1
Advertising on social media: the effects of
clandestine advertising on product attitudes through
authenticity
Author: Dominique Ham Student number: 10789197
University of Amsterdam
Faculty of Economics and Business
MSc Business Administration – Marketing track
Supervisor: Samuel Franssens Date of submission: June 22nd 2018 Version: Final Version
2
Statement of originality
This document is written by Student Dominique Ham who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this
document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3
Abstract
The purpose of this study was to examine the effect of clandestine advertising in social media posts on product attitude. Also, the possible cause for this effect, authenticity of the endorser, is researched. Clandestine advertising on social media is very common nowadays. The question is whether consumers gradually start to notice this kind of persuasive attempts, and whether the use of clandestine advertising affects the perceived authenticity of the endorser, and in turn product attitude. Furthermore, the differences in effects of clandestine advertising on perceived authenticity between celebrity endorser and social influencer endorser are researched. To answer these questions, an online experiment is conducted and 306
respondents from the US and the UK participated in this study. Respondents were randomly allocated to one of four experimental cells in a 2(celebrity vs. influencer) x 2(clandestine vs. clear) experimental design. The results show that the type of advertising has no effect on product attitude. Furthermore, perceived authenticity of the endorser does not mediate the effect of type of advertising on product attitude. There are no differences in perceived
authenticity when using either a celebrity endorser or a social influencer endorser. The use of clandestine advertisement has the same effect on perceived authenticity in case of using a celebrity endorser or a social influencer endorser.
Key words: clandestine advertising, Instagram, celebrity endorsement, social influencer, persuasion knowledge model, authenticity, envy, product attitude
4
Table of content
1 Introduction ... 5 2 Literature review ... 11 2.1 Online advertising ... 11 2.2 Social media ... 122.3 Persuasion Knowledge Model ... 13
2.4 Celebrity endorsement ... 14 2.5 Clandestine advertising ... 17 2.6 Authenticity ... 19 2.7 Research gap... 21 2.8 Conceptual model ... 22 3 Method... 23 3.1 Procedure ... 23 3.2 Manipulation ... 24 4 Results ... 25 4.1 Measures ... 25
4.2 Moderated mediation - Authenticity ... 27
4.3 Moderated mediation analysis with PROCESS – Authenticity ... 33
5 Conclusion ... 34
5.1 Moderated mediation – Envy ... 36
5.2 Moderated mediation analysis with PROCESS – Envy ... 40
6 Discussion ... 41
6.1 Discussion of variables and limitations ... 41
6.2 Managerial implications ... 45 6.3 Further research ... 46 7 References ... 48 8 Appendix ... 55 8.1 Appendix A – Questionnaire ... 55 8.2 Appendix B – Manipulation ... 60
5
1 Introduction
On a rainy day in December, Nola is chilling in her room and she is very bored. She just had a fight with her parents, so she will definitely not go downstairs anymore. She opens her laptop and watches a new vlog by her favorite YouTube star. Nola is now twenty-one years old and in the end of developing her identity. Sometimes she feels a little bit insecure about that. Her YouTube stars help her by talking about their everyday problems, showing their clothes and make-up routines and doing cool things all day long. At study Nola and her friends often talk about the most popular and beautiful vloggers. Actually, she loves one vlogger in particular, because she can identify herself with her. Like Nola, she is a small tiny brunette with the same skin problems and besides that she also fights a lot with her parents. Parents do not understand their lives and Nola is tired of it. Last Friday her favorite vlogger showed her whole make-up routine and it was so gorgeous that Nola went to the stores to buy the make-up herself. The day after, Nola’s friends were nice about her new make-up, and that gave Nola more confidence. She would not be surprised if her friends went to the store to buy the same make-up that day after school.
“Brands are relying on influencer marketing more than ever.” Companies can search for people who have more than a million followers on social media and target their consumers easily, through advertise via those profiles. Even companies like Amazon have started a social influencer program, where they are looking for the best social influencers to recruit (Fortune, 2017). In a Dutch survey among thirteen and twenty-four-year olds, 63% indicated that they would buy products that are recommended by influencers (Marketingfacts, 2016). Since late 2005 these ‘new influencers’ became popular by sharing their lives on social media. They share pictures of their everyday activities and make videos, named vlogs, about themselves and share these videos with the whole world. These influencers built a sizeable social network of followers (de Veirman, Cauberghe, Hudders, 2017). The highest number of followers of a
6 personal profile has now reached 132,5 million on Instagram (Statista, 2018). The use of such social media platforms grew phenomenally in a short time frame. For example, Instagram grew from approximately 90 million users in 2013 to 800 million users in 2017 (Statista, 2018). Jeroen van den Ven, social media manager of ABN AMRO: “Influencer marketing is the key to authentic human-to-human brand communication.” It is, of course, possible to deliver a message from brand to consumer, but social influencers are standing much closer to your target audience (Marketingfacts, 2017). Social influencers promote products and services to their followers as if they really like it and therefore recommend it. The recommendations that are made in the content, also known as ‘electronic Word of Mouth’, would be, according to earlier research, more effective than traditional media because the resistance to the message is lower. This is a result of higher authenticity and credibility of the persuasive message, because it seems that the social influencer really likes the product (de Vries, Gensler &
Leeflang, 2012). More perceived authenticity and credibility of persuasive messages is caused by the fact that half of the social media users is not aware of the sponsorships, according to a study by Harris Interactive (Langford & Baldwin, 2013). So, half of the social media users does not notice the sponsorship, which results in more perceived credibility and authenticity of the message, because people think that the message is not sponsored. I think that this finding changed in the last couple of years. Unclear sponsorships, also known as clandestine advertising, occurs so often nowadays that I expect that social media users do recognize the advertising, but because it is not mentioned, this will harm the authenticity of the social influencer. Much research focuses on the rise of social media marketing and how to use it in the most effective way (Gillin, 2009; Rohm & Crittenden, 2011; Kietzmann et al., 2011; Mangold, 2009). Social media give users the opportunity to create their own content and ‘ordinary people’ are able to become the ‘new social media stars’. Because of this phenomenon, marketing companies have shifted their focus from obtaining celebrity
7 endorsements to obtaining social influencer endorsements. We can ask ourselves if the
investment in social media influencers instead of celebrity endorsements is a smart strategy. Consumers could perhaps identify themselves more easily with social influencers than with celebrities because social influencers are ‘ordinary people’, like themselves. On the other hand, I can imagine that consumers are a little skeptical about social influencers, because it may not always be clear what their talents are. Many of them did not finish school and started with blogging and vlogging, because it seemed nice to them to make that their job. Some vloggers trouble with burn-outs, because they do not have a job from 9.00am till 5pm, but work all day long, according to themselves. People react incomprehensible and skeptical, because they do not believe that vlogging and blogging is so hard and they think that you do not need particular talents for it (rtlnieuws, 2018). In contrast, celebrities are often famous as a result of their talents like singing or acting. Much research has already been done
investigating the effectiveness of celebrity endorsement as a marketing tool and I can use that in predicting the effects of social media influencers.
