• No results found

EWOM Source, Source Credibility and Contextual Relevant Advertising on Facebook What are the effects on Brand Attitude?

N/A
N/A
Protected

Academic year: 2021

Share "EWOM Source, Source Credibility and Contextual Relevant Advertising on Facebook What are the effects on Brand Attitude?"

Copied!
75
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

EWOM Source, Source Credibility and Contextual Relevant

Advertising on Facebook

What are the effects on Brand Attitude?

By

Steven Wolma

(2)

2

EWOM Source, Source Credibility and Contextual Relevant Advertising on Facebook

What are the effects on Brand Attitude?

Master thesis, MscBA, specialization Marketing Management

University of Groningen, Faculty of Marketing

Words: 14.530

December 15, 2013

Steven Wolma Studentnumber: 1542176 A Jacobsstraat 64 9728 MD, Groningen Tel: +031(0)6-46592263 E-mail: stevenwolma@hotmail.com Supervisor: Dr. J.C. Hoekstra

(3)

3

Management summary

Technical developments have been rapidly changing the world over the last two decades. Marketing has also been influenced by these technological developments. The development of the internet is one of the most important technological developments in communication. Social media is another

phenomenon that changed the world as we know it today. The current leading social network site is Facebook with more than 950 million users around the world. By creating e.g. Facebook fan groups, using Twitter to inform consumers about the latest news and using famous bloggers to write about their newest innovations, companies are slowly realizing the value of social media for their marketing

strategies. Using social media as a marketing communications tool is a much cheaper and faster way to reach target groups compared to traditional marketing. When people have established relationships in social media, they are more prone to referrals, and so create electronic word-of-mouth (eWOM). Social media creates the possibility for eWOM. EWOM is considered to be more influential on consumers than any regular commercial. The internet is used more and more for searching recommendations and information by consumers. It therefore might provide the right tools for effective eWOM.

Banner advertisements generate the most of the income for social media sites. In this study a further investigation will be made to research factors that might influence banner advertising effectiveness on Facebook. Online effectiveness studies by academics and practitioners have been studied through two paradigms. In this study the focus will be on the paradigm that looks into brand attitude. Online targeting is a new trend that is developing in the online advertising market. This can be done in many ways, for example by Online Behavioral Targeting (OBA). Facebook offers a form of OBA known as contextual advertising, which is based on what users list in their profiles and what they like. Contextual advertising on Facebook is similar to contextual advertising offered by search engines. This, because it is based on direct input rather than inferred behavior.

This study will look at what matching an advertisement to eWOM, created by a consumer or a business on Facebook, does for brand attitude. Furthermore, this study will look at what important factors are that influence the role of eWOM on brand attitude like e.g. the source and the credibility.

The main problem in this study is: What effect does eWOM source, source credibility and contextual

relevant advertising on Facebook have on brand attitude? In this research the following variables are

(4)

4 This study is researched by conducting a field experiment. The results of this study show that the role of contextual advertising on brand attitude is very limited. Important factors influencing the role of eWOM on brand attitude are the source and its credibility. The test in this research if friend eWOM is more effective in raising brand attitude than business eWOM does not show large differences. The biggest influencer of eWOM on brand attitude is the credibility of the eWOM message. This research shows signs that, when you are allowed to post messages on a consumers Facebook timeline, this consumer sees you as a friend. Whether, you are a friend, a peer consumer or a company.

This research furthermore shows that an important factor in raising brand attitude, using eWOM and contextual advertising, is source credibility. Furthermore, it is recommended for companies to target consumers who are familiar with their product. When a company is able to create a connection with these consumers, a company is able to use Facebook, eWOM and contextual advertising in the most effective way to raise brand attitude. Creating strong bonds with a consumer on Facebook leads to two things:

(5)

5

Table of content

Contents

Table of content ... 5 1 Introduction ... 7 1.1 Background ... 7

1.2 Social media in marketing ... 7

1.3 Advertising in social media. ... 8

1.4 Problem statement and research questions ... 10

1.5 Structure of the study ... 11

2 Theoretical Background ... 12

2.1 Social Media ... 12

2.1.1 What are social media ... 12

2.1.2 Building relationships with consumers ... 13

2.2 Electronic Word-of-Mouth ... 13

2.2.1 EWOM ... 13

2.2.2 Benefits of eWOM ... 14

2.3 Advertising in Social Media ... 15

2.3.1 Social Media advertising ... 15

2.3.2 EWOM messages ... 16

2.3.3 Banners ... 17

2.3.4 Effectiveness of banner advertisements... 18

3 Conceptual model & hypotheses ... 19

3.1 Conceptual Model ... 19

3.2 Brand Attitude... 20

3.3 Source Credibility ... 21

3.4 EWOM Source ... 22

3.5 Banner Context Relevance ... 23

3.6 Product Involvement... 25

3.7 Daily Facebook Usage ... 27

(6)

6

3.7.2 Frequency of Internet usage ... 27

4 Empirical research design ... 28

4.1 Research design ... 28

4.2 Stimuli ... 29

4.3 Participants ... 31

4.4 Variables & scaling ... 32

4.5 Procedure ... 34

4.6 Plan of Analysis ... 36

5 Results ... 38

5.1 eWOM Source and Source Credibility ... 40

5.2 Product Involvement moderating eWOM Source and Source Credibility ... 40

5.3 Source Credibility and Brand Attitude ... 42

5.4 eWOM Source and Brand Attitude ... 43

5.5 Banner Context Relevance and Brand Attitude ... 43

5.6 eWOM Source combined with Banner Context Relevance and Brand Attitude ... 44

5.7 Product Involvement moderating Banner Context Relevance and Brand Attitude ... 47

5.8 Product Involvement moderating eWOM Source and Brand Attitude ... 48

5.9 Daily Facebook Usage moderating Banner Context Relevance and Brand Attitude ... 49

5.10 Complete model analysis for brand attitude plus the addition of the new variable brand previous purchases ... 52

5.11 Hypotheses support results summary ... 55

6 Conclusion and recommendations ... 56

6.1 Conclusion ... 56

6.2 Recommendations ... 58

6.3 Limitations and further research ... 59

References: ... 60

Appendix A: Facebook page eWOM Asics, banner Asics ... 68

Appendix B: Facebook page eWOM Asics, banner BMW ... 69

Appendix C: Facebook page eWOM Marcel Wolma, banner Asics ... 70

Appendix D: Facebook page eWOM Marcel Wolma, banner BMW ... 71

(7)

7

1 Introduction

1.1 Background

Over the last two decades, technical developments have been rapid and have changed the world. Technological developments have also influenced marketing. One of the most important technological developments in communication is the development of the internet and its endless possibilities (Stingl, 2009). Along with the rapid expansion of the web came another phenomenon that changed the world as we know it today called social media. According to Agichtein et al. (2008) social media are user

generated, so users are participating not only in consumption, but also in content creation. With social media one can publish and share information, discuss with others, network and much more. Examples of social media tools are Facebook, MySpace, YouTube, and Twitter. Facebook is the current leading social network site with more than 950 million users around the world (Facebook 2012(1)).

