Master’s Thesis
How Do Consumers Chat About Your
Brand on Facebook Versus Instagram?
A study on the effect of the type of social media platform on brand-related user-generatedcontent for hedonic and utilitarian brands.
M.G. van Barneveld, BSc | 11274557 MSc in Business Administration, Digital Business
University of Amsterdam Dr. J.Y. Guyt, supervisor
Statement of Originality
This document is written by Student Maartje van Barneveld who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Abstract
In a time where everyone is constantly online, consumers have more power than ever. It is highly relevant for managers to understand how consumers engage with different social media platforms, and how they can improve the effectiveness of their own online presence. This study examines the variances in brand-related user-generated content (UGC) between Facebook and Instagram for utilitarian and hedonic brands. It is the first study to include Instagram in a cross-channel content analysis, and the first to consider the influence of brand category.
The two questions that lead this study are “How does the type of social media platform affect brand-related user-generated content?” and “How is the effect of the type of social media platform on brand-related user-generated content moderated by brand category?”. A cross-channel content analysis is conducted of 500 posts for five utilitarian brands and five hedonic brands. The data is compared on five dimensions that are based on prior literature. The findings indicate that UGC on Facebook is more likely to feature brand centrality, marketer-directed communication, factual information, and negative brand sentiment. UGC on Instagram is more likely to feature promotional self-presentation and positive brand sentiment. Additionally, the results show that brand category indeed moderates the effect on promotional self-presentation, factual information, and brand sentiment. This study provides managerial implications that help business executives decide what platforms have most potential for the brand, how they can allocate resources accordingly, and how they can leverage these platforms in line with their objectives.
Acknowledgements
I would first like to thank my supervisor, Jonne Guyt for investing his time and effort in guiding me through the process of writing my Master’s thesis. I highly appreciated his calmness and positivity.
From my Bachelor’s thesis experience I can say that supervision has a strong impact on the entire process. Thank you Jonne for trusting me and making this a pleasant experience.
I would also like to thank my second reader for taking the time to read my Master’s thesis, and for providing valuable comments.
A warm thank you goes out to my close family, friends and roommates for reading my work, tolerating my thesis talk, coding my subsample, advising me on decisions and most importantly for
believing in me. I couldn’t have done it without your support.
This Master’s program has taught me a great deal about the digital future that lies ahead of us, and how this can enrich the business world. I can’t wait to put my knowledge into practice, and to always
keep learning along the way.
Contents
1. Introduction 6
2. Conceptual Framework 10
2.1 Brand-Related UGC Dimensions 10
2.2 Social Media Platforms and Brand-Related UGC 11
2.3 Brand Category and Brand-Related UGC 17
3. Methods 22 3.1 Sampling 22 3.2 Coding 24 3.3 Inter-Coder Reliability 25 4. Results 26 4.1 Promotional Self-Presentation 27 4.2 Brand Centrality 28 4.3 Marketer-Directed Communication 28
4.4 Factually Informative about the Brand 28
4.5 Brand Sentiment 30
5. Discussion 31
5.1 Validity 31
5.2 Effect of Social Media Platform on UGC 31
5.3 Moderating Effect of Brand Category 34
5.4 Limitations 37
5.5 Future Research 38
6. Conclusion 39
1. Introduction
With 68.3% of the internet users worldwide being active on social media in 2016, it is reasonable to say that social media penetration is in its full swing. The number of social media users worldwide grew from 0.97 billion in 2010 to 2.14 billion in 2015, and is expected to reach 2.95 billion in 2020 (Statista, 2016). In today’s world, social networking is one of the most popular activities online. With high user engagement rates and a large potential for mobile, social media has made its way to the top of the agendas of business executives (Statista, 2016; Kaplan & Haenlein, 2010). Social media is not only leveraged for one-way communication in the form of digital advertising, it is also a great tool to personally engage with consumers, handle customer service issues, and even realize a co-creation environment in which consumers can share ideas (Smith, Fisher, & Yongjian, 2012). Companies have access to a growing number of social media platforms, that are all unique in their format and use. Understanding how social media platforms vary can help business executives to optimally leverage each platform, and to decide how to allocate resources accordingly.
While the rise of social media comes with many opportunities, businesses are not always comfortable with the increasing empowerment of consumers. This new environment encourages the emergence of user-generated brand communications (Burmann & Arnhold, 2009). Previously, companies could control what information was available to consumers through their own marketing
Interesting for managers is that a large share of UGC is brand-related, and potentially influences the consumer’s perception of the brand (Smith et al., 2012). Even though companies no longer have full control of what information about them can be found, they do not have to stand on the sidelines and watch how consumers define their brand’s online existence. Managers can act upon findings in research that compares how consumers engage in different types of social media to increase the effectiveness of their own online presence.