With the increased reliance on the new social media stars for marketing purposes, it is important to know their influence on persuading consumers, and in particular in combination with the sort of sponsorship that is used in social media. There is a big difference between traditional marketing and social media marketing and legislations. For social media there are no strict advertising rules and therefore it is often unclear if social influencers are sponsored by a brand or not. Instagram recently developed the option to clearly indicate a sponsorship with the sentence: ‘Paid partnership with…’, above the post. For brands it is already obligated to use such clear advertising on their brand pages, but for social influencers it is optional. So for example, when Coca Cola post a picture on Instagram on their brand page, they have to mention clear that the post has persuasive purposes. But when Coca Cola sponsors an
8 not obligated to mention the sponsorship. When social influencers choose to use unclear advertising, we consider this as clandestine advertising, which is strictly forbidden in, for example, public, non-commercial television programs. The reason for this is that people are more malleable when they do not know about the sponsorships, in particular, younger people are protected by the ‘Commission for the media’. Think about no tobacco, medical treatments and alcohol advertising before 09.00 pm, and no product placement in children programs at all (Nederlandse Reclame Code, 2017). It could be assumed that vague advertising on social media is so common because companies expect that the effects of sponsoring are more positive when consumers do not know that the social influencers are sponsored. One study states that online advertising has become increasingly annoying and intrusive to internet users (Rohrer & Boyd, 2004). Therefore, companies search for more natural ways to advertise, for example through the sponsoring of social influencers. Companies think that in this way advertisings foster customer engagement and cause less intrusiveness. It is important to question this because I expect that people do notice the sponsorships without a clear description, because online clandestine advertising occurs very often. Besides that, people increasingly value authenticity in brands (NOS, 2016), and when brands or endorsers are not honest with their consumers, this could result in less perceived authenticity. The reason that people increasingly value authenticity is that consumers are confronted with increasing commercialization, an overflow of the fake, and the presence of meaningless market offers. Consumers want to overcome this meaninglessness and therefore look for brands that are original, relevant and genuine; they look for authenticity in brands (Morhart et al., 2015). Social influencers that are perceived authentic attract consumers that process the message or post deeper, than in the case of social influencers with less authenticity. When consumers are deeply processing the endorsed advertising, they are more likely to form strongly held attitudes that influence behavior (Kapitan & Silvera, 2016).
9 To investigate if there is a difference between the effects of clandestine and clear advertising on product attitude, and if this difference is influenced by the perceived authenticity of the endorser, the following research question is composed:
RQ: To what extent is there a difference in effect between clear and clandestine advertising in social media posts on product attitude and is this difference explainable through the perceived authenticity of the endorser?
Below, an Instagram post by a social influencer in the Netherlands is included, with a delivery service in the background (Image 1). The brand is not mentioned in the description of the social media post. People react negatively to this clandestine advertising, probably
because they notice the advertising but the sponsor is not mentioned clearly. In this way consumers may think that they are fooled by the brand and the influencer. A few people react with the brand of the delivery service and someone reacts: ‘Scary clandestine advertising.’ You also see another picture with clear advertising and people react very positively (Image 2). Because of these different marketing strategies and consumer reactions, it is very difficult for companies to decide which option to choose. There are limited studies conducted about this topic, and none of them included authenticity of the influencer or brand as a mediator. Therefore, I would like to fill in this gap with my study. The findings of the study not only provide academic insights, but also deliver valuable information for marketing managers and social influencers.
10
Image 1: Clandestine advertising ‘DHL’
11
2 Literature review
This chapter will provide a general introduction of online advertising and social media. Then, the persuasion knowledge model will be explained. Afterwards, celebrity
endorsements will be discussed and insights gained from this discussion will be applied to the phenomenon of social media endorsements by social influencers. Next, research about
clandestine advertising will be reviewed. Finally, the role of authenticity in endorsement is discussed followed up by the research gap.
2.1 Online advertising
Online advertising is advertising that is delivered over the Internet in many different forms. We can view this sort of advertising, according to Schumpeter (1942) as a gale of creative destruction. Creative destruction is the revolutionizing of the economic structure from within, incessantly destroying the old one, incessantly creating a new one (Schumpeter, 1942, p. 82-85). Traditional advertising sources, for example newspapers, have lost many users and readers to web media advertising sources. All aspects of the global advertising industry are disrupted by online advertising (Evans, 2009).
The effectiveness of online advertising is divided into two paradigms; direct response and brand building. Direct response means that consumers click on the advertisement when they notice it and proceed to the website of the product or service. Brand building means that consumers see the online advertising, which creates brand awareness, and that results in brand building. These two paradigms are not contradicting but complementary, and the two can be used to build brands and increase the probability of a purchase. Whether the advertising results in direct response or brand building depends on the immediate need for a product or service (Hollis, 2005). Online advertising is proven as effective, but generally refers to
12 but I research advertising in the context of social media. Instagram posts by social influencers are used, and according to Hollis (2005) this sort of advertising belongs to the brand building paradigm. Instagram is the most popular photo and video sharing application. Users can capture and share their life with friends and followers by uploading edited pictures and videos (Hu, Manikonda & Kambhampati, 2014). Until 2014, Instagram was not used as an
advertising platform, but companies found out that youth and adolescents are highly engaged with social media applications. Besides that, Instagram provides usage and engagement metrics that are extremely useful for brands (Barry et al., 2016). Instagram is used in this study, because of the popularity of the platform and the easiness of using clandestine advertising in the posts. Besides that, social influencers mostly influence via Instagram and YouTube.
