Exploring How National Culture Moderates
the Responses on Different Kinds of Brand
Posts for Facebook Fan Pages of Global
Brands
L.S.J. Kleine
Newcastle University & University of Groningen
Dual Masters Award in Advanced International Business Management
and Marketing and International Business and Management
16
thof December 2013
Exploring How National Culture Moderates the
Responses on Different Kinds of Brand Posts for
Facebook Fan Pages of Global Brands
By
L.S.J. Kleine
Dual Masters Award in
Advanced International Business Management and Marketing And
International Business and Management
Newcastle University Business School
5 Barrack Road, Newcastle upon Tyne, NE1 4SE, United Kingdom &
University of Groningen Faculty of Economics and Business
Nettelbosje 2, 9747 AE, Groningen, the Netherlands
Student number: 130026929 (Newcastle) and s1771361 (Groningen) Supervisors: Dr Mitchell Ness & Dr André van Hoorn
Confirmation
I confirm that this report is my own work and that all sources are fully references and acknowledged.
I give permission for this dissertation to be made available for viewing by future students via Blackboard.
Word count: 13602
i
Acknowledgements
ii
Contents
Page Acknowledgements i Contents ii List of Figures v List of Tables v Abstract viContent
1. INTRODUCTION ...1 2. LITERATURE REVIEW...5 2.1 Background ... 5 2.1.1 Social Media ... 52.1.2 Social Network Sites ... 8
2.2 Facebook ... 10
2.2.1 Facebook Facts ... 10
2.2.2 Facebook Brand Page ... 11
2.3 Culture ... 15
2.3.1 Culture defined ... 15
2.3.2 Global Brands on Facebook across cultures ... 16
3. THEORY AND HYPOTHESES ... 17
3.1 Form of the Brand Post ... 18
3.2 Substance of Brand Post ... 21
3.2.1 Brand and Non-Brand Related ... 22
3.2.2 Solicitation of Responses ... 25
3.3 Conceptual Model ... 28
4. DATA AND METHOD ... 29
iii 4.1.1 Dependent Variables ... 30 4.1.2 Independent Variables ... 31 4.1.3 Moderator ... 32 4.2 Method ... 33 5. RESULTS ... 35 5.1 Descriptive Statistics ... 35
5.2 Multiple Regression Results... 37
5.2.1 Form of Brand Post ... 37
5.2.2 Brand and Non-Brand Related Brand Posts ... 41
5.2.3 Solicitation of Responses ... 45
6. DISCUSSION & CONCLUSION ... 48
Managerial Implications ... 52
Research Implications ... 52
Limitations & Further Research ... 54
References ... 57
Appendices ... 66
Appendix 1 Theory and Hypotheses ... 66
Appendix 1a Brand Post Substance: classification and labels by other authors ... 66
Appendix 1b Brand Post Substance: this study its classification based on other authors ... 67
Appendix 2 Data ... 68
Appendix 2a Selected Countries ... 68
Appendix 2b Classification Scheme for Independent Variables ... 68
Appendix 2c Individualism in the Selected Countries... 69
Appendix 3 Method ... 70
Appendix 3a Dummy variables for form and substance of brand post ... 70
Appendix 3b Regression Equations ... 71
Appendix 4 Descriptive Results ... 74
Appendix 4a Brand Post Popularity across Countries ... 74
Appendix 4b Number of Brand Post across Countries ... 75
iv
Appendix 4d Brand and Non Brand Related Brand Post across Countries ... 78
Appendix 4e Brand Post with Solicitation of Responses across Countries ... 80
Appendix 5 Results ... 82
Appendix 4a SPSS output for Hypothesis 1a and 1b ... 82
Appendix 4b SPSS output forHypothesis 2a and 2b ... 88
v
List of Figures
Page
Figure 1 Conceptual model………..28
List of Tables
Page
Table 1 Types of Brand Posts.………..17 Table 2 Frequency of Brand Posts.……….………..35 Table 3 Multiple Regression Analysis Results for Form of Brand Post………..38 Table 4 Multiple Regression Analysis Results for Brand/Non-Brand Related
Brand Posts………...41 Table 5 Multiple Regression Analysis Results for Brand Posts with Solicitation
vi
ABSTRACT
This study explores if people from different national cultures seem to respond differently to different kind of brand posts on Facebook brand pages of global brands. This is done by looking at brand post popularity on brand posts that differ in form and substance across low-context/individualistic versus high-context/collectivistic cultures. The data consist of 3.000 brand posts from 10 countries and is analysed by using multiple regression. The study looked at if culture was moderating the effect of different brand posts on brand post popularity. The study finds that brand fans from different cultures do not seem to respond differently on the form of brand posts; however they tend to show different responses on some substance of the brand posts. Nevertheless, this study consists of serious limitations, so the results have to be interpreted with care.
Keywords
1
1. INTRODUCTION
People are turning away from traditional media to social media and use it more often as a source of information for purchasing decisions (Mangold & Faulds, 2009). Therefore, social media is becoming more important in the marketing of global brands and it seems it becomes the norm for global brands to use social media (Okazaki & Taylor, 2013) There are many types of social media which can be successful for global brands (Kaplan & Haenlein, 2010). One of them is the social network site Facebook, which is the most popular social network site in the world (Lipsman, et al., 2012; Hollenbeck & Kaikati, 2012; Nelson, Riebe & Sharp, 2012). Facebook has more than 1.18 billion monthly active users worldwide and is growing every day (Facebook, 2013). Even though this study does not argue that Facebook is the best social media platform for global brands to reach their target group or to communicate their brand message (Kaplan & Haenlein, 2010). It is worthwhile studying due to that so many global brands are active on Facebook and it is such a powerful tool for brand building, relationship building, and electronic word-of-mouth (Lin & Lu, 2011; Rauschnabel, Mau & Ivens. 2013).
2 Vianello, 2011; Goh, Heng & Lin, 2013; Smith, 2013). Therefore, it is important that fans like, comment on, and share the brand post. Another way to describe this is that global brands should try to increase brand post popularity (de Vries, Gensler, & Leeflang, 2012).
However, global brands deal with different cultures and cultures differ in how they use, behave and communicate on Facebook (Rosen, et al., 2010, Vasalou, et al., 2010; Chu & Choi, 2011; Pflug, 2011). Therefore, if brand chose to market their brand on Facebook fan pages, they should consider that culture has an effect on how people respond to their brand posts. There seem to be no study that looks at how fans across cultures respond to different brand posts, hence there is a literature gap. This study is trying to fill this gap by exploring how fans from different cultures respond to different brand posts on Facebook brand pages.