Previous research has analyzed social media platforms in isolation, but Smith et al. (2012) were the first to compare UGC on YouTube, Facebook, and Twitter within one study. They found significant differences in the level of self-promotion, brand centrality, marketer-directed communication, response to online marketer action, and factual information about the brand, as well as differences in brand sentiment across the platforms. These findings provide primary insights the variances in UGC across social media platforms, and have implications for managers who are concerned with social media. The article makes recommendations for marketers by describing for each social media platform what marketing strategies would be most effective, considering its functionalities. The authors suggest future research to confirm these insights, and to also extend into other types of social media and content dimensions (Smith et al., 2012).
Continuing this line of research, this study conducts a cross-channel UGC analysis for Facebook and Instagram. With 90 million monthly active users in January 2013 and 600 million in December 2016, Instagram is an increasingly growing social media platform (Statista, 2017). It has a high user engagement rate, which makes it an important social media marketing tool. As of 2014, two thirds of all luxury retail brands that used social media were also active on Instagram (Statista, 2017). Evidently, Instagram has proven to be an important platform for marketers. However, no previous research on UGC has included Instagram in a cross-channel analysis. This study aims to fill this gap
in the literature by comparing how UGC varies across Facebook and Instagram. As such, the first research question of this study is:
RQ1. “How does the type of social media platform affect brand-related user-generated
content?”
Another gap in the literature that has not been addressed yet is how brand category relates to differences in brand-related UGC. As adopted by Voss, Spangenberg and Grohmann (2003), consumer attitudes towards products or brands can be conceptualized in two dimensions: “The first dimension is a hedonic dimension resulting from sensations derived from the experience of using products, and the second is a utilitarian dimension derived from functions performed by products” (p. 310). As such, more sensational brands can be categorized as hedonic, and more functional brands as utilitarian. The body of research on the interaction between this categorization and UGC is still in its infancy. Pan and Zhang (2011) have investigated the helpfulness of a specific form of UGC: user-generated product reviews for customers. Different outcomes were found for hedonic and utilitarian products. The study suggests that hedonic products result in a lower perceived helpfulness of the review than utilitarian products. This finding is explained by the nature of both product categories. “Due to the nature of utilitarian products, reviews in this category are likely to be factual and objective, reflective of the functionality-driven consumer experiences. On the other hand, experiential reviews are likely to be
RQ2. “How is the effect of the type of social media platform on brand-related user-generated
content moderated by brand category?”
The contribution of this study is threefold. First, it improves our academic understanding of the interaction between social media platforms and UGC. Second, it reveals differences in UGC for hedonic versus utilitarian brands. Third, this study offers managerial implications. Understanding what factors cause variances in UGC can help managers of both hedonic and utilitarian brands decide how to allocate resources across different social media platforms to optimally leverage them. This study proceeds with a conceptual chapter in which the hypotheses are introduced. The following section elaborates on the research method. In the discussion section, the results of the content analysis are presented are interpreted, limitations of this study are discussed, and suggestions for future research are provided. The conclusion summarizes the findings of this study and addresses the managerial implications.
2. Conceptual Framework
In this section, the conceptual background of this study is explained. The dimensions that are used in this study to research brand-related UGC are identified, and the relationships between the core concepts are addressed. Lastly, the hypotheses are presented.
2.1 Brand-Related UGC Dimensions
To answer the research questions, five dimensions have been identified to serve as a framework for analyzing and describing UGC. They are based on a framework that allows for the comparison of brand-related UGC, developed by Smith et al. (2012). The first is ‘promotional self-presentation’, and measures whether the post self-promotes the author as well as the brand. The second is ‘brand centrality’, and measures whether the brand is central in the post. Thirdly, ‘marketer-directed communication’ tells us whether the post is directed towards the marketer of the brand. Fourthly, ‘factually informative about the brand’ measures whether the post presents or requests brand-related factual information (Smith et al., 2012). These four dimensions are binary, and are coded as either ‘yes’ or ‘no’. The fifth dimension ‘brand sentiment’ measures the overall sentiment of the poster towards the brand, and has four categories. It is coded as either ‘negative’, ‘neutral’, ‘positive’, or ‘unclear’. The five dimensions serve as the dependent variables in the data analysis.
2.2 Social Media Platforms and Brand-Related UGC
Kaplan and Haenlein define social media as “a group of Internet-based applications that build on the ideological and technological foundations of Web 2.0, and that allow the creation and exchange of User Generated Content” (2010, p. 61). The number of people getting access to social media is continuously growing, as well as the number of social media platforms available. Social media platforms are all unique and can be categorized in numerous ways, each highlighting different aspects. A well-established framework for defining social media platforms is the honeycomb of social media (Kietzmann, Hermkens, McCarthy, & Silvestre, 2011). This framework consists of seven building blocks that can each be used to examine a facet of a social media platform or its implication for a firm (Figure 1). “They are constructs that allow us to make sense of how different levels of social media functionality can be configured” (Kietzmann et al., 2011, p. 243).