2.2 Social media
With the rise of social media platforms like Facebook, Instagram and YouTube, it is now possible for one person to communicate with thousands of other people about companies and their products. Social media enable companies to talk and interact directly to their
customers, but social media also ensures that the content, timing and frequency of the conversations are outside managers’ control. Thus, for companies there are advantages of social media; direct contact with customers, but also disadvantages; the decrease of control. Besides this, it is often difficult for companies to be active on social media. They recognize the need of it, but do not truly understand how to use social media in an effective way (Hanna, Rohm & Crittenden, 2011). Social media have transformed the Internet from an informational platform to an influential platform. Many companies therefore view social media as a
mandatory element of their marketing strategy now. Only incorporating social media is not enough and companies have to consider both traditional and digital media. All elements of a
13 marketing strategy have to work together to a common objective to make the marketing mix work. Therefore, it is important for companies to understand the need to communicate a common objective, but more specific instructions are needed to lead companies in their decisions. A more specific approach of using social media for marketing purposes is the following one (Culnan, McHugh & Zubillaga, 2010). Companies are only able to gain full business value from social media when they develop strategies based on an approach with three elements: mindful adoption, community building, and absorptive capacity. Mindful adoption means weighing the expected business value with the risk before deciding to proceed. On social media mistakes can spread very fast, which can harm the brand image. Community building on social media means building a brand community with customers who identify with the community and stay involved (e.g. by following the brand on social media platforms). The greater the involvement of an individual with the community, the more likely the individual is to contribute like sharing posts or creating content about the brand by
themselves. Absorptive capacity means that companies need to recognize and acquire new knowledge and exploit the knowledge that is provided by their customers. For example, through commenting, liking and sharing consumers are able to express their opinions directly to companies. Companies can use these insights to adapt their marketing strategies.
Companies could use these three elements, but the recommendations for using social media remain vague, which results in huge mistakes in decisions. Companies need clear rules and instructions that they can follow up, to make their online marketing work.
2.3 Persuasion Knowledge Model
Consumers are the target of marketeers’ sales presentations and advertising. They have to develop knowledge to handle these persuasion tactics. The Persuasion Knowledge Model explains how people’s persuasion knowledge influences their responses to persuasion
14 attempts (Friestad & Wright, 1994). People learn about persuasion throughout their lives. For example, through interacting with friends and family, but also through receiving many
persuasive messages with marketing purposes. Logically, the effect of advertising on people’s attitude and behavior changes, because of the persuasion knowledge that develops constantly. The Persuasion Knowledge Model consists of two concepts, namely conceptual persuasion knowledge and evaluative persuasion knowledge. Conceptual persuasion knowledge refers to the recognition of a persuasion attempt. For example, in this study, the recognition of a sponsorship in an Instagram post. As soon as the persuasion is recognized as advertising the evaluative persuasion knowledge would be activated. The evaluative persuasion knowledge is used to judge the honesty of the content which can result in affection or criticism of the content by the consumer.
2.4 Celebrity endorsement
Celebrity endorsement refers to the use of celebrities in advertising as part of the marketing communication strategy. The goal of using celebrities is to support brand imagery. Firms believe that the qualities of celebrities are transferred to the brand and that this results in desirable campaign outcomes (Erdogan, 1999).
To predict the effect of social influencer endorsements, only literature of celebrity endorsement is used. The reason for this is that there is not much literature written about the influence of social influencer endorsement. After the literature review about celebrity endorsements, predictions are made for the effects of social influencer endorsements. The expectations are that in general, there will be more similarities between the effects of social influencer endorsers and celebrity endorsers, than there will be contradictions. The biggest contradiction that is expected is that consumers accept more dishonesty from celebrity endorsers rather than from social influencer endorsers, because of the experienced distance
15 between them and the celebrity.
Celebrity endorsements can result in more favorable advertising ratings and product evaluations (Dean & Biswas, 2001). In addition, it should be the attractiveness and credibility of the celebrity that make the endorsement work (McCracken, 1989). Celebrity endorsement is also often used to build Customer Based Brand Equity. Customer Based Brand Equity is defined as “the differential effect that brand knowledge has on customer response to brand marketing activity” (Keller, 2016, p. 3). Celebrity endorsement is a powerful mechanism for managing brand equity (Till, 1998). It is mostly used to create positive associations and it belongs to secondary brand elements. Secondary brand elements, like country of origin, events and endorsers, influence brand equity and vice versa (Keller, 2003). Most celebrities are famous as a result of their talents and their talents probably result in their positive image, and this positive image then positively influences the brand.
On the other hand, celebrity endorsement can also result in negative effects. When there is negative information available about a celebrity, for example the Tiger Woods scandal, this will result in a decline in attitude toward the endorsed brand (Till & Shimp, 1998). This means that the image of the endorser influences the attitude toward the brand. Therefore, brands have to be selective when choosing an endorser. According to the Persuasion Knowledge model (Friestad & Wright, 1994), consumers recognize celebrity endorsement as a persuasion tactic. This is the result of recognizing the persuasion (conceptual persuasion knowledge) and evaluate the honesty of the message (evaluative persuasion knowledge). Evaluative persuasion knowledge possibly influences the attitude toward the brand or endorser negatively (Boerman & Reijmersdal, 2016). Also, when
consumers assume that celebrity endorsers are being motivated by money, their evaluations of the brand become more negative (Bergkvist, Hjalmarson & Magi, 2016). This is probably the case because of the perceived dishonesty of the celebrity. However, a recent study concluded
16 that although the persuasion knowledge model states that consumers learn to recognize
persuasion, so they recognize the persuasion attempt of celebrity endorsement, this is not the case in the context of social media (Boerman, Willemse & Van Der Aa 2017).
In this study the research has been done in the context of Instagram. Even though there can be debated about the persuasion knowledge, we could expect in general, according to these articles, that both celebrity endorsements as well as social influencer endorsements result in favorable advertising ratings and brand associations, when they are attractive, credible and have a positive image. However as already mentioned in the introduction, there could be doubt about the positive image of social influencers, obtained by their specific talents. Social influencers are not famous as a result of for example writing, singing, acting or sports. With celebrities this image arises from their talents that are, most of the time, very obvious. Furthermore, as already mentioned, brand evaluations decrease when consumers assume that celebrities are only promoting a product because of the amount of money that they earn with it (Bergkvist, Hjalmarson & Magi, 2016). Our expectations are that with social influencers this effect is even stronger, because consumers consider them as their peers, that they can trust. Posts with money as the main purpose will result in strong disappointment in the social influencer.
Besides celebrity image, fit is also an important determinant of the impact that
celebrity endorsement has on brand image (Kamins & Gupta, 1994). The match-up hypothesis states that proper fit between the celebrity and the product results in a positive effect on product attitude. In this study, the effect of clandestine advertising in social influencer posts on product attitude is researched. I tried to keep image neutral or positive and I have chosen for fit between the social influencer and the product. The celebrity endorser and the social influencer are both attractive women that are promoting a make-up brand in the Instagram post.