Therefore the aim of the study is to explore if people from different national cultures seem to respond differently to different kind of brand posts on Facebook brand pages of global brands. In order to capture different kind of brand posts, a typology of form and substance was developed based on other authors’ typologies (Cvijikj & Michahelles, 2011; Kwok & Yu, 20102; Men & Tsai, 2012). Culture in this study especially refers to differences between low-context and high-context cultures (Hall, 1956) and individualism versus collectivism (Hofstede, 2001). Other cultural dimensions could also explain differences between the responses of fans across cultures; however this study assumes that the proposed dimensions are most relevant.
3 across cultures on their Facebook brand pages. Little is known about how cultures differ in their behaviour on Facebook, especially on a brand page. Therefore this study responds to many authors their call to perform cross-cultural analysis on Facebook, especially for global brands (de Vries, Gensler & Leeflang, 2012; Okazaki & Taylor, 2013).
In order to explore this, the study empirically analyses 3,000 brand posts of 6 global brands across 10 countries. Multiple regressions were used to find if culture indeed plays a role in the response to brand posts that differ in form and substance. Overall, the analysis did not find support that the form of the brand post show different responses across low-context cultures versus high-context cultures. However, the analysis seem to find some support for that individualistic cultures versus collectivistic cultures respond differently to some substance of the brand posts.
Therefore, this study is able to recommend managers of global brands that operate brand fan pages across cultures, what kind of brand posts they should post on their brand fan pages. Furthermore, this study supplements previous research on what type of brand posts enhance brand post popularity in general. Furthermore, this study sheds light on how culture might directly and indirectly affect brand post popularity. Since, no study has looked to brand post popularity across cultures, this study provides foundations for further research. Nevertheless, this study also contains a lot of limitations; therefore the results should be interpreted with care.
5
2. LITERATURE REVIEW
This chapter is concerned with how global brands can take advantage of Facebook brand pages across cultures. First, social media and social network sites are explained in section 2.1 in order to understand what Facebook actually is. In section 2.2 Facebook itself is described and the rational for global brands to adopt it. In the last section 2.3 it is discussed how global brands can take advantage of Facebook globally, which means the concept of culture is considered.
2.1 Background
This section explains what social media and social network sites are. Subsection 2.1.1 is concerned with social media. It answers questions about why social media can be beneficial and challenging for global brands by comparing it to traditional media. Subsection 2.1.2 is about social networking sites, which deliberates on how social network sites differ from other types of social media and why this might be beneficial for global brands. Overall, this section provides background information in order to understand Facebook.
2.1.1 Social Media
6 generated content could also be brand related, which can then be described as electronic word-of-mouth (Smith, Fischer, & Yongjian, 2012). This could include both positive and negative statements about the brand, which is published online and available to a multitude of people (Hennig-Thurau, et al., 2004).
Nevertheless, brands themselves can also create and exchange information online. Therefore social media is seen as the hybrid element of the promotion mix, which is a way of communicating the brand to the public (Mangold & Faulds, 2009). This is because it offers traditional one-way communication, but non-traditional two-way communication (Mangold & Faulds, 2009). It is argued that social media is especially valuable for global brands in highly competitive industries, especially for business-to-consumer industries (Plangger, 2012). Because, Plangger (2012) finds that there is a positive relationship between social media investment and firm value in these industries. However, the net benefit is unknown, because it is difficult to estimate the exact costs and benefits of the investment. This is however also the case with traditional media (Semenik, et al., 2012). Nevertheless, social media offers advantages for global brands, that traditional media cannot offer.
7 media is faster, cheaper, more direct, more efficient, and has a global reach (Kaplan & Haenlein, 2010; Okazaki & Taylor, 2013). It offers brand building capability to build and reinforce brand image, even stronger that traditional media (Bruhn, Schoenmueller & Schäfer, 2012; Okazaki & Taylor, 2013).
The two-way communication social media provides can offer benefits to the brand, but is also challenging. Electronic word-of-mouth has the potential to shape consumers brand perceptions (Smith, Fischer, & Yongjian, 2012). It is beneficial for the brand if the electronic word-of-mouth is positive, but not if it is negative. Therefore, it is also challenging for global brands (Mangold & Faulds, 2009). This is however not a reason to not use social media, because with or without the global brand present on social media, electronic word-of-mouth happens (Berthon, et al., 2012). On social media information about global brands rarely stays local or general due to the global reach (Berthon, et al., 2012). Therefore, it is advised to global brands to be present on social media, so they can take care of damage control or provide customer service by making use of social media (Dholakia & Vianello, 2011; Anderson, et al., 2012).
In summary, global brands are advised to implement social media into their marketing strategy, because consumers are turning away from traditional media to social media. Even though it is challenging, it offers benefits traditional media cannot offer. These benefits might be even stronger in social network sites, which is the topic of the next section.
8
2.1.2 Social Network Sites
On social network sites people can create a profile and link these profiles to other users, who then get access to the profile (Boyd & Ellison, 2008). Users can then interact, communicate and share information with other users (Lin & Lu, 2011). In this way social network sites are an ideal way to develop and maintain relationships (Lin & Lu, 2011). Most people use social network sites to keep in contact with people they know, so they present their real network (Boyd & Ellison, 2008). This is what makes social networking sites unique compared to other types of social media, because it is about people, relationships and networks (Boyd & Ellison, 2008). In addition, due to the personal relationships present on social network sites, electronic word-of-mouth is more effective on social network sites than on other social media types (Chu & Kim, 2011). Electronic word-of-mouth from somebody who is known and close to the user seems to be more credible and trustworthy (Chu & Kim, 2011).
Brands can take advantage of social network sites due to the efficient, easy and inexpensive way to create networks and weak ties with consumers, and interact with these consumers (Okazaki & Taylor, 2013). If users publish something about the brand on social network sites, it is a kind of “masspersonal communication” (Walther, et al., 2011), because the public and personal is blurred (Grimmelmann, 2010; Meikle, 2010).
9 1.15 billion monthly active users in June 2013 and 1.19 billion monthly active users in September 2012 (Facebook, 2013). Therefore, the reach of Facebook is enormous and increasing every day.
10
2.2 Facebook
This section explains what Facebook is, how it works and why it is beneficial for brands. In subsection 2.2.1 it is discussed how Facebook works and how far it reaches. In subsection 2.2.2 it is examined how global brands can take advantage of Facebook and why this might be beneficial for them.