Kietzmann et al. (2011) projected the honeycomb on LinkedIn, Foursquare, YouTube and Facebook, and determined the relevance of each construct per platform. They presented this with honeycomb figures in which the most important constructs have a dark grey color, the medium important have a mid grey color, and the least important are colored white. This framework has not been applied to Instagram in previous literature. To conceptualize the differences in functionalities between Facebook and Instagram, we have projected the honeycomb framework on Instagram (Figure 2). This conceptualization has been reviewed and agreed upon by two objective co-readers with professional knowledge about social media.
Facebook is a social networking website that was launched in 2004 by Mark Zuckerberg. In terms of the honeycomb of social media, Facebook’s focus lies on relationships. Identity takes an important role as users can create a profile with personal information such as their name, profile picture, profession and preferences. Amongst numerous other possible actions, they can post status updates, digital photos and videos, and links that are displayed in a timeline. Users can ‘check in’ at physical locations and share their availability, which makes presence another important facet for Facebook. Conversation and reputation are also relevant functions. Users can add other users as ‘friends’ with whom they can interact through ‘liking’ or ‘commenting’ on other users’ posts, or through the private messenger function. The number of likes and comments is considered as an important indicator for the user’s online reputation. Users can also like business pages of brands, events, and celebrities. These pages provide information, posts and the possibility for users to posts messages in the “posts from visitors” section. Over the past years, Facebook has implemented several updates to its business pages that enable brands to better interact with their customers, intended to transform business pages from promotion tools into full customer service platforms. For example, a ‘contact us’ button was implemented, businesses can set their own average response rate, they can add annotations to private messages that include business information such as order numbers, and they can track a customer’s page interaction history (Templeman, 2015).
Instagram is a photo-sharing website that was founded by Kevin Systrom and Mike Krieger in 2010. For Instagram, the most important facets of the honeycomb are sharing and identity. Users can create either a public or a private profile with limited personal information including their name, a profile photo, a biography with a maximum of 150 characters, and a link to their website. They can share photos that can be edited within the application and add a short description, a so called ‘caption’. Users can tag a location in their posts, which adds to the presence construct. These posts appear in a timeline that give other users an overview of the user’s identity. Users can follow other users and ‘like’ or ‘comment’ on their posts. Just like on Facebook, the number of likes and comments is perceived as
an important indicator of the user’s reputation. Users can also use the ‘direct messaging’ function, which is intended to privately share existing posts with other users.
Promotional Self-Presentation
Based on the two different platform descriptions, a couple of assumptions can be made that are tested in this study. Sharing and identity are Instagram’s two key constructs, and they are significantly stronger represented than on Facebook. Facebook users can mention brands in their preferences and list of followed pages, and can mention brands in their posts. However, the self-promotion culture by means of brand tagging appears to be more developed on Instagram. Here, users often tag the brand in the image or caption, which is intended to illustrate the user’s identity. Therefore, brand-related UGC on Instagram is expectedly more likely to feature promotional self-presentation than on Facebook.
H1A. Brand-related UGC on Instagram is more likely to feature promotional self-presentation
Brand Centrality
In line with Facebook’s stronger focus on relationships and conversation, it offers sophisticated business functions to its business pages, making the platform an important customer service tool (Templeman, 2015). Customers make serious use of this opportunity, and post messages on the page in which they aim to interact with the brand. These posts mostly appear to focus on the brand. A post can also appear in the ‘posts from visitors’ section when a brand is tagged in a post on a user’s own profile. In this case, users sometimes tag multiple brands. On Instagram, users mostly post images to a general audience or to their friends. Brand-related posts are not aggregated on one page like on Facebook, which is why it is less likely that posts are focused on one brand.
H1B. Brands are more likely to be central in related UGC on Facebook than in
brand-related UGC on Instagram.
Marketer-Directed Communication
In line with the previous argumentation, one could hypothesize that UGC on Facebook is more likely to be marketer-directed. The relationship and conversation constructs are more emphasized on Facebook than on Instagram. The functionalities of Facebook’s business pages empower brands to build relationships with their customers, and allow customers to directly communicate with marketers.
H1C. Brand-related UGC on Facebook is more likely to be marketer-directed than brand-related
Factually Informative about the Brand
Whereas customers used to email of call a brand for questions and remarks, the number of people using social media for customer service has increased over the years (Templeman, 2015). Facebook brand pages have functions that facilitate the interaction between the brand and the customers. It is a convenient channel to quickly get responses to informative questions and remarks. Instagram has less developed possibilities for functional interaction with brands. Whenever factual information is requested or presented on Instagram, the response rate of the brand is low, which provides little incentive for users to use the channel for factual communication.