1.2 Social media in marketing

(8)

8 Utpal & Durham (2010) researched the effects of Facebook for a café chain, and their results showed how it changed consumer behavior. Consumers who were members of the Facebook group generated positive eWOM more easily, than consumers who were not in the group (Utpal & Durham., 2010). Being visible in the right places in social media helps organizations to interact with consumers in real time, and this way encourages eWOM. Visibility in social media can be enhanced by joining several networks, by investing in good quality photos, by tailoring the sent messages to fit the current mission, being interactive, up to date and prompt when answering consumer questions (Noella, 2009).

Consumers use the internet more and more for searching recommendations and information. It therefore might provide the right tools for effective eWOM (Li et al., 2006)

So, with a growing group of consumers exploring social media and showing interest towards companies in that environment, it is a good time to start exploring the possibilities that lie in social media for companies to use this eWOM and combine it with advertising to raise brand attitude.

1.3 Advertising in social media.

(9)

9 A new trend that is developing in the online advertising market is called online targeting. Online

targeting can be done in many ways for example by Online Behavioral Targeting (OBA). Facebook itself offers a form of OBA known as contextual advertising, which is based on what users list in their profiles and what they like. Contextual advertising on Facebook is similar to contextual advertising offered by search engines, as it is based on direct input rather than inferred behavior.

Facebook offers businesses the opportunity to promote their products, services and brands to their fans using services such as Facebook Pages and Facebook Ads. Facebook pages allow fans the opportunity to engage with a business and its brands in a variety of ways, whereas Facebook Ads can tell you about how ads can be targeted at very specific groups of users. In this article we will concentrate on Facebook Ads. Facebook’s database provides a company to segment customers by location of customer, by language, according to demographic characteristics and according to some personal factors (education, workplace, interests, likes…) (Facebook., 2012 (2)). This research targets either a contextual relevant or irrelevant banner to a eWOM message from either a company or a users’ friend.

This narrow segmentation of targeting customers increases advertising efficiency and effectiveness. Furthermore, according to Goldfarb and Tucker (2011) matching an ad to website content and increasing an ad’s obtrusiveness independently increases the success of the advertisement. People can however become skeptical when being targeted. If the tactic will be perceived as

manipulative, it will have a negative effect on consumer perceptions of the product being advertised. Given that deception is particularly easy online, consumer awareness of manipulation is higher online too (Boush et al. 2009). There is however, a relative high consumer tolerance to targeted ads because the information is perceived as useful (e.g., Cho and Cheon 2003, Edwards et al. 2002, Wang et al. 2008). Gruen at al., (2005) argue that eWOM is more credible, emphatic and relevant for consumers than other sources on the web. It is essential that the received information is perceived to be credible. This helps maintaining trust between the organization and the customer. So, in order for eWOM to be successful the source has to be credible and the argumentation strong. When these aspects are in place, eWOM can influence the customer purchase decision (Cheung et al., 2009).

(10)

10

1.4 Problem statement and research questions

In scientific literature there is still a lot to gain from research into the possibilities of social media for marketing. Therefore, this article is encouraging researchers to dig on the topic deeper in the near future. Besides helping to market the company, eWOM is also more and more important for a company’s reputation management (Kane et al., 2009). Targeting advertisements based on a users’ profile information and interests is happening more and more. The combination of eWOM and targeting advertisements to the contents of a eWOM message is something relative new. The effects of the contextual relevancy of advertisements on brand attitude are a linkage that did not get much attention yet. The scientific relevance of this study aims at giving more insight into the usefulness of advertising and eWOM on brand attitude. The practical relevance focuses on giving new directions for companies marketing strategy, and helps them find the right tools to increase brand attitude using social media. In this study the main problem statement is: What effect does eWOM source, source credibility and

contextual relevant advertising on Facebook have on brand attitude?

To be able to understand the phenomenon of social media, eWOM, advertising and brand attitude the following research questions should be answered:

1. How does the source of eWOM messages affect brand attitude? 2. How does source credibility affect brand attitude?

3. How does contextual relevant advertising affect brand attitude?

4. How does contextual relevant advertising influence the effect of the source of eWOM messages on brand attitude?

(11)

11

1.5 Structure of the study

In chapter 2 the theoretical background of this study is presented. Chapter 3 presents the conceptual model, giving a graphical image of the relationships found in literature. And furthermore shows how, by reviewing existing literature on the topic, hypotheses are derived. Following, in chapter 4, the research design describes the methods used, the data collection and plan of analysis. After this, the results will be presented in chapter 5. Finally, in chapter 6, the problem statement is being answered by giving

(12)

12

2 Theoretical Background

In this section social media, electronic word-of-mouth and advertising on social media are being defined and explained more thoroughly. This is done by making use of existing scientific literature on these subjects.

2.1 Social Media

Since the use of social media in marketing is considerably new and companies are still learning and experimenting with it, its research is still in its starting point (Harridge-March & Quinton, 2009). The possibilities of social media are vast, but companies should not jump into it blindly and use everything they can. Companies should know how their consumers are using social media and how they respond to it (Cheung & Lee, 2010). Therefore, companies should begin at the basics, and find out thoroughly what social media entails.

2.1.1 What are social media

Kaplan & Haenlein (2010) define social media as: digital: real time and traceable online tools that lack geographical boundaries and give the possibilities to connect with other people. They bring the consumer closer to the company and at the same time give more power to the consumer. Kaplen & Haenlein (2010) furthermore classified social media, by using two dimensions: 1) social presence/media richness and 2) self-presentation/ self-disclosure. Here, social presence is the acoustic and visual contact that can be achieved. The goal of having a high social presence is that it can lead to a large social

influence on the behavior of the participants. Media richness is the amount of information that is transmitted in a certain time. The higher the amount of information transmitted, the more effective ambiguity and uncertainty are resolved. Self-presentation refers to the self- image one desires to give to others and whether it is consistent with ones’ personal identity. Self-disclosure relates to

(13)

13 2.1.2 Building relationships with consumers

An interesting feature (for companies) of social media is that consumers who use social media tools can connect with the company via several tools (e.g. Facebook group, Blog, YouTube videos) which can make consumers feel part of the companies community (Kane et al., 2009). This gives better possibilities for companies to advertise, promote products and services, and transfer knowledge with the possibility of filtering information. Furthermore, there has been prove of the fact that often, the connection between current and potential consumers places high significance on the direct marketing results; so companies can use social media networks to learn about current customer behavior and this way plan new

strategies to attract others (Hill et al., 2006) So, well planned social media marketing enables companies to be close to their customers around the world, to respond to their questions faster, to promote new products and services and to transfer information directly.

With the right social media strategies companies can guide consumers’ social media behavior towards electronic word-of-mouth creation. Due to the bi-directionality and the fact that social media function in real time, consumers can instantly react and communicate with the company desiring further details (Kane et al., 2009). With frequent updates of the products and services, consumers can keep up with the latest trends, and this way might be more willing to continue making purchases with the company. So, clearly social media, when integrated with marketing strategy, can be used for electronic word-of-mouth creation.

2.2 Electronic Word-of-Mouth

2.2.1 EWOM

(14)

14 not know the person making the recommendation. Yet this electronic word-of-mouth (eWOM) is getting more and more popular (Cheung et al., 2009).