17 2.5 Clandestine advertising
Social influencers often make use of clandestine advertising. They are sponsored by a brand, but do not mention the sponsorship in their posts. The reason for this is that the brand and the influencer hope that the consumer thinks that the influencer really likes the product. It is difficult for consumers to distinguish whether the influencer honestly likes the product or pretend that he or she does. The reason for this is that when influencers are not yet popular, they show products that they honestly like, buy and use, and they honestly tell their followers what they think about the product. However, when their number of followers increase, the influencers become attractive for firms, because via social influencers firms can easily reach their target group. In this way, the lines between actually liking a product and pretend liking a product by influencers are blurred. On social media, the social influencers mostly share their personal life. They show their clothes, their favorite spots for dinner, their furniture etcetera. Companies pay social influencers to promote the company’s product in their Instagram posts or YouTube vlogs. So, in many cases, it is not clear for the consumer if the social influencer actually likes the product or if the social influencer is sponsored to pretend to like the product. Earlier research focused on the effect of sponsorship disclosure on television, on brand memory and product attitude (Boerman, Reijmersdal & Neijens, 2012). The results show that a sponsorship disclosure of six seconds activates attitudinal persuasion knowledge, increases brand memory, but decreases favorable brand attitudes. A following study concluded that a sponsorship disclosure results in critical thinking through the activation of the evaluative persuasion knowledge. So, consumers have to recognize the sponsorship through the
sponsorship disclosure, and this results in critical thinking because the consumers are aware of the persuasion. However, according to Evans et al. (2017), transparency of sponsorships results in more positive attitudes. Consumers are nowadays familiar with clandestine advertising and recognize it, and therefore value transparency of sponsorships.
18 Findings of previous literature about clear or clandestine advertising are divergent. According to the persuasion knowledge model, consumers develop defense against
advertising as soon as they notice a persuasive message. We have to take into account that the persuasion knowledge model is developed in 1994, and the recent study of Evans et al. (2017) states that transparency of sponsorships does results in more positive attitudes of consumers. I go along with this conclusion and also expect that clear advertising results in a more positive attitude of the consumer, than clandestine advertising. Consumers are confronted with
clandestine advertising more than ever through the increased use of social media, and I expect that most of the time they notice the clandestine advertising.
Hypotheses 1: Clear advertising in social influencer posts has a more positive effect on product attitude than clandestine advertising.
In this study, research about celebrity endorsement is used to make predictions, but it is possible that the effects of clandestine advertising in combination with celebrity
endorsements are different than the effects in combination with social influencer
endorsement, because of the reason that consumers feel more distance between them and a celebrity. Celebrities have always been ‘media figures’ and perhaps consumers accept their thoughtful interactions with companies. Social influencers started filming their daily lives and consumers could follow them from the start, when the social influencers were not famous at all. They recommend products to viewers, without a sponsorship with a brand, or in contrary discourage products that they do not like. Only when their Instagram followers and YouTube followers started to rise, did they receive attention from media and companies. Consumers consider social influencers more as their peers or equals and trust them more, also because of earlier honest recommendations. This trust in the social influencer and authenticity of the
19 social influencer can result in stronger disappointment in case of clandestine advertising, in comparison with celebrity endorsers. Therefore, the second hypothesis of this study is conducted.
Hypothesis 2: Clandestine advertising in social influencer posts has a more negative effect on product attitude, than clandestine advertising in celebrity posts.
When social influencer endorsers make use of clear advertising, followers would appreciate this honesty and this trust will lead to more credibility, and credibility results, according to McCracken (1989), in working celebrity endorsement. I assumed that when social influencers make use of clandestine advertising, their authenticity would be harmed and therefore consumers evaluate the sponsored product less positive, than when they mention the sponsorship clear in their posts. When celebrity endorsers use clandestine advertising, I expect that this would not harm their authenticity, because of the distance and ‘media figure’ image, and thus the product attitude stays the same. Because of these expectations, celebrity endorsement will function as the control condition in this study.
2.6 Authenticity
Media and social sciences often consider people who express themselves
spontaneously as more authentic than others, and therefore as a potential source of truth on which to base judgements and actions (Van Leeuwen, 2001). An impulsive answer or expression is seen as more valuable and truthful. For brand authenticity it is quality commitment, sincerity and heritage that correspond with a higher order brand authenticity construct (Napoli et al., 2014). Brand authenticity could also be operationalized as consisting of three factors: brand consistency, brand customer orientation and brand congruency. Brand consistency and congruency foster brand trust (Eggers et al., 2013). This indicates that
20 authenticity result in trust. We can relate this knowledge to social influencers, because social influencers often build their own brands, with a team behind them. Another definition of authenticity is “the unobstructed operation of one’s true- or core-self in one’s daily enterprise” (Kernis & Goldman, 2006, p. 294). The authors developed the Authenticity Inventory to measure the authenticity of individuals. The Authenticity Inventory features four components: awareness, unbiased processing, behavior, and relational orientation. Awareness means understanding your feelings, motives and desires. So, being aware of who you are and what you want. Unbiased processing means being aware of the strengths and weaknesses of oneself. Behavior refers to acting like the real self without acting fake to acquire rewards or avoid punishment. The last concept is relational orientation what refers to real and close relationships with others. We could use the measurement of authenticity of individuals to predict the authenticity of the social influencer, but we could also use brand authenticity, since social influencers could also be viewed as a brand on itself. To determine brand authenticity, there are five dimensions to look at (Morhart et al., 2015). The first one is continuity, what stands for the stability of the brand over time. The second dimension is credibility. Credibility refers to the transparency and honesty of the brand. Thirdly, for high perceived brand authenticity brands have to be integer and sincere. Besides the economic agenda, brands have to be led by people who are motivated by deeply held values. Lastly, authenticity is determined by symbolism. It is important that consumers can easily identify who the brand is and what it stands for.
When we look at the context of social media, the need for popularity is a significant predictor for online authenticity. Bloggers with a higher need for popularity have lower ‘sense of real me’, and lower ‘expression of real me’ than those with a lower need for popularity. Gender and age were both negatively correlated with the need for popularity. Young bloggers have a stronger need for popularity than older bloggers, and male bloggers have a stronger
21 need for popularity than female bloggers. The bloggers with a lower need for popularity are thus more authentic than the ones with a higher need for popularity. In this study gender and age are kept as similar as possible.
In this study, I expect that authenticity has a mediating role in the effect of the sort of advertising on product attitude. When clear advertising is used, the expectations are that consumers value the honesty of the social influencer or celebrity, what result in more perceived authenticity. The perceived authenticity of the social influencer or celebrity has then a positive effect on product attitude.
Hypothesis 3: Authenticity has a mediating role in the effect of advertising on product
attitude, and this role is moderated by sort endorser.
The mediating role of authenticity is expected to be different in the case of a celebrity endorser than in the case of a social influencer endorser. The distance between a celebrity and a consumer is bigger than the distance between a social influencer and a consumer. Social influencers are actually ‘normal people’ showing their everyday lives, and consumers could assume them as their peers. Celebrities are famous as a result of their talents and are often well-known media figures. Therefore, clandestine advertisement is expected to be more accepted in the case of a celebrity endorser than in the case of a social influencer endorser. Thus, the effect of clandestine advertising on perceived authenticity of the endorser is expected to be more negative in case of a social influencer than in case of a celebrity.