2.2.1 Facebook Facts
Facebook was founded in 2004, where it was firstly exclusive to Harvard student only. It became accessible to everyone in 2006, also for corporations (Boyd & Ellison, 2008). Facebook became popular due to the communication tools they offer, which are all designed to increase information sharing among networks. This is also showed in Facebook its slogan: “Facebook helps you connect and share with the people in your life” (Facebook, 2013). Facebook users can post a message on their own profile (wall), a message on their “friends” profile (wall), or send private messages to each other (Ellison, et al., 2011). On the first two activities, users can interact by liking, commenting on or sharing the post (Kwok & Yu, 2012). If users post a message on their own wall, friends’ wall or if they interact with a post, this is showed on all their friends their homepage (newsfeed) in chronological order (Wilson, Gosling, & Graham, 2012). In this way people know what is going own in their network (Facebook, 2013). The newsfeed is an important feature, which distinguished Facebook of other social network sites and made it popular (Shih, 2011).
11 2012). Approximately 80% of the daily active users are outside the United States and Canada (Facebook, 2013). In 2013, Facebook has grown to the most popular social network site in most countries. Even in Japan Facebook has reached more active daily users than the local social network site Mixi (Espinosa, 2012), which was not expected (Barker & Ota, 2011; Berthon, et al., 2012). The only countries where it is not the biggest social network site is in China, Russia, Iran and a few other countries (Vincos, 2013). This is due to that it is banned in China, censored in Iran and in Russia they prefer more local social network sites (Berthon, et al., 2012; Vincos, 2013). Overall, Facebook offers an enormous global reach, which is increasing every day. Global brands can and should take advantage of this. They can do this by creating Facebook brand pages, which is explained in section 2.2.1
2.2.2 Facebook Brand Page
Brands can create a profile on Facebook, called brand page. Just as other users they can invite users to become friends (fans) and post messages (brand posts). These brand posts are shown in the newsfeed of all their fans, fans can interact by liking, commenting, and sharing the brand posts, which is again showed on their friends’ newsfeed (Kwok & Yu 2012; Nelson, Riebe & Sharp, 2012; Rauschnabel, Mau & Ivens, 2013). Therefore, brand posts not only reach current fans, but also friends of fans (Lipsman, et al., 2012).
12 acquisitions of new fans and potential customers in a cost-effective way (Chu & Kim, 2011; Dholakia & Vianello, 2011; Goh, Heng & Lin, 2013; Smith, 2013).
Research shows that if fans interact with a brand posts, they become more knowledgeable and aware of the brand (Gummerus, et al., 2012). Furthermore, they feel that the relationship between them and the brand is stronger than before (Gummerus, et al., 2012). When fans interact with the brand post, they also claim that they are three times more likely to do a purchase funnel (prefer the brand over its competitors) than fans that do not interact with the brand (Smith S. , 2013). Finally, it may even increase their purchase expenditures of the brand (Goh, Heng, & Lin, 2013).
Thus, if the fan interact with the brand this might strengthen the relationship between the fan and brand and it automatically leads to electronic word-of-mouth, which means greater market exposure and the possibility of acquiring new customers or fans (Chu & Kim, 2011; Dholakia & Vianello, 2011; Gummerus, et al., 2012; Lipsman, et al., 2012; Nelson, Riebe & Sharp, 2012; Goh, Heng & Lin, 2013; Rauschnabel, Mau & Ivens, 2013; Smith, 2013). Therefore, global brands should try to improve the interaction of fans with their brand posts. This can be done by improving the likes, comments, and shares on their brans posts, which is called brand post popularity (Kwok & Yu, 2012; de Vries, Gensler & Leeflang, 2012). Brand post popularity can be seen as a success metric for fans engagement, participation and content effectiveness of the brand post (Treadaway & Smith, 2012).
13 Spiegler, & Michahelles, 2011; Kwok & Yu 2012; de Vries, Gensler, & Leeflang, 2012). The advice of the authors seem to be contrasting, nobody really knows how to increase responses on Facebook brand pages. Furthermore, these study do not take into account that some of their advice might be only applicable in some cultures, while they can be inappropriate in others. This is because they do not consider that cultures might differ in their respons to different brand posts. For global brands culture is important, because they sell products across various countries.
Even though culture is important, most studies on Facebook are single nation/culture only. Comparisons of samples across countries or cultures are still rare (Wilson, et al, 2012; Okazaki & Taylor 2013). This is a shame, because Facebook offers a lot of behavioural data of users that are well suited to explore cultural differences (Wilson, Gosling, & Graham, 2012). The reasons for this are according to Okazaki & Taylor (2013) that firstly, it might seem too early for studies across countries and cultures. Studies about social media just appeared recently even in the context of single country studies. Secondly, they think that it could be due to the difficulty in coordinating data collection across countries, which might be too enormous and overwhelming. Thirdly, it could be that social media is seen as very personalized medium, instead of a global medium, which might result into a focus of internal aspects, rather than external aspects. Finally, it might be due to “a lack of clear theoretical perspectives from which research hypotheses can be formulated and tested by an empirical study” (Okazaki & Taylor, 2013: p.58).
15
2.3 Culture
Since global brands are sold in various cultures, the concept of culture is considered in this section. In section 2.3.1 culture is defined and in section 2.3.2 it is argued how it might influence behaviour in general and on social network sites such as Facebook specifically. The section ends with the research question.
2.3.1 Culture defined
16
2.3.2 Global Brands on Facebook across cultures
Global brands are sold in multiple cultures, but this does not imply that people across the world consume and perceive the brand the same (de Mooij, 2010). This is because consumer behaviour and communication are influenced by culture (de Mooij, 2010; Semenik, et al.). This also applies to Facebook, which is also a global brand. Even though Facebook is used in most countries in the world, users across the world have different motives to join Facebook (Madpu & Cooley, 2010; Vasalou, et al., 2010), behave differently on Facebook (Rosen, et al., 2010, Vasalou, et al., 2010; Chu & Choi, 2011; Goodrich & de Mooij, 2013), and communicate differently on Facebook (Rosen, et al., 2010; Chu & Choi, 2011; Pflug, 2011). Therefore, if brand chose to market their brand on Facebook fan pages, they should consider that culture has an effect on how people use Facebook. Thus, it might be not appropriate to apply the same strategy across cultures, if the goal is to increase responses to brand posts.