H1D. Brand-related UGC on Facebook is more likely to be factually informative about the
brand than brand-related UGC on Instagram.
Brand Sentiment
Sharing and identity are Instagram’s two pillars. Users carefully choose what content they share on the platform, because it influences their desired online identity. This content appears to more often have a positive sentiment than negative. Facebook is used to express one’s identity as well, but it also focuses strongly on conversation and relationships. Interaction with brands on Facebook does not necessarily need to be positive, as its purpose is less often to build a desirable online reputation for the user, but rather to have informative communication with the brand or with other users.
2.3 Brand Category and Brand-Related UGC
As explained by Batra and Ahtola (1991), consumer attitudes are bidimensional, because consumers show consumption behavior for two reasons: “(1) consummatory affective (hedonic) gratification (from sensory attributes), and (2) instrumental, utilitarian reasons concerned with “expectations of consequences” (of a means-end variety, from functional and nonsensory attributes)” (p. 159). In other words, while hedonic attitudes are driven by a desire for emotional and affective outcomes, utilitarian attitudes are driven by a desire for task- and utility-oriented outcomes (Eggert & Uluga, 2002). We can categorize brands based on consumer attitudes towards them. This study analyzes brand-related UGC of five hedonic and five utilitarian brands. Hedonic and utilitarian consumer attitudes towards a brand or product need not to be mutually exclusive; some are approached for hedonic as well as utilitarian reasons. For example, whitening toothpaste makes the teeth whiter (hedonic motivation), but also prevents cavities (utilitarian motivation). Perceptions of how hedonic or utilitarian a brand is may also vary per individual. Keeping these variances in mind, the brands in this study have been selected with the intention to have minimal disagreement on the brand character.
As these two brand categories have different consumer attitudes towards them, one could say that consumers communicate differently about these categories on social media platforms, leading to variances in UGC. Therefore, this study hypothesizes that brand category, in terms of hedonic and utilitarian, has a moderating effect on the relationship between social media platform and brand-related UGC. Literature has shown that different attitudes are linked to different behavior on social media (Lin & Rauschnabel, 2016). Users with utilitarian motivations tend to process and produce content in a more functionally oriented way. For example, they are likely search and ask for information about brands and products, look for advice from friends and other consumers, and share their own experiences. On the other hand, users with hedonic attitudes take on a more experiential approach when they are active on social media. These users are more likely to look for enjoyment, relaxation,
escape and other affective outcomes next to the utilizing the content. Content that appeals to their hedonic motivations can have an enhancing effect on the emotional responses towards the product or brand (Lin & Rauschnabel, 2016).
Promotional Self-Presentation
When promoting the self, users try to express a desirable identity. Users can mention or tag brands to support this process. What kind of brands are referred to has impact on the way the user is perceived by others. While being involved with utilitarian brands might not lead to strong emotional responses amongst other users, being involved with hedonic brands can cause others to link the user to affective outcomes such as pleasure, luxury and entertainment. As the mentioning of hedonic brands is more likely to be beneficial for a desirable identity, UGC of hedonic brands is expected to feature more promotional self-presentation than UGC of utilitarian brands.
H2A. Brand-related UGC of hedonic brands is more likely to feature promotional
self-presentation than brand-related UGC of utilitarian brands.
Brand Centrality
Users can have different reasons to focus their content on one brand. For utilitarian brands, this can be a result of their functional use of social media. Focusing their posts on one brand can increase
brand centrality, this study hypothesizes that utilitarian and hedonic brands are equally likely be central in UGC.
H2B. Brands are equally likely to be central in brand-related UGC of utilitarian and hedonic
brands.
Marketer-Directed Communication
As the number of consumers approaching social media for their customer service issues has increased over the past few years, it is logical that marketer-directed interaction has become an important part of UGC. With utilitarian brands serving functional outcomes, their reputations strongly depend on the functionality and user-friendliness of their products. It is therefore important for these brands to continuously be involved with their customers to handle customer service issues. As users with utilitarian motivations approach social media with a focus on functional outcomes, it can be hypothesized that UGC about utilitarian brands is more often marketer-directed than UGC about hedonic brands.
H2C. Brand-related UGC of utilitarian brands is more likely to be marketer-directed than
Factually Informative about the Brand
In line with the argumentation of the last hypothesis, this hypothesis is also based on literature showing that users with utilitarian motivations approach social media in a more functionally oriented way (Lin & Rauschnabel, 2016). Presenting or requesting factual information is an example of how consumers can use social media to reach specific functional outcomes, such as making a well-considered product choice or sharing a complaint. Content related to hedonic brands is expected to feature more expressions of emotion rather than factual information. Therefore, one could hypothesize that UGC of utilitarian brands is more likely to request or present factual information than UGC of hedonic brands.
