Effectiveness of social media effectiveness of social media : an application of the UTAUT model

(1)

Effectiveness of Social Media

Influencers on purchase intention:

an application of the UTAUT Model

Daphne Elyze Veelers

11926228

22-06-2818

Final version

MSc. Business Administration – Digital Business

University of Amsterdam

(2)

STATEMENT OF ORIGINALITY

Statement of originality

This document is written by Daphne Elyze Veelers 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)

4.3 Moderating variables 34 4.3.1 Gender 34 4.3.2 Age 36 4.3.3 Product segment 38 5. Discussion 40 5.1 Practical implications 40 5.1.1 Performance expectancy 40 5.1.2 Effort expectancy 41 5.1.3 Social influence 42 5.1.4 Attitude 44 5.1.5 Behavioural intention 45 5.2 Theoretical implications 47 5.3 Limitations 48 5.4 Future Research 48 6. Conclusion 49 7. References 50 8. Appendices 53 8.1 Questions questionnaire 53

(4)

LIST OF FIGURES AND TABLES

Page

1. List of figures

Figure 1 Conceptual framework: Revised UTAUT model 18 Figure 2 Q-Q Plots independent and dependent variables 30

2. List of tables

Table 1 Definition of independent variables 16

Table 2 Definition of dependent variables 16

Table 3 Summary of survey constructs 26

Table 4 Demographics 27

Table 5 Cronbach’s alpha 29

Table 6 Skewness and kurtosis for normality testing 29

Table 7 Mean and standard deviation 31

Table 8 Pearson Correlation Coefficients 32

Table 9 Multicollinearity statistics 32

Table 10 Model summary 33

Table 11 ANOVA 33

Table 12 Multiple regression analysis 34

Table 13 Regression analysis for gender 35

Table 14 Descriptive statistics gender 36

Table 15 Regression analysis for age 36

Table 16 Descriptive statistics for age 37

Table 17 Regression analysis for product segment 38 Table 18 Descriptive statistics for product segment 39

(5)

5

ABSTRACT

The purpose of this research is to find out how performance expectancy, effort expectancy, social influence and attitude towards social media influencers are related to customer’s behavioural intention to use the recommendations of the influencers in their purchase decision. A revised version of the UTAUT model is therefore applied. The UTAUT model looks at user acceptance of a new technology. For the purpose of this research the users are potential customers and the new technology is the videos on YouTube from social media influencers as a new source of product information. An online questionnaire was developed, with questions based on performance expectancy, effort expectancy, social influence, attitude and behavioural intention. The sample size of this research is 179. After conducting a number of statistical tests in SPSS it can be concluded that performance expectancy doesn’t have significant effect on behavioural intention. The other three remaining independent variables effort expectancy, social influence and attitude, do show a significant relationship with behavioural intention. Furthermore, when taking a look among three different product segments, fashion, make-up and games, it can be concluded that the highest behavioural intention influenced by social media influencers is for the product segment make-up.

(6)

6

1. INTRODUCTION

Social media has been one of the fastest developing phenomena during the past years. In 2017 there was an estimate of around 2.46 billion social media users, which is expected to grow to around 3.02 billion in 2021 (retrieved on 24-01-2018 from statista.com). Moreover, Facebook alone grew from 1 billion active users in the third quarter of 2012 to 2.07 billion active users in the third quarter of 2017 (retrieved on 24-01-2018 from statista.com). Lastly, around 82% of US inhabitants have a profile on one of the social media platforms (retrieved on 24-01-2018 from statista.com). These enormous numbers show, that due to the fact that so many people make use of social media, it has become crucial for brands to be present online, especially on social media platforms. In 2015 more than 50 million businesses worldwide already had a Facebook page, and each month there were, on average, more than 2.5 billion comments made on these pages (Chaykowski, C. 2015). Next to Facebook, other often used and well-known social media platforms are Instagram, Snapchat and YouTube. Social media platforms were originally used as a communication tool between people, and it was a place where people would share their thoughts and actions. However, social media has turned to Business-to-Consumer platforms as well since many brands have their own social media pages and they use these pages as part of their digital marketing strategy.

A new development on social media platforms is the presence of social media influencers, who often promote a certain brand or product online. These social media influencers can be celebrities, like movie stars and artists, or ‘ordinary’ people who got famous due to their online activities. Social media influencers often have millions of followers worldwide, which is a very interesting statistic for businesses. Most of these followers are Millennials or people from Generation Z. According to a research conducted by Google, 40% of millennial YouTube subscribers state that they feel that their favourite YouTube personalities understand them better than their own friends (O’Neil-Hart, C. &

(7)

7 Blumenstein, H. 2016). The focus of this research is on social media influencers who

promote products via videos on the social media channel YouTube. There are different ways products and brands are presented in videos. This can be in the form of an actual review, in which the pros and cons of the product are explained. Furthermore, influencers may just mention the product that they are using or wearing, without outlining the pros and cons of the product. And lastly, they do not explicitly mention the brand, but it just appears in their video without any special attention towards it.

One of the latest digital marketing trends is influencer marketing, in which social media influencers become part of a digital marketing strategy. With influencer marketing, businesses work together with influencers to represent or promote their product or brand online (Bulik, B. S. 2018). In order to execute a successful and effective influencer marketing campaign, it is essential that the voice of the influencer is credible and that people can relate to him/her. Social media influencers are seen as trustworthy sources, who have a positive impact on the attitude and purchase intentions for a brand (Boerman, S. 2018). Traditionally businesses used celebrity endorsement to promote a product or service in an advertisement, however, a new development on social media shows that businesses start working together with ordinary people who became influencers due to their online activities via their social media platforms (Bulik, B. S. 2018). According to Amy Osmond Cook, CEO of Osmond Marketing, companies that don’t use sponsored blog and social media posts to build their brand visibility, will most likely fall behind in their competition (Cook, A. O., 2017).

Nowadays brands pay a lot of money to social media influencers, who in return post pictures or videos in which the brand is shown and mentioned. In the research conducted by Google, 70% of the respondents said they feel more related to influencers who became famous through the Internet, than to traditional celebrities. (O’Neil-Hart, C. & Blumenstein, H. 2016). This shows that businesses that want to tap into the world of social media influencers

(8)

8 should focus on influencers who became famous through social media and not just

celebrities. Furthermore, statistics show that 58% of surveyed respondents (enterprise brand strategists and marketers) say that the use of influencer marketing will be present and

integrated in all future marketing activities and that being part of a digital marketing strategy of a brand will become the main focus of social media influencers (retrieved on 31-01-2018 from statista.com). Moreover, according to research conducted by Linqia, 92% of marketers who made use of social influencer marketing in 2017 concluded it as being a success and an effective marketing strategy (Linqia, 2017). Next to all these statistics, big advantages of using influencer marketing is that it’s not disturbed by ad blockers, which was used by 17% of the Dutch Internet population in 2017, and usage grew globally with 30% (Cortland, M. 2017) and that brands have the ability to reach a younger generation that hardly watches any TV anymore (retrieved on 24-01-2018 from marketingland.com).

