BRANDED FACEBOOK CONTENT: LESS “MESS” WHILE MATCHING THE MESSAGE AND THE MESSENGER?

(1)

1

BRANDED FACEBOOK CONTENT:

LESS “MESS” WHILE MATCHING THE MESSAGE AND THE

MESSENGER?

By

Boaz Disselkoen

(2)

2

BRANDED FACEBOOK CONTENT:

LESS “MESS” WHILE MATCHING THE MESSAGE AND THE

MESSENGER?

December, 2013

By

Boaz Disselkoen

S1637266

Rijksuniversiteit Groningen Faculty of Economics and Business

Master Thesis Business Administration Marketing Management

1

st

_{Supervisor: prof. dr. J.C. (Janny) Hoekstra}

2

nd

_{Supervisor: MSc. L. (Lisette) de Vries}

Nassaukade 378-3

1054AD Amsterdam

0626977939

(3)

3

Management Summary

During the past three decades a series of technological communication innovations rapidly reshaped and globalized communication. Especially social network sites accelerated the innovation of mass communication. By successfully introducing company fan pages, the social network site Facebook became one of the most important communication channels for companies. This social network site became more important to marketers, serving as a new communication channel. Therefore, brands, products, services and their respective marketers were faced with these changes (Thompson et al., 2008). As a result, Facebook became no longer a short term investment for marketers, but rather a long term strategic channel that allowed brands to connect with consumers over time.

Companies that are using a “Company fan page” of Facebook are able to transmit various forms of brand or product related “Content” to their consumers and their online friends. This branded Facebook content could be a (combination of) text, picture, sound, video or location (Trusov et al., 2009). The content may include information about functions and cognitive features, called utilitarian content, or emotions in the form of entertainment, exploration, and self-expression, called hedonic content (Yang, 2010; Adjei et al, 2010; Hennig-Thurau et al., 2010; Brown et al., 2007; Chiu et al, 2007). Social network sites efforts valuable in the ability to grow brand awareness and increases dialogue with consumers (Hennig-Thurau et al., 2010; Yang, 2010; Trusov et al., 2009). In addition, to increase the return on investment and engage the audience, marketers need to know how to design and use the content on Facebook.

This research is elaborating on these subjects, divided into two parts. First it elaborates on the actual effect of Facebook content on the representing brand attitude and purchase intentions. Second, the possible effect of the content type on content attitude, brand attitude and purchase intentions is determent by testing the possible difference between hedonic and utilitarian content. In addition to the content, possible moderating drivers (Facebook involvement, product involvement and intention to respond) for content attitude, brand attitude and purchase intention are determined.

Results show that Facebook content has no influence on brand attitude or purchase intentions. However, the consumers’ attitude towards hedonic content is higher than the attitude towards utilitarian content. Likewise, the increase of brand attitude after being confronted with hedonic content is higher (compared to the same effect) after being confronted with utilitarian content. This result occurs when the intention to respond on the content increases.

(4)

4

Preface

Beste Lezer,

Om een goede scriptie te schrijven moet men de tijd nemen. Dat heb ik dus ook gedaan. Na een jaar van inspanning ligt eindresultaat inmiddels voor u. Gedurende het schrijven en zelfs het nalezen van deze, zevenenveertig pagina’s tellende toevoeging aan de wetenschap, heb ik altijd rekening

gehouden met een Nobelprijs voor de economie. Helaas is dat het niet geworden, maar toch ben ik tevreden met het eindresultaat. Beginnend bij Grolsch in Amsterdam/Enschede, later herschrijvend bij Vrumona in Bunnik om het uiteindelijk af te maken in de avonduren en weekenden bij Heineken was niet gelukt zonder de hulp en steun van een aantal mensen. Daarom wil ik via deze weg een aantal van hen in het bijzonder bedanken:

Van Grolsch:

Edwin Blom (Sales Directeur) en Jan Nales (Marketing Directeur): Voor de kans om een scriptie te schrijven voor het Grolsch merk.

Michal Rabiej (Brand Develop Manager): Voor de inzichten, kennis en kunde en in de marketing, trademarketing, sales en biercatergorie. Maar bovenal voor de waanzinnig leuke gesprekken over deze onderwerpen.

Van Vrumona:

Gerrit van Loo (Algemeen Directeur) en Sandy de Koning (Marketing Directeur): Voor het creëren en geven van ruimte bij Vrumona voor een afstudeerstudent, die van de een op de andere dag komt aanwaaien en binnen vier maanden klaar moet zijn.

Marlou Lancee (Brandmanager Crystal Clear): Voor de begeleiding die zij van de een op de andere dag als extra taak kreeg toegewezen en dit op geheel eigenwijze heeft ingekleed.

Van Heineken:

Lieke Westendorp (Trade Marketing Manager): Voor het stimuleren en inspireren van de naderhand onmogelijke combinatie fulltime werken en studeren.

Het Heineken Horeca Trade Marketing team: Voor het op geheel eigenwijzen aanmoedigen van mijn afstuderen.

Nicoline Braat: Voor het lezen en verbeteren van mijn scriptie, midden in de zomer vakantie. En ten slotte natuurlijk mijn ouders voor hun slapeloze nachten die ik heb bezorgd gedurende de gehele studie, maar die na een aantal overtuigende woorden van mijn kant, toch, net als ik, in een goede afloop bleven geloven.

(5)

5

1. Introduction

Mark Zuckerberg, founder and chief executive officer of Facebook:

“Advertising works most effectively when it's in line with what people are already trying to do. And people are trying to communicate in a certain way on Facebook - they share information with their friends, they learn about what their friends are doing - so there's really a whole new opportunity for a new type of advertising model within that” (Locke, 2007).

D

uring the past three decades a series of technological communication innovations rapidly reshaped commercial communication. Especially social network sites accelerated the innovation of

commercialized mass communication. Through social network sites the consumers have become enabled to interact worldwide, day and night, leaving a wealth of information for the taking. Several companies are taking advantage of the interactions that have become known as social media (Trusov et al., 2009). Assessable by their smartphone, tablet, laptop or personal computer, consumers became both socially and commercially easy to reach (Yang, 2010). From a commercial perspective, potential consumers became one-click-away “fans”. However, as the quote above implies, a well-chosen online approach needs to be applied by the companies to reach the online consumers (Trusov et al., 2009). Moreover, before choosing an online communication strategy, it may be of importance to understand the actual business opportunities that arise from social network sites.

This research explores the phenomenon of Facebook content posted by companies on Facebook as communication messages towards their (potential) consumers. To understand the principals of this research, the first section (1.1) explores the background, development and opportunities of social media, social network sites and Facebook. Second, the usage and the opportunities of the different types of branded Facebook content are introduced (1.2). Third, the possible moderating effects of the consumer’s nature are explained (1.3). Fourth, section 1.4 introduces the problem statement. Fifth, the managerial and academic relevance of exploring this statement is given. The last section shows the construct of this report.