2.7 Research gap
As mentioned before, the lines between actually liking a product and pretend liking a product are blurred. At the moment, there is no existing research that examines what the effect
22 is of clandestine advertising in social media on product attitude and what the role of
authenticity plays in this process. Since social media became a popular marketing tool for companies, the blurred lines became a more important topic. I expect that consumers often notice that social influencers are sponsored by brands despite clandestine advertising, and that they are more disappointed in that, than when celebrity endorsers use clandestine advertising. Not only for companies is it relevant to know the effect of clandestine advertising and the mediation of authenticity of the social influencer. Also for influencers themselves it is important what the effect is of clandestine and clear advertising, because many of them want to stay authentic. They often do not want to be dishonest in relation to their true and faithful followers. Besides that, social media grows so fast, and many influencers are rising up. When consumers do not trust an influencer anymore, the influencer is easy offside. Social
influencers are easily replaceable, and with the fact that many of them are professional social influencer, they will profit from the outcomes of this study.
23
3 Method
The current research is examined by an online experiment. A 2 (social influencer vs. celebrity) x 2 (clandestine sponsorship vs. clear sponsorship) between-subjects experimental
design is used. Celebrity and clear sponsorship are both used as control condition. Every
participant is randomly assigned to one of the two conditions. In total, 320 respondents participated in this study, but 14 respondents were excluded from the analyses because they did not finish the experiment. 306 (64,4% female) respondents finished this study with an average age of 25 (M = 25.16, SD = 3.60).
3.1 Procedure
Participants between 25 and 35 years, from the United States and the United Kingdom, were recruited through Prolific, a website that connect researchers with participants. This age range is chosen, because Instagram is mostly used by this age group. Besides that, consumers older than 35 years are not always familiar with social influencers. These specific countries are chosen, because I wanted to prevent that the English questionnaire would not be well understood. The experiment consisted of the following components. Firstly, the participants were asked to sign an informed consent. After that, they were asked to fill in some
demographical questions. Then, the participants were told that they would see an Instagram post in which a woman shows off her new make-up. Also, they were informed that the post would remain on the screen for 30 seconds, to ensure exposure to the stimuli, and were asked to take a careful look at the Instagram post that would follow. The participants were randomly assigned to one of the four conditions (social influencer x clear advertising, social influencer x clandestine advertising, control condition x clear advertising, control condition x clandestine advertising). After the exposure of the Instagram post, participants continued to the
24
measurement scale. After that, ‘authenticity’ of the endorser was measured. This variable is measured, because in this study there is expected that type of advertising (clandestine versus clear) has an effect on perceived authenticity of the endorser. Expectations are that
clandestine advertising results in a more negative endorser authenticity, and this effect is expected to be stronger in case of a social influencer endorser. Then, the variable ‘envy’ is measured, because there is suspected that when authenticity does not have a mediating effect, envy could be an alternative cause of the effect of type of advertising on product attitude. Finally, a few questions were asked about the manipulation material; make-up brand names, the names of the endorsers, whether the endorser was a celebrity, and whether the endorser was sponsored by a brand or not. In the end the participants were debriefed and thanked for participating in the experiment.
3.2 Manipulation
Four Instagram posts were developed for the experiment; a social influencer post with clear advertising, a social influencer post with clandestine advertising, a celebrity post with clear advertising and a celebrity post with clandestine advertising. From both the social influencer and the celebrity I used an existing post. The social influencer and the celebrity have the same appearance and are both holding a make-up product with the brand on it. Also, I kept the position of the social influencer or the endorser, and the number of likes as similar as possible. In the social influencer condition the brand name was ‘Narsissist’, and in the control condition the brand name was ‘Honest Beauty’. In the clear advertising condition the sponsorship is disclosed by the sentence ‘Paid partnership with..’, similar to the current disclosing practices on Instagram (Instagram, 2018). Besides that, in the clear advertising condition the original post description is used, with the brand name of the make-up product and the hashtag #ad. In the clandestine advertising condition a different text description is used, without the brand name of the make-up product and without the hashtag #ad. The
25
stimuli can be found in the appendices (Appendix A, Appendix B).
4 Results
4.1 Measures
Out of the results of the normality of variables and reliability of the scales and the correlations between the variables it can be concluded that all items were normally distributed (Field, 2013).
The dependent variable ‘product attitude’ is measured using a six-item semantic differential (Perkins & Forehand, 2011), seven-point bipolar Likert scale (‘What do you think about the make-up that is used in the picture? Bad/good; negative/positive; useless/useful, worthless/valuable; unpleasant/pleasant’). The product attitude scale has high reliability, with Cronbach’s Alpha = .940. The corrected item-total correlations indicate that all the items have a good correlation with the total score of the scale (all above .30). Also, none of the items would substantially affect reliability if they were deleted.
The variable ‘authenticity’ is measured using a four-item semantic differential (Ilicic & Webster, 2016) seven-point bipolar Likert scale (‘Please answer the following questions about the person in the picture; Tries to act in a manner that is consistent with her held values, even if others criticize or reject her for doing so; Cares about openness and honesty in close relationships with others; In general, places a good deal of importance on people
understanding who she truly is; People can count on her being who she is regardless of the situation’). The authors composed a measurement scale with twenty-one questions, but the factor analysis in their study indicated that four questions were strongly measuring
authenticity. I did the same factor analysis and came to the same conclusion as Ilicic and Webster (2016). Therefore, four of the twenty-one questions are used to compute the variable
26 ‘authenticity’. The authenticity scale has high reliability, with Cronbach’s Alpha = .822. The corrected item-total correlations indicate that all the items have a good correlation with the total score of the scale (all above .30). Also, none of the items would substantially affect reliability if they were deleted.
The dependent variable ‘envy’ is measured using an eight-item semantic differential
(Smith et al., 1999), seven-point bipolar Likert scale (‘Please answer the following questions about yourself: I feel envy every day; The bitter truth is that I generally feel inferior to others; Feelings of envy constantly torrent me; It is so frustrating to see some people succeed so easily; No matter what I do envy always plagues me; I am troubled by feelings of inadequacy; It somehow doesn’t seem fair that some people seem to have all the talent; Frankly the
success of my neighbors makes me resent them’). The envy scale has high reliability, with Cronbach’s Alpha = .902. The corrected item-total correlations indicate that all the items have a good correlation with the total score of the scale (all above .30). Also, none of the items would substantially affect reliability if they were deleted.