In recent years a lot of research has been done to social media influencers, especially to who they are, what they do, and how they can become part of a digital marketing strategy of a brand. Also, case studies have been conducted to important social influencers, like Lucie Fink, and there have been studies to the use of influencer marketing for specific brands. Social media influencer marketing is one of the newest methods of digital marketing, and it’s seen as a hot topic for 2018. A lot of research has already been conducted concerning the effectiveness of social media influencers, and it can be concluded that the use of social media influencers is very effective. However, the focus of this research is to look at social media influencers as a new source of product information and recommendation, which in the end might affect consumer’s purchase intention. According to customer psychologist Patrick Wessels, social media influencers are ‘your new neighbour or friend’ concerning word-of-mouth recommendations (retrieved on 19-05-2018 from nos.nl). In an interview with Forbes, Rachelle Croft, co-founder of the company Ju-Ju-Be, mentioned that social media influencers

(9)

9 are the new source of information for consumers to find product. She stated that: We

(consumers) don’t need to go shopping and search for the perfect bag or dress or makeup because our favourite bloggers gather all of that information (Cook, A. O., 2017). This study applies a revised version of the UTAUT Model. The UTAUT model is applied in studies that investigates user acceptance to new technology. In this research the users are consumers and the new technology are the online videos on YouTube of social media influencers that serve as a new source of product information. This research takes into account four different independent variables, namely performance expectancy, effort expectancy, social influence and attitude. The first two are more practically oriented, whereas the last two are more socially oriented. The dependent variable is behavioural intention, which focuses on whether people will use and follow the recommendations of social media influencers to buy certain products. Furthermore, this research includes three moderating variables, namely age, gender and product segments. The research gap that is addressed in this research is that it takes into account four independent variables, both practically and socially oriented, and that it looks at social media influencers as new sources of product information and recommendation. By applying the UTAUT model the social media influencers are seen as a new online

technology, that in the end might affect consumers’ purchase intentions.

The research question that is answered with this study is: ‘In what way do performance expectancy, effort expectancy, social influence and attitude of social media influencers impact the behavioural intention of consumers? In order to collect data an online questionnaire was distributed including questions based on the independent variables and dependent variable of the revised and applied UTAUT model. This paper continuous with a literature review including a detailed description concerning social media influencers and their link to the social identity theory of Henri Tajfel and the explanation of the conceptual framework. Afterwards the methodology of this research is explained, followed by the results

(10)

10 of the questionnaire. The last part discusses whether the hypotheses can be accepted or need to be rejected, future research opportunities are provided, as well as the limitations of this research.

2. LITERATURE REVIEW

2.1 Social Media Influencers

The so called social media influencers represent a new group of independent advocates, whose goal is to shape the attitude of their audience and to give their audience a look in their lives through vlogs, blogs, tweets, Instagram posts and other uses of social media (Freberg, K., Graham, K., McGaughey, K. & Freberg, L. 2011). They are individuals or groups, who can be celebrities or popular persons due to social media, who spread information about a specific brand or product through their social media platforms, and therefore provide marketers with opportunities to make social influencers part of their digital marketing strategy (Snijders, R. & Helms, R.W. 2014). Social media that are frequently used by influencers include Instagram, Facebook, Twitter and YouTube. While traditional social media focuses on the conversation between people or between companies and customers (Hoffman, E., Khanfar, N.M., Harrington, C. & Kizer, L.E. 2016), social media influencers focus mainly on the interaction between the influencers and their followers (Uzunoglo, E. & Kip, S.M. 2014). Not every influencer is the same or has the same goal, therefore Ranga & Sharma identified four different kinds of influencers, namely:

1. ‘Traditional influencers: These are individuals, mainly celebrities, in a specific area of subject expertise.’

2. ‘Emerging (digital) influencers: ‘These are individuals that are emerging as influencers, who have a large audience and drive through leadership in a specific space.’

(11)

11 3. ‘Influencers by connection: These are individuals who have thousands of followers on

their Facebook, Instagram, Twitter, YouTube and/or other social media platforms, who represent and promote a brand on their social platforms.’

4. ‘Influencers by topic: These are individuals who are opinion leaders for specific topics.’ (Ranga, M. & Sharma, D. 2014, p. 18).

Besides the existence of various kinds of social media influencers, Ranga & Sharma also identified that there are different ways social media influencers can have effect on one’s business, for example:

1. Influencers can write a blog post/article or make a video/vlog.

2. Influencers can share brand- or product-related information on their social media platforms.

3. ‘Influencers can ask or permit to guest post on their website.’ (Ranga, M. & Sharma, D. 2014, p. 17).

These three ways can also be combined in order to reach an even greater audience and to increase the effectiveness of implementing social media influencers on an organization’s marketing strategy. The focus of this research is on influencers by connection, who provide product information and recommendation via their videos on YouTube.

There are different ways to assess how influential a person is on his/her social media platform. One of the most well-known and used metrics was the Klout score. This score was used as a metric for measuring the influence of users on online (social media) platforms (Rao, A., Spasojevic, N., Li, Z. & DSouza, T. 2015). ‘The Klout score measures a user’s overall online influence with a score ranging from 1 to 100, with 100 being the highest amount of possible influence’. ‘Klout analyses more than 25 variables, also offering the possibility to combine the score from multiple social media platforms’ (Anger, I. & Kittl, C. 2011, p. 2). People with high Klout scores are for example Barack Obama (99), Justin Bieber (92) and

(12)

12 Zooey Deschanel (86). The website of the Klout score described the concept of influence as follows: ‘Influence is the ability to drive action.’ ‘When you share something on social media or in real life, and people respond, that’s influence.’ (Retrieved on 25-01-2018 from

www.klout.com). As of May 25, 2018, it was decided to shutdown Klout, since it wasn’t seen as an effective long-term strategy (Leswing, K. (2018).

2.2 Social Identity Theory

This section will provide an introduction to the Social Identity Theory and will explain why people want to identify with a specific social group, and why this is one of the reasons for the success of social media influencers. According to the Social Identity Theory, originally formulated in 1979 by Henri Tajfel and John Turner, ‘people tend to classify themselves and others into various social categories, such as organizational membership, religious affiliation, gender and age cohort’ (Tajfel, H. & Turner, J.C. 1979, p. 20). Compared with many other social psychological theories, the focus of the social identity theory lays in the willingness of belonging to a certain social group (Tajfel, H. & Turner, J.C. 1985). A social group consists of some individuals who feel and perceive that they belong to that group, because of matching personal characteristics. (Tajfel, H. & Turner, J.C. 1985). ‘Research on identification suggests that people tend to make social classifications because it provides a systematic way to define others and to locate oneself in the social environment’ (Cornwell, T. B., Weeks, C. S. & Roy, D. P., 2005, p. 28).