1.1 The Birth and Use of Online Content Communication

(8)

8 environment for the realization of social media together with content produced by users on the one hand and communities on the other (Kangas et al., 2007). So, the applications developed for social media are either completely based on user generated content or in which user generated content and the actions of users play a substantial role in increasing the value of the application or service (Kangas et al., 2007).

One of the most important platforms of social media is the online communication application in the form of social network sites. Social network sites are defined as: web-based services that allow individuals to (1) construct a public or semi-public profile within a bounded system, (2) articulate a list of other users with whom they share a connection, and (3) view and traverse their list of connections and those made by others within the system (Boyd and Ellison, 2007).

One of the biggest and most important social network sites is Facebook. The platform was launched in 2004 as an application for Harvard students. In September 2005, the site was extended to include all individual consumers. Unlike other social network sites, Facebook users are able to publically display personal profiles to all fellow users. Furthermore, Facebook was one of the first of social network sites that included the possibility for external, private developers to build applications (“apps”) for the site. These apps allow users to personalize their profiles and perform other tasks, such as the comparison of brand preferences (Boyd and Ellison, 2007).

In 2006, the site introduced the “Company Fan Pages”, a platform specially focused on commercial use. Companies were able to create a business profile, on which they could interact with their (potential) consumers. Before the creation of this innovation, Facebook users could only be linked to each other by becoming “Friends” on Facebook. After the company fan page introduction,

consumers were allowed to become friends with a company by liking their page. When a Facebook user “Likes” a company fan page he or she becomes a “Fan” of that company. When this user is a fan, he/she becomes a member of the company’s Facebook community. The company that owns the fan page is able to transmit various forms of brand or product related “Content” to their fans and the friends of their fans, through the company fan page. The branded content appears among other messages on the “Newsfeed” (also known as Wall) of the fan. Newsfeed is described as a continues updated list of reports from people or company pages, a user follows (www.facebook.com/help). When Facebook users scroll down their personal newsfeed they browse through all the different Facebook content messages of their online friends or of the content from the companies they have liked. While browsing they are being confronted with the commercialized content of the companies. This (branded Facebook) content could be a (combination of) text, picture, sound, video or location (Trusov et al., 2009).

(9)

9 the company’s brand or product. Moreover the company may use the interaction to build a closer relationship with the fans (Trusov et al., 2009). These opportunities are found in the ease of reaching consumers when a brand is introducing new products or services, notifying promotions or events, or educating about a brand or product elements. Moreover, various researchers conclude that online communities contain suitable ingredients for generating and magnifying brand or product

involvement (Ahuja and Medury, 2010; Hennig-Thurau et al., 2010). A positive attitude towards a brand and a higher rate of involvement may result in an increase of purchase intentions according to Tuten (2008).

Before the company fan page introduction, organizations mostly used broadcast or print media to inform, persuade, or remind present and potential customers of their offerings or about the

organization itself (Berthon et al, 2008). The traditional consumer, has been a passive recipient of the communication. He/she could react by just becoming aware, by being swayed to do something they might otherwise not have done, or by having their memories jogged and reinforced (Berthon et al, 2008). Berthon (2008) presumes that the advertisement was ignored most of the time. Now the online fans are the new receivers of the branded communication when they browse through the content on their newsfeed. They have become a fan of the brand, simultaneously allowing or even desiring to be confronted with the advertisements in the form of content. Moreover, they have the opportunity to respond to the content. By doing this, they create their own new messages and simultaneously increase the reach of the branded content, enhancing its value (Trusov et al., 2009). Therefore, Facebook content seems to be an effective tool to communicate and change the

consumer’s attitude and behavior. The efficiency of content is being explored in this research

1.2 The Different Content Types

The impact and importance of social network sites is reflected in the expansion of social network communities. On June 2013 Facebook was the most popular site, with 1,11 billion users

(www.investor.fb.com), followed by Twitter and Google+ respectively with around 554 million users in June 2013 (statisticbrain.com) and 500 million users in December 2012 (www.business

(10)

10 common, easiest methods to communicate with the fans is by posting branded content on the company fan page (Trusov et al., 2009).

However, the fan on their turn, is able to speak freely about companies, brands or products on the same platform. Therefore companies may perceive increasingly less control over the information that is available about them on social network sites (Kaplan and Haenlein, 2009). Consequently, it is of importance that a fan understands the message, as was intended by the company, to gain a higher return on investment of the content (Tan and Chia, 2007).

A key element to cover this issue is to use the right type of content (Hennig-Thurau et al., 2010; Yang, 2010; Trusov et al., 2009). A fan may show a different attitude towards different content types resulting in different behavior (Yang, 2010; Thackeray et al., 2008; Tuten, 2008). Content can be divided into two broad types. A company could transmit information about functions and cognitive features, which is called utilitarian content. This company could also transmit emotions in the form of entertainment, exploration, and self-expression, which is called hedonic content ( Yang 2010, Chiu et al, 2007; Edwards, 1990). The informative (utilitarian) and emotional (hedonic) elements of the content are creating informational or emotional value within the message (Yang, 2010). For example, a fan that has been browsing through branded content with entertaining elements could have formed a specific brand attitude. However, if this same content was purely informative, the brand attitude could have been different compared to the hedonic content (Yang, 2010). So, one specific type of content may be more effective than the other (Berthon et al, 2008). Therefore it may be of interest to subsequently search for the relationship between different types of branded Facebook content and the assumed difference in attitude and behavior of the fans. By understanding the impact of branded content and the most effective type of content, a company could be more effective within its attempt to increase the consumer’s brand attitude and purchase intentions on Facebook (Voss et al., 2003).

1.3 The Nature of the Fan as Different Moderating Effects

Previously it was assumed that companies could alter the attitude and behavior of the fans by just showing content on Facebook. However, not all the fans may behave in a same way. Some fans may be more susceptible for content than others (Berthon et al, 2008).

(11)

11 Moreover, fans may become a member of a company fan page for different reasons. A fan may overtly search for relevant information because he/she is highly involved with the company’s product (Zaichkowsky, 1986). Whereas a fan that is less interested in the product, may fall for the emotional and symbolic (hedonic) aspects of the brand (Yang, 2010). As was previously mentioned, by

responding on content the reach of the message and therefore its value is enhanced (Trusov et al., 2009). Fans that perceive a high urge to respond on content may therefore be more valuable to the company. However, a fan’s intention to respond may have a different outcome while being

confronted with different kinds of content (Berthon, 2008; Chiu, 2007).

In conclusion, the nature of the fan could cause a difference within the assumed effect between content and the fan’s attitude/behavior. It may therefore be of interest to explore the possible moderating effects of Facebook involvement, Product involvement and the Intention to respond on content.

1.4 Problem Statement

Two primary issues are addressed. First, the actual impact of branded Facebook content compared to no Facebook content on the (potential) consumer’s attitude and purchase intentions are tested. The second issue contains the different effects between hedonic and utilitarian content. The impact difference of content type will be tested for content attitude, brand attitude and purchase intentions.