Social media can work as a vehicle to eWOM, since users are more sensitive to peoples’ opinions and influence (Cheung & Lee, 2010). Particularly online social networks can bring business value and opportunities, but the members should be kept involved. Companies can enhance positive eWOM by being part of the different communities in social media and contribute by answering consumers’ questions and creating value by keeping them updated. This can result in positive eWOM, and in new consumer acquisitions.

2.2.2 Benefits of eWOM

Word-of-mouth marketing has been even called the world’s most effective marketing strategy (Trusov et al., 2009). WOM strategy is overcoming consumer resistance, with low costs and fast delivery. Besides helping to market the company, eWOM is also more and more important for a company’s reputation management (Kane et al., 2009). So, the benefits of eWOM are: decrease in costs, fast delivery, diminishing consumer resistance and further developing ones’ business by using the word in the web. After understanding the benefits of eWOM, it is important for companies to know what the antecedents of eWOM are and what motivates consumer participation.

According to Engel et al. (1969) reasons why consumers generate WOM are because they want to help other consumers in their decision making, they want to warn about bad service or quality and

recommend good ones, or they just want to show their knowledge on the subject.

(15)

15

2.3 Advertising in Social Media

2.3.1 Social Media advertising

Social networks enable users to be part of a large, international community and to share information, opinion and content with other members of the network. This sharing process companies see as their opportunity to advertise and hope to find a new way to use these communities (Zeng et al, 2009, p 1). Social networks as a medium of direct marketing help companies target individual consumers, by using databases companies adjust personalized offer to customers. Advertisers pay attention to social networking sites and are using the content from users’ profiles to target ads individually tailored to consumer. These ads can be based on general demographic, geographic, psychographic information to deliver personal (individual) ads (Grubbs Hoy & Milne, 2010, p 28). Even though, marketers are still too afraid that they intrude users’ privacy and their personal space, so they do not take advantage of this medium as much as they could. But social networking platforms encourage users to post and share personal information as part of their online social interactions (Grubbs Hoy & Milne., 2010, p 29). The core of social networking sites consists of personalized user profiles. Individual profiles are usually a combination of users’ images, posts and lists of interests (hobby, music, book, and movie preferences) which makes useful source for database marketers. Different sites impose different levels of privacy in terms of what information is revealed through profile pages to non-affiliated visitors and how far “strangers” versus “friends” can traverse through the network of a profile’s friends [Trusov et al, 2009, pp 92- 93].

Facebook offers businesses the opportunity to promote their products, services and brands to their fans using services such as Facebook Pages and Facebook Ads. Facebook Pages allow fans the opportunity to engage with a business and its brands in a variety of ways, whereas Facebook Ad can tell you about how ads can be targeted at very specific groups of users. When a Facebook users ‘likes’ a company’s

(16)

16 2.3.2 EWOM messages

EWOM messages on Facebook are often product reviews, or promotions of products, services or brands. As explained before these messages can originate from a Facebook user’s friend or a company, and are placed on a users’ Timeline. Examples of business eWOM messages and friend eWOM messages are shown in figures 1, 2 and 3 below.

Figure 1 example of business eWOM on Facebook

(17)

17 Figure 3 example of friend eWOM engaged with business eWOM on Facebook

2.3.3 Banners

(18)

18 Figure 4 Examples of banner advertisements

2.3.4 Effectiveness of banner advertisements

Measuring banner click-through rates has become important for both the advertiser and the host website. By tracking click-through rates, they are able to determine the effectiveness of their banner advertisements (Cho, 2003). The effectiveness of banner advertising is however an ongoing debate in literature. Despite initial success, there are arguments that banners fail to engage consumers. According to Fischler (1998), the number of people who ‘never’ look at banners jumped from 38% in 1997 to 48% in 1998, while the number of people who ‘(very) often’ look at banners dropped from 16% to 9% respectively. Research from Benway and Lane (1998), Pagendarm and Schaumburg (2001), Norman (1999), and Bayles (2000) concluded that the tendency to ignore or avoid banner ads (‘banner blindness’) is becoming an important issue in the online advertising industry. Further, research from MediaPost (2001) showed that the click-through rate dropped from 2% in 1998 to 0.33% in 2001. This indicates that the effectiveness of banner ads is decreasing. However, a high click-through rate might not be the only indication of a banner’s success. The goal of banner ads are twofold: (1) generating direct responses (e.g. drawing consumers into a company’s site, purchasing) and (2) obtaining brand recognition through banner exposure (Cyberatlas, 1999; Maher, 1999). In other words, the banner ad itself does a significant amount of brand enhancement even without being clicked (Cho, 2003).

However, the second goal of banners is a concept that rarely gained research attention yet. Therefore, this study aimed to fill this gap.

(19)

19

3 Conceptual model & hypotheses

In this chapter the conceptual model of this study is presented. Followed by the explanation of the variables and how the hypotheses are build up.

3.1 Conceptual Model

The conceptual model below (See figure 5) graphically represents all variables and their relationships that are researched in this study. It is expected that the source of the eWOM message, in this study business or friend, influences source credibility and brand attitude. Friend eWOM is expected to be more credible. Being more credible is expected to result in a higher brand attitude score as well. Having a contextual relevant banner next to the eWOM feed enhances the effect of the feed on brand attitude. Product involvement is expected to play a mitigating role between; eWOM source and source credibility, eWOM source and brand attitude, banner context relevance and brand attitude. Furthermore, it is expected that the more time consumers spent on Facebook per day, the lower the effects of the contextual relevancy of banners will be.

(20)

20

3.2 Brand Attitude

As already mentioned in chapter 2, the goal of banner ads are twofold: (1) generating direct responses (e.g. drawing consumers into a company’s site, purchasing) and (2) obtaining brand recognition through banner exposure (Cyberatlas, 1999; Maher, 1999). In other words, the banner ad itself does a significant amount of brand enhancement even without being clicked (Cho, 2003).

One important reason for considering brand attitudes to be a brand association is that they can vary in strength (Farquhar, 1989). “Attitude strength has been measured by the reaction time needed to

evaluative queries about the attitude object”(Fazio et al., 1986). Individuals with a quick evaluation of an attitude object are assumed to have a highly accessible attitude. Research has shown that attitudes formed from direct behavior or experiences are more accessible than attitudes based on information or indirect forms of behavior (Fazio and Zanna, 1981). Highly accessible brand attitudes are more likely to be activated spontaneously upon exposure to the brand and guide subsequent brand choices (Berger and Mitchell, 1989; Fazio, Powell, and Williams, 1989). Next to that, it is investigated that consumer with a strong, favorable brand attitude should be more willing to pay premium prices for the brand (Starr and Rubinson, 1978). Similarly, a positive image should result in increased consumer search (Simonson, Huber, and Payne, 1988) and a willingness to seek out distribution channels for the product or service. In general, attitudes can be stable evaluations (attitudes for a long time) and attitudes can be formed and influenced by situational factors. Petty, Cacioppo, and Schumann (1983) suggested that social and consumer psychology theories emphasize one of two distinct routes to attitude change: One, called the ‘central route’, views attitude change as resulting from a person's diligent consideration of information that (s)he feels is central to the true merits of a particular attitudinal position. A second group of theoretical approaches to persuasion emphasizes a more ‘peripheral route’ to attitude change. Attitude changes that occur via the peripheral route do not occur because an individual has personally