H2D. Brand-related UGC of utilitarian brands is more likely to be factually informative about
the brand than brand-related UGC of hedonic brands.
Brand Sentiment
The sentiment of UGC is determined by the user’s emotions towards the brand. For utilitarian brands, users tend to take on more functional approaches when creating content. This is expected to lead to higher levels of neutral sentiment. For hedonic brands, UGC is more likely to be an expression of affective and emotional gratification, which is expected to be paired with more positive sentiment.
In the figure below the conceptual model of this study is illustrated. The arrow from ‘social media platform’ to ‘brand-related user-generated content’ stands for the hypotheses that directly help answer RQ1. The dotted arrow from ‘brand category’ to ‘brand-related user-generated content’ is briefly discussed in the results section, but does not directly answer RQ2. The arrow from ‘brand category’ to the other arrow stands for RQ2 (the moderation effect), and can be answered by looking at the interaction of the two independent variables. In following section the methods that are used to test these hypothesized relationships are explained.
Figure 3: Effect of social media platform on brand-related user-generated
content, moderated by brand category
3. Methods
The overall design of this research was an archival research design. A cross-channel content analysis was conducted to systematically compare data from five hedonic and five utilitarian brands, on two social media platforms. Content analysis is an established research method that is suitable for systematical comparison of content (Kolbe & Burnett, 1991). This type of analysis was suitable for this research setting because it offers objectivity and structure to the process of addressing a large sample of data.
3.1 Sampling
The unit of analysis was a set of individual brand-related UGC posts on Facebook and on Instagram. A post included text (the ‘caption’ for Instagram) and/or an image. Facebook posts were manually selected from the brand pages, under the section ‘posts from visitors’. Instagram posts were also manually selected, by using the brand hashtags (Table 1 and 2). These two data streams yielded lists of posts from all over the world, in chronological order. The posts in the sample were checked if they were produced by consumers, if they were not produced with commercial interests, and if the brand hashtag referred to the brand in question and not to another concept. To ensure the analyst’s accurate understanding of the text in the post, only Dutch and English posts were selected. To keep the sample representative yet manageable, the 25 latest posts per brand were selected, resulting in a
A total of ten brands was included in the cross-channel analysis, of which five were utilitarian, and five hedonic. The first criterion for a brand to be selected for the sample was that it had to have a clear utilitarian or hedonic character. Brands of which it was likely that people could disagree on the nature of the brand were avoided. The second criterion was that the brand had to have a Facebook page with a ‘posts from visitors’ section with enough posts, as well as usable Instagram hashtag that led to brand-related content. Lastly, the stream of brand-related content had to include enough English or Dutch written posts. As a control procedure, the compiled list of brands was reviewed by two independent people who had no content knowledge this study, but had been explained the previously described definitions of utilitarian and hedonic brand characters. The ten brands that were eventually used in this study are presented in Table 1 and 2.
Table 1: Social media information utilitarian brands
Utilitarian brands Facebook page Instagram hashtag
Duracell @Duracelluk #duracell
Post-it @postitbenelux #postit
Rubbermaid @RUBBERMAID #rubbermaid
Tefal @TefalNederland #tefal
Swiffer @swiffer #swiffer
Table 2: Social media information hedonic brands
Hedonic brands Facebook page Instagram hashtag
Milka @Milka.NL #milka
Starbucks @starbucks #starbucks
Ben & Jerry’s @benenjerrys #benandjerrys
Nespresso @nespresso.nederland #nespresso
3.2 Coding
After the sample was selected, the posts were manually coded for each dimension by the author in SPSS version 22. When doing so, the entire post was considered (text and/or image).
Promotional Self-Presentation
The post was coded as ‘yes’ when the poster was explicitly mentioned, referenced or featured in the text and/or the image with a self-promotional purpose. If this was not the case, it was coded as ‘no’. For example, if a post included a smiling selfie with a Starbucks cup, it was coded as ‘yes’. If a post included an image of a Milka chocolate bar and a cup of tea, it was coded ‘no’.
Brand Centrality
Content was coded as ‘yes’ when the main focus was on the brand. If the focus was on multiple brands, or not on the brand at all, it was coded ‘no’. For example, if a post read ‘Is there a loyalty program at Starbucks?’, it was coded ‘yes’. If a post included an image of multiple chocolate bars from different brands including Milka, it was coded as ‘no’.
Marketer-Directed Communication
Posts were coded ‘yes’ when they were clearly directed towards the brand. If not, it was coded as ‘no’. For example, a post that read ‘Have you ever considered making a lactose-free chocolate bar?’,
Amsterdam central station open on Sundays?’ was coded as ‘yes’. ‘I’ll never buy Milka again’ was coded as ‘no’ because it is an opinion.