The self-categorization theory, founded by Turner et al., further elaborates on the social identity theory, by saying that individuals who identify themselves with the in-group, have a great chance to internalize the values, norms and attitudes of the group as their own, and follow the actions taken by the group (Turner, J.C., Hogg, M.A., Oakes, P.J., Reicher, S.D. & Wetherell, M.S. 1987). ‘This is supported by over two decades of social identity studies on collective action, which demonstrated that higher levels of group identification

(13)

13 with the in-group are related to greater likelihood of engagement’ (Chan, M. 2016). This could also account for the audience of the social media influencers, when they identify themselves with the influencer, there is an increase in the possibility that they follow their actions, which could lead to purchasing similar products as the influencer promotes. A study done by Gashi, L. showed that: ‘A social media influencers’ ability to provide, content, expertise, attractiveness and social identity shows evidence of how the influence of social media influencers play an important role in each and every stage of the purchase decision process of consumers’’ (Gashi, L. 2017, p.1).

2.3 Social Media Influencer Marketing

Companies are more often allocating greater parts of their marketing spending to social media activities (Hudson, S., Huang, L., Roth, M.S. & Madden, T.J. 2016). One of the newer forms of digital marketing is social media influencer marketing. Through this social media influencers become part of an organization’s digital marketing strategy, where the social media influencers promote and provide information about a certain product or brand. As identified by Ranga & Sharma, social influencer marketing includes four stages. First of all, social media influencers should be identified and graded/ranked according to their significance. Secondly, an organization can start marketing to influencers, in which

awareness of the firm will be created amongst the influencer’s audience. The next step could be marketing through influencers, in which influencers are used to increase market awareness of the firm amongst specified target markets. Lastly, a firm can start to completely align their marketing strategies with the influencer, by which influencers will be turned into advocates of the firm (Ranga, M & Sharma, D. 2014). Compared to other marketing techniques, social media influencers are being perceived as authentic and trustworthy, which is crucial for competing to maintain your audience (Kedzior, R., Allen, D.E. & Schroeder, J. 2016).

(14)

14 When influencer marketing is used in a social media context it can be seen as a viral marketing technique (Ferguson, R. 2008). ‘Viral marketing is the tactic of creating a process where interested people can market to each other’. (Subramani, M.R. & Rajagopalan, B. 2003, p. 300). Social media platforms are increasingly being seen as important sources

through which customers can receive product information and therefore be influenced in their purchasing decisions (Hoffman, E., Khanfar, N.M., Harrington, C. & Kizer, L.E. 2016). This used to be done mainly through electronic word-of-mouth (EWOM) from family and friends, however nowadays there is an increasing importance of recommendations of social media influencers. This also has to do with the fact that traditional advertising techniques like television and radio ads are too general and inefficient, and that more attention is needed for targeted advertising since that is seen as more efficient (Aaker, J. L, Brumbaugh, A. M & Grier, S. A. 2000). Therefore, a more holistic and comprehensive relationship marketing strategy must be utilized to be able to target specific people (Hoffman, E., Khanfar, N.M., Harrington, C. & Kizer, L.E. 2016). Organizations are more and more starting to realize the value of using social influencers in their digital marketing campaigns, since they are able to reach their niche audiences. This is also supported by the work of Alshawaf & Wen as they show that brands can benefit from the reach that social influencers have (Alshawaf, E. & Wen, L. 2015). When consumers are online they are constantly exposed to numerous advertisements from different organizations. ‘There is no doubt that influencers assist in meeting marketing objectives on social media, but the challenge lies in selecting the most effective influencers (Hobson, R. 2016, p. 14), and who have the best fit with the target audience (Booth, N. & Matic, J. A. 2011).

(15)

15

2.4 Conceptual Framework 2.4.1 UTAUT Model

The theoretical framework that is used in this research is an updated and extended version of the UTAUT Model. The UTAUT Model was found and developed by Venkatesh in 2003 and has since been applied to many studies concerning the acceptance of

(information) technology. ‘Explaining user acceptance of new technology is often described as one of the most mature research areas in the contemporary information systems (IS)

literature’ (Venkatesh, V., Morris, M.G., Davis, G.B. & Davis, F.D. 2003, p. 426). The model focuses on the users and a new kind of technology. During the past years many research models in the field of user acceptance of technology have been found, with specifications of psychology, sociology and information systems. Venkatesh et al. used the results of these studies in order to create the UTAUT model. The original model consists out of four constructs, which are seen as the independent variables that directly influence behavioural intention, which in turn influences actual usage behaviour (Venkatesh, V., Morris, M.G., Davis, G.B. & Davis, F.D. 2003). The four original constructs are performance expectancy, effort expectancy, social influence and facilitating conditions. ‘The labels used for the constructs describe the essence of the construct and are meant to be independent of any particular theoretical perspective’ (Venkatesh, V., Morris, M.G., Davis, G.B. & Davis, F.D. 2003, p.447). Next to the four independent variables, the UTAUT model also includes four different moderating variables, namely gender, age, experience and voluntariness of use (Venkatesh, V., Morris, M.G., Davis, G.B. & Davis, F.D. 2003). These variables have the intention to moderate the relationship between the independent variables and the dependent variables. Table 1 below provides the definitions of the four independent variables of the original model.

(16)

16 Table 1

Definition of independent variables

Construct Definition

Performance expectancy ‘The degree to which an individual believes that using the

system will help him or her to attain gains in job performance’.

Effort expectancy ‘The degree of ease associated with the use of the system’.

Social influence ‘The degree to which an individual perceives that important others believe he or she should use the new system’.

Facilitating conditions ‘The degree to which an individual believes that an

organizational and technical infrastructure exists to support use of the system’.

(Venkatesh, V., Morris, M.G., Davis, G.B. & Davis, F.D. 2003, p. 447, 450, 451 & 453) Furthermore, the model consists out of two dependent variables, namely behavioural intention and use behaviour. Behavioural intention is influenced by performance expectancy, effort expectancy and social influence, whereas use behaviour is influence by behavioural intention and facilitating conditions (Venkatesh, V., Morris, M.G., Davis, G.B. & Davis, F.D. 2003). Table 2 below provides the definitions of the two dependent variables of the original model.

Table 2

Definition of dependent variables

Construct Definition

Behavioural intention ‘Gives an indication about an individual’s readiness to perform a

specific behaviour’.

Use behaviour ‘Observable response in a given situation with respect to a giver target’.