Some studies indeed show that brands use different strategies on Facebook, by using different kind of brand posts (Men & Tsai, 2012; Waters & Lo, 2012). However, they do not look at how fans response to those brand posts. There seem to be no study that looks at how fans across cultures respond to different brand posts, hence there is a literature gap. This study is trying to fill this gap by exploring how fans from different cultures respond to different brand posts on Facebook brand pages. Therefore, the research question of this paper is:
RQ: How do fans from different cultures respond to different brand posts on Facebook
17
3. THEORY AND HYPOTHESES
In order to understand how fans from different cultures respond to different brand posts on Facebook brand pages, first different kind of brand posts have to be identified. This study identified that brand posts can differ on their form and substance (Cvijikj & Michahelles, 2011; de Vries, Gensler & Leeflang, 2012; Kwok & Yu, 2012; Men & Tsai, 2012). Both characteristics are explained in the following sections and overview of them is presented in Table 1. To understand how culture might affect responses on these different kind of brand posts, this study looks especially to Hall (1956) concept of low-context and high-low-context cultures and Hofstede’s (2001) individualism and collectivism dimension. What this means and how it affects the brand posts is explained in the same section as the different brand posts. Section 3.1 looks at the form of brand post and Hall (1956) concept of low-context and high-context cultures. Section 3.2 looks at the substance of brand posts and Hofstede (2001) individualism versus collectivism dimension. Finally, section 3.3 presents the conceptual model.
Table 1 Types of Brand Posts
Type of Brand Posts Description
Form Text
Link Photo Video
Substance Brand / Non-brand Related
18
3.1 Form of the Brand Post
Looking at what influences brand post popularity, many scholars argue that the form of the brand post is determinative (Cvijikj & Michahelles, 2011; Cvijkj, Spiegler & Michahelles, 2011; Kwok & Yu, 2012; de Vries, Gensler & Leeflang, 2012). A brand post can have different forms which can include the following characteristics:
Text
Link
Photo
Video
Most of the studies that investigate the influence of form on brand post popularity, find that brand posts with text receive the most comments and with photos or videos most likes. A brand post with a link seems to be negatively related to both likes and comments (Cvijikj & Michahelles, 2011; Cvijkj, Spiegler & Michahelles, 2011; Kwok & Yu, 2012; de Vries, Gensler & Leeflang, 2012). Nevertheless, these studies do not look at brand posts across cultures, which might lead to different results. Some studies have looked at brand posts across cultures, but not at brand post popularity (Men & Tsai, 2012; Waters & Lo, 2012). Nonetheless, these studies find that brands initiate different forms of brand posts across cultures.
19 Hall (1976) argues that some cultures use low-context and others high-context communication. This is because cultures differ in how much context is needed in the message, which is the information that surrounds an event and produces a given meaning (Hall & Hall, 1990). In low-context communication, most of the information is in the message and not in the context. That is why in low-context communication the message has to be coded and explicit, preferably by direct and verbal communication, where the meaning is in words (Hall, 1976). Examples of low context-communication countries are the Anglo-Saxon world and Northern Europe, while all other countries show more high-context communication (de Mooij, 2014). High-context communication is the opposite of low-context communication; most of the information is in the person or context, and not in the message (Hall, 1976). Therefore, the messages in high-context communication countries show very little coded and implicit messages (Hall, 1976). People from high-context cultures prefer indirect and nonverbal communication, where meaning is derived from context and not words (Hall, 1976).
Looking at the form of brand posts it can be argued that some forms are more popular in low-context cultures, while others are more popular in high-context cultures. Studies find that on websites and social networking sites, brands use more text and links in low-context cultures, while they use more photos and videos in high-context cultures (Cho & Cheon, 2005; Hermeking, 2006; Würtz, 2006; Men & Tsai, 2011; Waters & Lo, 2012). Since brands use more text and links in low-context cultures and more photos and videos in high-context cultures, it can be argued that these forms are also more popular in these cultures.
20 words (Cho & Cheon, 2005; Hermeking, 2006; Würtz, 2006; de Mooij, 2010). In addition, a link can provide more information in case this cannot be included in the brand post itself. Ko, Roberts & Cho (2006) also find that people from low-context cultures are also more willing to click into deeper links on websites than people from high-context cultures. Therefore, brand posts with only text and links could be more popular in low-context cultures than in high-context cultures. Brand posts with only photos or videos have little coded meaning and the meaning can be derived from the context. For that reason brand posts with only a photo or video might be more popular in high-context cultures than in low-context cultures (Cho & Cheon, 2005; Hermeking, 2006; Würtz, 2006; de Mooij, 2010). This leads to the following hypotheses:
H1a: In low-context cultures brand posts with only text or a link will be more popular
than brand posts with only a photo or video
H1b: In high-context cultures brand posts with only a photo or video will be more
21
3.2 Substance of Brand Post
Another determinant of brand post popularity is the substance of the brand post, which represents what and how the brand communicate to its fans. Authors that have tried to capture brand post substance, all used different labels and classification schemes (Cvijikj & Michahelles, 2011; Kwok & Yu, 2012; Men & Tsai, 2012; de Vries, Gensler & Leeflang, 2012). The different labels and classifications schemes for brand post substance are presented in Appendix 1a. This study its classification scheme is mainly based on Men & Tsai (2012) their classification scheme, because it seems to capture all the classifications of the other authors. Appendix 1b shows how the classification of this study captures the classification of the other authors.
22
3.2.1 Brand and Non-Brand Related
Brand and non-brand related brand posts represent what the brand communicates in its brand post. Brand related brand posts consists of brand posts about products, the company, promotions or can be brand entertainment or educational (Men & Tsai, 2012). Looking at the literature some studies have looked how these types of brand posts can enhance brand post popularity. For example, Kwok & Yu (2012) find that conversational brand posts are more popular than sales / marketing brand posts. De Vries, Gensler & Leeflang (2012) find that brand related informational brand post leads to more brand post popularity than non- informational or entertaining brand posts. Even though it is not clear what kind of brand posts are more popular, this might also be different across cultures. Specifically, because the role and purpose of marketing communication differs across cultures, especially between individualistic and collectivistic cultures (de Mooij, 2010; de Mooij & Hofstede 2010).
in-23 group goals priority, emphasize relatedness, and norms determine social behaviour (Triandis, 1995; Triandis & Gelfland, 1998). In individualistic cultures the role and purpose of marketing communication is to persuade the individual, which often involves facts and claims about products (de Mooij, 2010; de Mooij & Hofstede 2010). In collectivistic cultures marketing communication is more used for relationship building and creating trust in order to establish in-group-like relationships with the customer (de Mooij, 2010; de Mooij & Hofstede 2010). Therefore, providing information about the product would not persuade collectivistic consumer, but rather trying to induce positive feeling about the brand might be persuasive (de Mooij, 2010; de Mooij & Hofstede 2010).
