Celebrity vs Influencer: Which Endorser Type has the biggest influence on Post Performance in Tiktok?

(1)

1

Celebrity vs Influencer: Which

Endorser Type has the biggest influence on Post Performance in Tiktok?

Author: Finian Barke-Asuni Student number: 12192333 Date of submission: 30/06/22

Program: BSc. Business Administration Track: Management in the Digital Age Thesis supervisor: dr. R. de Bliek

ABSTRACT

There has been continuing growth in usage of social media networks, hence there has been an explosion in the money firms are willing to spend marketing their. Therefore, this research looked at the relationship between the type of endorser used in a post and the performance of that post. The focus of this research was the Tiktok of the successful athleisure brand Gymshark, which

successfully leveraged social media to grow the brand. To obtain results, a regression and an ANOVA were performed to assess the significance of the relationship. The data for this was collected from the Gymshark Tiktok page, using a social media analytics tool: Tagger. Results indicated that endorser type was not a significant cause for variance in post performance – values of both regressions and ANOVAs indicated as much. This highlighted that the lines between celebrity and influencer are becoming less clear. Hence, there are other factors that need to be assessed as the causes for variance in post performance.

Keywords: Tiktok, Influencers, Celebrities, Post Performance

(2)

2

collaborations can leave an indelible mark on society. There is plenty of research into how the image of these celebrities can successfully transfer to a brand and consequently improve brand image through association. McCraken’s (1989) Meaning Transfer Model (MTM) highlights the typical path of cultural meaning in consumer societies. Further explaining that advertising is a key instrument in transferring cultural meaning from its beginning in a culturally constituted world to the consumer’s life. Celebrities can help to expedite this cultural transfer.

For example, a brand that utilises this effect is the one this essay will be focused on: Gym shark. The Brand was valued at $1.3 billion and was a pioneer in the social media marketing industry (Cook, 2020). Until recently an independent brand, their success would not have

(4)

4

been possible in a world without social media. Brands can now thrive because Social media’s growth has led to the democratization of advertising and instant access to markets. For

Gymshark, TikTok has been a huge growth frontier and has been changing the landscape of advertising. They have utilized endorsements by both traditional celebrities and influencers to spread their message and cause this growth. Hence, it is important to explore the

techniques and posts that have led brands, such as Gymshark, to have such large success.

Although the influence of this area of marketing continues to grow, it is an area still with relatively little research (Schouten et al., 2019). Schouten et al (2019) explore the

effectiveness of traditional celebrity versus influencer marketing on SNS. They suggest that influencers add value in comparison to traditional celebrities, but also argue product-endorser fit was a more significant, despite being a moderating factor. Whilst this is a similar area of research to this one, being written previously it did not explore Tiktok as a platform. It is, however, still useful as a tool to better understand what the significant characteristics between the two categories of person are making the post. Sabate et al(2014) explore how “richness”

and “Time Frame” impact the popularity of a post. Whilst , De Vries et al (2012) use

comments and likes as a tool to measure the success and popularity of posts. They found that vivid and bright posts were key to having a high number of likes and that interactive posts help to increase the number of comments. This pre-existing research is useful and provides theoretical background, however, it only further demonstrates the need for this research.

Firstly, none of the above-mentioned research is focused on TikTok. This is because it is still a relatively new platform, despite its size. The nature of the platform may lead to different results in comparison to Facebook or Instagram, where most research is based. Furthermore, there is a lack of research examining the type of influencer as a feature of a post. For

example, Schouten et al (2020) do focus on this comparison, but more on the characteristics of celebrities or influencers that may lead to advertising effectiveness – which were the

(5)

5

mediators. Contrastingly, this research will focus on endorser type as a characteristic of a post. This enables future comparison against other post characteristics in terms of biggest impact on post performance. In addition, most of this research explores categories primarily relevant to images, which are not available on Tiktok as a platform. This essay will first provide insight into the celebrity endorsement phenomena and how it is effective on social media. Then provide some theoretical background as to how this relationship has been previously explored. The method will then go on to describe how the data was collected and analyzed. Finally, there will be some discussion of the results in order to reach a conclusion and highlight possible areas of future research.

2.0 -Theoretical Framework

2.1- Celebrity and Influencer Endorsement

To establish the theoretical basis for this research it is important to establish how

endorsements and advertising on SNS have become an important area of research. A key aspect of advertising that has been a successful and enduring method to enhance marketing communications has been celebrity endorsements (Bergkvist & Zhou, 2016).

McCraken’s (1989) Meaning Transfer Model (MTM) highlights the typical path of cultural meaning in consumer societies. This is one of the earliest and most significant pieces of research into this area. McCraken first offered criticism of other available models and then responded through the creation of his own. This work is key to understanding the basic principles of how and why celebrity endorsements are successful. He explains that advertising is a key instrument in transferring cultural meaning from its beginning in a culturally constituted world to the consumer’s life. This transfer of meaning involves transference from consumer goods into the consumer’s life. Within this framework, advertising operates as a means to facilitate these transfers of meaning.