Moreover, the proposed effect of Facebook content could be affected by different moderating variables. Three moderating variables are proposed within this research: first Facebook involvement, second product involvement and third the fan’s intention to respond. Only the first two moderators are used to test the influence of content in general. The intention to respond is added as a third moderating variable to explore the proposed effect difference of hedonic and utilitarian content. The main question of this research is in twofold:

What is the effect of branded content on brand attitude and purchase intentions; and what is the difference in effect of hedonic branded content and utilitarian branded content; on content attitude, brand attitude and purchase intention, within a specific FMCG-branded Facebook community?

1.5 Managerial and Academic Relevance

(12)

12 (Adjei et al., 2010). Therefore the interest of managing a company fan page is growing among

organizations (econsultancy.com). From a managerial perspective, the study is trying to give an insight into an effective way of investing in potential consumers on Facebook, in order to gain a higher return on investment. Within the context of Facebook, this specific branch has hardly been examined. Remarkably, almost all FMCG’s make use of social media to reach their consumers. Within the top ten company Facebook pages three FMCG companies are present: Coca Cola is third, Red Bull sixth and Oreo tenth (http://www.ignitesocialmedia.com). These organizations are using their Facebook communication as part of the whole marketing strategy. The online content strengthens the overall campaign.

The main stream of current social media studies are focused on the viral effect of social media, its word of mouth opportunities, the influence of user generated content, or the effect of

user-generated content in online communities (Hennig-Thurau et al., 2010; Adjei et al., 2010; Thompson et al., 2008). The most notable recurring examined element is the consumer-to-consumer

communication on company fan pages and social media. For example, Zhang et al (2010) conducted a content study on Facebook in which they partly searched for the success or failure of a brand page. However, their research did not reveal the actual impact on content attitude, consumer brand attitude and their purchase intentions. Furthermore, social and viral impact of consumer communications on social network about companies is a common issue discussed in marketing literature. These current studies, however, nearly examine the communication strategy of branded Facebook content to influence consumers.

While current research predominantly finds positive results of social media, less than a decade ago the few studies that were conducted on social media and e-commerce were moderate. Teerling and Huizingh (2004) for example, concluded that online activities had no direct influence on online purchases. This change of perspective indicates the speed at which Social Media has evolved (Thompson et al., 2008). Due to the constant and fast developments of social media, continues examination of this evolutionary subject may be needed.

1.6 Structure of the report

(13)

13

2 Theoretical Framework

This chapter introduces the two models that are central in this research. Both models consist of different constructs. The base is the manipulation of a fan after (not) being confronted with

Facebook content, or after being confronted with different content types. The difference in outcome caused by this manipulation is tested. Moreover moderating variables are added to test additional influences on the manipulation. The assumed effects of the manipulation are translated in different hypotheses.

2.1 Introduction of the Constructs

As was described, the problem statement was divided into two parts: the effect of branded content compared to no content and the effect of hedonic content compared to utilitarian content.

Therefore two conceptual models were composed (see figure 1 and Figure 2). Both models are based on several studies that seek for the causal relation between the attitudinal and behavioral change of a consumer and the interaction within an (online social) brand community (Yang, 2010; Adjei et al, 2010; Hennig-Thurau et al., 2010; Brown et al., 2007). The models try to further extend on these studies, by constructing factors (independent variables) that may affect the attitudinal change towards the content, towards the brand and the change of purchase behavior. The base of the models consists of the “Branded content messages” (from now referred to as “Content”) sent by the company.

First, the influence of content is going to be examined to test the impact of the Facebook content in general (figure 1) on the fans. This first model provides insights into the potential utility of content for an FMCG company. The main dichotomous independent variable, “Content Influence”, contains of: content and no content. Fans that were not provided with branded content could not have formed a content attitude. Therefore the independent variable is only manipulated to test the effect of brand attitude and purchase intentions.

In the second model the type of content is the main subject. The tone of voice within the content will be manipulated through the main dichotomous independent variable “Content type”. The possible behavioral difference after being faced with hedonic or utilitarian content is tested. It is assumed that both types of content have a positive effect on content attitude, brand attitude and purchase intention. However, one type of the two types of content could be more effective in increasing the: content attitude, brand attitude and purchase intentions.

(14)

14 consideration within the first model (figure 2). Fans that have not seen any content are not able to form an intention to respond.

Figure 1: Content Influence

Figure 2: Content Type

2.2 Introducing the Hypotheses

Two models were composed to test the difference of two main effects. The first main effect investigates the impact difference between content and no content. The second main effect investigates the impact difference between hedonic and utilitarian content. Third and fourth, the assumed effect of Facebook involvement and product category involvement in both models is described. Finally the effect of respond intentions on content type is introduced.

2.2.1 The Influence of Content

One goal of the FMCG’s communication is to create positive awareness (or a positive content attitude), consideration (or brand attitudinal change) and purchase intentions (Adjei, 2010; Ahuja and Medury, 2010; Hennig-Thurau et al., 2010; Yang, 2010; Mangold and Faulds, 2009; Brown et al., 2007). The company informs, persuades, or reminds present and potential customers of its offerings or of the organization itself. This was traditionally conducted by broadcasting or printed media

H4b,c H5a, b H3b H2 H1c,d,e Content Type: Utilitarian (U) Hedonic (H) Moderator 1: Intention to Respond (IR) Moderator 3: Product Category Involvement (PCI) Moderator 2: Facebook Involvement (FBI)

Behavioral Change of: Content attitude (CA) Brand attitude (BA) Purchase Intention (PI) H4a H3a H1a,b Content Influence: Content (C) No content (N) Moderator 3: Product Category Involvement (PCI) Moderator 2: Facebook Involvement (FBI)

(15)

15 (Berthon et al, 2008). The traditional consumer could react by just becoming aware, by being swayed to do something they might otherwise not have done, or by having their memories jogged and reinforced (Berthon et al, 2008).

Nowadays, FMCG-companies are using Facebook for the same purposes. The branded online interaction on company fan pages is supplied by the company. Packed within the message, the company transmits brand- and product information to their fans. These companies assume that online communication allow the fans to get involved, obtain data, information and experience (Steffes and Burgee, 2009). So, when fans within the online community are confronted with relevant information their attitude towards the brand and their intention to purchase are assumed to be increased (Mangold and Faulds, 2009; Tuten, 2008).

 H1a: Branded content posted on a company fan page results in a higher brand attitude than no content.

 H1b: Branded content posted on a company fan page results in higher purchase intentions than no content.

Using common sense, the first model could be experienced as an open door. However, the volatile nature of Facebook content and the variance of messages perceived by the fan may be not the same as traditional media. Competing messages on the newsfeed of a Facebook user, within the volatile online media, could alter the influence of a single message (Eichenwald, 2013). This model brings perspective throughout the research method and the conclusions. (Not) accepting model 1 may have consequences for the relevance of model 2 (see sections 5.3 and 5.4).

2.2.2 The effect of Content Type

(16)

16 to do” the branded content may receive its intended behavioral change. Therefore, the type of content value within the message needs to match or trigger the expectations of the browsing fan. Understanding the expectations of the fan and adjusting the content of the message towards that expectations may therefore be the key of an effective way to change behavior.