(21)

21

3.3 Source Credibility

Gruen at al., (2006) argue that eWOM is more credible, emphatic and relevant for consumers than other sources on the web. It is essential that the received information is perceived to be credible. This helps maintaining trust between the organization and the customer. Word-of-mouth also has an effect on customer value. Customers that are acquired via WOM tend to be more valuable on the long term when compared to customers acquired via traditional marketing induced methods (Villaneuva et al., 2008) So, in order for eWOM to be successful the source has to be credible and the argumentation strong. When these aspects are in place, eWOM can influence the customer purchase decision (Cheung et al., 2009). Cheung et al., (2009) say that eWOM connects different consumers and widens and opens up the WOM network to the entire Internet world. They also say that the sheer amount of information compromises its validity. This is why source credibility is very important for eWOM to work (Cheung et al., 2009). Doh & Hwang (2009) proved that a few negative messages among the positive ones could actually help the message to be perceived as more credible. They explain this by saying that if consumers only see the positive messages about the company, they might start doubting the credibility, since no company is perfect. Here it is assumed that source credibility has an effect on eWOM creation. So important factors in eWOM creation are source credibility and mutual trust; without these two factors consumers don’t see any reason to believe the recommendations (Cheung et al., 2009).

Following the more ‘peripheral route’ to attitude change we can say that brand attitude can be changed by reading a message about a brand. A positive or negative eWOM message will thus have a positive or negative effect on brand attitude. In this research we make use of a positive eWOM message. Therefore, the following hypotheses are conducted:

(22)

22

3.4 EWOM Source

Consumer Generated Media (CGM) is a gift to advertisers because we've been struggling to find -- and in some cases develop -- compelling content that both captivates attention and harmonizes with ad messaging. Consumers are literally creating it themselves (Blackshaw., 2006).

Blackshaw (2006) continues that, CGM also brings to the table several coveted building blocks that advertisers let slip a long time ago: trust, credibility, authenticity and, often, restraint. CGM thrives because consumers trust other consumers more than advertisers, period. But jumping on the CGM bandwagon in this environment of consumer control presents real risks and liabilities -- and could needlessly alienate consumers. It's unfamiliar terrain and credibility is at stake if marketers and agencies jump into the space recklessly. It's about listening to real and authentic conversations and using that learning and insight to be become a better, more effective marketer.

Interpersonal communication about products and services between consumers – is one of the most influential sources of marketplace information for consumers (Arndt 1967; Alreck & Settle 1995). It is so influential because consumers generally trust peer consumers more than they trust advertisers or marketers (Blackshaw 2006; Sen & Lerman 2007).

It is assumed that the source of the eWOM message has an effect on brand attitude. Whereas the effect of the eWOM message on brand attitude will be stronger for friend eWOM, than for business eWOM. Hence the following hypotheses are:

H2: Friend eWOM is perceived to be more trustworthy than business eWOM.

(23)

23

3.5 Banner Context Relevance

Consumers are exposed to advertisements in the context of editorials or ad vehicles (websites) rather than as stand-alone messages. It is concluded that the editorial environment of a vehicle significantly influences the effectiveness of advertisements placed on the vehicle (Cho, 2003). Before presenting literature about the subject of contextual relevance of online advertisements and its expected influence on brand attitude, it is important to get an understanding of the phenomenon of ‘context relevance’. Jeong and King (2010) stated that “contexts might refer to any number of form or content qualities associated with specific advertising mediums that are of interest to researchers”. Advertising relevancy is defined as “the degree to which information in the ad contributes to the identification of the primary message communicated by the number of form or content qualities associated with a website. Hereby, banners can be divided into two types: (1) contextual relevant banners and (2) contextual irrelevant banners, based on the context of websites.

The effects of the context relevancy of advertisements on brand attitude is a linkage that did not get much attention yet. However, some researchers have investigated how brand attitude changes due to the relevance of the context in which an (online) advertisement is exposed. In general, information that is consistent with existing brand associations should be more easily learned and remembered compared to unrelated information (Keller, 2013). This is supported by Palanisamy (2004) who stated that

(24)

24 However, the unexpectedness of information that is inconsistent with the brand, sometimes can lead to more elaborate processing and stronger associations than even consistent information (Houston, Childers, and Heckler, 1987; Myers-Levy and Tybout, 1989; Wyer and Srull, 1989). Also Moore et al. (2005) found that consumers produce more favorable attitudes towards contextually irrelevant

advertisements. Furthermore, customer appreciation of the informativeness of contextual relevant ads is reduced by privacy concerns (Turow et al., 2009; Wathieu and Friedman, 2009). When privacy concerns are more important, individuals are more likely to have a prevention focus (Van Noort, Kerkhof, and Fennis, 2008). This means that they are more sensitive to negative outcomes than positive outcomes.

From previous research there is not an unequivocal understanding about how context relevance

influences brand attitude, the following hypothesis is partly based on common sense. It is expected that, in general, a contextual relevant banner has a stronger effect on the relationship between banner advertising and brand attitude, compared to a contextual irrelevant banner. The expectation is based on the findings of Jeong and King (2010); Rodgers (2003/2004); and Becker-Olsen (1998), who investigated that relevant associations improve brand attitude. Hence, the following hypothesis is determined; H4: Contextual relevancy of banners positively influences brand attitude.

Furthermore, it is expected that combining the two elements of eWOM source and contextual relevant advertising will have a strengthening effect of both on brand attitude. Thus the following hypothesis is made:

(25)

25

3.6 Product Involvement

Several involvement constructs are conceptualized and operationalized in consumer research (Muehling, Laczniak, and Andrews, 1993). For example, Batra and Ray (1983) stated that the term involvement is mostly used to describe one of two phenomena: involvement with a product class

(Zaichkowsky, 1985), or involvement with an advertising message (Greenwald and Leavitt, 1984; Petty et al., 1983). In both concepts, personal relevance seems to be a crucial factor to determine consumers’ involvement level with products and/or advertisements (Petty and Cacioppo, 1986). Although

involvement can be used in several concepts, Zaichkowsky (1985) developed one general definition of involvement that recognizes past definitions of involvement: “A person’s perceived relevance of the object based on inherent needs, values, and interests”. Or, as O’Cass (2000) more recently defined “involvement can be viewed as a construct linked to the interaction between an individual and an object”, and it refers to the relative strength of a consumer’s believe concerning the product. When looking at consumers’ involvement of a product in particular, Taylor (1981) identifies product

involvement as the identification of a particular product category to be “more or less central to peoples’ lives, their sense of identity, and their relationship with the rest of the world”.

(26)

26 When combining the product involvement concept with online advertising, it is found that product involvement affects consumers’ recall, attitude, and click-through of online ads. For example, research suggests that high-involvement products are recalled better compared to low-involvement products in online advertisements (Briggs and Hollis, 1997). Furthermore, Cho and Leckenby (2000) found that consumers with a high level of product involvement, compared to those with low product involvement, are more likely to click through banners, which, in turn, result in more favorable attitudes towards the banner ad and brand. These findings are supported by Cho (2003), who state that “when consumers are highly involved with a product, they tend to be very receptive to most information related to that product and thus pay more attention to the content of that product’s ads”. This may even suggest that product involvement has a higher moderating effect in a contextual relevant situation.