24
H2a: In individualistic cultures brand related brand posts will be more popular than
brand posts with brand entertainment or educational and non-brand related substance.
H2b: In collectivistic cultures non-brand related brand posts will be more popular than
25
3.2.2 Solicitation of Responses
Brand post that include a solicitation of responses, asks the fan to like, comment on or share the brand post. Some studies examine how these brand posts increases brand post popularity. For example, Cvijkj, Spiegler & Michahelles (2011) find that if brand post contains a question it receives more comments, while statements receive the least. De Vries, Gensler & Leeflang (2012) find the same, in addition they find that if fans are asked to like something or have to choose between alternatives, this increases likes and comments on the brand posts as well. Therefore, it seems that brand posts with a solicitation of responses increases brand post popularity. Nevertheless, this might be more appropriate in some cultures than others. Brand posts with the solicitation of responses ask fans for their opinion by liking, commenting on or sharing the brand post. Expressing ones opinion is more preferred in some cultures than in others, especially in individualistic and indulgent cultures (Hofstede, 2001, 2011; Hofstede, Hofstede & Minkov, 2010)
26 willing to express their opinion, because it might lead to direct confrontation, which should always be avoided (Kim & Sherman, 2007; Hofstede, 2001).
In addition, most individualistic cultures use low-context communication, while collectivists use more high-context communication (Hofstede, 2001; de Mooij, 2010). As explained in section 3.1 low-context cultures prefer messages which are direct and implicit, whereas high-context cultures prefer more indirect and explicit messages (Hall, 1976). Thus, brand posts with the solicitation of responses might not only be more popular in individualistic cultures, because it allows them to express their opinion. Also, because the direct of way asking fans for a response is more appropriate and might be even necessary in order to get a response. For high-context cultures, these brand posts might not be popular, because it can lead to direct confrontation and the direct nature of asking for a response might not be necessary or appropriate.
27 In summary, it seems that brand posts with a solicitation of responses might be more popular in individualistic, than in collectivistic cultures. Due to that the former appreciates the expression of opinions, while the latter seems to avoid this. Therefore, the following hypotheses are developed:
H3a: In individualistic cultures brand posts with the solicitation of responses will be
more popular than brand posts without the solicitation of responses
H3b: In collectivistic cultures brand posts without the solicitation of responses will be
28
3.3 Conceptual Model
In Figure 1 the conceptual model is showed, which summarizes all hypotheses. The different kind of brand posts are based on the work of Cvijikj & Michahelles (2011), de Vries, Gensler & Leeflang (2012), Kwok & Yu (2012), and Men & Tsai (2012). This will influence brand post popularity, which is a term that is introduced by de Vries, Gensler & Leefland (2012). They however do not include shares on brand posts. This is because the share button is introduced in 2012 (Darwell, 2012), while they obtained their data in 2010/2011 (de Vries, Gensler, & Leeflang, 2012). This study argues that culture is moderating the effect of different kind of brand post on brand post popularity, which means that some kind of brand posts might be more popular than others across cultures. The cultural dimensions are based on Hall (1956) and Hofstede (2001).
Figure 1 Conceptual Model
Brand Post Popularity
Likes
Comments
Shares
On Average per brand fan
Different Brand Posts
Form
Substance
Culture
Low / High Context
29
4. DATA AND METHOD
This chapter describes how this study is going to give an answer on the research question. It is divided into two sections, data and method respectively. In section 4.1 the sample is explained and the variables. In section 4.2 the method is explained.
4.1 Data
In this study the unit of analysis are brand posts on Facebook brand pages of global brands across countries. The sample consist 3,000 brand posts of 6 global brands for 10 countries. The countries that are selected are based on three criteria, linguistic capabilities of author and Facebook usage in the countries. Concerning linguistic capabilities, the author is native Dutch, fluent in English and has good knowledge of German, French and Spanish. From these criteria a few countries were available. India, South-Africa and the Philippines are also selected, because their Facebook fan page is in English. Then the second criteria is applied, Facebook usage. The 8 countries that have the highest score on Facebook usage are selected, based on the data from SocialBakers (2013), except the Netherlands and South-Africa. The Netherlands was added, because the Facebook pages are in the author native language. South-Africa was added, because it is regionally distant from all other countries. Overall, the sample shows variety between regions, countries, and culture. Furthermore, Facebook is popular in the countries and the author is able to research the countries.
30 global brands from Brandirectory (2013) are considered if they have pages for all selected countries. Then it was checked if the fans on the brand page also have the nationality of the brans page. For example, fans on Facebook brand pages of most global fast-food chains in the United States are from somewhere else, only a small percentage of fans are actually from the United States. The reasons for this could be that there is no brand page for their own country or because they did not know there was one. Therefore, these brands are not selected. An overview of the selected countries and brand is presented in Appendix 2b.
These two levels might contain a selection bias. However, because the brand posts are the real unit of analysis, this might be not a problem. Nevertheless, the results cannot be generalized for other countries or other global brands. Finally, at the brand pages the last 50 brand posts are analysed. Facebook shows the newest posts on top and the oldest on the bottom, therefore the first selected brand post is on the 29th of November 2013. The second goes back into time. The time between the first and last selected message depend on the frequency of messaging of the brand. On the 9th of December 2013 all likes, comments and shares are checked again, so that they have at least 10 days to generate responses. This ensures that the brand posts all receive the same time to generate likes, comments and shares. Thus, that the brand posts which were analysed firstly are not disadvantaged.
4.1.1 Dependent Variables
31 activeness, because fans publish it on their own wall, liking and commenting have the same contribution (Muntinga, Moorman, & Smit, 2011). The number of fans differs across time; therefore the numbers of fans are checked on the date the brand post was released. Socialbakers.com is a website that keeps track of the number of brand fans across time; hence this website is used to obtain the number of brand fans present when the brand post was released. Socialbakers is also called the heart of Facebook statistics (Al Omoush, Yaseen, & Alma'aitah, 2012)
4.1.2 Independent Variables
The independent variables are the form and substance of brand posts. Form is derived from the analysis of Cvijikj & Michahelles (2011) and Kwok & Yu (2012). They both look at the relationship between text, link, photo and video and brand post popularity. De Vries, Gensler & Leeflang (2012) also look to these form, but classify it differently. The substance of the form is mainly based on Men & Tsai (2012) their classification. How and why was explained in section 3.2. It exists of two independent variables, namely brand or non-brand related substance and the solicitation of responses. Even though this categorization seems to capture different brand posts and how this might differ across cultures, it might be not the most optimal categorization. It could be that other factors are more important in brand posts, especially across cultures. Since, no study has looked at brand post popularity across cultures this is hard to tell.