On the one hand, a Facebook fan participates within a company fan page and browses through online content (see section 1.2 The Different Content Types) to decrease decision making time, lower purchase risks, feel related or part of the brand or form and share opinions about (unfamiliar) brands (Lake, 2011; Zaichkowsky, 1986). Information within the content may therefore be of importance to change behavior.

On the other, Nadkarni and Hofmann (2012) stated that people actively use Facebook for the need to belong and the need for self-presentation. This is often performed in combination of a romanticized reality in which one’s life seems to be more beautiful. Emotions within the content, such as humor, fear, sadness, or inspiration, may therefore be of importance to change behavior (Chiu et al., 2007). Moreover, the majority of Facebook users became a member of Facebook to fulfill the need to belong and the need for self-presentation (Nadkarni and Hofmann, 2012).

As was introduced in section 1.2, informational value within the content is called utilitarian content. Utilitarian content is primarily instrumental, containing tools or information about functions, and cognitive features. It provides consumer value as the means to an end. Therefore it is more focused on the product information itself rather than the brand values (Chiu et al., 2007). Examples of utilitarian content are: product introductions, ingredient information or price information. Utilitarian content provides product information that serves the consumer in reducing the route to purchase and overall purchasing risks (Kusum et al., 2001).

Emotional value within the content is called hedonic content. Hedonic content is an outcome related to spontaneous responses that are more subjective and personal (Chiu et al., 2007). Hedonic content is focused on aspects such as entertainment, exploration, and self-expression. It can be characterized by joy, amusement and exploration (Edwards, 1990). The hedonic content allows the browsing fan to directly perceive the brand values in advertising activities. He or she will perceive the capacity to become part of the brand or play a certain role within the brand (Yang, 2010). Examples of hedonic content are messages primarily based on emotions such as: humor, fear, sadness, or inspiration (Chiu et al., 2007).

(17)

17 to do on Facebook”. Utilitarian content may lower purchase risks and decrease the route to

purchase, and may therefore increase the fan’s purchase intentions (H1e).

Hedonic content on the other hand is more capable of supplying the (actively) browsing fan with the “entertainment, exploration, and self-expression” part of “what people are already trying to do on Facebook”. Moreover, people became Facebook user primarily for the need to belong and the need for self-presentation, matching hedonic value within the content. The fan perceives the capacity to become part of the brand or play a certain role within the brand. Hedonic content may therefore increase the emotion within the message and the company’s brand elements, and therefore increase the fan’s content attitude and brand attitude (H1c and H1d).

The following hypotheses examine the previously described effects.

 H1c: Hedonic content results in a more positive content attitude than utilitarian content does.  H1d: Hedonic content results in a more positive brand attitude than utilitarian content does.  H1e: Utilitarian content results in higher purchase intentions than hedonic content does.

After utilitarian or hedonic messages are posted on Facebook, fans are able to browse through that content on their newsfeed. While browsing through different content, a message may be somehow relevant to a fan. In contrast to traditional media, if fans within the online community are confronted with this relevant information, their motivation to participate in a discussion within the community may increase (Mangold and Faulds, 2009). These responses are online actions on Facebook in the form of: “likes” (button to press if you like the content as a consumer), “shares” (button to post the company’s content with additional text, as a consumer, on one’s page to share it with friends) and “comments” (consumer responses often related to the content posted by the company).

Content that spark receivers emotions, such as humor, fear, sadness, or inspiration, (hedonic) are more likely to be shared (Chiu et al., 2007). So, compared to informational content, it may be assumed that fans are more sensitive towards emotional content such as humorous, entertaining, positive, explorative, and self-expressive posts. So hedonic content matches the moderate fan’s positive emotions on Facebook, increasing the intention to respond (Nadkarni and Hofmann, 2012; Chiu et al., 2007).

 H2: Hedonic content results in a higher willingness to respond than to utilitarian content.

2.2.3 The Moderating Effect of Facebook Involvement

(18)

18 exist as well. Some fans are naturally more active on social networks than others, caused by the, afore mentioned, different “need to belong” and different “need for self-presentation” (Nadkarni and Hofmann, 2012). This might also explain why fans, with the same attitude towards a message, may have different intentions to respond to that same message as was described in section 2.2.2. Lake (2011) categorized Facebook fans into different groups. Based on their involvement on Facebook and their participation on company fan pages, they were categorized from low involved Facebook fans to highly involved Facebook fans. Highly active fans perceive a match between the communication channel and their preferred way of communication, while low active fans are not. In other words, highly active Facebook users may be more “in line with what people are already trying to do on Facebook” than less active users. The degree of involvement may therefore alter the effect of the content. An increase of Facebook involvement may stronger increase the behavioral change caused by the branded content:

H3a: The moderating effect of Facebook involvement on brand attitude or purchase intentions is greater while browsing through content compared to not browsing through content.

The degree of Facebook involvement was explained by the fan’s need to belong and the need for self-presentation. So the emotional value of Facebook matches the reason to become more active on the social network. As hedonic content is focused on aspects such as entertainment, exploration, and self-expression, it shows more similarities with the expectations of the higher involved Facebook user (Chiu et al., 2007). This type of content is comparable to the motivation of being active on Facebook (Nadkarni and Hofmann, 2012). An increase of Facebook involvement may stronger increase the behavioral change caused by hedonic content compared to utilitarian content:

 H3b: The moderating effect of Facebook involvement on content attitude, brand attitude or purchase intentions is greater while browsing through hedonic content compared to browsing through utilitarian content.

2.2.4 The Moderating Effect of Product (Category) Involvement

Product involvement is one of the elements that is taken into consideration while persuading consumers (Hass, 1981). Celsi and Olson (1988) point out that an enhanced level of consumer involvement increase both the attention and comprehension to a brand or product. Consumers who are involved within a product category may have a different approach towards (commercial)

(19)

19 the purchase was seen as important, the information and reduce uncertainty through: mere

willingness to perceive information the consumer is subjected to, and extensive, overt search for relevant information. Product involvement is based on three factors according to Zaichkovsky (1994): the characteristics of that person, the characteristics of the stimulus, and the characteristics of the situation. This implies that people, who are involved, may show different behavior towards both the content and the content type. For example, someone who wants to buy a car will pay more attention to a Volks Wagen message than people who are not willing to buy a car. If this person is interested in the features of the car he or she could pay more attention to informative content. This perspective of involvement is including: 1. the content of the communication, 2. the personal value towards the content and, 3. the situation or type of medium (Zaichkowsky, 1986).

In this research, the type of situation is the communication through social media (Facebook). The content of the communication was split into utilitarian, hedonic or no content. The previous described definition assumes that the personal value of the content depends on the rate of involvement. Petty et al (1983) stated that commercial messages can persuade through emotional and informational messages. Highly involved consumers exhibit more positive evaluations of informational (utilitarian) content, while low-involved consumers tend to accept more emotional (hedonic) communication. This is caused by the thought that content on Facebook is more likely to fall within an acceptable range of an explicit attitude when a person is involved (Petty et al, 1983). Therefore, it may be concluded that utilitarian content will be evaluated more positively by a high involved consumer, whereas, hedonic content will be evaluated more positively by low involved consumers.