(Tarhini, A., El-Masri, M., Ali, M. & Serrano, A. 2016, p. 837)

Furthermore, the original model includes four moderating variables, namely gender, age, experience and voluntariness of use (Venkatesh, V., Morris, M.G., Davis, G.B. & Davis, F.D. 2003).

(17)

17

2.4.2 Revised UTAUT Model

The UTAUT model is used in order to research user acceptance of new technologies. The model has been revised for the scope of this research in order to study user acceptance of treating the new technological phenomenon social media influencers as a new source of product information in which consumers are influenced in their purchase intention. The videos of social media influencers are seen as a new technology for information concerning new products and brands. This study includes four independent variables in order to research their effects on the dependent variable behavioural intention. In this study a number of constructs of the UTAUT model are changed and added in order to make it properly fit with this research. This research looks at the effects of social media influencers, as new sources of product information and product awareness, on future purchase intentions. The updated model includes four independent variables as well, namely performance expectancy, effort expectancy, social influence and attitude. The first two independent variables focus mainly on practical matters concerning the videos of social media influencers, whereas the last two independent variables are more socially oriented. Compared to the original model, the dependent variable actual usage is not taken into account in the revised model in this research. This is due to the fact that it was out of the scope for this research to measure whether social media influencers actual lead to purchases, without other possible variables like traditional advertising or word-of-mouth from friends/family. Actual usage in the context of this research means that people bought products due to the recommendations of social media influencers. The behavioural intention in the context of this research is considered as using the videos of social media influencers to make a purchase decision. The figure below shows the revised version of the UTAUT model used in this research.

(18)

18 Figure 1

Conceptual framework: Revised UTAUT model

Performance expectancy (PE)

The first independent variable is performance expectancy (PE). This research focuses on regarding social media influencers as new sources of product information and awareness. In this study, performance expectancy focuses on the content of the videos that social media influencers produce and post on YouTube. In these videos social media influencers give their personal opinions about a product, as well as factual information about the characteristics of the product. What is expected is that whenever social media influencers provide valuable information about certain products, it influences the behavioural intention of consumers to use social media influencers in their purchase decision. Whenever social media influencers are very positive about the product and they provide valuable information, it may affect consumers to buy the product as well. This leads to the following hypothesis:

(19)

19

H1: Performance expectancy will significantly affect customers’ behavioural intention to

purchase products promoted in the videos of influencers.

Furthermore, for all the four independent variables the moderating variables gender and age are included as well. Venkatesh et al. state that women are more sensitive to other’s opinions (Venkatesh, V. & Morris, M.G. 2000). This may impact the fact that women are more influenced by others’ views to buy certain products and brands. This means for this research that women are more likely to use the videos of social media influencers to make a purchase decision, since these videos contain opinions from others. Therefore, it is expected that in this research the relationship between the independent variables and the dependent is strongest for women, what leads to the following hypothesis for performance expectancy:

H1a: The effect of performance expectancy on behavioural intention is different by gender,

in such a way that the effect will be stronger for women.

Furthermore, the fact that internet usage is biggest amongst younger people, the impact of social media influencers will most likely be bigger. The younger generation has a stronger presence online and it’s becoming more important to show a certain social status on these platforms (Evers, C. W., Albury, K., Byron, P. & Crawford, K. 2013). Therefore, it is expected that in this research the relationship between the independent variables and the dependent is strongest for the youngest age group, what leads to the following hypothesis for performance expectancy:

H1b: The effect of performance expectancy on behavioural intention is different by age, in

such a way that the effect will be stronger for the youngest age group.

Effort expectancy (EE)

The second independent variable is about effort expectancy. For this research, it is applied to investigate how much effort consumers have to take in order to clearly understand the given product information and take away the key message. It also includes whether it is

(20)

20 clearly mentioned to consumers that the video is sponsored or not. Nowadays many social media influencers get paid to mention a brand in a video, and this might decrease the trustworthiness of the message of the influencer. What is expected is that when the video is very clear, that not much effort has to be taken in order to understand the message and that it is clearly mentioned when the video is sponsored, it positively impacts the purchase intention for products promoted by influencers. This leads to the following hypothesis:

H2: Effort expectancy will significantly affect customers’ behavioural intention to purchase

products promoted in the videos of influencers.

The moderating variables and their effects, as explained for the independent variable performance expectancy, are the same for effort expectancy. This leads to the following hypotheses:

H2a: The effect of effort expectancy on behavioural intention is different by gender, in such

a way that the effect will be stronger for women.

H2b: The effect of effort expectancy on behavioural intention is different by age, in such a

way that the effect will be stronger for the youngest age group.

Social influence (SI)

The third independent variable is about the social influence that impacts the purchase intention. For this research, it is first of all seen as the social influence that makes people want to follow and behave the same as certain well-known influencers. Furthermore, social influence can be found in the environment of customers. Social media influencers mainly make videos about products and brands that follow the latest trends. When your environment uses social media influencers to stay up-to-date with the newest trends and products, this might influence you to do the same as well. Consumers might experience the social pressure within their environment to follow famous influencers and follow their recommendations so that it’s noticed by friends and family that they follow the latest trends.

(21)

21 What is expected is that when influencers have a lot of followers, are socially seen as important influencers that many people know, and when your environment looks at

influencers as well, it positively impacts the purchase intention for products promoted by these influencers. This leads to the following hypothesis:

H3: Social influence will significantly affect customers’ behavioural intention to purchase

products promoted in the videos of influencers.

The effects of the moderating variables age and gender, are the same as for performance and effort expectancy, which thus leads to the following hypotheses:

H3a: The effect of social influence on behavioural intention is different by gender, in such a

way that the effect will be stronger for women.

H3b: The effect of social influence on behavioural intention is moderated by age, in such a

way that the effect will be stronger for the youngest age group.

Attitude (ATT)

The fourth and final independent variable is about the attitude towards influencers. This independent variable is added, compared to the original model. The inclusion of attitude in the UTAUT has been widely discussed. Venkatesh et al. debated that the effect of attitude on behavioural intention only emerges when performance expectancy and effort expectancy effort is removed from the model (Venkatesh, V., Morris, M.G., Davis, G.B. & Davis, F.D. 2003). This is due to the fact that attitude will not provide enough unique information which is not already been provided by performance expectancy and effort expectancy (Venkatesh, V., Morris, M.G., Davis, G.B. & Davis, F.D. 2003). However, in this research performance expectancy and effort expectancy mainly focus on practical matters concerning the content of videos of the influencers, while social influence and attitude focus more on social matters and the social connection with the influencer. Due to this, it is expected that there will be enough unique information concerning attitude towards the videos of social media influencers, in

(22)

22 order to research the effects of videos from influencers on purchase intention. What is

expected is that when consumers have a positive attitude towards an influencer, they will more likely watch their videos and follow their recommendations, which in the end might lead in buying products that are promoted by influencers. This leads to the following hypothesis:

H4: The attitude towards influencers will significantly affect customers’ behavioural

intention to purchase products promoted in the videos of influencers.