(6)

6

In the instance of celebrity endorsements, McCraken (1989) describes celebrities as differing from normal “anonymous” actors in that they convey a deeper and more powerful meaning via their presence. Similarly, to other examples of the MTM, celebrities are tools that can be used to effectively transfer meaning to the consumer products and then to the consumer’s life.

However, these effects can be deeper and more meaningful than what can be provided by other endorsers. For example, a celebrity can offer a clearer association and identification to a given demographic group or lifestyle than a traditional model. On this basis, McCraken (1989) produces the seminal definition of a celebrity endorser: “any individual who enjoys public recognition and who uses this recognition on behalf of a consumer good by appearing with it in an advertisement”.

Whilst this definition was a key one, it does not appear to be fully appropriate for modern conditions (Bergkvist & Zhou, 2016). Bergkvist and Zhou (2016) highlight this through their literature review and research agenda. This research was based around six key areas of

celebrity endorsements. It indicated that there was a need for increased research into celebrity persuasion and understanding of brand transference - especially under the new conditions that the market operates under. This literature review was key to understanding the current

position of research in an ever-changing field. The paradigm created by social media and the advent of the internet have blurred the lines of celebrity endorsement. Firstly, the definition of what an advertisement is changing thanks to social media - as adverts are increasingly becoming integrated into content. Furthermore, some celebrities are developing platforms larger than the brands themselves (Wood & Burkhalter, 2013). Hence, whilst McCraken (1989) offers an effective basis for the theoretical background and provides explanations for why the broad concepts within advertising and celebrity endorsements work but does not explain how the phenomena has evolved as a result of SNS.

(7)

7

The growth of users on SNS has been both rapid and exponential. Kepios reports (2022) indicate that 4.65 billion people currently use some form of SNS and an annualized growth rate of 7.5%. It is, therefore, vital for firms to access this market through advertising and its importance will only grow. Companies have of course taken note of this and heavily invested into the industry. A report produced in collaboration with IAB UK and PWC (2021) indicated that the digital advertising market had experienced 41% growth in that. Hence, studying this form of advertising will only continue to grow in its significance.

2.2- Endorsement and engagement: important factors

In the last decade there has been increasing research into how different SNS have been used as advertising tools. Voorveld et al. (2018) produced a study highlighting how consumers’

engagement with SNS drove engagement with advertising across eight different SNS. The study was based on a survey asking participants about their experiences using SNS. This was important to study as understanding how to engage with consumers on SNS will continue to be a key driving force in successful advertising. They delineate that the reception of an advertisement is dependent upon the context that it is seen in. Hence, the same advert received in a traditional form of advertising will have a different effect to the same advert received via social media.

Similarly, Tuten and Solomon (2020) explain that the customer’s ability to interact and engage with a brand is increased through social media marketing. This is through an educational textbook, which is designed to provide expert insight into social media marketing. They argue for this increase because social media provides customers with greater access to branded information and can interact with other consumers through comment sections and other means. This research can demonstrate some of the potential benefits and reasons for companies engaging in advertising on SNS. In combination with an

(8)

8

understanding of McCraken’s (1989) ideas surrounding the MTM and celebrity endorsements a picture can start to be built of how firms are using SNS as the next frontier to transfer meaning to the consumer.

SNS prevalence led to individuals developing large followings on these platforms. They have become their own kind of celebrity: Influencers. The definition of a social media influencer is a person who has built a large social network of people following them (Hudders et al., 2020). Contrastingly, traditional celebrities have acquired their fame and influence as a result of their personal achievements (sports, media and entertainment) (Harnish, & Bridges, 2016 as cited by Brooks et al., 2021) Their following enables them to sell products like a

traditional celebrity. Ypulse research indicates that in 2020 young consumers’ consumption, the primary users of SNS, are as influenced by influencers as ever (Ypulse, 2020, as cited by Taylor, 2020). As a result, the line between traditional celebrity and influencer is becoming increasingly blurred. Celebrities are now leveraging their social media following to advertise and influencers are beginning to increasingly advertise on more traditional forms of media (Strugatz, 2016; Hobbs, 2015, as cited by Gräve & Bartsch, 2021). Hence, there is no more prescient time than now to compare the success of celebrity and influencer branded posts.

This will enable an understanding of the different conditions through which both are successful and the different characteristics of their endorsements (Gräve & Bartsch, 2021).

2.3 – Other Determinants of Social media Engagement

Although this research is focused on the relationship between endorser type and post

performance, there are other factors that can impact engagement. Sabate et al(2014) explore how “richness” and “Time Frame” impact the popularity of a post. A model, describing the link between a post’s “richness” and “Time Frame” and its likes and comments, was produced. This model was then tested using multiple linear regressions. This research was

(9)

9

important as it investigates how much the features of the post influence engagement. Through this firms can understand the degree to which what, and not who, you post is important. They suggest that there is no correlation between “Time Frame variables” and popularity but that some of the “richness variables” were significant explanatory factors. They stressed the importance of the use of images as being key to the success of a post. Sabate et al (2014), offer interesting conclusions and have a similar post-based focus; their focus is on different characteristics of this research. Their Framework for measuring popularity is similar to the one used in this research but is focused on different features of a post.