According to the definition a high involved person “sees the purchase or consumption as personally relevant or important” and “overtly searches for relevant information”. The involved person knows, or has the perception to know the facts. He or she is therefore more critical, taking the time to look for relevant informational aspects of the message, before coming to an attitude about the content, the brand or the intention to purchase (Yang, 2010).

The non-involved person, in contrast, may fall for the emotional and symbolic (hedonic) aspects of the message (Yang, 2010). So, the conviction towards attitude and purchase results from the emotional content of the message rather than the informational aspects. Therefore it may be concluded that a consumer who does not know or care much about a product, will base its content attitude, brand attitude, and purchase intention on emotional values.

(20)

20

 H4a: The moderating effect of product involvement on brand attitude and purchase intentions is the same while browsing through content compared to not browsing through content.

 H4b: Product involvement positively moderates the effect of utilitarian content on content attitude, brand attitude or purchase intentions.

 H4c: Product involvement negatively moderates the effect of hedonic content on content attitude, brand attitude or purchase intentions.

2.2.5 The Moderating Effect of the Intention to Respond

As was described in 2.2.3, different types of content and different types of fans may positively influence the amount of online consumer responses. However the quantity of those reactions may not correlate with the quality of the responses. So, the impact of the intention to respond to hedonic or utilitarian content, on content attitude, brand attitude and purchase intentions needs to be examined.

The degree of credibility towards content evaluated by the consumer is a fundamental component of the message value according to Brown et al (2007). Moreover, the perceived credibility of the

(21)

21 with a low intentions to respond.

On the other hand, Chiu (2007) also concluded that fans may provide reactions on content because they want to show their joy or positive emotions, corresponding to hedonic content. If consumers intend to respond to hedonic content, they may be more sensitive to the emotional aspects, corresponding with hedonic content (Yang, 2010). Therefore, it may be expected that fans with a high response rate, who have browsed through hedonic content may perceive a more positive content attitude and brand attitude (and lower purchase intention) compared to the consumers with low intentions to respond.

 H5a: The intention to respond positively moderates the effect of utilitarian- and hedonic content on content attitude, brand attitude or purchase intentions.

(22)

22

2. Research Design

3.1 Experimental Design

Three types of stimuli were used to form four different groups: Hedonic, Utilitarian, Content and No content. The participants were either part of the hedonic group, the utilitarian group or the no content group. The hedonic group and the utilitarian group were both stimulated by (different) content; these groups were serving as the content group. The first group was stimulated by hedonic content, a second (different group) was stimulated by utilitarian content and a final third group was stimulated by a photograph of just the product (without any emotions or information). Both the hedonic and utilitarian content types contained the same product, with the same message within the content. However this same message was given with an emotional tone of voice for the hedonic group and an informational tone of voice for the utilitarian group. After the groups were stimulated with (no) content, a survey followed. The flowchart of this survey is shown in figure 3.

As an incentive to take part of the survey, potential respondents could win a “Crystal Clear Summer Package”, worth 50 euro’s after completely filling in the survey (see appendix 1). First some

background information about the experiment was given, followed by two sample questions. Next the respondents were confronted with a picture randomly containing either: utilitarian content, hedonic content or a Crystal Clear-pack shot with a Crystal Clear brand image (see appendix 2). The respondents were asked to thoroughly examine the picture and keep it in mind while completing the survey.

The respondents who were confronted with the Facebook posts were asked their opinion about the just viewed posts to measure the advertising (content) attitude. These questions allowed the respondents to simultaneously think about the content for a second time, reprocessing the information. These same respondents were asked about the intention to respond on the post they just viewed, in order to measure the “intention to respond”. Fourth, all the respondents were asked the same questions to measure respectively: brand attitude, purchase intentions, product

(23)

23

Figure 3: Flowchart of the experiment

3.2 Data Collection

The population consists of fans (mainly women) who are active on Facebook. They were approached via a Crystal Clear online medium (Facebook, email and/or the website). In order to complete the survey, the respondents needed to have a registered Facebook profile. They had one week to fill in the survey. In total 926 respondents started the survey, 703 surveys were completed. After the outliers were deleted 545 respondents remained (N=545). Appendix 3 gives more information about the composition of the total sampling group.

The surveys were distributed via the Crystal Clear Facebook page (2.400 fans), the Crystal Clear website (www.crystalclear.nl) and the Cyrstal Clear newsletter (35.000 receivers).

The survey was hosted by Qualtrics, an online survey hosting page, the research was fielded from the 25th of June 2013 until the 2nd of July 2013. The questionnaire consisted of seven-point semantic differential scales, seven-point Likert scales and open questions. The semantic differential scales were scaled from “totally disagree”, “disagree”, “more or less disagree”, “neutral”, “more or less agree”, “agree” to “totally agree”.

The soft drink Crystal Clear is a relatively homogeneous product. The low calorie flavored

(carbonated) water is available in several different tastes. The brand is primarily focused on highly educated woman in the age of eighteen to thirty-five. The message within the content was focused on the low calorie, flavored and feminine features of the product.

The goal of the Crystal Clear Facebook page is to engage (potential) consumers, by posting content on the company fan page. The aim is to create a viral (snowball) effect with the posted content. The assumption is that more reactions in the form of “likes”, “shares” and “comments” receive a greater reach among actual and potential consumers than less responses. The attitude and purchase intentions may thereby increase, corresponding with the goal of most FMCG-companies (Chaney, 2011).

Hedonic Group N=164 Hedonic Content Questions about:

(24)

24

3.3 Measurement and Composing Construct

This survey contains six different constructs that were measured by existing scale items. The

(25)

25

Table 1: Measurement of the Constructs

Constructs and Cronbach’s Alpha Measures Factor Loadings Literature Brand Attitude (BA) Cronbach’s a: 0,72

How would you best express your feeling towards the Crystal Clear brand?” “Crystal Clear is:”

1. A brand that fits me A Brand that does not fits me

2. An attractive brand An Unattractive brand

3. A brand that makes me happy A brand that does not make me happy

4. (removed) A dull brand An Exciting brand

5. A refreshing brand Not a refreshing brand

6. A confident brand An insecure brand

7. Different than other brands The same as other brands

8. An old fashion brand A contemporary brand

0,34 0,56 0,67 removed removed 0,74 0,66 0,73 Chang and Thorson (2004) and Spears and Sing (2004). Voss et al (2003), Matrixlab Product Purchase Intention (PI) Cronbach’s a: Total: 0,77

Rate the following statements from definitely disagree (1) to definitely agree (7): 1. I will buy Crystal Clear Sparkling Lemon in the near future

2. It is very likely that I will purchase Crystal Clear Sparkling Lemon

3. If I may buy a soft drink in the supermarket, the changes are great it will be Crystal Clear Sparkling Lemon

4. I will buy a flavor of Crystal Clear in the near future 5. When I go to buy a soft drink, it is always Crystal Clear 6. I never buy Crystal Clear

0,84 0,89 0,91 0,32 removed removed

Spears and Singh (2004) Attitude Towards Content (CA) Cronbach’s a: 0,83

Take a look at the following statements with respect to the 2 posts and please indicate how they relate to you.