Based on this discussion, it is suggested that product involvement moderates the relationship between the independent variable banner context relevance and dependent variable brand attitude. In particular, it is expected that higher product involvement strengthens the effect of context relevance on brand attitude.

Therefore, this assumption is hypothesized as follows:

H6: Product involvement enhances the effect of banner context relevance on brand attitude. EWOM is created in social media communities among those consumers who participate to the discussion in that community (Harridge-March & Quinton, 2009). Within these social media

communities consumers create strong relationships among each other, which make them more loyal to each other and more likely to trust each other’s opinions. Social media sites are good for increasing eWOM, due to Internet’s accessibility, transparency and reach (Kozinets et al., 2010). In different social media communities, creating strong bonds leads to two things: 1) giving direct recommendations about a brand (eWOM) and 2) creating barriers (Harridge-March & Quinton, 2009). Barriers are created due to the social bonding and the feeling of loyalty. This can be beneficial for the companies because it makes consumers more loyal. Although consumers can be involved with a company or a brand, it is arguable that this bond is stronger than the same bond between two friends. So, the following hypotheses are: H7: Product involvement enhances the effect of eWOM source on source credibility

(27)

27

3.7 Daily Facebook Usage

3.7.1 Facebook Usage

The remarkable growth in Internet access has increased the number of users dramatically since its introduction. Prior to 1995, when the Internet began to permeate the society, less than 20 million individuals worldwide used the Internet regularly (Weiser, 2000). However, in the year 2000, about 200 million people worldwide were using the Internet (Weiser, 2000). This rapid increase in the number of Internet users indicates that online users are becoming more diverse regarding online behavior, like usage/experience. Therefore, it is interesting to test whether the effectiveness of banners, and in particular the contextual relevance of banners, differs among the level of Internet usage of individuals. According to Teo, Lim, and Lai (1999), previous research on computer usage derived three indicators of Internet usage. One of them was used in this study, namely frequency of Internet usage. This indicator is used as a separate moderating variable in this study.

3.7.2 Frequency of Internet usage

In general, research suggests that the more people are using the Internet, the less affective an online ad will be. This is reflected by the findings of Benway and Lane (1998) and Cho (2003), who found that heavy Internet users tend to ignore or avoid online banners while surfing the web (‘banner blindness’). Similarly, an eye-tracking experiment of Dreze and Hussherr (2003) resulted that experienced users, compared to inexperienced users, have a significantly lower attention to banners and remembered the advertised brands fewer. The more experienced an Internet user is, the more (s)he is focused in their usage session (Hoffman and Novak, 1996). Therefore, experienced Internet users tend to react less to unexpected stimuli (Dahlén, 1997; Bruner and Kumar, 2000), such as banners. This tendency results in the finding that inexperienced users exhibit increased brand awareness and brand attitude than experienced users (Dahlén 2001).

Following these studies, it can be assumed that banner advertisements result in a higher brand attitude for low Internet users than for heavy Internet users. Thus, it is expected that the more time consumers spend per day on Facebook, the less banner context relevance will affect brand attitude. From this expectation, the following hypothesis is derived:

(28)

28

4 Empirical research design

In this chapter the method of data collection and analysis are explained. Section 4.1 contains the design of the research. Next, the stimuli are described in section 4.2. In section 4.3 the participants of this study are described. Section 4.4 gives an explanation about the variables and scaling that are used in this study, and section 4.5 explains the procedure of this research. Finally, the plan of analysis is presented in section 4.6.

4.1 Research design

The hypothesis are tested with a web experiment in which the participants are exposed to a Facebook page on which a banner ad is displayed and a eWOM post. An experiment is the process of manipulating one or more independent variables and measuring their effect on one or more dependent variables, while controlling for the extraneous variables (Malhotra, 2010). In this study, a 2 (eWOM Source: Business (Asics) vs. Friend (Marcel Wolma)) x 2 (Banner Relevance: Contextual Relevant Banner (Asics) vs. Contextual Irrelevant Banner (BMW)) between-subjects factorial online experimental design is used. The experiment existed out of four experimental groups, who are all exposed to a different treatment of the independent variables. Therefore, the independent variable ‘Banner Context Relevance’ is

manipulated by showing two groups a banner in a contextual relevant setting and the other two groups are shown a different banner that is not relevant to the context. The independent variable ‘eWOM Source’ is manipulated by showing two groups a eWOM Facebook post from a company (Asics) about running shoes and the other two groups are shown a eWOM Facebook post from a friend (Marcel Wolma) about running shoes. In table 1 below, a graphical representation of the 2x2 design is presented.

Table 1 2x2 experimental design.

Business eWOM Friend eWOM

Contextual Relevant Banner Asics eWOM & Asics Banner

Marcel Wolma eWOM & Asics Banner

Contextual Irrelevant Banner Asics eWOM & BWM Banner

(29)

29 This type of research is chosen since it is aimed to find differences between two levels of contextual relevance and two levels of eWOM by giving various experimental treatments to different groups. This between-group experiment is developed by using the online software of Qualtrics, which is offered by the University of Groningen.

4.2 Stimuli

In this experiment, two eWOM posts are exposed to four groups in a different setting; (1) eWOM post from Asics and (2) eWOM post from a friend (Marcel Wolma). In other words, on a Facebook page a message is placed from either a company (in this case Asics), or a message from a friend (in this case Marcel Wolma) about their new running shoes.

Furthermore, the two banners are exposed to four groups in a different setting; (1) contextual relevant banner and (2) contextual irrelevant banner. In other words, a contextual relevant banner is either shown on a Facebook page or a contextual irrelevant banner is shown on a Facebook page. This product category has been selected because of personal preference and because of the increasing popularity in healthy lifestyles and running being part of it. However, running shoe brands with a low to average degree of awareness needed to be selected, since it is expected that respondents will have a high brand awareness and - attitude anyways for strong brands like Nike and Addidas. Asics is used in this

experiment and a banner advertisement of this brand is made. (See figure 8).

The independent variable eWOM Source messages are either originating from the company Asics (See figure 6) or the friend Marcel Wolma (See figure 7).

(30)

30 Figure 7 eWOM message source Marcel Wolma

Figure 8: Banner advertisement Asics brand

(31)

31 In order to control for ‘position’ effects, both banners and eWOM messages, were placed on the same position at all four Facebook pages. The banners are placed on the position were Facebook normally places their banners. The eWOM messages are placed on the middle of the Facebook page together with some other messages. In Appendix A, B, C and Appendix D, a graphical representation of all Facebook pages including banners and eWOM messages are displayed.