Just as for social influence, gender and age may play a moderating role on the effectiveness of attitude on purchase intention. Therefore, the independent variable attitude contains the following hypotheses as well:

H4a: The effect of attitude on behavioural intention is different by gender, in such a way that

the effect will be stronger for women.

H4b: The effect of attitude on behavioural intention is different by age, in such a way that the

effect will be stronger for the youngest age group.

Behavioural Intention (BI)

The dependent variable in this research is behavioural intention. The behaviour that is expected in this research is that consumers start letting videos of social media influencers be part of their purchase decision. This means that it is expected that the four independent variables influence the possibility that consumers buy products in the future after watching a video of a social media influencer. In this research the videos of social media influencers are seen as the new technology through which products are promoted and advertised.

Behavioural intention is the only dependent variable in this study, compared to the original UTAUT model that also includes actual usage of the new technology. It is out of the scope of this research to be able to check for actual purchases done by consumers after they watched a video from an influencer.

(23)

23 The dependent variable behavioural intention also includes a moderating variable namely product segments. This research includes three different product segments, namely fashion, make-up and games. The different product segments are added to research if there are any differences in the effect on behavioural intention and so that respondents of the online questionnaire could fill out the questions based on their own preferences. What is assumed is that make-up has the biggest effect on behavioural intention. This is due to the fact that for make-up there is an enormous choice of products, with a very large price-range. Due to this price range consumers want to make sure that they buy the right product, especially when it concerns a large amount of money. Videos in which influencers give reviews on the working of different brands of make-up, can be an influential method to persuade consumers to buy this product. For fashion it is more about personal taste whether you like it or not.

Furthermore, most videos for games contain a review as well, however it is expected that whether you like or dislike a game is very personal as well. Some people prefer racing games, whilst other prefer more violent games. Expecting that make-up has the highest results for behavioural intentions leads to the following hypothesis:

H5: Behavioural intention will be different amongst the product segments, in such a way that

it will be highest for make-up.

Moderating variables

The original UTAUT model includes three moderating variables, namely gender, age, experience and voluntariness of use (Venkatesh, V., Morris, M.G., Davis, G.B. & Davis, F.D. 2003). The last two are not considered for this research, since the phenomenon is still new so all respondents will not have lots of experience yet and since it is for personal use and it’s not the only source of product information it is taken into account that usage is voluntarily. This research does take into account three moderating variables, namely gender, age and product segment. Gender and age are used as moderating variables on the relation between the

(24)

24 independent variable on the dependent variable, whereas product segment is used to research whether the behavioural intention is different amongst product segments.

3. METHODOLOGY

3.1 Research design

The purpose of this research is to find out how performance expectancy, effort expectancy, social influence and attitude towards influencers are related to customer’s purchase intention. This research is a correlational study since it explores the relationship between four independent variables and one dependent variable. In order to collect data to investigate this relationship, an online survey was developed using Qualtrics survey software. The questionnaire can be found in Appendix 1.

3.2 Data collection

The questions in the online survey were based on a revised version of the UTAUT model of Venkatesh. For each variable a number of items were constructed, which were answered based on a 5-likert scale. Furthermore, the survey consisted out of three general descriptive questions namely based on gender, age and product segment. Gender was asked based on male and female. Age was based on three different age groups, namely 11-17, 18-24 and 25-31. The first age group mainly includes respondents currently at high school, the second group of 18-24 mainly includes students, and the last group mainly includes post-students whose main activity is working. The last descriptive question was based on a choice between product segments. This question was included to let respondents fill out the survey based on one product segment of which they would most likely watch a video on YouTube. Through this, respondents were better able to fill out the survey based on their own

experiences. The product segments which respondents could choose from included fashion, make-up and games. These segments were chosen, because they are three very different

(25)

25 product categories and all of the segments are very popular amongst social media influencers. After respondents chose a product segment, they were given the opportunity to watch a short video of an influencer in this chosen segment. This was done in order to let respondents remind them of what influencers do, who they are, and what the survey is based on. The survey was distributed to target Dutch inhabitants, so the included influencers are all Dutch. For fashion both a male and female influencer were included, in order to reflect on both male and female fashion styles. For male, Milan Carvalho was chosen. He has a weekly fashion talk on YouTube in which he, together with another influencer, reflects on the latest trends in the fashion world. On the 7th_{of June 2018 he had 43,656 subscribers on his YouTube channel}

(retrieved on 07-06-2018 on youtube.com). For female Yara Michels was chosen who regularly reflects posts videos in which she provides information about the latest trends for women in fashion. On the 7th_{of June 2018 she had 34,840 subscribers on her YouTube}

channel (retrieved on 07-06-2018 on youtube.com). For make-up NikkieTutorials was chosen. She posts different kind of makeup videos, including tutorials and reviews and she has millions of followers. The included video for the questionnaire was a review concerning a number of different new makeup brands. NikkieTutorials is known worldwide and she has recorded videos with celebrities like Kim Kardashian and Nicole Richie. On the 7th_{of June}

2018 she had 10,033,965 subscribers on her YouTube channel (retrieved on 07-06-2018 on youtube.com). Lastly, for games the influencer Gregor was included. He makes videos that review and provide information concerning the latest video games and appliances. In the included video he reviews the newest hardware of Nintendo. On the 7th_{of June 2018 he had}

48,352 subscribers (retrieved on 07-06-2018 on youtube.com).

The remaining part of the survey included questions on the four independent variables and one dependent variable. For every variable a number of items were developed, which were answered based on a 5-likert scale. The 5-likert scale was based on: strongly disagree,

(26)

26 disagree, neutral, agree, strongly agree. Using parametric statistics for Likert scales have been widely discussed, however according Norman ‘parametric statistics can be used with Likert data, with no fear of coming to the wrong conclusion’ (Norman, G. 2010, p. 8). The table below a summary is given on the 5 constructs, the definition, the number of items that were used and the measurement scale. The questionnaire containing all the questions can be found in Appendix 1.

Table 3

Summary of survey constructs

Construct Definition Items Measurement

Performance expectancy The performance of the videos of social media influencers, including valuable and useful information about the product.

4 Ordinal

Effort expectancy The effort that needs to be taken to in order to watch a video of social media

influencers, including whether the review is clear and mentioned when sponsored.

3 Ordinal

Social influence The influence that consumers experience within their environment to follow famous influencers and follow their

recommendations so that it’s noticed by friends and family that they follow the latest trends.

4 Ordinal

Attitude The attitude towards an influence, including whether the consumer like or dislike the social media influencer and whether he or she trusts him/her and feels related to him/her.