The two posts were, in my opinion:

1. Interesting Uninteresting 2. Unpleasant Pleasant 3. Good Bad 4. Positive Negative 0,78 0,70 0,91 0,88 Yang et al (2010), Tan and Chia (2007) Product-category Involvement (PCI) Cronbach’s a: 0,91

“Keep the product soft drink in the back of your mind while answering the following statements”, “soft drink:”

1. Is important to me Is unimportant to me

2. Is not related to me Is not related to me

3. Is meaningless to me Is meaningful to me

4. Has much value to me Has low value to me

5. (removed) Intrigues me Can I find boring

6. Is unattractive to me Is attractive to me

7. (removed) Is nice to talk about Not nice to talk about

0,88 0,91 0,87 0,80 removed 0,81 removed Zaichkowsky (1985) and Zaichkowsky (1994) The consumers’ intention to respond on content (IR) Cronbach’s a: 0,85

If I saw one of these two posts on my Facebook wall, I:

1. would see it as an important message if it is often shared by others 2. would have "liked" it

3. (removed) would find it credible, only if there are many positive reactions from others 4. (removed) certainly leave a negative reaction about the post

5. would have shared the post with friends

6. would have left a positive response in the heading 7. was willing to recommend this post to friends 8. liked it if it is liked by many others

0,62 0,75 removed removed 0,84 0,83 0,87 0,62 Harrison-Walker (2001) Anderson (1998) Intensity of Facebook use (FBI) Cronbach’s a: 0,82

1. Facebook is part of my everyday activity 2. I am proud to tell people I’m on Facebook 3. I am nearly using Facebook

4. I feel out of touch when I haven’t logged onto Facebook for a while 5. I feel I am part of the Facebook community

6. I would be sorry if Facebook shut down

(26)

26 In order to form the different constructs within the total model, first the inter-correlation among all of the scale items were analyzed through a principal component analysis. The results can be found in appendix 3. This analysis gave 6 linear combinations (the constructs) of the covariates to avoid multicollinearity. Moreover it gave an indication of the amount of factors that may be extracted to compose a construct.

Second, the composing items within the six separate constructs were analyzed. The factor loadings of each different item within the six constructs were measured. These factor-analyzes show the

contribution of each relevant factor to a construct by the height of the factor loading (table 1). A smaller number of linear combinations of the original factors were produced in a way that captures most variability in the pattern of correlations (Pallant, 2007). All Items with a factor loading between 0,3 and -0,3 were deleted. The remaining items contain a significant contribution to one of the six different constructs (Pallant, 2007).

Finally the Cronbach’s α was used as a bubble check to estimate the internal consistency associated with the scores that are derived from a scale or composed score. The α gives the internally consistent reliable variance. This is an indication of the relationship between the different factors that were measured to indicate a construct. For a reliable indication of the items, the Cronbach’s α needed to be > 0,7 (Nunally, 1978). Eight items were removed, thirty-one items remained, composing the six different constructs.

To measure the attitude towards the overall Crystal Clear brand, an eight-item, seven-point semantic differential scale was adopted. Besides the different attitude items that were compiled by Voss (2003), Chang and Thorson (2004) and Spears and Sing (2004), three additional items were added. These items were extracted from the official research bureau “Matrixlab” that tracks the positioning for Crystal Clear.

Second, a six-item, seven-point semantic differential scale from Spears and Singh (2004) is used to measure the respondents’ intention to purchase Crystal Clear (Sparkling Lemon). The respondents were asked to answer the statements from totally disagree to totally agree.

Third, the construct “attitude towards content” was measured. From the scales of Yang (2010) and Tan and Chia (2007) a four-item, seven-point semantic differential scale was composed.

Fourth, in this study “product-category involvement” is regarded to soft drink as a product category. The scale to measure the respondent’s involvement consists of a seven item seven-point semantic differential method designed by Zaichkowsky (1985 and 1994). The bipolar statement was proceeded by: “Keep the product soft drink in the back of your mind while answering the following statements”, “soft drink is:”.

(27)

seven-27 point Likert scale, composed by Harrison-Walker (2001) and Anderson (1998). Respondents were asked to rate statements that started with: “If I saw one of these two posts on my Facebook wall, I:” Finally, the intensity of Facebook usage was measured with help of six statements rated on a seven-point Likert scale composed by Ellison et al (2007).

3.4 Plan of Analysis

After careful consideration it was decided to use different regression models. This has been considered as regression is normally applied to continuous variables. Content influence (CI) and Content type (CT) however are categorical variables.

The regression analyses examine the difference of the two dichotomous independent variables: Content influence (CI): Content=C or No content=N and Content type (CT): Hedonic content =H or Utilitarian content =U. Both the effects of CI and CT are tested on brand attitude (BA) and purchase intentions (PI) as Dependent variables. The effect of CT is tested on content attitude (CA) as well. Moreover, the effects of the moderating factors: Facebook involvement (FBI), Product category involvement (PCI) are examined in both models. The third moderation effect, Intention to respond, (IR) is added to the CT model as well.

The different statistical analyses are described in this chapter. The analysis were conducted using the software program SPSS.

The conceptual models will be tested in two steps. First, the two main effects (H1) and the effect of CT on IR (H2) are tested through a simple regression. Second, the moderator effects of both models are tested by multiple regressions H3, H4 and H5.

This regression analysis predicts the value of the dependent variable caused by the observed value of the independent variables (Pallant, 2007). The simple regression tests the causal relation between the dependent and independent variables. Because the C, N, H and U are categorical variables, dummies are used to test the effect on CA, BA and PI. The simple regression equations are written as:

H1a BA = α + β1N + ε H1b PI = α + β2N + ε H1c CA = α + β3H + ε H1d BA = α + β4H + ε H1e PI = α + β5H + ε H2 IR = α + β6H + ε

 CA, BA, PI and (here) IR are the values of the dependent variables. These are the values that are being explained by the independent variables.

 H, U and N are the values of the independent variables (X). These variables are explaining the Y variables.

(28)

28

 β is the coefficient of X. This value explains the slope of the regression line. It indicates the change of Y for each one-unit change in X.

 ε is the error term. This value in predicting the value of Y, given the value of X.