4.3 Participants

To measure the effects of the contextual relevance of banner advertisements and eWOM source on brand attitude, this study concentrated on Dutch people between 16 and 65 years old. For data collection, family, friends, fellow students, and other relatives are contacted via social media Facebook and by e-mail, to participate in the web experiment. In doing so, nonprobability sampling technique is used. Convenience sampling is used first, since it is easy and free to access people via social media and e-mail. Next, snowball sampling is used in order to achieve a sufficient number of respondents to perform the experiment. As a rule of thumb at least 40 subjects per group are needed when performing an experiment (Malhotra., 2010). In order to create a reliable comparison between the four

(32)

32

4.4 Variables & scaling

An overview of the scales that are used in this research is presented in table 2. Continuing, all constructs are explained in more detail in this section. The homogeneity of the scales was tested with a Cronbach’s alpha analysis. The results are shown in table 2. By averaging the scores on the separate items the constructs are created.

Table 2 Constructs

Construct Items Source Alpha

Brand Attitude

7 point Likert scale Meyers-Levy, Joan and Laura A. Peracchio (1995)

.921

I would not purchase this product – I would purchase this product

Mediocre – Exceptional product Not at all high quality- Extremely high quality

Poor value – Excellent value Boring – Exciting

Not a worthwhile product – A worthwhile product

Unappealing product – Appealing product

Source Credibility

7 point Likert scale Ohanian, Roobina (1990), .934

Undependable – Dependable Dishonest – Honest

Unreliable – Reliable Insincere – Sincere

(33)

33 Table 2 Constructs continued

Product Involvement

7 point Likert scale Beatty, Talpade (1994), .821

In general I have a strong interest in this product category

This product category matters a lot to me

I get bored when other people talk to me about this product category.

Brand Attitude

To measure the dependent variable brand attitude, scales of Meyers-Levy (1995) are used. The scale of Meyers-Levy (1995) consists of seven items in which participants are asked to give their attitude towards the brand on a 7-point Likert scale, ranging ‘ would purchase-would not purchase’, ‘mediocre-exceptional product’, ‘not at all high quality-extremely high quality’, ‘poor value-excellent value’, ‘poorly made-well made’, ‘boring-exciting’, ‘not a worthwhile product-a worthwhile product’, ‘unappealing product-appealing product’.

Source Credibility

(34)

34 Product Involvement

The moderating variable product involvement is measured using scales from Beatty and Talpade (1994). The scale from Beatty and Talpade (1994) consists three items in which participants are asked about their involvement with the product category (running shoes). On a 7-point Likert scale respondents were asked to answer, ranging from ‘totally disagree-totally agree’, giving their opinion on three statements. These statements were: ‘In general I have a strong interest in this product category’, ‘this product category matters a lot to me’, ‘I get bored when other people talk to me about this product category. ‘.

4.5

Procedure

Extraneous variables are not controlled while the participants are subjected to the web experiment. In practice, people are exposed to banner ads while surfing the Internet in their own environment with many extraneous variables. Therefore, the effect of contextual relevant banners and eWOM Source is measured without controlled elements.

Before this research was conducted a pre-test has been done. This pre-test showed that the original eWOM message had to be increased in size, due to the low level of participants noticing the eWOM message. Furthermore, some of the questions relating Facebook usage were changed.

In the introduction of the web experiment, the participants were told that this research is about the social-network site Facebook, and how they use it, so that they are distracted from the actual goal of the research. First, a few short demographic questions about the participant are asked, like age and gender. Second, questions to measure the respondents’ product category involvement were asked as well as their brand preference. These questions are followed by the questions to measure their Facebook involvement. The entire questionnaire for the eWOM source Asics and Asics banner is presented in appendix E.

(35)

35 Furthermore, Robin ‘liked’ the following things: Coca-Cola, Asics, a restaurant in town, favorite pub, Bob Marley and a few other artists. Robin goes for a run, three times a week, with Robins’ good friend Marcel Wolma and participates annually in several running competitions.

After the participants read the scenario, they were randomly exposed to one of the four Facebook pages for one minute.

After being exposed to one of the Facebook pages, the participants were asked about their attitude towards the brand Asics and how trustworthy they found the sender of the message about the Asics running shoes was.

Finally, two manipulation check questions were asked to control if the participants actually noticed the stimuli that were put into the test. One question to test for the source of the eWOM message and one for the Facebook banner. The questions were asked as followed. On the Facebook page you have just seen was: A message about running shoes from: Asics, Your friend Marcel Wolma or different. An advertisement from Asics: Yes/ No, The results are shown in table 3, 4 and 5 below. The value 1 equals a correct answer and the value 2 equals an incorrect answer. Failing to answer, whether they saw the right banner or sender of the eWOM message could be partly accredited to the ‘banner blindness’ that is found (Chapter 5) to play a role in this research. It is therefore chosen to keep all the respondents in the analysis.

Table 3 Manipulation check eWOM message correct Table 4 Manipulation check banner correct Frequency Percentage

Valid 1 139 86.3

2 22 13.7

Total 161 100

(36)

36 4.6 Plan of Analysis

In SPSS basic descriptive statistics are used to explore the data and to get a general view of the results. The hypotheses are tested by using, tests like Independent sample T-Test, Correlation and Regression in SPSS. In table 6 it is shown which test are used for each hypothesis and which variables are used in the test.

Table 6 Plan of Analysis

Hypotheses Test in SPSS Variables used

H1: Source credibility will have a positive effect on brand attitude.

Regression Source Credibility, Brand Attitude H2: Friend eWOM is perceived to be more

trustworthy than business eWOM

Regression eWOM Source, Source Credibility

H3: Friend eWOM has a more positive effect on brand attitude than business eWOM.

Regression eWOM Source, Brand Attitude

H4: Contextual relevancy of banners positively influences brand attitude.

Regression Banner Context Relevance, Brand Attitude

H5: A friend eWOM message combined with contextual relevant advertising will have the most positive effect on brand attitude.

Two-Way Anova eWOM Source, Banner Context Relevance, Brand Attitude

H6: Product involvement enhances the effect of banner context relevance on brand attitude.

Regression Product Involvement, Banner Context Relevance, Brand Attitude

H7: Product involvement enhances the effect of eWOM source on source credibility

Regression Product Involvement, eWOM Source, Source Credibility

H8: Product involvement enhances the effect of eWOM source on brand attitude.

Regression Product Involvement, eWOM Source, Brand Attitude

H9: Time spent per day on Facebook mitigates the effect of banner context relevance on brand attitude.

(37)

37 The constructs eWOM source and banner context relevance are made as follows. First for each of the four test groups (see table 1) a group-page number variable was constructed. Than this group-page variable was used to split the groups into either of the two possibilities of each constructs. EWOM source is coded as 0 being business eWOM and 1 being friend eWOM. Banner context relevance is coded as 0 being contextual relevant and 1 being contextual irrelevant.

First a correlation analysis is made to test for correlations between the variables (See Table 7). Second, the hypotheses are tested using the methods mentioned in table 6.