32 scheme is produced, which is presented in Appendix 2b. After the data collection, 200 cases were again analysed and compared to the original results. Only 2 % was different and mainly this was about product or company related brand posts. Therefore, it seems the interpretation was valid.
Even though, it is tried to minimize interpretation biases, the validation of the interpretation is still problematic. This is even more problematic because the brand posts are in different languages. If necessary the translation option, which Facebook offers on their brand posts, is used. To minimize all interpretation biases, the brand posts are carefully interpreted and if doubts exist another person will be asked to interpret it. The brand post is accepted if both agree, otherwise it is going to be not coded.
4.1.3 Moderator
33 individualism, such as in-group and institutional collectivism of House, et al. (2005) and embeddedness versus intellectual and affective autonomy of Schwartz (1994). This study chose Hofstede, because most psychology, sociology, marketing or management studies are using this framework (Sondergaard, 1994; Steenkamp, 2001; Soares, Farangmegr & Shoham, 2007). The selected countries and their scores on the dimensions are presented in Appendix 2c.
4.2 Method
This study uses multiple regression analysis to understand if culture is moderating the relationship between the form and substance of brand posts and brand post popularity. This is done by using SPSS (2012). First the data is cleaned data by using descriptive analysis. It might be that the data contains missing values or out of range values, these might be typing errors (Malhotra, 2010). Then the original message is checked again and the values are replaced. Outliers are not retained, since the numbers represent true values. Furthermore, there are not many outliers.
34 1991). Then, the dummy variables are multiplied by centered culture, which represent the moderator effect. A three step approach is used to test if the moderator effect is significant (Hair, Black, & Anderson, 2010). Firstly, the original unmoderated equation is estimated. Secondly, the moderation affect is added and estimated. The final step is to assess the R² change, and see if it is statistically significant. If this is the case, then a moderation effect is present (Hair, Black, & Anderson, 2010).
35
5. RESULTS
This chapter describes the results of the multiple regression analysis for the moderating effect of culture on form and substance of brand posts on brand post popularity. Section 5.1 examines the descriptive statistics, while in section 5.2 the results of the multiple regressions are presented.
5.1 Descriptive Statistics
In this study 3000 brand posts were analysed for 10 countries. Thus for every country 300 brand posts were analysed. This was done for 6 global brands. On average brand posts received 0.0025 likes, 0.002 comments, and 0.003 shares per brand fan. Looking at the form and substance of brand posts, Table 2 presents the frequency of the brand posts. Most brand posts included text and a photo, were product related or brand entertaining or educational related and included not solicitation of responses.
Table 2 Frequency of Brand Posts
36
Form Text and Photo 1671
Only Text 101
With Link 808
Only Photo 197
With Video 223
Substance Product Related 1028
Company Related 207
Promotion Related 242
Brand Entertaining or Educational Related 1211
Non-Brand Related 312
Solicitation of Likes 74
Solicitation of Comments 782
Solicitation of Shares 55
No Solicitation* 2115
* please note that brand posts with a solicitation of responses do not sum to 3000, because some brand posts include a solicitation of likes, comments and shares together.
37 brand posts with the solicitation of shares receive most shares in Argentina. Thus it seems that there is a lot of between country variations.
5.2 Multiple Regression Results
This section consists of three subsections, which deals with hypothesis 1, 2 and 3, respectively. First an overview of the results is provided followed by an interpretation o f the results and a conclusion.
5.2.1 Form of Brand Post
38 Table 3 Multiple Regression Analysis Results for Form of Brand Post
n = 3000 Likes Comments Shares
Model 1
Text and Photo (baseline) –– –– ––
Only Text -.002277*** -.000069 -.000219 With Link -.000958*** -.000125** -.000041 Only Photo -.000865** -.000186** -.000012 With Video -.000871** -.000055 .000429*** LC Centered .000000 .000004*** .000001 Adj. R² .010 .006 .007 Model 2
Text and Photo (baseline) –– –– ––
39 Numbers show unstandardized beta coefficients
* p < 0.10, ** p < 0.05, *** p < 0.01
Number of Likes
Looking at the first model, it shows that compared to brand posts with text and a photo, all the other forms of brand post are significantly and negatively related to likes. Brand posts with only text (β = -.002277, p = 0.000) or with a link (β = -.000958, p = 0.000) are even more significantly and negatively related to likes than brand posts with only a photo (β = -0.000865, p = 0.029) or with video (β = -.000871, p = 0.020) compared to brand posts with text and a photo. Finally, low-context or high-context cultures seem not to be related to the number of likes.
40 slightly more brand posts with links are liked compared to brand posts with text or photo than in high-context cultures.
Number of Comments
Looking at the variables that influence the number of comments in model 1, especially low-context cultures are significantly and positively related to the number of comments (β = 0.000004, p = 0.000). Also, all forms of brand posts seem to be negatively related to the number of comments compared to brand posts with text and a photo. However, only brand posts with a link (β = -0.000125, p = 0.012) or only a photo (β = -0.000186, p = 0.033) are significant. The second model shows, that the R² change is 0.001, but again this is not significant (F = 0.0475, p = 0.754). Therefore, hypothesis 1a and 1b for comments are also not supported.
Number of Shares
41 model 2, the R² change is 0.0003600 and it is not significant (F = 0.271, p = 0.897). Therefore, hypotheses 1a and 1b for share is also not supported.
Conclusion
In summary, for hypothesis 1a and 1b the analysis found no support, hence they should be rejected. It seems that brand posts with text and a photo leads to more brand likes and comments across cultures. Only brand posts with a video seem to be more shared. Finally, fans from low-context cultures are commenting more than fans from high-context cultures.
5.2.2 Brand and Non-Brand Related Brand Posts
In this subsection hypothesis 2a and 2b are tested. The hypotheses assumed that in individualistic cultures brand related brand posts will receive more likes, comments and shares than brand posts which are non-brand related brand posts. For collectivistic cultures it was the other way around. In order to test the hypotheses, non-brand related brand posts were selected as the baseline. An overview of the results are presented in Table 4 Appendix 4b shows the SPSS output of the analysis, which are used for constructing Table 4 In the next subsections the results for likes, comments and shares are interpreted.