Second, a multiple regression analysis is used to predict both the effect of CI on BA and PI when the PCI and FBI are added and the effect of CT on CA, BA and PI when the PCI, FBI and IR are added. The multiple regression is conducted in three steps:

1. the effect of CI or CT on CA, BA or PI when FBI, PCI or IR is zero 2. the effect of FBI, PCI and IR when CI or CT is zero on CA, BA or PI

3. the change in effect of CI or CT on CA, BA or PI as FBI, PCI or IR goes from 0 to 1

The multiple regression equations for H3, H4 and H5 are written as:

BA (CI) H3a H4a α + β1N +β2PCI + β3FBI + β4N*FBI + β5N*PCI + ε

PI (CI) H3a H4a α + β6N +β7PCI + β8FBI + β9N*FBI + β10N*PCI + ε

CA (CT) H3b H4b,c H5a,b α + β11H + β12IR + β13PCI + β14FBI + β15H*IR + β16H*PCI + β17H*FBI + ε

BA (CT) H3b H4b,c H5a,b α + β18H + β19IR + β20PCI + β21FBI + β22H*IR + β23H*PCI + β24H*FBI + ε

PI (CT) H3b H4b,c H5a,b α + β25H + β26IR + β27PCI + β28FBI + β29H*IR + β30H*PCI + β31H*FBI + ε

 CA, BA and PI are the values of the dependent variables. These are the values that are being explained by the independent variables

 α is a constant value that equals the value of the dependent Y variables when the value of X = 0

 β1,2,3… are the beta coefficients or slopes for the independent variables. These values explain

the slope of the regression line. It indicates the change of the dependent variables for each one-unit change in independent variables.

 X1,2,3…. =(N), (H), (IR), (PCI) or (FBI) Independent variables that are explaining the variance in

the dependent variables

 ε is the error of coefficient β1,2,3…. This error in predicting the value of the dependent

variables, given the value of the independent variables

(29)

29

4. Results

This chapter presents and discusses the results of the regressions, in order to present the results of the rejections or acceptation of the hypotheses.

All the regression models are conducted separately for the outcomes: content attitude, brand attitude and Purchase intentions. The sequence of the results is changed, compared to the structure of this report (explained in chapter 5). First the main effects of the content type, the content

influence and on the intention to respond were tested in a simple regression (4.1). Second, the moderating effects were tested separately for the three different outcomes of content type in three multiple regressions (4.2). Third, moderating effect was tested on the content influence outcomes (4.3).

4.1 Simple Regression

4.1.1 Content Influence

It was expected that people who browsed through content would immediately generate a more positive brand attitude compared to the no content group. The analysis confirmed that a statistic mean difference (β=0,096) was found on brand attitude of people who did not browsed through content and those who were browsing through content (ρ=0,025). This relatively small Beta had an effect size of 9,2% (ηρ²=0,092). However, this finding concluded that people, who have not browsed through the content, experienced a slightly higher brand attitude than people who have browsed through content (Table 2) Therefore H1a was rejected, although significance was found.

Second, it was expected that people who browsed through content would generate more purchase intentions compared to the group who did not browsed through content. However, analysis found that there was no statistic difference (β =0,047) between purchase intention of people who browsed through content and those who were not browsing through content (ρ=0,273) (Table 2). Therefore, H1b was rejected as well.

Table 2: Descriptives of Content Influence H1a and H1b

Dependent variable Stimulus N Mean Std. Deviation Std. Error Brand Attitude No content 214 5,819626 ,815223 ,055727

Content 331 5,654834 ,849005 ,046665

Total 545 5,719541 ,839023 ,035939

Purchase Intentions No content 214 5,6152 ,89884 ,06144

Content 331 5,5292 ,89156 ,04900

Total 545 5,5630 ,89459 ,03832

4.1.2 Content type

(30)

30 content would generate a more positive content attitude compared to utilitarian content. Hedonic and utilitarian content are together contributing for a total of only 1,2% of the content attitude within this model (ηρ²=0,012) (Table 3).

Second, it was expected that hedonic content would generate a more positive brand attitude compared to utilitarian content. However no statistical significance was found specifically on the mean difference within the variance of content type (U and H) to confirm H1d: β=-0,007, ηρ²=0,003, ρ=0,939 (Table 3).

Third, no statistic main difference (β =0,111, ηρ²=0,005) of purchase intentions was found for difference in content (ρ=0,247). It was expected that utilitarian content would generate higher purchase intentions compared to hedonic content. However, no statistical significance was found to confirm H1e (Table 3).

Finally the influence on content type on the intention to respond was tested. No statistic main effect (β =0,04, ηρ²=0,002) on the intention to respond was found for difference between hedonic or utilitarian content (ρ=0,468). It was expected that fans that browsed through hedonic content would generate a higher intention to respond, compared to utilitarian content. H2 was therefore rejected.

Table 3: Descriptives of Content type H1c, H1d H1e

Dependent Variable Stimulus N Mean STD Deviation STD Error

Content Attitude Hedonic Content 164 5,6966 ,90997 ,07106

Utilitarian Content 167 5,4895 ,95507 ,07391

Total 331 5,5921 ,93734 ,05152

Brand Attitude Hedonic Content 164 5,6512 ,90997 ,06374

Total 331 5,7195 ,93734 ,03594

Purchase Intention Hedonic Content 164 5,4665 ,85885 ,06706

Total 331 5,5630 ,89468 ,03832

4.2 Multiple Regressions

This section shows the results of the multiple regressions. The moderating effects of Facebook Involvement, Product Category Involvement and the Intention to Respond on the before mentioned relationships were tested by the multiple regression models. First, the difference between browsing through content and not browsing through content was tested. Within this part, just product involvement, Facebook involvement, brand attitude and purchase intentions were included (4.2.1). Second, the content type models including the moderator variables were tested on the three different outcomes (4.2.2).

4.2.1 Content Influence

(31)

31 have an effect on the brand attitude.The amount of variance explained by the type of content and the moderating variables added up is 20,5%.

The hypothesis 1a, difference between no content and content within this brand attitude multiple regression model is now rejected. Not browsing through content, has no significantly different effect on brand attitude compared to browsing through content (β=0,059, t=1,580, p>0,05).

The H3a and H4a were tested within this brand attitude model. The results are displayed in table 4. The results show that both the moderating effect of Facebook and product involvement are

statistical insignificant, rejecting H3a and H4c within the first model.

The second regression shows that the purchase intentions model for content influence is significant. (adj. ηρ²=0,102, F=12,306, p=.000). This means that a change of the independent variables has an effect on the purchase intentions. The amount of variance explained by the type of content and the moderating variables are added to 10,2%.

The hypothesis 1b, difference between no content and content the multiple regression purchase intention model is again rejected. Not browsing through content, has no significantly different effect on purchase intentions compared to browsing through content (β=0,021, t=,502, p>0,05).

H3a and H4a were tested within this purchase intention multiple regression model. The results are again displayed in table 4. The results show that both the moderating effect of Facebook and product involvement are statistical insignificant, rejecting H3a and H4a within this second model.

Figure 4: Multiple regression Content Influence H3a and H4a

Dependent Variable Stimulus B T Sig.