Brand Sentiment
Content was coded ‘negative’, ‘neutral’, or ‘positive’ based on the overruling sentiment of the post towards the brand. For example, a post that read ‘The caramel layer is my chocolate bar is missing! What a disappointment’ was coded ‘negative’. ‘Where can I buy the Oreo Milka bars in The Netherlands?’ was coded ‘neutral’. ‘My dog loves a good Puppuccino!’ was coded ‘positive’. In case the sentiment was ambiguous or undefinable, it was coded as ‘unclear’. ‘Just wanted to let you know that the S’mores Frappuccino is my new favorite! Too bad that the barista was rude to me though’ would be coded as ‘unclear’, because it features positive as well as negative brand sentiment.
3.3 Inter-Coder Reliability
Before analyzing the data, a subset of 50 posts was coded by an independent coder and tested for inter-coder reliability. The Krippendorff’s alpha is considered an accurate measure for inter-coder reliability, and is therefore chosen in this case (Hayes & Krippendorff, 2007). The KALPHA macro for SPSS by Hayes was used to calculate the alphas. This procedure was executed for each of the dimensions separately. An alpha of 0.8 is often seen as the norm for a good reliability, with a minimum of 0.6 or 0.67. When a variable is extremely easy to code, such as ‘presence of visual content’, one should raise the standard (De Swert, 2012). As the dimensions in this study might leave some room for interpretation, an alpha of 0.8 was considered acceptable. For all dimensions the alpha fell above this norm (promotional self-presentation α = 0.8355, brand centrality α = 0.8479, marketer-directed communication α = 0.8393, factually informative about the brand α = 0.8416, and brand sentiment α = 0.8812). Following coding of the data, frequency tables were generated to get an overview of the data scores. The hypotheses were then tested by performing regression analyses.
4. Results
The analysis of the coded data has led to interesting results. Table 3 and 4 illustrate the score frequencies on each of the dimensions, for the independent variable social media platform and for the moderator brand category. An initial analysis of the data suggests that Facebook scores higher on brand centrality, marketer-directed communication and factually informative about the brand. Additionally, it suggests that Facebook scores higher on negative and neutral sentiment than Instagram. Regarding brand category, the results suggest that hedonic brands score higher on promotional-self presentation, brand centrality, and factually informative about the brand. The data also suggests that posts about hedonic brands are much more likely to feature positive brand sentiment.
Table 3: Coding frequencies for Facebook and Instagram
Code Social Media Platform
Facebook (N=250) Instagram (N=250) Total Across Sites
Promotional Self-Presentation Yes No 22 228 61 189 16.6% 83.4% Brand Centrality Yes No 245 5 184 66 85.5% 14.2% Marketer-Directed Communication Yes No 166 84 1 249 33.4% 66.6%
Table 4: Coding frequencies for utilitarian and hedonic brands
Code Brand Category
Utilitarian (N=250) Hedonic (N=250) Total Across Brands
Promotional Self-Presentation Yes No 24 226 59 191 16.6% 83.4% Brand Centrality Yes No 210 40 219 31 85.8% 14.2% Marketer-Directed Communication Yes No 84 166 83 167 33.4% 66.6%
Factually Informative about the Brand
Yes No 75 175 88 162 32.6% 67.4% Brand Sentiment Negative Neutral Positive Unclear 51 73 75 51 34 71 137 8 17.0% 28.8% 42.4% 11.8%
To analyze the relationships as depicted in the conceptual model (Figure 3), four binary regression analyses were performed for the binary dimensions (Table 5), and one multinomial regression analysis for brand sentiment (Table 6). Each of the analyses is discussed in the following sections.
4.1 Promotional Self-Presentation
The first binary regression analysis assessed the effects of the platform, the type of brand, and their interaction term ‘social media platform * brand category’ on the likelihood that posts featured promotional self-presentation. The model is significant as a whole and has a Nagelkerke R2 of 0.153.
No significant effect of social media platform on promotional self-presentation was found (B = 0.372, p = 0.393). No significant effect of brand category on the dependent variable was found either (B = 0.200, p = 0.656). However, a significant moderation effect was found between the interaction term
and the dependent variable (B = 1.364, p = 0.015). Posts about hedonic brands were 3.912 times more likely to feature promotional self-presentation on Instagram than on Facebook.
4.2 Brand Centrality
A second binary regression analysis was conducted to calculate the effects of the platform, the type of brand, and their interaction term on the likelihood that posts scored on brand centrality. The model is significant with a Nagelkerke R2 of 0.252. A significant relationship was found between
social media platform and brand centrality (B = -4.029, p = 0.018). Posts on Instagram were far less likely (55 times) to score on brand centrality. No significant effects were found for brand category (B = -1.411, p = 0.210) and for the interaction term (B = 1.909, p = 0.100).