Table 4 Multiple Regression Analysis Results for Brand/Non-Brand Related Brand Posts
n = 3000 Likes Comments Shares
Model 1
Non Brand Related (baseline) –– –– ––
Product Related .002121*** -.000051 .000228***
Company Related .001550*** -.000088 .000170
Promotion Related .000794* -.000124 .000121
42
IND Centered -.000001 .000004*** .000000
Adj. R² .014 .003 .001
Model 2
Non Brand Related (baseline) –– –– ––
Product Related .002132*** -.000050 .000231***
Company Related .001556*** -.000089 .000172
Promotion Related .000784* -.000119 .000112
Brand Ent / Edu Related .001136*** -.000066 .000186**
IND Centered .000012 .000011*** -.000000
INDxProduct Related -.000002 -.000008** .000005 INDxCompany Related -.000016 -.000009* .000002 INDxPromotion Related -.000008 -.000008 .000004 INDxBrand Ent / Edu Related -.000026 -.000007** -.000003
Adj. R² .015 .004 .003
∆ R² .002 .002 .003*
Numbers show unstandardized beta coefficients * p < 0.10, ** p < 0.05, *** p < 0.01
Number of Likes
The first model shows that product related (β = 0.002121, p = 0.000), company related (β = 0.001550, p = 0.001), and brand entertainment or educational related (β = 0.00129, p = 0.001) brand are positively and significantly related to the number of likes compared to non-brand related brand posts. Promotion related brand posts appear to be marginally and positively related to the number of likes compared to non-brand related brand posts (β = 0.000784, p = 0.077). Individualism seems not to be related to the number of likes.
43
Number of Comments
Model 1 in Table 4 shows that individualism has a strong significant and positive effect on the number of comments (β = 0.000004, p = 0.000). The different brand posts are not significantly related to the number of comments compared to brand posts with non-brand related substance, but have a negative slope. Model 2, shows that the change in R² is 0.002 and that it is not significant (F = 1.326, p = 0.258). Nevertheless the moderators are negatively and significant for product related brand posts (β = -.000008, p = 0.026) and brand entertaining or educational related brand posts (β = -.000007, p = 0.048). Company related brand posts seem to be marginally and negatively related (β -.000009, p = 0.091). Thus it seems that collectivistic fans comment more on product related and brand entertaining or educational related brand posts than on non-brand related brand posts compared to individualistic fans. They also marginally comment more on company related brand posts. This is contrary to what was hypothesized, hence H2a and H2b for comments is rejected.
Number of Shares
44 beta coefficients of the moderators, it seems that individualistic fans seem to share more brand posts which are product, promotion, and company related than non-brand related posts. Collectivistic fans seem to share more brand entertaining or educational related brand posts than non-brand related posts compared to individualistic fans. This seems to partially support hypothesis 2a and 2b for shares, however the moderators are not significant, so they cannot be accepted.
Conclusion
45
5.2.3 Solicitation of Responses
This section covers the analysis for hypotheses 3a and 3b, which considers brand posts with the solicitation of responses. For the analysis, the baseline of the dummy variable was no solicitation of responses. The result of the analysis is presented in Table 5 and are interpreted in the following subsections for likes, comments an shares, respectively. The data output of SPSS is accessible in Appendix 4c.
Table 5 Multiple Regression Analysis Results for Brand Posts with Solicitation of responses
n = 3000 Likes Comments Shares
46 No Solicitation (baseline) –– –– –– Solicitation Like .001501** .000058 -.000186 Solicitation Comment -.000047 .000366*** -.000085 Solicitation Share -.000130 -.000063 .001162*** IND Centered -.000000 .000002 .000001 INDxSolicitation Like .000034 .000003 .000009 INDxSolicitation Comment -.000011 .000008*** .000001 INDxSolicitation Share -.000007 .000008 -.000023*** Adj. R² .001 .026 .019 ∆ R² .001 .003** .003**
Numbers show unstandardized beta coefficients * p < 0.10, ** p < 0.05, *** p < 0.01
Number of Likes
The first model shows that the solicitation of likes also has a positive and significant effect on the number of likes (β = 0.001533, p = 0.015). It seems that the more the brand posts asks for comments and shares, the less likes the brand posts receives. In the second model the change in R² is 0.001, which is not significant (F = 0.799, p = 0.506). Thus individualism is not moderating the effect of the solicitation of likes and the number of likes the brand post receives. This means that the analysis does not find enough support for hypothesis H3a and H3b concerning the likes, hence it is rejected for likes.
Number of Comments
47 not significant. Individualism is positively and significantly related to the number of comments (β = 0.000003, p = 0.002), which was already known. By adding the moderator effect of individualism in model 2, the R² change is 0.003, which is significant (F = 3.381, p = 0.018). The moderator effect of individualism on solicitation of comments is positively significant (β = 0.000008, p = 0.000). This means that individualistic fans comment even more on brand post with a solicitation for comments than collectivistic fans. Nevertheless, the variance of the dependent variable seem not to be explained well (R² = 0.28, Adj. R² = 0.023). This would suggest that collectivistic fans comment more on brand posts that do not include the solicitation of comments. However, looking at the histograms that were produced in Appendix 4e, it seems that on averaged all countries comment more on brand posts with a solicitation of comments. Thus probably a mistake is made in the interpretation of the moderator.
Number of Shares
48
Conclusion
H1a and H1b are rejected for likes and shares, but there seems to be support for comments. For likes it appears to be that brand posts with a solicitation of likes, receive more likes across cultures. Brand posts with the solicitation for comments seem to lead to more comments in individualistic cultures than in collectivistic cultures. Finally, if the brand post asks for shares, collectivistic fans share the brand post more often than individualistic fans, which is contrary to the hypotheses.