Brand Attitude No Content 0,059 1,580 0,125

Facebook Involvement 0,210 4,156 0,000

Product Cat. Involvement 0,352 7,054 0,000

Facebook involvement*N (H3a) -0,016 -0,312 0,755 Product Cat. Involvement*N (H4a) 0,040 0,800 0,424 Purchase Intentions No Content 0,021 0,502 0,616

Facebook involvement*N (H3a) 0,024 0,440 0,660 Product Cat. Involvement*N (H4a) 0,015 0,276 0,783

4.2.2 Content Type

The first regression shows that the content attitude model is significant. (adj. ηρ²=0,264, F=17,944, p=.000). The amount of variance explained by the type of content and the moderating variables is 26,4%. A change of the independent (moderating) variables have an effect on the content attitude. H1c is supported again, in this model the hedonic content, has a significantly stronger positive effect on content attitude compared to utilitarian content (β=0,138, t=2,900, p<0,05).

(32)

32 involvement level is not significantly influencing the positive effect of hedonic content on the level of content attitude. It was expected that Facebook involvement would have a stronger positive effect when a fan was browsing through hedonic content. For that reason H3b (p=0,391) is rejected within this model.

Moreover, no significant evidence was found to accept H4b,c (p=0,449) or H5b (p=0,148) on content attitude. Neither an increase of the consumer’s product category involvement, nor an increase of the consumer’s intention to respond does have a moderating effect on the relationship between content type and content attitude.

The second regression shows that the brand attitude model is significant (adj. ηρ²=0,257, F=17,346, p=.000). A change of the moderating variables have an effect on the brand attitude. The amount of variance explained by the type of content and the moderating variables, added to 25,7%.

The hypothesis 1d, difference between hedonic and utilitarian content on brand attitude is again rejected within this model. The Hedonic content, has no significantly stronger positive effect on brand attitude compared to Utilitarian content (β=0,025, t=0,517, p>0,05).

In addition, H3b, H4b,c and H5b were tested within the brand attitude model. The results are displayed in table 5. The results show that the moderating effect of Facebook involvement is marginal significant with a Beta of -0,130 (H3b). It indicates that the amount of brand attitude (compared to utilitarian content) is less increasing if a fan’s Facebook involvement increases, while browsing through hedonic content. So, in contrast to the assumption, utilitarian content has a stronger positive effect, rejecting H3b within the brand attitude model.

No significant evidence was found to accept H4b within the brand attitude model (p=>0,05). An increase of the consumer’s product category involvement has no effect on the relation between content type and brand attitude. A p-value of 0,080 was found for the moderating effect of the intention to respond (β=0,121). H5b may therefore be marginal significant. This could mean that (compared to utilitarian content) an increase of the intention to respond may result in a stronger increase of brand attitude while browsing through hedonic content.

The third and final regression shows that the purchase intentions model is significant (adj. ηρ²=0,215, F=12,638, p=.000).. This means that a change of the moderating variables have an effect on the purchase intentions. The amount of variance explained by the type of content and the moderating variables added contain 21,5%.

(33)

33 In addition, H3b, H4b,c and H5b were tested within the purchase intentions model. The results are displayed in table 5. The results show that the moderating effect of Facebook involvement is statistically insignificant. Therefore H3b is rejected within the purchase intentions model. Finally, no significant evidence was found to accept H4b,c or H5b within the purchase intentions model. An increase of the consumer’s product category involvement or intention to respond has no effect on the relation between content type and purchase intentions.

Table 5: Multiple regression Content Type H3b, H4bc, H5b

Dependent Variable Content Type B T Sig.

1. Content Attitude Hedonic Content 0,138 2,900 0,004

Intention to Respond 0,377 5,482 0,000

Facebook involvement*H (H3b) -0,058 -0,859 0,391 Product Cat. Involvement*H (H4b,c) 0,049 0,757 0,449 Intention to Respond*H (H5b) 0,099 1,449 0,148 2. Brand Attitude Hedonic Content 0,025 0,517 0,605

Facebook involvement*H (H3b) -0,130 -1,917 0,056 Product Cat. Involvement*H (H4b,c) 0,010 0,152 0,879 Intention to Respond*H (H5b) 0,121 1,755 0,080 3. Purchase Intentions Hedonic Content -0,051 -1,030 0,304

Facebook Involvement 0,036 ,520 0,603

(34)

34

5. Discussion

The goal of this study was to test the possible causal relationship between the absence and presence of branded Facebook content and between the types of branded Facebook content on the level of content attitude, brand attitude and purchase intention. The main question was:

What is the effect of branded content on brand attitude and purchase intentions; and what is the difference in effect of hedonic branded content and utilitarian branded content; on content attitude, brand attitude and purchase intention, within a specific FMCG-branded Facebook community? The intention to respond, the Facebook involvement and the product category involvement were proposed to somehow moderate the effect of this relationship.

After an extensive literature study the fundaments and possible connections of the independent variables, the dependent variables and the moderators were selected. Based on both the literature study and the empirical study, this section will discuss the main research question. Afterwards the limitations and future research will be discussed.

The discussion of the proposed conceptual models can be divided into two parts: the main effects and the moderating effects. The first part contains the main relationship of (no) content that result in an attitude towards the content, brand attitude and purchase intentions. Because only H1c was significant, this hypothesis is taken as starting point for the discussion of the first part. Second H1d and H1a about brand attitude are discussed. Third H1e and H1b about purchase intentions are discussed.

The second part describes the effect of the relationship of the moderating variables.

5.1 The effect of content

The answer of the second main research question (H1c,d,e) was divided into three different hypotheses: H1c, H1d and H1e. The answer of the question includes the following scope: The difference of influence between hedonic and utilitarian content only influences the attitude towards the content (H1c).

BRANDED FACEBOOK CONTENT: LESS “MESS” WHILE MATCHING THE MESSAGE AND THE MESSENGER?

BRANDED FACEBOOK CONTENT:

LESS “MESS” WHILE MATCHING THE MESSAGE AND THE

MESSENGER?

By

Boaz Disselkoen

BRANDED FACEBOOK CONTENT:

LESS “MESS” WHILE MATCHING THE MESSAGE AND THE

MESSENGER?

December, 2013

By

Boaz Disselkoen

S1637266

Rijksuniversiteit Groningen Faculty of Economics and Business

Master Thesis Business Administration Marketing Management

1

Supervisor: prof. dr. J.C. (Janny) Hoekstra

2

Supervisor: MSc. L. (Lisette) de Vries

Nassaukade 378-3

1054AD Amsterdam

0626977939

Management Summary

Preface

Table of Contents

1. Introduction

D

1.1

The Birth and Use of Online Content Communication

1.2

The Different Content Types

1.3

The Nature of the Fan as Different Moderating Effects

1.4

Problem Statement

1.5

Managerial and Academic Relevance

1.6

Structure of the report

2 Theoretical Framework

2.1

Introduction of the Constructs

2.2

Introducing the Hypotheses

2. Research Design

3.1

Experimental Design

3.2

Data Collection

3.3

Measurement and Composing Construct

3.4

Plan of Analysis

4. Results

4.1

Simple Regression

4.2

Multiple Regressions

5. Discussion

5.1

The effect of content

_{Supervisor: prof. dr. J.C. (Janny) Hoekstra}

_{Supervisor: MSc. L. (Lisette) de Vries}