4.3 Marketer-Directed Communication
For the dimension marketer-directed communication the same test was conducted. The model was significant, with a Nagelkerke R2 of 0.636. A significant relationship was found between social
media platform and this dimension (B = -5.501, p = 0.000). Posts on Instagram were far less likely (250 times) to be directed towards a brand marketer. The effect of brand category on this dimension was insignificant (B = 0.000, p = 1.000), just like the interaction term (B = -16.383, p = 0.996).
4.4 Factually Informative about the Brand
significant effect on this dimension (B = -1.740, p = 0.008). Posts related to hedonic brands were 5.714 times less likely to present or request factual information when they were on Instagram than on Facebook.
Table 5: Results of the binary regression analyses per dimension
Variable B Sig. Exp(B)
Promotional Self-Presentation
Constant
Social Media Platform Brand Category
Social Media Platform * Brand Category
-2.442** 0.372 0.200 1.364* 0.000 0.393 0.656 0.015 0.087** 1.450 1.221 3.912* Brand Centrality Constant
Social Media Platform Brand Category
Social Media Platform * Brand Category
4.820** -4.029** -1.411 1.909 0.000 0.000 0.210 0.100 124.000** 0.018** 0.244 6.747 Marketer-Directed Communication Constant
Social Media Platform Brand Category
Social Media Platform * Brand Category
0.681** -5.501** 0.000 -16.383 0.000 0.000 1.000 0.996 1.976** 0.004** 1.000 0.000
Factually Informative about the Brand
Constant
Social Media Platform Brand Category
Social Media Platform * Brand Category
0.048 -2.386** 0.669* -1.740* 0.788 0.000 0.010 0.008 1.049 0.092** 1.953* 0.175* Note. *p<.05 **p<.01
4.5 Brand Sentiment
As the dimension brand sentiment has multiple categories, a multinomial regression analysis was performed to examine the effects of social media platform, brand category, and their interaction term on brand sentiment (Table 6). The category ‘unclear’ was excluded from the analysis, because it was meant to identify posts with ambiguous or undefinable brand sentiments, and does not provide useful insights.
A significant effect of social media platform on negative brand sentiment was found (B = -2.855, p = 0.000). Posts on Instagram were 17 times less likely to feature negative sentiment. The effect of brand category on negative brand sentiment is significant (B = -0.815, p = 0.009). UGC related to hedonic brands was 2.3 times less likely to feature negative brand sentiment. Lastly, there is a significant effect between the interaction term and positive brand sentiment (B = 2.306, p = 0.000). Posts regarding hedonic brands were 10 times more likely to feature positive brand sentiment on Instagram than on Facebook.
Table 6: Results of the multinomial regression analysis for brand sentiment
Sentiment B Sig. Exp(B)
Positive
Constant
Social Media Platform Brand Category -0.059 0.159 -0.470 0.808 0.630 0.151 . 1.172 0.625
5. Discussion
5.1 Validity
This study aimed to gain a deeper understanding of how UGC is affected by the type of social media platform, and whether this effect is moderated by brand category. The research was set up as a cross-channel analysis of 500 posts. After coding all posts on the five dimensions and analyzing the data, it was possible to determine which hypotheses could be accepted and which had to be rejected. This research method successfully led to answers to the research questions. We can therefore conclude that this study has satisfactory internal validity. For the sample, posts of five utilitarian and five hedonic brands were included. Even though these the brands were selected on their evident brand categories, there is variance in UGC between the brands that cannot be fully explained by the brand category or type social media platform. The findings of this study provide valuable preliminary insights. However, a higher level of external validity could be achieved in the future by researching a larger sample. In the following two sections, each of the two research questions and their hypotheses are discussed.
5.2 Effect of Social Media Platform on UGC
With the findings of this study we are able to answer the first research question “How does the type of social media platform affect brand-related user-generated content?”. The results show that UGC differs significantly on four out of the five dimensions; brand centrality, marketer-directed communication, factually informative about the brand, and brand sentiment.
H1A states that UGC on Instagram is more likely to feature promotional self-presentation than
UGC on Facebook. No significant effect was found, which means that this hypothesis is rejected. However, when we look at the score frequencies, 61 post on Instagram featured promotional
self-presentation, while only 22 did on Facebook (Table 3). This statistically insignificant difference in scores can be explained with Instagram’s stronger focus on identity and sharing. Evidently, users on both platforms show promotional self-presentation to some extent. The honeycomb model illustrates that Instagram as well as Facebook have functionalities that support the identity construct (Kietzmann et al., 2011).