6. DISCUSSION & CONCLUSION
The aim of the study was to explore if people from different national cultures seem to respond differently to different kind of brand posts on Facebook brand pages of global brands. In order to capture different kind of brand posts, a typology of form and substance was developed based on other authors’ typologies (Cvijikj & Michahelles, 2011; Kwok & Yu, 20102; Men & Tsai, 2012). Then the study analysed the responses on brand posts by counting the average number of likes, comments, and shares on brand posts for 10 different countries, which differed along low and high context communication and individualism versus collectivism. Finally, the mediation effect of culture was measured by using multiple regression analysis. The regression equations tested if responses on different kind of brand posts differed across cultures.
high-49 context cultures. Both cultures seem to like and comment more on brand posts with text and photo than the other forms. These findings seem to contrast the findings of Cvijkj, Spiegler & Michahelles (2011) and de Vries, Gensler & Leeflang (2012). They find that fans like more videos and comment more on text. Furthermore, it seems that a brand post with a video receives the most shares. Finally, there seems to be a small change that brand posts with a link receive more likes in low-context cultures, compared to high-context cultures, which was hypothesized. Overall, the study found no support for that low-context and high-context cultures respond different to brand posts if they have different forms. Looking at the descriptive results in Appendix 4c it seems that fans from Argentina and France respond more to brand posts with only a photo or a video compared to only text or with a link. Whereas Mexican fans seem to respond more to only text or a link compared to brand posts with only a photo or video. Thus, even though there is variation between countries, this is not captured by low-context or high-context cultures.
50 Looking at brand posts with the solicitation of responses, the findings suggest that especially for comments and shares individualistic and collectivistic cultures differ. For likes, both cultures seem to like more brand posts that ask for likes. For comments, brand posts with the solicitation of comments, receive more comments from individualistic fans than from collectivistic fans. In all analysis it was found that fans from individualistic cultures seem to comment more than collectivistic fans. Thus, commenting seems to suit individualistic cultures, which is not that surprising. Commenting is a direct way of expression in words; this seems to suit individualistic cultures because they use more often low-context communication (Hall, 1956; Hofstede, 2001; de Mooij, 2010). If individualists are asked for an opinion in the brand post by commenting, they will provide this more often than if no opinion is asked. This is also no surprise, because it was argued that individualists have a positive attitude towards expressing their own personal opinion (Hofstede, 2001; Kim & Sherman, 2007).
51 information in collectivistic societies (de Mooij & Hofstede, 2010; Goodrich & de Mooij, 2013). Therefore, it might be not a surprise that when a collectivistic fan agrees with the brand that they share this with their friends. It also appeared that in more collectivistic cultures, the solicitation for shares seem to be communicated differently. For example, on collectivistic brand posts, the brand said more often: “do not forget to share this message with your friends”, while on individualistic brand posts the brand said more often: “share this message if you agree”. The former seems to better suit the option of sharing.
Overall, the analysis did not find support that the form of the brand post and if the brand post is brand or non-brand related show different responses across individualistic/low-context cultures versus collectivistic/high-context cultures. However, the analysis seem to find support for that individualistic/low-context cultures versus collectivistic/high-context cultures respond differently to brand posts with the solicitation of responses for comments and shares.
52
Managerial Implications
This section deals with recommendations for managers of global brands that operate brand fan pages about how to increase brand post popularity on their brand posts across individualistic/low-context cultures versus collectivistic/high-context cultures. First the recommendation for likes are presented, followed by comments, and lastly for shares.
In order to enhance the number of likes on brand posts, a brand post that consist of text and a photo, is brand related and have a solicitation for likes would generate most likes compared to other brand posts. In low-context cultures it is advised to add a link to the brand post.
For comments, brand posts that consist of text and a photo and is non brand related seems to generate most comments compared to other brand posts. Furthermore, it is recommended to include a solicitation for comments in the brand post, specifically for individualistic cultures. In general, individualistic fans comment more than collectivistic fans, thus keeps this in mind when the effectiveness of brand posts are considered across cultures.
Finally, to increase the number of shares on brand posts, it is advised to include a video in the brand posts and to keep it brand related. In addition, in collectivistic cultures it is recommended to include a solicitation of shares into the brand post.
Research Implications
53 has to be carried out on what causes responses to brand posts. This also concerns the typology of different kind of brand posts. This study offers a typology, which can be used in other studies. Nevertheless, the form might not be that appropriate and the non-brand or non-brand-related non-brand posts should be modified.
This study also contributes to studies that look at cultural behaviour of people on Facebook. It seems that people from individualistic cultures seem to comment significantly more that people from collectivistic cultures. This could be explained by that stating your opinion in words suits individualistic cultures better. If brands want to increase conversations with their fans, they have to consider if this is worthwhile doing in collectivistic cultures. Finally, this study seems to show that different kind of brand posts lead to different responses across countries. Even though individualism versus collectivism might not explain this, other cultural dimensions might do this.
54 research hypotheses can be formulated and tested by an empirical study” (Okazaki & Taylor, 2013, p. 58). This study has tried to contribute to the literature by comparing cultures, however found indeed the problems which are described above. Facebook is very interesting for its behavioural data, but it is difficult to obtain a valid sample.
Limitations & Further Research
This study suffers from several limitations. First, the data has limitations. Not enough countries were selected to test the real effect of culture. Furthermore, it would have been better if the study included East-Asian countries, because these represent more collectivistic values than the countries that are now included. Moreover, some countries used Facebook in English even though this is not the national language of the countries. This might influence fans their response to brand posts. In addition, the number of brands is too small to generalize finding for other brands. Also product category is a problem, only cars and beauty products are selected. In addition to this all, the sample might also concern a selection bias.
Second, the classification of brand post might not be the best solution. For brand form it might be better to look at message length and the purpose of the photo. For brand related / non-brand related brand posts a combination of brand / non brand related entertainment and brand / non brand related information might have been better. The solicitation of responses seems to be appropriate.
55 Fourth, the selected variables might not be sufficient only. For example, the dependent variable comments does not consider if the comments are negative or positive. Furthermore, it could be that even though fans do no respond to the message it will influence their perception of the brand positively, maybe even more than with responding to the brand post. Moreover, the moderation effect of culture only exists of individualism versus collectivism. It is better to look at other cultural dimensions of Hofstede as well, like masculinity, power distance, uncertainty avoidance, long-term orientation, and indulgent versus restraint. In general, using Hofstede dimensions and looking to national culture offers limitations. Not only because Hofstede’s data and method have received many criticism and that the cultural dimensions represent values that are not related to Facebook or brand post popularity, but also because culture does not apply to nations, but societies. It might be the case that fans that do respond to brand posts seem to deviate from the norm. Finally, using Hofstede’s dimension individualism might not be a valid proxy for low-context cultures.
56 Six the interpretation of the data might not be valid, due to a lack of statistical knowledge capabilities of the author. The author had especially problems with interpreting the interaction effect of culture. It would have been better to select another methodology, so that the author was able to interpret the results. However, the methodology seems to fit the study Therefore, the discussion and conclusion should be interpreted with care. Furthermore, the results cannot be generalized for other countries, brand, non-fan pages, and other social network sites.