Similarly, De Vries et al (2012) use comments and likes as a tool to measure the success and popularity of posts. Their research focused on the characteristics of a post that were drivers for brand post popularity. Data was collected through analysis of 11 different brand pages, across multiple industries. Based on this data OLS regressions were performed, that assessed the different relationships. They found that vivid and bright posts were key to having a high number of likes and that interactive posts help to increase the number of comments. These findings align with those of Sabate et al.(2014) – in that post “richness”, or the particular visual features, of a post were key to its success. This research into alternative factors is key to improving overall knowledge of the social media space, but still highlights the need for focus on endorser type.

2.4 Endorsement and Engagement on emerging platforms: TikTok

Social media, just like an ecosystem, is continually evolving. The next, potentially paradigm shifting social media evolution has been the advent of Tiktok’s entry into the mainstream.

Tiktok describes itself as “A personalised video feed based on what you watch, like and share”. It is a platform that, unlike others such as YouTube, is primarily focused on short form video – no longer than one minute. Whilst other SNS, such as Vine, have focused on

(10)

10

short-form video – no other company has been able to do so to the same massive degree of success. Tiktok’s parent company ByteDance have valued the company at $50 Billion (Wang et al., 2020). The success of TikTok is in no small part down to the uniqueness of its

platform. The landing page for TikTok is its “For-You” page, which uses artificial

intelligence to display videos specified to the user’s tastes. This can lead to users spending longer on the app than the user may have intended (Shao, Z. 2018, as cited by Montag et al., 2021). This being because they are able to get more lost in the experience of using the app.

Furthermore, TikTok can provide instant gratification through the ability to reject or approve of content by either liking or scrolling past it (Montag et al., 2021). All of these factors have led to a high potential for advertising on TikTok. However, this area of research is relatively unexplored. TikTok’s popularity boomed during the Covid-19 period, when people were spending their time in lockdown(Su et al., 2020). Hence, a lot of research is exploring TikTok phenomena with the pandemic being considered as a key factor in that. This research would like to explore advertising on TikTok moving forward - with its already established large user base. There will be discussion on how advertising and branded posts on the platform will continue to develop.

There is some literature exploring similar relationships to that found in this study. For example, Schouten et al. (2020) explored the relationship between endorser type and Ad attitude, Product attitude and Purchase intention (as a means to measure Ad effectiveness).

These are mediated numerous variables, such as: Trustworthiness and Expertise. This was done through two studies: one focused on food and fashion and the other beauty and fitness.

Both studies involved participants using a likert scale to determine the degree to which they agreed with a statement. This research is important because it is one of the few studies making a very direct comparison between influencers and celebrities. Whilst these dependent variables do not directly relate to post performance, they are an extrapolation of where post

(11)

11

performance as a variable goes. In the context of branded posts, the aim is to in the end create Purchase intention. In this case this will be measured via post performance in a more

quantitative manner. In both versions of their study, Schouten et al. (2020) found that the only area where there was a significant difference in advert effectiveness was in Product Intention. In this case it was found that Influencers created a significant impact on Product Intention versus a traditional celebrity.

Whilst, this research is useful and provides insight into a similar relationship to that being investigated in this research - there are still some factors missing. Schouten et al. (2020) are not focused on posts on social media but rather how endorser type can impact performance of an advert. Hence, it can provide insight into how endorser type might potentially impact the performance of a branded post - but is not specifically in the same context. Consequently, they use a different means to measure Ad effectiveness. In one of their studies using a Likert scale and in the other five 7-point semantic differential scales. Hence, although it provides useful insight, it is not in the same social media context as this research.

A SNS-based comparison of endorser type on consumer brand engagement is produced by Marques et al. (2021). Focussed on Instagram, they explore how “micro influencers” and

“macro influencers” (celebrities) can produce followers for a page, as well as the consumer engagement via clicks, comments and likes. This was done through analysis of a jewellery page over a month-long period. Marques et al. (2021) found that both the influencer and celebrity increased the following of the account, but that the celebrity increase was greater.

Contrastingly, the post-engagement metrics found higher performance for influencer posts.

This was despite the celebrity having significantly more followers than the influencer on their personal pages. This supports the suggestion that the greater closeness and identification that a consumer may feel to an influencer leads to more consumer engagement (Schouten et al., 2020). The research of Marques et al. (2021) is especially useful and important because it is

(12)

12

within a social media context and utilises similar quantitative measures to assess post performance. Nevertheless, this research does not have TikTok as a focus, as this research will. This is significant as the uniqueness of the TikTok short-video format may impact how post engagement works. Hence, it can be said that whilst similar versions of the relationship assessed in this research has been explored to some degree, there is still a need to explore in the specific context of TikTok and Gymshark.