Product attitude - There was a positive main effect of endorser on how positive participants’ attitude was toward the product (F(1, 305) = 5.13, p = 0.024), but no main effect of partnership (F(1, 305) = 3.67, p = .056) and no interaction between endorser and
partnership (F(1, 305) = 0.16, p = .689). Figure 1 shows that when participants saw a celebrity (Jessica Alba) promoting the make-up product with clear advertising, the attitude toward the product was more positive (M = 4.99, SD = 1.15) than with promoting the make-up product with clandestine advertising, but this effect was not significant (M = 4.69, SD = 1.10, est = 0.29, 95% CI [-0.06; 0.65], t(302) = 1.6, d = 0.27, p = .104). Also in the social influencer condition (Negin Mirsalehi), the attitude toward the product was more positive when clear advertising was used (M = 5.22, SD = 1.07), than when clandestine advertising was used, but this effect was not significant either (M = 5.03, SD = 1.12, est = 0.19, 95% CI [-.16; .01],
27 t(302) = 1.08, d = 0.17, p = 0.282). The results confirm that clear advertising has a positive effect on product attitude in comparison with clandestine advertising (H1), but these results are not significant. The results do not support the hypothesis that clandestine advertising in the social influencer post has a stronger negative effect on product attitude, than clandestine advertising in the celebrity post (H2).
Figure 1: Means of sort endorser + sort advertising
4.2 Moderated mediation - Authenticity
There is tested whether the effect of type of advertising on product attitude is mediated by authenticity and whether the mediation depends on the type of endorser (Figure 2). In other words, I test for moderated mediation.
4.69
4.99 5.03
5.22
Jessica clandestine advertising
Jessica clear advertising Negin clandestine advertising
Negin clear advertising
Product attitude
28
Figure 2: moderated mediation model
The expectation is that advertising (X) has both a direct effect on product attitude (Y), and an indirect effect through authenticity (M). In other words, the effect of advertising on product attitude will be mediated by authenticity. For example, when clandestine advertising is used, this could harm the authenticity of the social influencer or endorser, and therefore the product attitude of the consumer could be lower. The indirect effect, as well as the direct effect are expected to be moderated by endorsement. For example, when the advertising is clandestine, the effect of that advertising on authenticity depends on the sort of influencer. I expect that in the case of the social influencer as endorser, the effect of a clandestine advertising on
authenticity is more negative than when a celebrity is used as endorser. The reason for this is that celebrities have always been media figures and are more related to money driven actions than social influencers. Social influencers are often started without earning any money for what they do, but from the moment that their followers increase and they become more popular, they earn money with promoting products. As mentioned in the literature review, consumers are overwhelmed with advertisements and many of them will notice clandestine advertisement, even when it is not mentioned (Evans et al., 2017). Therefore, they will be more disappointed when the social influencer, where consumers can identify themselves more easily with, uses clandestine advertising, than when a celebrity uses clandestine advertising.
29 In case of a clear advertising, consumers will be not disappointed in one of the endorsers, which results in an overall more positive effect of clear advertising than clandestine
advertising. Thus, the expectations are that the negative effect of clandestine advertising on authenticity is smaller in the case of a celebrity than in the case of a social influencer.
The direct effect of X on Y is also expected to be moderated by sort endorser. The use of a different sort of endorser can result in a different product attitude of the consumer. The results are very useful for companies that are shifting from celebrity endorsers to social influencer endorsers. In this way, we can conclude which type of endorser is more effective to use when increasing the product attitude is the main purpose.
Firstly, the mediating effect of authenticity (M) between advertising (X) and product attitude (Y) is measured, with the moderation of endorser (W). The three steps of Baron and Kenny (1986) are taken to research this. The first step is to look at the effect of sort
advertising (X) on authenticity (M), a in Figure 3. So, there tested whether the variations of advertising result in variations of authenticity.
Figure 3: Effect X on M (a)
30 The parameters estimates table with intercept (B) = 4.14 showed that the effect of
partnership on the mediator authenticity was a = 0.23 for celebrity (Jessica) and a = -0.01 for social influencer (Negin). Thus, the effect of advertising on authenticity is 0.23 higher in case of using a celebrity as endorser than using a social influencer as endorser.
The second step is to look at the effect of M on Y, b in Figure 4. So, there is looked if the variations of authenticity result in variations in product attitude. The parameters estimates table showed that the effect of authenticity (M) on product attitude (Y) is b = 0,531, p < 0.001.
Figure 4: Effect M on Y (b)
The third step is to look at the effect of advertising (X) on product attitude (Y), taken into account the moderated effect of sort endorser, in Figure 5 (c).
The parameters estimates table showed that the direct positive effect of using Jessica as endorser is smaller (c = 0.17) than the direct positive effect of using Negin as endorser (c = 0.19).
31
Figure 5: X on Y (c)
To summarize the previous findings, two figures are made to clarify the effects of using Jessica (celebrity) as endorser and using Negin (social influencer) as endorser.
32
Figure 7: total social influencer endorser
Now, to establish mediation the total effect of X on Y is measured by the following formula (Zhao et al., 2010): c’ = (a x b) + c. With this formula the sum of the indirect path (a x b) and the direct path (c) is given.
With W (celebrity endorser) c’ = (0.23 x 0.53) + 0.17, c’ = 0.29
With W (social influencer endorser) c’ = (-0.01 x 0.53) + 0.19, c’ = 0.19 The indirect path (a x b) with W (celebrity endorser) = 0.122
The indirect path (a x b) with W (social influencer endorser) = -0.005 The indirect path celebrity endorser – indirect path social influencer = 0.127
The effect of a is bigger for Jessica (a = 0.23) than for Negin (a = -0.01). This means that participants found Jessica more authentic than Negin, but this difference is not significant (p = 0.164). The direct path c is bigger for Negin (c = 0.19) than for Jessica (c = 0.17). This means that Negin has a more positive effect on product attitude than Jessica, but this difference is not significant (p = 0.282).
33
4.3 Moderated mediation analysis with PROCESS – Authenticity
PROCESS is used to test if the differences in results are significant. For the moderated mediation analysis in PROCESS, model 8 is used:
The interaction term (Appendix C, Table 1, Column 1, W*X) means how much the effect of partnership on authenticity differs between celebrity and social influencer. In this study this effect is not significant, and negative, meaning that the negative effect of partnership on authenticity does not depends on the sort endorser.