Before trying to answer the hypotheses in complete models, first various analyses are made to explore the data and to get a feeling for possible links and relationships between the variables. To get a feeling for the differences between the four scenarios, the mean scores of each scenario are analyzed for both dependent variables. Second, a correlation analysis is made to test for any correlations between the variables (See Table 7). Third, the relationships regarding the dependent variable source credibility are tested. Fourth, the relationship between source credibility and brand attitude is tested. Fifth the relationships regarding brand attitude are tested. After that a total model for the dependent variable brand attitude is tested. In this model a new variable is added to the equation. In the questionnaire questions were built in so a possible base line for brand attitude could be established. In hindsight these questions were inadequate to establish a proper base line. However during analyses of the data they showed a new variable that played a role on brand attitude namely previous purchases. Two total model analyses on brand attitude were conducted. The original model and one with the new variable. These analyses showed that previous purchases did not affect the other variables or relationship in any way. So by adding this new variable the theory underlying the original model was not affected, only the adjusted R2 was increased. Because of the reason that the addition of the new variable did not change the results of the original model, it is chosen to only present the results of the model with the additional variable in this thesis.

(38)

38

5 Results

In total 161 respondents filled in the questionnaire, of which 57.1% was male and 42.9% was female. The average age is 24 years old. The average time spent on Facebook is high with an average of 50 minutes per day. On average the respondents have 288 Facebook friends.

One of the four experiments was shown at random to the respondents. 27.3% of the respondents have been shown the Facebook page containing a friend eWOM message and a contextual relevant banner. 26.1% of the respondents have been shown the Facebook page containing a business eWOM message and a contextual relevant banner. 24.8% of the respondents have been shown the Facebook page containing a friend eWOM message and a contextual irrelevant banner. 21.7% of the respondents have been shown the Facebook page containing a business eWOM message and a contextual irrelevant banner. In total, 47.8% of the respondents have been shown a business eWOM message and 52.2% a friend eWOM message. 53.4% of the respondents have been shown a contextual relevant banner and 46.6% have been shown a contextual irrelevant banner. The average age in each group was similar as well as the ratio males to females between each group.

The mean scores for each group on brand attitude and source credibility are shown below in table 7.

Table 7 between group mean scores on brand attitude and source credibility Group_page Brand attitude Mean Source Credibility

(39)

39 For a first indication of the relationships between each variable a correlation test is conducted. The results are shown in the correlation matrix below in table 8.

Table 8 Correlation Matrix

Gender Age Product

Involvement Number of Facebook friends Daily Facebook Usage Brand Attitude Source Credibility Banner Context Relevance EWOM Source Gender 1 Age -.012 1 Product Involvement .175** .022 1 Number of Facebook friends -.052 -.278*** .009 1 Daily Facebook Usage .257*** -.195** .069 .181** 1 Brand Attitude .124 -.111 .348*** -.006 .153* 1 Source Credibility .008 .071 .108 -.066 .054 .299*** 1 Banner Context Relevance -.029 .028 -.111 .040 .047 -.113 .058 1 EWOM Source -.075 .024 -.234*** -.128 -.088 -.004 .146* .022 1

(40)

40

5.1 eWOM Source and Source Credibility

To get a feeling for the hypothesis: Friend eWOM is perceived to be more trustworthy than business eWOM an independent sample T-test is performed. The average score of friend eWOM and business eWOM on source credibility was 4.5 and 4.2 respectively. The answers were given on a scale from 1-7, one being “untrustworthy” and seven being “trustworthy. The T-test showed a 2-tailed significance level of .060, thus marginally significant.

5.2 Product Involvement moderating eWOM Source and Source Credibility

In order to answer the hypothesis: Product involvement enhances the effect of eWOM source on source credibility a hierarchical linear regression is made. In order to account for moderating effects of product involvement on the relationship between eWOM source and source credibility first an interaction variable must be made to account for the moderating effect. This is done by multiplying both independent variables, but detracting the mean of each variable for each of its score. This to reduce possible multi-collinearity effects. The formula for the new interaction variable is (eWOM Source –mean (.5217)) * (Product Involvement – mean (3.0083)). EWOM source is coded as 0 being business eWOM and 1 being friend eWOM.

Second a linear regression analysis is made with source credibility as dependent variable and eWOM source and product involvement as independent variable. For the second model the new interaction variable for eWOM source and product involvement is added to the regression. The first model has an adjusted R2 of .031 and thus accounts for 3.1% of the dependent variable source credibility and is significant at the level p< .05.

The second model has an adjusted R2 of .045 and thus accounts for 4.5% of the dependent variable source credibility and is slightly more significant but still at the level of p< .05.

(41)

41 Table 9 Moderating effect of product involvement on eWOM source and source credibility

Model Standardized Coefficients Adjusted R

Square F VIF Beta 1 (Constant) .031** 3.530** eWOM Source .181** 1.058 Product Involvement .151* 1.058 2 (Constant) .045** 3.486** eWOM Source .177** 1.059 Product Involvement .145* 1.059

Interaction eWOM Source x Product Involvement

.140* 1.002

*Correlation is significant at the 0.10 level (2-tailed). **Correlation is significant at the 0.05 level (2-tailed). ***Correlation is significant at the 0.01 level (2-tailed).

The above mentioned model depicts the total model for the dependent variable source credibility. Based on these results, there must be concluded that Friend eWOM is perceived to be more trustworthy than business eWOM and thus hypothesis 2 is supported. Furthermore, because the F-change value is marginally significant we must conclude that there is a moderating effect of product involvement on the relationship between eWOM source and source credibility. Based on these result, there must be

(42)

42

5.3 Source Credibility and Brand Attitude

In order to test the hypothesis: Source credibility will have a positive effect on brand attitude. A regression analysis was made.

A linear regression was used to measure how much source credibility explains the dependent variable brand attitude. Based on the adjusted R2, source credibility accounts for 8.3% of the variation in brand attitude (See table 10). It can be concluded that the regression model is significant at p= .000 and that source credibility has a significant effect on brand attitude. Based on these results, it can be concluded that hypothesis 1 is supported.

Table 10 the effect of source credibility on brand attitude

Model Standardized Coefficients Adjusted R Square F Beta 1 (Constant) .083** 15.553** Source Credibility 0.299** *Correlation is significant at the 0.05 level (2-tailed).

(43)

43

5.4 eWOM Source and Brand Attitude

To get a feel for the hypothesis: Friend eWOM has a more positive effect on brand attitude than business eWOM, an in Independent sample t-test is made. The mean score for friend eWOM and business eWOM on brand attitude are 4.35 and 4.36 respectively. The answers were given on a scale from 1-7, one being “unappealing” and seven being “appealing”. The t-test showed a significance level of .96, meaning the differences between friend eWOM and business eWOM on brand attitude are marginally significant with an Alpha < .10. These results indicate a possible support for hypothesis 3.

5.5 Banner Context Relevance and Brand Attitude

(44)

44

5.6 eWOM Source combined with Banner Context Relevance and Brand

Attitude

This research wants to test which combination of eWOM source and banner context relevance results in the highest value for brand attitude. To test this, a Two Way Anova analysis was performed using brand attitude as dependent variable and the Group_page variable as fixed factor. The mean scores on brand attitude for each group are shown in Table 11 below.

Table 11 Group mean scores on Brand Attitude

Group page Mean Std.