3.0 -Methodology

My research method consisted of quantitative data collection from the Gymshark TikTok page. Gymshark is an Athleisure and apparel company that was founded in 2012 by Ben Francis and Lewis Morgan. In 2021 the company was valued at $1.3 billion and reached sales of $350 million and an increase of 50% from the previous year. Their adroit use of social media has been key to drawing attention to the brand and creating its success (Faithfull, 2021). Hence, it is a very exciting and innovative brand to analyze in this research.

Gymshark was chosen due to a variety of factors. First, the Gymshark TikTok has a large sample of videos – featuring both celebrities and influencers. This makes making an effective comparison between the two categories easier and makes it more possible to control certain variables. For example, it eliminates the impact that differing follow counts and

demographics of pages could have on the results. Furthermore, choosing to focus on

Gymshark was because the company operates in the Athleisure industry. This is an industry where the marketing of a product plays a big role in its success. For example, Nike were not an established player in the Basketball sneaker market until the arrival of Michael Jordan in 1984. Prior to that, Adidas and New Balance have dominated thanks to endorsements with Magic Johnson and James Worthy. Thanks to Gymshark’s membership in this industry, it too relies on endorsements and social media marketing for its success. Additionally, Gymshark’s

(13)

13

TikTok page itself features regular uploads; featuring a wide variety of videos that engage audiences. This means that there is an appropriately large set of data from which a sample can be derived. Hence, Gymshark is a good company for this research to be based on.

3.1- Data &Variables

The data was extracted using an Influencer Marketing Tool named Tagger, which allowed me to download the entirety of Gymshark’s online content, including views, likes and comments.

This method reduced error and aided reliability. From this data, I was able to assess the dependent variable post performance by combining the views, comments, shares and likes together. The data initially didn’t come with an endorser-type classification. Therefore, it was required to manually go through each TikTok video and categorise them as either influencer or celebrity. It was on this basis that I was able to perform analysis using various regression, Anova and T-tests. These were used to assess the significance of the variables.

To successfully assess the impact of endorser type on post-performance, it is key to first set out the variables involved. The independent variable is the endorser type of the post.

Endorser type has two categories: celebrity and influencer. Celebrity endorsers are defined as an individual who has some public recognition that is used by a product/service through appearing in adverts. This is one of the six traditional endorser types (Erfgen, Zenker &

Sattler, 2015; McCracken, 1989, as cited by Schimmelpfennig, 2019). Contrastingly,

Influencers are defined as third-party independent endorsers who affect audience perceptions through various means on social media(Freberg et al., 2011). Through McCraken’s (1989) MTM, celebrity association with a brand can change that brand image (Torres et al., 2019).

Brands take the idea of this model and then apply that to influencers – using their image to improve their own. Hence, Endorser type (independent variable) involves determining whether the post involves endorsement by a celebrity or influencer. This distinction still

(14)

14

remains the same in the TikTok context, as the endorsements involve the same categories of person as any other form of SNS.

The dependent variable in this essay is Post Performance. Post-performance is a numerical variable. It is composed of four different measures: views, likes, comments and shares. Likes and comments are forms of more active engagement. These forms of interaction are

behavioural manifestations of consumer engagement (van Doorn et al., 2010 as cited by Menon et al., 2019). Hence, in analyzing the behaviour of this data in the presence of the independent variable can demonstrate the strength of customer engagement on influencer versus celebrity brand posts. Post Performance is important to assess as engagement with posts provides value to both consumers and firms through the sharing of information and reviews regarding a product (Dolan et al., 2015).

(15)

15

Figure 1

Conceptual Model for Regression Analysis

Note. This figure demonstrates the conceptual model for the regression analysis performed in this

research. It shows that Post Performance (OV) is the combination of views, likes, comments and shares.

3.2 - Control Variables

In order to ensure a high level of internal validity in this research it was necessary to have some variables that were controlled. The key variables that were controlled was the industry of the firm and the follower count of the firm. First, the industry was controlled by focusing on the Gymshark, which operates within the Athleisure industry. This aimed to limit the potential differences that could occur in different categories of products. Influecners might

(16)

16

not sell products, which are a luxury, as well due to a lower status than celebrities.

Controlling industry would control this. Additionally, the follower account of the Tiktok account that the advert was posted to. This was done by collecting the data on a singular date and only using one account (Gymshark). This helped to eliminate possible variation caused by a large audience.

3.3 -Sampling

At the time of data collection (13 June 2022), the sampling frame was 538 TikTok videos. All of which had been uploaded to the same TikTok account and worked within the confines of the platform. From this, the most recent 200 videos were selected as the sample. This was done to minimize the impact that Gymshark’s ever-rising follower count could have. This is necessary to ensure that as best as possible each post would have access to the same potential viewership. All the videos in the sample were then categorized by endorser type. If a video did not fall into either category they were labelled ‘n/a’ and removed from the dataset. In total there were 31 cases removed from the data. This left 137 influencers and 32 celebrities, for a total of 169 videos. This, along with information on other variables, can be seen through Descriptive statistics table found in Table 1. Although most of the sample contains

Influencers, there is still enough celebrity videos to draw conclusions about their impact on Post Performance.