4 Ordinal

Behavioural intention Whether consumers intend to purchase products in the future after receiving information and recommendation via the videos of social media influencers.

(27)

27

3.3 Sample

In total 185 respondents started the online survey. These respondents were reached, mainly by convenience sampling via spreading the survey through close contacts and fellow students. After data screening six respondents were found to provide incomplete responses, which were therefore excluded from this research. This lead to a sample size of 179. The gender distribution for this research was 30,2% male and 69,8% female. The biggest number of respondents, namely 46,9% fall between the 11-17 age group. And the smallest group with 10,1% fall in the 25-31 age group. Respondents had a choice between three different product segments, and they had to choose one that was most suited to them. The biggest group chose for fashion, with 41,3%. The two remaining product segments lay very close to each other, with 29,6% for make-up and 29,1% for games. All of the respondents were from the Netherlands, since the survey was distributed in Dutch.

Table 4

Demographics

Measure Item Frequency Percentage

Gender Male Female 54 125 179 30,2% 69,8% 100% Age group 11-17 18-24 25-31 84 77 18 179 46,9% 43,0% 10,1% 100% Product segment Fashion

Make-up Games 74 53 52 179 41,3% 29,6% 29,1% 100%

(28)

28

3.4 Data analysis

In order to analyse the data obtained from the survey, all the data was exported to the statistical program SPSS. A number of analysis were performed in order to find out how performance expectancy, effort expectancy, social influence and attitude towards influencers are related to customer’s purchase intention. Furthermore, separate tests were performed in order to look for differences amongst gender and age on relation between the independent variables and the dependent variable. Lastly, a separate analysis was conducted in order to look for differences amongst the three product segments. The validity of this research is guaranteed, due to the fact that the questions are all based on scientific research on the UTAUT model and the ‘UTAUT model has been proven to be a valid research instrument’ (Tarhini, A., El-Masri, M., Ali, M. & Serrano, A. 2016, p. 832). However, when conducting a factor analysis and re-examining all the items, it is decided that the second item of

behavioural intention is removed. It is concluded that, within the context of this research, this item does not measure the behavioural intention to use social media influencers as new sources of product information.

In order to check for internal consistency for the variables, a reliability analysis via Cronbach’s alpha is conducted. All of the variables show a reliability measure above 0,5. Values above 0.8 are seen as having a high reliability and values below 0.5 are seen as having insufficient reliability (McCrae, R.R., Kurtz, J.E., Yamagata, S. & Terracciano, A., 2010). In general, values above 0.7 are seen as reliable. In this research, performance expectancy, attitude and behavioural intention have a value of around 0.7. Social influence has the highest value of 0.8, which is therefore seen as highly reliable. Effort expectancy has a value of just 0.54, however since it’s not below 0.5 it will still be considered for this research.

(29)

29 Table 5

Cronbach’s alpha

Variable Number of items Cronbach’s a

Performance expectancy 4 0,70

Effort expectancy 3 0,54

Social influence 4 0,80

Attitude 4 0,71

Behavioural intention 2 0,70

In order to check for normal distribution of the data, a look is taken at skewness, kurtosis and QQ plots plot for every variable. What can be seen in table 6 is that all of the variables have values around zero for skewness, indicating normal distribution. Desired values for kurtosis are normally around three, however, the results in table 6 show smaller kurtosis variables. This indicates a lightly tailed normal distribution. However, when taking a look at figure 2, which included the QQ plots of the independent variables, we may assume that the data is fairly normally distributed since a lot of the data falls along the line.

Table 6

Skewness and kurtosis for normality testing

Variable Skewness Kurtosis

Performance expectancy -0,402 -0,106

Effort expectancy -0,817 1,187

Social influence 0,290 -0,503

Attitude -0,202 -0,469

(30)

30 Figure 2

Q-Q Plots independent and dependent variables

4. RESULTS

This part will discuss the findings of a number of statistical tests performed in SPSS. These findings will be used in order to find out how performance expectancy, effort expectancy, social influence and attitude towards influencers are related to customer’s purchase intention. First of all, table 7 gives a summary of descriptive statistics including the mean and standard deviations of the variables. A more elaborated table of descriptive

statistics including the mean and standard deviations for every item can be found in Appendix 2.

(31)

31 Table 7

Mean and standard deviation

Construct Min Max Mean Std. Deviation N

Gender 1 2 1,70 0,460 179 Age group 1 3 1,63 0,661 179 Product segment 1 3 1,88 0,832 179 PE 1 5 2,75 0,749 179 EE 1 5 3,95 0,653 179 SI 1 5 2,31 0,852 179 ATT 1 5 2,69 0,711 179 BI 1 5 2,69 0,945 179 4.1 Correlation

The correlation analysis is used in order to investigate if there is any correlation between the variables, and especially for every independent variable on the dependent variable. The higher the correlation coefficient, the stronger the correlation between

variables. A Pearson correlation test is performed and is measured on a standard scale with a range of -1 and +1. According to Cohen a correlation coefficient of 0.10 is perceived as weak, a coefficient of 0.3 is considered moderate and a coefficient of 0.5 is large (Cohen. J. 1988). In table 8 the results of the Pearson correlational test can be seen. All of the

correlation coefficients are significant at the 0.01 level. It can be concluded that all of the independent variables, performance expectancy, effort expectancy, social influence and attitude correlate with the dependent variable behavioural intention. The strongest correlation effect on behavioural intention is for social influence, namely with 0,704, followed by

attitude with 0,587. Performance expectancy and effort expectancy are moderately

(32)

32 Table 8

Pearson Correlation Coefficients

Variable 1 2 3 4 5 1. Performance expectancy (0,70) 2. Effort expectancy 0,336* (0,54) 3. Social influence 0,454* 0,251* (0,80) 4. Attitude 0,558* 0,261* 0,664* (0,71) 5. Behavioural intention 0,344* 0,310* 0,704* 0,587* (0,70) *. Correlation is significant at the 0.01 level (2-tailed)

4.2 Multiple regression analysis

Before conducting the multiple regression analysis, a multicollinearity test is

performed. This is done in order to check for possible linear relationships between any of the independent variables. If there are linear relationships between the independent variables, it would be very difficult to assess the effects of the independent variables on the dependent variables. The table below shows the results of the test, and it indicates that there is no problem with multicollinearity, since the variables of tolerance are higher than 0,1 and VIF is below 9.

Table 9

Multicollinearity statistics

Model Variables Tolerance VIF

1 PE 0,641 1,559

EE 0,874 1,144

SI 0,546 1,833

ATT 0,476 2,103

a. Dependent variable BI

Since there is no problem with multicollinearity, a multiple regression analysis can be performed. This analysis will investigate whether there is statistical evidence in this research for the relationship between the independent variables performance expectancy, effort expectancy, social influence and attitude and the dependent variable behavioural intention.