The interaction term (Appendix C, Table 1, Column 2, W*X) means how much the effect of partnership on product attitude is different between celebrity and social influencer. In this study this effect is not significant, and positive, meaning that the effect of partnership on product attitude does not depend on the type of endorser. Thus, the results indicate that the effect of sort advertising on product attitude is not conditioned on the sort endorser, since the interaction is not statistically significant (c’4=0.03, p = 0.901). A closer inspection of the conditional effects indicates both direct and indirect effect not to be conditioned by sort endorser. Firstly, the direct relationship between sort advertising and the product attitude
34 appears as not significant within both the social influencer (effect= -0.199, SE = 0.16, LLCI= -0.506; ULCI=0.110), and the celebrity group (effect= -0.172, SE = 0.159, LLCI= -0.485; ULCI=0.142). Secondly, the indirect relationship between sort advertising and the product attitude via authenticity also appears as not significant within both the social influencer (effect= 0.006, SE = 0.09, LLCI= 0.178; ULCI=0.176), and the celebrity group (effect= -0.129, SE = 0.08, LLCI= -0.286; ULCI=0.042). In other words, advertising does not affect product attitude by differences in endorser, what does not affect authenticity.
5 Conclusion
In this study, three hypotheses were tested. In the first hypothesis I expected that clear advertising has a more positive effect on product attitude, in comparison with clandestine advertising. The reason for this was that consumers experience clandestine advertising via social media more and more while at the same time they increasingly care about authenticity of a brand or endorser. There was a more positive effect of clear advertising than of clandestine advertising on product attitude, but this effect was not significant. The lack of significance is possibly due to the absence of clarity of the given sponsorship in the manipulation material. The sentence ‘Paid partnership with…’ is not very clear and does not stand out (see Appendix B). However, this is the actual option of Instagram to indicate the sponsorships and I tried to make this study as applicable as possible to the real world.
In the second hypothesis I expected that clandestine advertising in the social media post has a stronger negative effect on product attitude, than clandestine advertising in the celebrity post. The reason for this was that the distance between consumers and celebrities is bigger than the distance between consumers and social influencers, since social influencers are just ‘normal people’ like consumers themselves. Sometimes people follow social influencers from the
35 beginning, in their non-professional period, and follow them for years. For these people, social influencers feel like a friend or a peer. Therefore, I expected that the disappointment of noticing a clandestine advertising in a social media post was bigger than in a celebrity post. In contrast, the difference between effects of clandestine advertising and clear advertising on product attitude was bigger in the celebrity endorser condition. However, the dissimilarity between the differences in effect between social influencer endorser and celebrity endorser were not statistically significant.
In the third hypothesis I expected that authenticity has a mediating role in the effect of advertising on product attitude, and that this role is moderated by sort endorser. The reason for this was that the expected negative effect of clandestine advertising, or the expected positive effect of clear advertising on product attitude would be caused by the mediation of authenticity. Thus, it was expected that sort advertising affects authenticity, and authenticity affects product attitude. This effect was expected to be different in case of different endorsers. The expected negative effect of clandestine advertising on authenticity was expected to be stronger in case of the social influencer than in case of the celebrity endorser. The tested mediating effect of authenticity between sort advertising and product attitude was not significant, and thereby the moderating effect of sort endorser also not. The reason for the lack of mediation could be that there is another variable that mediates the effect of sort advertising on product attitude. The possible mediator could be ‘envy’. Envy can be defined as: ‘A desire for something that another person has.’ (Varian, 1974). With celebrities and social influencers, it is possible that envy affect consumers’ attitude toward the product. For example, when clear advertising is used, consumers could think something like “I would also have all that products or clothes or go on holiday that often.” When clear advertising is used, consumers know the perfect pictures are just for their job and are not a representation of their normal life. Therefore, I expect that when clear advertising is used, consumers experience less
36 envy and less envy result in a higher product attitude. The expectation is that there will be more feelings of envy in case of a social influencer endorser than in case of a celebrity endorser. Consumers have probably the feeling that social influencers’ lives are more
attainable than celebrities’ lives. They possibly assume that social influencers are famous out of nowhere, and celebrities have passed a long way to be where they are. It may be feels more reachable and therefore I expect that consumers desire the lives of social influencers more than the lives of celebrities.
5.1 Moderated mediation – Envy
I test whether the effect of type of advertising on product attitude is mediated by envy, in other words I test for mediation (Figure 8).
Figure 8: The mediating effect of envy between sort advertising and product attitude
Firstly, the mediating effect of envy (M) between advertising (X) and product attitude (Y) is measured. The three steps of Baron and Kenny (1986) are taken to research this. The first step is to look at the effect of advertising (X) on envy (M), a in Figure 9. So, there is looked if the variations of advertising result in variations of Envy.
37
Figure 9: Effect X on M (a)
The parameters estimates table with intercept (B) = 2.87 showed that the effect of partnership on the mediator envy was a = 0.06 for Jessica and a = 0.26 for Negin. Thus, the effect of advertising on envy is 0.20 higher in case of using Negin (social influencer as endorser than using Jessica (celebrity) as endorser.
The second step is to look at the effect of M on Y, b in Figure 10. So, there is looked if the variations of envy result in variations in product attitude.
Figure 10: Effect envy (M) on product attitude (Y)
38 The parameters estimates table showed that the effect of envy (M) on product attitude (Y) is b = 0,032, p = 0.531. But this effect is not significant.
The third step is to look at the effect of advertising (X) on product attitude (Y), taken into account the moderated effect of sort endorser, in Figure 11 (c).
Figure 11: X on Y (c)
This effect stays the same for authenticity as for envy. The parameters estimates table showed that the direct positive effect of using Jessica as endorser is smaller (c = 0.17) than the direct positive effect of using Negin as endorser (c = 0.19).
To summarize the previous findings, two figures are made to clarify the effects of using Jessica (celebrity) as endorser and using Negin (social influencer) as endorser.
39
Figure 12: total celebrity endorser
Figure 13: total social influencer endorser
Now, to establish mediation the total effect of X on Y is measured by the following formula (Zhao et al., 2010): c’ = (a x b) + c.
With W (celebrity endorser) c’ = (0.06 x 0.03) + 0.17, c’ = 0.17
With W (social influencer endorser) c’ = (0.26 x 0.03) + 0.19, c’ = 0.20 The indirect path (a x b) with W (celebrity endorser) = 0.002
The indirect path (a x b) with W (social influencer endorser) = 0.008
The indirect path social influencer - indirect path celebrity endorser = 0.006
40 participants perceived more feelings of envy in the social influencer (Negin) condition than in the celebrity condition (Jessica), but this effect is not significant (p = 0.201) The direct path c is bigger for Negin (c = 0.19) than for Jessica (c = 0.17). This means that Negin has a more positive effect on product attitude than Jessica, but this effect is not significant (p = 0.282).
5.2 Moderated mediation analysis with PROCESS – Envy
PROCESS is used to test if the differences in results are significant. For the moderated mediation analysis in PROCESS, model 8 is used.