(17)

17 Table 1

Descriptive Statistics table for variables

Variable M SD Min Max

Endorser Type 0.81 0.393 0 1

Post Performance 199,998.45 548,308.75 21,823 6,844,548

Likes 19,164.43 58,087.18 483 725,300

Views 180,472.78 489,309.995 21,100 6,100,000

Shares 199.60 884.54 1 10,900

Comments 161.64 708.30 6 8348

Note. This Table highlights the descriptive statistics of the different variables used in this study. It includes basic statistics to give an overview of the nature of the data

3.4 -Analysis:

In order to assess the relationship between endorser type and post performance, it is

necessary to perform analysis. The tool for performing this analysis was SPSS. The analysis will assess the degree to which either endorser type has a positive correlation to Post

Performance (or any of its components). To demonstrate the nature of the relationship there will be Regression tests performed on the data. The primary relationship being analysed will be that between endorser type and post performance. This relationship will be further divided into the different components of post performance ( likes, views, comments and shares) to see the potential relationships between endorser types and the more specific components of post performance. The Endorser type variable was recoded to be a dummy variable – to make the regressions possible. The potential relationships can seen through the model highlighted in Figure1. A conclusion will be reached on the basis of R²values, which indicate the proportion of variance that can be explained via the model. Alongside this, an ANOVA will be used to

(18)

18

produce F-values and P-values. These figures will highlight whether the endorser type is a significant factor in the variance of post performance between posts.

4.0 - Results & Discussion:

The results found regarding the relationship between endorser type and post performance can be seen in Table 2. The results are highlighted through the R²value, alongside the P-value and F-value produced within an ANOVA. It was on the basis of these figures that there can be said to be no significant relationship between endorser type and post performance.

Indicating that the model (figure1) is not effective for explaining variance in post performance.

Table 2

Regression and ANOVA combined results table

Measure F η2 R²

Post Performance 0.605 0.438 0.004

Likes 0.577 0.448 0.003

Shares 0.466 0.496 0.003

Comments 0.537 0.465 0.003

Note. This Table shows the results for the different measures of the OV. This includes the R- squared values for the different regressions that were performed. Additionally, it features the F-values and Significance values for the ANOVA.

**p < .05.

4.1 - Regression Results:

(19)

19

Regressions were used to not only test the primary model, but also to look at the relationships between the PV and the different components of the OV. These different results are

highlighted in Table 2.

In the primary model the regression produced an R² =.004. This indicated that 0.4% of the variance in Post performance could be explained by endorser type. This would indicate that the model is not very successful at explaining variance. This means that the type of endorser used on a post does not strongly impact how that post performs. Furthermore, with p >.05, F(1,167) = .605 it indicates no statistically significant difference between the model and the

constant. Hence, it can be concluded that the type of endorser is not a significant cause for variance in post performance on the Gymshark Tiktok posts.

Additionally, the regressions from the other relationships produced similar results to this, as can be seen in Table 2. The R² = .003 for the relationship between likes and endorser type was 0.003, indicating that 0.3% of the variance was due to endorser type. Similarly, this indicates that the model was not a successful one. This in combination with a p > .05 and F(1,167) = .577 demonstrated that the model was not significant.

The R² = .004 for the analysis of the relationship between views and endorser type. This indicated that 0.4% of the variance in views was due to endorser type. This in combination with a p > .05 and F(1,167) = .607, that were found by the ANOVA, proved that the model was not significant. Hence, endorser type is not a strong predictor of the views on a post.

An R² = .003 was produced by the regression assessing the relationship between shares and endorser type. This shows that 0.3% of the variance in the amount a post was shared was due to endorser type. This demonstrates that this model was not a predictor of . This in

combination with a p > .05 and F(1,167) = .577 demonstrated that the model was not significant.

(20)

20

The R² =.003 was found via a linear regression for the relationship between comments and endorser type. This highlights that 0.3% of the variance in comments on a post was due to endorser type. This highlights the poor explanatory power endorser type has for why a post might perform well, in terms of comments. There was also p > .05 and F(1,167) = .537 found through the ANOVA. Both of these figures only serve to support the conclusion that this model was not successful. Overall, it can be seen that endorser type is not a successful predictor for any aspect of post performance.

4.2 - Discussion:

The results highlighted through the regression can be compared with existing research.

Schouten et al.(2020) found that, due to the greater trustworthiness and relatability of influencers, they perform better only in terms of product intention. Additionally, results indicated the strong mediating factor that was product-endorser fit.

For Schouten et al. (2020), the relationship betweeen endorser type and the Outcome variables product and advertising attitudes both possessed p>0.05. This indicates that the relationship between these variables did not provide significant explanatory power for

insignificant relationship. However, with the relationship between endorser type and product intention the p-value(0.01) indicated that influencers were more effective for altering product intention. Although this singular result differs, the majority of the data aligns with this research. Hence, in the main both studies found limited reason to conclude that endorser type had significant impact on post performance or advert effectiveness.