Error 95% Confidence Interval Lower Bound Upper Bound Business eWOM + Relevant

Context Banner

4.534 .160 4.218 4.850

Friend eWOM + Relevant Context Banner

4.403 .156 4.094 4.712

Friend eWOM + Irrelevant Context Banner

4.300 .164 3.976 4.624

Business eWOM +

Irrelevant Context Banner

4.155 .175 3.809 4.502

(45)

45 Table 12 Two Way Anova Pairwise Comparison

(I) Group_page Mean

(46)

46 Table 12 Two Way Anova Pairwise Comparison continued

Friend eWOM + Irrelevant Context Banner Business eWOM + Relevant Context Banner -.234 .229 -.687 .219 Friend eWOM + Relevant Context Banner -.103 .227 -.551 .345 Business eWOM + Irrelevant Context Banner .145 .240 -.330 .619 Business eWOM + Irrelevant Context Banner Business eWOM + Relevant Context Banner -.379 .238 -.848 .090 Friend eWOM + Relevant Context Banner -.247 .235 -.712 .217 Friend eWOM + Irrelevant Context Banner -.145 .240 -.619 .330

*Correlation is significant at the 0.10 level (2-tailed). **Correlation is significant at the 0.05 level (2-tailed). ***Correlation is significant at the 0.01 level (2-tailed).

From these results we can conclude that hypothesis 5: A friend eWOM message combined with contextual relevant advertising will have the most positive effect on brand attitude, is not supported.

(47)

47

5.7 Product Involvement moderating Banner Context Relevance and Brand Attitude

In order to get a feeling for the answer of the hypothesis: Product involvement enhances the effect of banner context relevance on brand attitude a hierarchical linear regression analysis is made. This is done to get an indication for the stand alone effect of product involvement as a moderator between banner context relevance and brand attitude. The hypothesis will however be answered when the entire model is tested.

In order to account for moderating effects of product involvement on the relationship between banner context relevance and brand attitude first an interaction variable must be made to account for the moderating effect.

This is done by multiplying both independent variables, but detracting the mean of each the variable for each of its score. This to reduce possible multi-collinearity effects. The following formula is used to create this new interaction variable: (Product Involvement –mean (3.0083 )) * (Banner Context Relevance – mean (.4688))

Second a linear regression analysis is made. The dependent variable is brand attitude and for the first model the independent variables are banner context relevance and product involvement. The new interaction variable is entered afterwards to create the second model. Banner context relevance is coded as 0 being contextual relevant and 1 being contextual irrelevant.

The first model has an adjusted R2 of .115 and thus accounts for 11.5% of the dependent variable brand attitude and is significant at the level p= .000.

(48)

48 Table 13 moderation effect of product involvement on banner context relevance and brand attitude.

Model Standardized Coefficients adjusted R Square F VIF Beta 1 (Constant) 0.115*** 11.442*** Banner Context Relevance -.075 1.012 Product Involvement 0.33*** 1.012 2 (Constant) 0.129*** 8.887*** Banner Context Relevance -.079 1.013 Product Involvement 0.325*** 1.025 Interaction Banner Context Relevance x Product Involvement -.137* 1.012

*Correlation is significant at the 0.10 level (2-tailed). **Correlation is significant at the 0.05 level (2-tailed). ***Correlation is significant at the 0.01 level (2-tailed).

5.8 Product Involvement moderating eWOM Source and Brand Attitude

As in section 5.7 to get a feel for the moderating effect of product involvement on the relationship between eWOM source and brand attitude a hierarchical linear regression analysis is made. The depending variable in this case is brand attitude and the independent variables are eWOM source and product involvement. For the second model the interaction variable between eWOM source and product involvement (See 5.2) is added. EWOM source is coded as 0 being business eWOM and 1 being friend eWOM.

The first model has an adjusted R2 of .116 and thus explains 11.6% of the dependent variable brand attitude and is significant with p= .000.

(49)

49 The F-change value is 2.887 and is marginally significant with a significance level of .091. The VIF values are well below the cutoff point of 10 indicating no multi-collinearity problems. The beta’s for each variable can be found in table 14.

Because the F-change value is marginally significant we must conclude that product involvement does play a moderating effect on the relationship between eWOM source and brand attitude. These results indicate a possible support for hypothesis 8.

Table 14 moderating effect of product involvement on eWOM source and brand attitude.

Model Standardized Coefficients Adjusted R Square F VIF Beta 1 (Constant) .116*** 11.517*** eWOM Source .082 1.058 Product Involvement .367*** 1.058 2 (Constant) .127*** 8.732*** eWOM Source .078 1.059 Product Involvement .362*** 1.059

Interaction eWOM Source x Product Involvement

.126* 1.002

*Correlation is significant at the 0.10 level (2-tailed). **Correlation is significant at the 0.05 level (2-tailed). ***Correlation is significant at the 0.01 level (2-tailed).

5.9 Daily Facebook Usage moderating Banner Context Relevance and Brand Attitude

The Hypothesis: Time spent per day on Facebook, mitigates the effect of banner context relevance on brand attitude is tested by the use of regression models. Unlike with the previous moderating effects this hypothesis is answered directly after this test and not after the total model. This because this hypothesis is meant to show, that banner blindness plays a role on the effect between banner context relevance and brand attitude. Because the total model regression might be significant does not mean that banner blindness did not play a role in the model. Therefore this effect is tested with this

regression.

(50)

50 Regression analysis showed that the model account significantly for 2.7% (based on the adjusted R2) of the brand attitude (p< .05). The Fvalue is 3.201 and beta is .159 for daily Facebook usage (p<.05) and -.125 for banner context relevance (p>.05).

In order to account for moderating effects of daily Facebook usage on the relationship between banner context relevance and brand attitude first an interaction variable must be made to account for the moderating effect. This is done in the same way as in 5.7. The following formula is used to create this new interaction variable: (Daily Facebook Usage – mean (50.48)) * (Banner Context Relevance – mean (.4688)).

(51)

51 Table 15 Moderating effect of daily Facebook usage on the relationship between banner context relevance and brand attitude Model Standardized Coefficients Adjusted R Square F VIF Beta 1 (Constant) .027* 3.201*

Daily Facebook usage 0.159* 1.002

Banner context relevance -.125 1.002

2 (Constant) .027 2.124

Daily Facebook usage 0.158* 1.031

Banner context relevance -.125 1.002

interaction Daily Facebook Usage x Banner Context Relevance

.008 1.029

Referenties

GERELATEERDE DOCUMENTEN

Next to that, the analysis based on survey data provided no evidence for the presence of a significant effects of source availability and promotion of feedback-seeking behavior of

Bij de attitude ten opzichte van reclame spreken we van een trend dat jongeren negatiever zijn dan de ouderen ten opzichte van zowel experts als leken.. Dit ongeacht de ‘sexe van

This implies that eWOM messages with different degrees of persuasiveness inherent to the eWOM platforms as identified in paragraph 2.6.1, impact brand awareness,

In the current study it is hypothesized that the effect of the independent variables (the presence of demographic/ psychographic characteristics attached to an OCR)

Source credibility → Cognitive trust → higher eWOM adoption → More positive attitude. 05/07/2018

Next to this, unfavorable cognitive responses in an online review context — contrary to the source credibility literature regarding advertising — are negatively moderating the effects

•   A positive consumer response (disconfirming response) compared to a negative consumer response (confirming response), decreases the impact of source credibility on

This research shows that when people are confronted with a negative OCR and subsequently with a disconfirming (positive) response, the influence of source credibility on