(33)

33 The results of the multiple regression analysis can be found in table 10, 11 and 12 below. First of all, when looking at R Square, it can be concluded that 73,5% of the variance of behavioural intention is explained by performance expectancy, effort expectancy, social influence and attitude.

Table 10

Model summary

R R Square Adjusted R Square Std. Error of the

Estimate

0,801a 0,735 0,529 0,64887

a. Predictors: PE, EE, SI, ATT

Furthermore, when looking at the ANOVA results in table 11 we can conclude that the model is significant, since the p-value is below 0.01.

Table 11 ANOVAa

Sum of Squares

df Mean Square F Sig.

Regression 85,841 4 21,460 50,971 0,000b

Residual 73,259 174 0,421

Total 159,101 178

a. Dependent variable: BI b. Predictors: PE, EE, SI, ATT

Lastly table 12 shows the coefficient results of the multiple regression analysis. What is striking is that for performance expectancy no statistical evidence can be found. Performance expectancy shows a p-value far above 0,05. This means that the performance expectancy of the videos of social media influencers do not have a statistically significant effect on

behavioural intention. The remaining three independent variables, effort expectancy, social influence and attitude do show a statistical evidence. Effort expectancy shows a p-value

(34)

34 below 0,05, and a Beta of 0,139. Both of remaining independent variables, social influence and attitude have a p-value below 0.01. The highest Beta, with 0,555 is for social influence, what means that this independent variable has the biggest effect on behavioural intention. The Beta of attitude is a bit lower with 0,229, but it does show a statistically significant effect on behavioural intention.

Table 12

Multiple regression analysis

Variable β Sig. (Constant) 0,857 PE -0,082 0,201 EE 0,139 0,012 SI 0,555 0,000 ATT 0,229 0,003 a. Dependent variable: BI 4.3 Moderating variables

Since the overall multiple regression analysis did not find any statistical evidence for one out of four independent variables, a more detailed analysis is done to find possible explanations for this. Therefore, a separate regression analysis is performed for gender, age groups and product segments.

4.3.1 Gender

First of all, a regression analysis is conducted for the differences between male and female. The results are presented in table 13 below. It can be concluded that for both male and female no statistical evidence is found for the independent variable performance expectancy. For effort expectancy, statistical evidence for the relationship on behavioural intention is only found for male, with a p-value just below 0,05 and a Beta of 0,199.

(35)

35 the highest Beta of 0,534 for female. Lastly, for attitude only statistical evidence is found for female, with a p-value below 0,01 and a Beta of 0,282.

Table 13

Regression analysis for gender Male Variables β Sig. Constant 0.874 PE_Male -0,087 0,421 EE_Male 0,199 0,049 SI_Male 0,518 0,000 ATT_Male 0,256 0,070 a. Dependent variable: BI Female Variables β Sig. Constant 0.841 PE_Female -0,085 0,308 EE_Female 0,097 0,158 SI_Female 0,534 0,000 ATT_Female 0,282 0,003 a. Dependent variable: BI

For all of the moderating variables a closer look is also taken to the descriptive statistics for behavioural intention. In table 14 below the results can be seen for gender. What can be concluded is that the mean for male is slightly higher than for female. This indicates that the behavioural intention to use social media influencers as a new source of product information to make a help making a decision concerning purchase intention is slightly higher for male, with a mean of 2,84, compared to a female with a mean of 2,63.

(36)

36 Table 14

Descriptive statistics for gender

Gender N Mean Std. Deviation

Male 54 2,84 0,98

Female 125 2,63 0,93

4.3.2 Age

Next to a separate regression for gender, this part shows the different regressions based on age. For this research, three different age groups are used, namely 11-17, 18-24 and 25-31. The results can be found in table 15 below. Amongst all the different age groups there is again no statistical evidence found for how performance expectancy is related to

customers’ behavioural intention. Furthermore, what is interesting to see is that for the oldest age group of 25-31, there is also no statistical evidence for effort expectancy, social influence and attitude, since the p-values are all higher than 0,05. This means that for the age-group of 25-31 years old no statistical evidence is found that the independent variables are related to behavioural intention. When looking at the youngest age group it can be concluded that for the three remaining independent variables, statistical evidence is found, with the highest Beta of 0,543 for social influence. The middle age group, shows statistical evidence for social influence and attitude. This age group furthermore, shows compared to the other age groups, the highest Beta of 0,578 for social influence and the highest Beta of 0,252 for attitude of all. Table 15

Regression analysis for age groups 11-17 Variables β Sig. Constant 0.434 PE_11-17 -0,069 0,426 EE_11-17 0,158 0,043 SI_11-17 0,543 0,000 ATT_11-17 0,247 0,021 a. Dependent variable: BI

(37)

37 18-24 Variables β Sig. Constant 0.835 PE_18-24 -0,094 0,375 EE_18-24 0,120 0,156 SI_18-24 0,578 0,000 ATT_18-24 0,252 0,034 a. Dependent variable: BI 25-31 Variables β Sig. Constant 0.702 PE_25-31 0,012 0,965 EE_25-31 0,038 0,881 SI_25-31 0,257 0,370 ATT_25-31 0,400 0,202 a. Dependent variable: BI

Furthermore, when taking a look at the descriptive statistics of behavioural intention for the variable age in table 16, it can be seen that the highest mean is for the youngest age group. It has a mean of 3,056 and a standard deviation of 0,960. The age group of 18-24 follows it and the age group of 25-31 has the lowest mean with 2,607. These results indicate that the behavioural intention to purchase products promoted by social media influencers is highest for the age group of 11-17.

Table 16

Descriptive statistics for age

Age N Mean Std. Deviation

11-17 84 3,056 0,960

18-24 77 2,701 0,919

(38)

38

4.3.3 Product segment

Lastly a multiple regression analysis is conducted for the different product segments, in order to investigate whether there are differences in statistical relationships between the independent variables and the dependent variable amongst fashion, make-up and games. The results can be seen in table 17 below. What can be concluded is that if we split the data between the three different product segments, also no statistical evidence is found for performance expectancy, since again the p-values are above 0,05. For effort expectancy, statistical evidence is only found for the category games. For social influence statistical evidence is found for every product segment since all p-values are even below 0.01. For social influence the by far the highest Beta is for fashion, namely with a value of 0,686, compared to 0,475 for make-up and 0,423 for games. For attitude, statistical evidence is found for the segments fashion and make-up, with make-up having the highest Beta of 0,475. Table 17

Regression analysis for product segment Fashion Variables β Sig. Constant 0.450 PE_Fashion -0,139 0,167 EE_Fashion 0,011 0,895 SI_Fashion 0,686 0,000 ATT_Fashion 0,244 0,028 a. Dependent variable: BI Make-up Variables β Sig. Constant 0.693 PE_Make-up -0,197 0,155 EE_Make-up 0,181 0,115 SI_Make-up 0,475 0,000 ATT_Make-up 0,399 0,007 a. Dependent variable: BI

(39)

39 Games Variables β Sig. Constant 0.994 PE_Games -0,080 0,543 EE_Games 0,281 0,024 SI_Games 0,423 0,007 ATT_Games 0,151 0,349 a. Dependent variable: BI

Next to seeking the different effects for the product segments on the relationship between the independent variables and the dependent variable, also a closer look is taken whether there are differences in behavioural intention between the product segments. Table 18 below shows the mean and standard deviations for every product segment for behavioural intention. It shows that the highest mean, of 2,859, is resulted for make-up, which indicates that according to the obtained data people will most likely be influenced to buy products promoted by influencers for the segment make-up, closely followed by fashion with having a mean of 2,750. The lowest mean is for games, of 2,442, which indicates that consumers are least likely to purchase products in the segment of games when they received the information and recommendations from social media influencers.