This differing result could be because product intention could not be sufficiently captured through the post performance measures used in this research (views, likes, comments and shares). Hence, this relationship could not be borne out in this research. Additionally, the

(21)

21

differing platforms that the studies were completed on may have impacted the results.

Instagram, whilst having incorporated short form video, is primarily a photo sharing SNS.

The shorter attention span required from Tiktok, alongside the quick hits of dopamine that it creates, makes it more difficult to get the audience to focus on one product (Darmatama &

Erdiansyah, 2021). This may provide some explanation as to why a significant relationship between endorser type and post performance was not seen. It may be the case that the photo- based nature of Instagram leads to increased focus on the category of endorser being used on a post.

The nature of this research’s result, when compared to that of Schouten et al. (2020),

suggests the need for additional research. The data in both research likely suggest that there is not a strong relationship directly between endorser type and post performance. However, due to the significant nature of influencers in product intention (Schouten et al.,2020.), there could be interesting research to be done into which aspects of advertiser effectiveness might be impacted by endorser type. This could be done across SNS, in order to compare how the different formats impact the relationship.

Additionally, research assessing the impact of macro and micro-influencers on Instagram posts and account performance by Marques et al. (2021) can be compared with this research.

They found that whilst both endorser types did produce some benefits, the larger effect on the posts and follower counts came from celebrity endorsers. This contrasts my research in that it finds a relationship between endorser type, which my research did not. There are numerous possible reasons for the difference in the results.

First, the two studies were conducted in different industries. Marques et al.(2021) focused on a jewellery firm’s Instagram page in their research. It may be the case that the endorser type may be more important in the jewellery industry. Jewellery is more of a premium product

(22)

22

than the Athleisure that Gymshark sells. This may result in the image of the endorser being a more important factor in determining how well a post performs. Karasiewicz and Kowalczuk (2014) conducted an analysis of the effect of celebrity endorsement on different product categories. In the watch industry, one similar to the jewellery industry, that celebrity

endorsers had a statistically significant impact on the perceived attractiveness of the advert.

However, this was not the case for adverts in the juice industry. Karasiewicz and Kowalczuk (2014) suggest that product categories, like jewellery and watches, that allow for the image and social status of a celebrity to be better transferred to the consumer. This may explain why the study of Marques et al.(2021), focused on jewellery, found contrasting results to the Gymshark study. Although,Karasiewicz and Kowalczuk (2014) do not focus on influencer versus celebrity endorsers conclusions regarding the effectiveness of celebrity endorsements can be drawn. One can argue that the research suggests that premium products may work better with celebrity endorsements as they have an established and high-brow image unlike the, typically, more relatable influencer. The nature of jewellery as a premium industry may be more conducive to specifically celebrity endorsements. This could explain why

Karasiewicz & Kowalczuk (2014) found a significant difference in using a celebrity endorser.

Overall, a comparison between the results of this research and those found by others (Karasiewicz & Kowalczuk, 2014; Schouten et al., 2020) provides interesting context. It mainly highlights that because of the relatively new nature of SNS, in particular Tiktok, greater research is still needed. A lot of the currently available has been focused on very different industries and in different specific contexts. Hence, there ought to be a more complete set of research which greater comparisons can be made.

4.3 -Limitations of Research:

(23)

23

Although this research produced interesting results, there were some limitations to the research conducted. The primary issues being the difficulty categorizing the type of endorser in the post. There were many occasions where an endorser could potentially fall into either category. This is because the line between celebrity and influencer is now more than ever being blurred. There still remains certain celebrities whose fame clearly originates outside of social media, such as, footballers and singers. However, there is a category of celebrities whose fame whilst being outside of social media is heavily dependent on how their social media perform. Equally, there are many influencers, such as KSI the English Youtuber, who have transitioned to more traditional forms of media. Hence, in both these instances it would be difficult to categorize the endorser type. This potential issue with ongoing research in this area may lead to insignificant results. Furthermore, there were also potential issues regarding the sample used. The sample of 200 Gymshark videos was 68% influencer videos and just 32 of the videos feature a celebrity. The difference in sample size may have had negative impact on the reliability of the results and generalizability of them. In future research, it might be interesting to look at a more evenly distributed sample.

5.0 Conclusion:

This study investigated the relationship between endorser type (PV) and post performance (OV). It attempted to assess whether either an influencer or celebrity as an endorser proved to be more successful at inducing high comments, views, likes, or shares. The context of this research was an analysis of the Gymshark Tiktok page. This was important to study because of the growing place Social Media is playing, and will continue to do, in the future of advertising. Tiktok, with its short video format, is skewing extremely young and claims to have users using the app for 90 minutes per day (Kafka, 2022). This provides advertisers with

(24)

24

the opportunity to reach and communicate with the next generation of consumers in a way that television can no longer. This is because Television audiences are based in much older demographics(Kafka, 2022). Hence, understanding how to operate on Tiktok’s platform and the role that it can play in the future of advertising is key.

