F. (2010). e‐ WOM Scale: word‐ of‐ mouth measurement scale for e‐ services context. Canadian Journal of Administrative