Table 18

Descriptive statistics product segment

Product segment N Mean Std. Deviation

Fashion 74 2,750 0,857

Make-up 53 2,859 0,906

(40)

40

5. DISCUSSION

5.1 Practical Implications

The data obtained via the online questionnaire have shown some interesting findings and have some interesting results for the hypotheses. This research contains in total thirteen different hypotheses, of which seven are accepted and six are rejected. All of the hypotheses that include the relationship between performance expectancy and behavioural intention are rejected. For effort expectancy, social influence and attitude part of the hypotheses are accepted. The final hypothesis looked at the behavioural intention for different product segments and is accepted. The practical implications of this research will be structured in such a way that every variable will be discussed separately. Afterwards, the theoretical implications will be discussed.

5.1.1 Performance expectancy

The first hypothesis for performance expectancy is ‘Performance expectancy will significantly affect customers’ behavioural intention to purchase products promoted in the videos of influencers’. It can be seen from the multiple regression analysis, that there is no statistical evidence for this relationship. This means that we cannot conclude that the level of performance expectancy is a good predictor for behavioural intention. Even when people value a high-performance expectancy for videos of social media influencers, this does not mean that this also positively impacts consumers’ behavioural intention to let these videos be part of their decision process to buy a product or not.

The other two hypotheses included gender, ‘The effect of performance expectancy on behavioural intention is different by gender, in such a way that the effect will be stronger for women’ and age-group, ‘The effect of performance expectancy on behavioural intention is different by age, in such a way that the effect will be stronger for the youngest age group’. Also, for these hypotheses no statistical evidence was be found. This means that for both men

(41)

41 and women, and for all of the different age groups, performance expectancy is not a

significant indicator for behavioural intention. These are striking results, since Venkatesh et al. developed the model in which performance expectancy is an important indicator for user acceptance and behavioural intention towards a new technology. The fact that only for

performance expectancy no statistical evidence is found is interesting. The reason for this can be that consumers mainly use social media influencers due to the social influence, rather than for actual reviews. Consumers, right now, might still trust their friends/family’s opinion more than the opinions from social media influencers. Furthermore, consumers might still use companies’ website to search for product information, rather than watching social media influencers provide this information.

5.1.2 Effort expectancy

The second hypotheses were based on effort expectancy. The first hypothesis is: ‘Effort expectancy will significantly affect customers’ behavioural intention to purchase products promoted in the videos of influencers’. After analysing the data this hypothesis can be accepted, which means that there is statistical evidence for the relationship between effort expectancy and behavioural intention. The less effort that has to be taken to watch and understand a video from social media influencers, the more likely that consumers follow the recommendations from these influencers. This will in the end affect consumers purchase intention to buy products that are promoted via videos from social media influencers.

Furthermore, when looking at the corresponding hypothesis for gender: ‘The effect of effort expectancy on behavioural intention is different by gender, in such a way that the effect will be stronger for women’, it can be concluded that this hypothesis is rejected. There is statistical evidence found for male, but not for female. This shows that there is a relation for effort expectancy on behavioural intention for male consumers, but not for female.

(42)

42 Lastly, when looking at the differences in gender, the corresponding hypothesis is: ‘The effect of effort expectancy on behavioural intention is different by age, in such a way that the effect will be stronger for youngest age group’. After analysing the results this hypothesis can be accepted. So, the youngest age group shows the strongest relationship between effort expectancy and behavioural intention.

So, in the case of social media influencers this result show that the effort taken of watching videos, of social media influencers, and to be able to understand them and take away the key message, is an important factor for future purchase intentions. So, when videos are very clear, easy to find and clearly state when it is sponsored this may lead to a

behavioural intention to follow the recommendations made in the videos of social media influencers and let this be an important factor for future purchase intentions.

5.1.3 Social influence

The third hypothesis took into account the social influence of the customer’s environment that might affect behavioural intention, this lead to the following hypothesis: ‘Social influence will significantly affect customers’ behavioural intention to purchase products promoted in the videos of influencers’. What can be seen in the results, is that for this relationship statistical evidence is found. This means that when customers experience social pressure of the surrounding environment to follow the latest trends and to buy the newest products, this may lead to the behavioural intention that customers use social media influencers as their source of product information and to take their recommendations into account for future purchase intentions. Social influence mainly focuses on the social pressure of belonging to a certain group and to have products that are conform the latest trends. So, in people’s purchase intentions they find it important to follow the recommendations of social

(43)

43 media influencers who have a lot of followers and who are seen as important information sources of the newest products and trends.

For social influence there is also a corresponding hypothesis that compares male and female: ‘The effect of social influence on behavioural intention is different by gender, in such a way that the effect will be stronger for women’. From the results we can conclude that we can accept this hypothesis. The statistical evidence is both significant for both male and female, however the Beta for female is higher, which indicates that this effect is stronger. This means that women find the social influence from their environment to use social media influencers as the new source of product information more important in their purchase intention, than this is the case for men. Furthermore, women find it more important to follow influencers who have a lot of followers and who promote the latest products, compared to men.

The third hypothesis concerning social influence looked at the differences between the different age groups, which lead to the following thesis: ‘The effect of social influence on behavioural intention is different by age, in such a way that the effect will be stronger for the youngest age group’. From the results it can be analysed that this hypothesis is rejected, since the strongest effect is found for the age group of 18-24. This second age group of

respondents, which mainly includes students, show that for them the social pressure to follow social media influencers and purchase what they promote is highest compared to the other age groups. However, difference in Beta compared to the youngest age-group is very small. For, the age-group of 25-31 there is no statistical evidence for the relationship between social influence and purchase intention. This means that we cannot conclude that there is even an effect of social influence on purchase intention in this age group, so they don’t experience any pressure to do same as what influencers do or what their environment expects, and

Effectiveness of social media effectiveness of social media : an application of the UTAUT model