To assess the aforementioned relationship data was collected from the Gymshark Tiktok page. A sample of 200 of the latest videos released on the page was selected. The data was extracted using the social media analytics tool Tagger and each post was categorized as featuring either a celebrity or an influencer. From this data, using recoded dummy variables, a regression was performed. This was to assess whether there was a significant relationship between endorser type and post performance. From these results, p-values and R2 values were used to demonstrate this.

Overall, the results did not indicate a significant relationship between endorser type and post performance. This means that whether a post featured an endorsement by a celebrity or an influencer it was a not a significant factor in the performance of that post. This was demonstrated through and R² = .04 – indicating that only 0.4% of variance could be

explained by endorser type. Furthermore, the p = >.05 meaning that the relationship was not of statistical significance. Similarly, when analyzing the individual components of the OV (likes, comments, shares and views) low R² values were found and p = > 0.05 indicated non- significant relationships. The most important result was the one regarding the primary model – highlighting a not statistically significant relationship between endorser type and post performance. This is because it demonstrated that the model was not successful.

There are various lessons that can be taken away from these results. From the data collection, it can be seen how difficult it is becoming to categorize celebrities and influencers. The line between these groups is becoming ever more blurred. This may have impacted the results of

(25)

25

the regression. There is a need for a re-assessment of the definitions of the two categories are.

With social media becoming an increasingly vital component of the business of being a celebrity – discussing the boundaries between celebrities and influencers is a key message to come out of the research. Furthermore, the results produced by the regression indicated a non-significant relationship. This meant that the endorser type did not factor in post

performance. Hence, this further alludes to the diminishing importance of the categorizations of endorsers. From this, marketeers can learn that the individual merits of an endorser, and how they may connect to an audience, are more important than their influencer type.

On the basis of these conclusions, the need for additional research presents itself. First, there is a need to investigate Tiktok more as a platform. As the popularity of Tiktok only really exploded during the pandemic, it is a social media platform that has not been widely explored. This could build on the research of Schouten et al. (2020), which was focused on Instagram. Furthermore, there is a clear need to investigate the boundaries between celebrities and influencers. This research highlights the increased difficulty in categorizing endorsers.

Consequently, it could be interesting to look at whether those categories will continue to be relevant in marketing.

(26)

26

Statement of Originality

This document is written by Finian Barke-Asuni who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document are 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.

(27)

27

References:

3 stats that show influencers are as influential as ever. YPulse. (n.d.). Retrieved March 18, 2022, from https://www.ypulse.com/article/2020/09/01/3-stats-that-show-influencers- are-as-influential-as-ever/

Bergkvist, L., & Zhou, K. Q. (2016). Celebrity endorsements: A literature review and research agenda. International Journal of Advertising, 35(4), 642–663.

https://doi.org/10.1080/02650487.2015.1137537

Cook, J. (2020, August 18). How gymshark became a $1.3 billion brand, and what we can learn. Forbes. Retrieved March 18, 2022, from

https://www.forbes.com/sites/jodiecook/2020/08/17/how-gymshark-became-a-13bn- brand-and-what-we-can-learn/

de Vries, L., Gensler, S., & Leeflang, P. S. H. (2012). Popularity of brand posts on Brand Fan Pages: An investigation of the effects of social media marketing. Journal of Interactive Marketing, 26(2), 83–91. https://doi.org/10.1016/j.intmar.2012.01.003

Darmatama, M., & Erdiansyah, R. (2021). The influence of advertising in Tiktok social media and beauty product image on Consumer Purchase Decisions. Advances in Social Science, Education and Humanities Research.

https://doi.org/10.2991/assehr.k.210805.140

Digital advertising to exceed 60% of global adspend in 2022. Zenith. (2022, March 4).

Retrieved March 18, 2022, from https://www.zenithmedia.com/digital-advertising-to- exceed-60-of-global-adspend-in-

2022/#:~:text=Social%20media%20adspend%20will%20rise,%25%20and%20televisio n%20at%2021.0%25

Dolan, R., Conduit, J., Fahy, J., & Goodman, S. (2015). Social Media Engagement behaviour: A uses and Gratifications Perspective. Journal of Strategic Marketing, 24(3- 4), 261–277. https://doi.org/10.1080/0965254x.2015.1095222

Faithfull, M. (2021, December 16). Gymshark to open first brick-and-mortar store as CEO downplays IPO. Forbes. Retrieved June 22, 2022, from

https://www.forbes.com/sites/markfaithfull/2021/12/15/gymshark-to-open-first-brick- and-mortar-store-as-ceo-downplays-ipo/?sh=5c768c74dc68

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. https://doi.org/10.1016/j.pubrev.2010.11.001

Gräve, J.-F., & Bartsch, F. (2021). #instafame: Exploring the endorsement effectiveness of influencers compared to celebrities. International Journal of Advertising, 1–31.

https://doi.org/10.1080/02650487.2021.1987041

Hudders, L., De Jans, S., & De Veirman, M. (2020). The commercialization of Social Media Stars: A Literature Review and conceptual framework on the strategic use of Social Media influencers. International Journal of Advertising, 40(3), 327–375.

https://doi.org/10.1080/02650487.2020.1836925

(28)

28

Kafka, P. (2022, June 22). It's TikTok's world. can TV live in it? Vox. Retrieved June 29, 2022, from https://www.vox.com/recode/2022/6/22/23177051/tiktok-tv-streaming- peter-kafka-media

Karasiewicz, G., & Kowalczuk, M. (2014). Effect of celebrity endorsement in advertising activities by product type. International Journal of Management and Economics, 44(1), 74–91. https://doi.org/10.1515/ijme-2015-0010

Kepios. (n.d.). Global Social Media Statistics - DataReportal – global digital insights.