The results indicate that the effect of sort advertising on product attitude is not
conditioned on the sort endorser, since the interaction is not statistically significant (c’4=0.03, p = 0.901). A closer inspection of the conditional effects indicates both direct and indirect effect not to be conditioned by sort endorser (Appendix C, table 2). Firstly, the direct
relationship between sort advertising and the product attitude appears as not significant within both the social influencer (effect= -0.184, SE = 0.18, LLCI= -0.537; ULCI=0.168), and the celebrity group (effect= -0.293, SE = 0.181, LLCI= -0.649; ULCI=0.064). The indirect relationship between sort advertising and product attitude via envy appears as not significant within both the social influencer (effect= -0.008, SE = 0.02, LLCI= -0.060; ULCI=0.020), and the celebrity group (effect= -0.002, SE = 0.01, LLCI= -0.035; ULCI=0.020). In other words, clear advertising does not affect product attitude by differences in endorser, what does not affect envy.
41
6 Discussion
The purpose of this study was to examine the effect of clandestine advertising in social media posts on product attitude. Also, the possible cause for this effect, perceived authenticity of the endorser, is researched. Clandestine advertising on social media is very common
nowadays. The question was whether consumers gradually start to notice this kind of persuasive attempts, and whether this increased awareness influences the perceived
authenticity of the endorser, and in turn product attitude. Expectations were that this possible effect of clandestine advertising on authenticity of the endorser differs for sort endorser. The distance from a consumer to a celebrity is greater than the distance from a consumer to a social influencer. Celebrities have always been media figures and does not feel as friends or peers. However, social influencers are just ‘normal people’, started blogging or vlogging to share their lives. Consumers followed their path from being totally unknown till being well know or sometimes famous. Therefore, the expectations were that clandestine advertising causes more disappointment when used by social influencers than used by celebrities. The increased disappointment was expected to result in a lower perceived authenticity of the endorser.
For this study, an experiment is developed with three measurement scales of earlier research to examine the dependent variables. In total, 306 participants finished the experiment and they were randomly assigned to one of the four conditions.
6.1 Discussion of variables and limitations
The results show that in consistency with the expectations, clear advertising causes a more positive product attitude than clandestine advertising. However, this finding was not statistically significant. A possible explanation for the lack of significance is the somewhat
42 unclear and minimalistic sign for clear advertising. In the manipulation material the sentence ‘Paid partnership with…’ is used to indicate the advertisement, because this is the original sentence that Instagram provides to clarify the sponsorship. I wanted to check whether participants saw the indication sentence, so I included the question: ‘In the picture you saw someone endorsing a brand of make-up. Do you think this was a paid partnership or not?’. 88.2% of the participants answered ‘Yes’ to this question, regardless to the clear or
clandestine advertising condition. This is probably due to my expectation that even when clandestine advertising is used, consumers assume social media posts with recommended products as persuasive attempts. So unfortunately, I could not filter participants on the requirement that participants in the clear advertising condition noticed the sponsorship indication, because the measurement would be erroneous. Namely, almost all participants thought that the make-up product was sponsored, regardless of the sponsorship indication. According to the Persuasion Knowledge Model (Friestad & Wright, 1994) consumers are able to the judge the persuasion attempt from the moment that they recognize the attempt. With the majority of participants who thought that the endorser was sponsored by the brand, there could be assumed that most of them were able to judge the Instagram post. Yet, no significant difference could be found between clear and clandestine advertising.
Another explanation from the lack of statistical significance between clear and clandestine advertising on product attitude is that consumers do not care about advertising indication. This would contradict the study of Evans et al. (2017), that states that transparency of sponsorships results in more positive attitudes of the consumer. Maybe because consumers expect already that successful endorsers enter into cooperation’s, they do not judge it.
According to Bergkvist, Hjalmarson and Magi (2016), the evaluations of a brand become more negative when consumers assume that endorsers of the brand are only
43 that endorsers are in some way driven by money. Although, as already mentioned there is no difference found in product attitude between clear and clandestine advertising. It is possible that this is caused by the fit between the endorsers and the products. Both women are attractive, and it seems like they value beauty products and according to the match-up hypothesis proper fit between the endorser and the endorsed product has a positive effect on product attitude (Kamins & Gupta, 1994).
The results show that in contrast to the expectations, the difference of use of
clandestine versus clear advertising has a bigger effect in case of a celebrity endorser (Jessica) than in case of a social influencer (Negin). Thus, the difference in product attitude was more divergent between using clear or clandestine advertising in the celebrity post, than in the social influencer post. Unfortunately, this effect was not significant. A possible explanation for the lack of significance could be that consumers assume celebrities and social influencers both as people that are successful, and feel a great distance between them and both type of endorsers. Also, in this study I used Negin Mirsalehi as social influencer endorser, and with 4.8 million followers on Instagram she can be assumed as a real celebrity. Maybe therefore participants also felt a great distance with her, what resulted in no significant difference between the celebrity endorser and the social influencer endorser. The choice for Negin Mirsalehi and Jessica Alba was taken because of the similarities in looks, hair color, number of followers, number of likes, pose in the picture and the promotion of a make-up product in the post to rule out external factors. On the one hand, this is correct for the reliability of this study. On the other hand, this can have caused that the results do not significantly differ between the celebrity endorser and the social influencer endorser, because of the perceived distance of the participant with both endorsers. This assumption also arises from the data. The question: ‘Do you think the person in the picture is a celebrity?’ is asked to the participants and only 43 out of 306 participants answered ‘The person is not a celebrity’, regardless of the
44 social influencer condition or the celebrity condition.
The results show that there is no statistically significant difference of the moderation of endorser between sort advertising and product attitude. This means that the use of a celebrity endorser or a social influencer endorser does not make a difference in product attitude of the consumer.
The means of the effect of both sort endorsers on product attitude were high (Figure 1). According to McCracken (1989) this is the result of the attractiveness and credibility of the endorser. Both endorsers are therefore probably seen as reasonable attractive and credible. When there is negative information available about an endorser this can influence the brand image in consumers’ heads. The negative information about an endorser is then transferred to the endorsed brand (Till & Shimp, 1998). With these findings we can conclude that of both endorsers have a positive image and therefore have a positive effect on product attitude. The results show that there is no statistically significant difference between both endorsers in the indirect effect of advertising on product attitude via authenticity. This means that participants did not perceive one of the endorsers as more authentic than the other. I expected that in case of clandestine advertising, the authenticity of the social influencer endorser would be more damaged than of the celebrity endorser. The mean of total
authenticity was 4.3 out of 7, so above average. The perceived authenticity of the celebrity endorser was a little bit higher, but this difference was not significant. The lack of
significance could be derived from the fact that both Jessica Alba and Negin Mirsalehi are independent women who fight for their own career. They are brands on themselves and built this up on their own. To achieve that you must have a strong personality and maybe therefore they both score above average on authenticity.
The results show that there is no statistically significant difference between both endorsers in the indirect effect of advertising on product attitude via envy. This means that