DataReportal. Retrieved June 15, 2022, from https://datareportal.com/social-media- users

Marques, I. R., Casais, B., & Camilleri, M. A. (2021). The effect of macrocelebrity and Microinfluencer endorsements on consumer–brand engagement in Instagram. Strategic Corporate Communication in the Digital Age, 131–143. https://doi.org/10.1108/978-1- 80071-264-520211008

McCracken, G. (1989). Who is the celebrity endorser? Cultural Foundations of the endorsement process. Journal of Consumer Research, 16(3), 310.

https://doi.org/10.1086/209217

Menon, R. V., Sigurdsson, V., Larsen, N. M., Fagerstrøm, A., Sørensen, H., Marteinsdottir, H. G., & Foxall, G. R. (2019). How to grow brand post engagement on Facebook and Twitter for airlines? An empirical investigation of design and content factors. Journal of Air Transport Management, 79, 101678.

Montag, C., Yang, H., & Elhai, J. D. (2021). On the psychology of TikTok use: A first glimpse from empirical findings. Frontiers in Public Health, 9.

https://doi.org/10.3389/fpubh.2021.641673

Published by Statista Research Department, & 28, J. (2022, January 28). Daily Social Media Usage Worldwide. Statista. Retrieved March 18, 2022, from

https://www.statista.com/statistics/433871/daily-social-media-usage-worldwide/

PwC. (2021). IAB Digital adspend. Adspend | IAB UK. Retrieved June 15, 2022, from https://www.iabuk.com/adspend

Sabate, F., Berbegal-Mirabent, J., Cañabate, A., & Lebherz, P. R. (2014). Factors influencing popularity of branded content in Facebook Fan Pages. European Management Journal, 32(6), 1001–1011. https://doi.org/10.1016/j.emj.2014.05.001

Schimmelpfennig, C. (2019). The use of different endorser types in advertising: a content analysis of magazine advertisements. Journal of Global Marketing, 32(3), 139-153.

Schouten, A. P., Janssen, L., & Verspaget, M. (2020). Celebrity vs. Influencer endorsements in advertising: The role of identification, credibility, and product-endorser fit.

International Journal of Advertising, 39(2), 258–281.

https://doi.org/10.1080/02650487.2019.1634898

Shao, Z. (2018). Analysis of the characteristics, challenges and future development trends of Tik Tok. Mod Educ Tech, 12, 81-7.

Su, Y., Baker, B. J., Doyle, J. P., & Yan, M. (2020). Fan engagement in 15 seconds:

Athletes’ relationship marketing during a pandemic via TikTok. International Journal of Sport Communication, 13(3), 436–446. https://doi.org/10.1123/ijsc.2020-0238

(29)

29

Taylor, C. R. (2020). The urgent need for more research on influencer marketing.

International Journal of Advertising, 39(7), 889–891.

https://doi.org/10.1080/02650487.2020.1822104

Torres, P., Augusto, M., & Matos, M. (2019). Antecedents and outcomes of Digital Influencer Endorsement: An exploratory study. Psychology & Marketing, 36(12), 1267–1276. https://doi.org/10.1002/mar.21274

Tuten, T. L., & Solomon, M. R. (2020). Marketing Communication: From Top-down to Bottom-up. In Social Media Marketing. essay, Pearson.

Voorveld, H. A., van Noort, G., Muntinga, D. G., & Bronner, F. (2018). Engagement with social media and social media advertising: The differentiating role of platform type. Journal of Advertising, 47(1), 38–54.

https://doi.org/10.1080/00913367.2017.1405754

Wang, E., Wu, K., & Zhu, J. (2020, July 29). Exclusive: ByteDance investors value TikTok at

$50 billion in takeover bid - sources. Reuters. Retrieved March 18, 2022, from https://www.reuters.com/article/us-bytedance-tiktok-exclusive-idUSKCN24U1M9 Wood, N. T., & Burkhalter, J. N. (2013). Tweet this, not that: A comparison between

brand promotions in microblogging environments using celebrity and company- generated tweets. Journal of Marketing Communications, 20(1-2), 129–146.

https://doi.org/10.1080/13527266.2013.797784

By Finian Barke-Asuni

(30)

30

(31)

31

Celebrity vs Influencer: Which Endorser Type has the biggest influence on Post Performance in Tiktok?

Celebrity vs Influencer: Which

Endorser Type has the biggest influence on Post Performance in Tiktok?

Table of Contents