H1B states that brands are more likely to be central in UGC on Facebook than on Instagram. A
significant relationship was found that supports this hypothesis. Instagram posts are 55 times less likely to feature brand centrality. This is in line with the honeycomb model of Kietzmann et al. (2011), in which relationships and conversation are more relevant constructs for Facebook than for Instagram. Also, Facebook offers several functionalities that enhance it as a customer service platform. Users who want to share their experiences or ask for information can effectively do so by posting content with a focus on the brand in question.
H1C states that UGC on Facebook is more likely to be marketer-directed than UGC on
Instagram. This hypothesis is supported with a strong significant effect. Instagram posts were 250 times less likely to be directed towards marketers. Just like for brand centrality, this finding can be explained by the relationships and conversation constructs from the honeycomb model. Users respond to the behavior of brands on Facebook. When brands are active and responsive on the channel, users
H1D states that UGC on Facebook is more likely to present or request factual information than
UGC on Instagram. As the results show that Instagram posts are almost 11 times less likely to be factually informative, this hypothesis is also supported. Again, this result can be explained with Facebook’s business pages that function as customer service platforms. Facebook has become a convenient online channel for functional communication that was previously done over the phone or by email. As Instagram focuses on different constructs of the honeycomb (Kietzmann et al., 2011), it is less suitable for factual content with functional value, and more suitable for expressing the self by means of content with emotional value.
Lastly, H1E states that UGC on Instagram is more likely to feature positive brand sentiment
than UGC on Facebook. No significant effect on positive sentiment was found. However, a significant relationship between social media platform and negative sentiment was found. Instagram posts are 17 times less likely to feature negative brand sentiment. When looking at the score frequencies, 146 Instagram posts featured positive sentiment, while only 66 did on Facebook. This finding is supported by the honeycomb of social media (Kietzmann et al., 2011), as well the fact that Facebook functions as a full-service customer hub (Templeman, 2015). Instagram is a platform where users express themselves by sharing visual content that has a positive effect on their perceived identity. Logically, this is more effective when the content has a positive sentiment. Facebook on the other hand, is used as the online reporting room for all kinds of brand-related questions, stories and remarks. The data revealed that the level of complaints is much higher on Facebook than on Instagram. Instagram illustrates a carefully created reflection of someone’s life, while a business page on Facebook is more of a functional two-way communication channel where users can share anything, from negative to positive.
5.3 Moderating Effect of Brand Category
The data of this study allows us to answer the second research question “How is the effect of the type of social media platform on brand-related user-generated content moderated by brand category?”. We can conclude from the findings that brand category indeed moderates the effect on promotional self-presentation, factually informative about the brand, and brand sentiment.
H2A states that UGC of hedonic brands is more likely to feature promotional self-presentation
than UGC of utilitarian brands. No significant direct effect of brand category on UGC was found, but the results did show a significant moderation effect. UGC about hedonic brands is 3.912 times more likely to feature promotional self-presentation on Instagram than on Facebook. This is in line with Lin and Rauschnabel’s (2016) finding that content that appeals to a user’s hedonic motivations can increase emotional responses. Getting emotional response is more important to Instagram users than Facebook users, because Instagram focuses on the sharing and identity constructs (Kietzmann et al., 2011). Users who try to express a desirable identity on Instagram have more chance of succeeding when their followers see content about hedonic brands, rather than utilitarian brands, that appeal to them.
H2B states that utilitarian and hedonic brands are equally likely to be central in UGC. As no
significant relationship between brand category and brand centrality, nor a significant moderation effect of brand centrality was found, the hypothesis is supported. As argued in section 2.3, there can
H2C states that UGC of utilitarian brands is more likely to be marketer-directed than UGC of
hedonic brands. This hypothesis is rejected, as no significant direct effect nor a significant moderation effect was found. Just like for brand centrality, the score frequencies of marketer-directed communication are very similar for utilitarian and hedonic brands. This is conflicting with the expectation that users of utilitarian brands have a more functional approach to social media, and are therefore more likely to direct their content to marketers. Evidently, users of hedonic brands are equally motivated to do so. Too little academic theory exists to explain what motivates consumers to direct their content towards marketers, and whether these motivations vary per brand category. Based on logical reasoning one could argue that marketer-directed UGC increases when the brand is active on social media and has a high response rate. Consumers who want to pose questions or complaints about the brand then get the impression that the platform is suitable for direct communication with the brand, where they would have used the customer service line or email in the past.
H2D states that UGC of utilitarian brands is more likely to present or request factual information
than UGC of hedonic brands. Interestingly, significant effects were found in the opposite direction. UGC of hedonic brands is almost twice as likely to be factually informative. Additionally, brand category moderates this effect. UGC relating to hedonic brands has a 5.7 times smaller chance to be factually informative when it is on Instagram than when it is on Facebook. This means that the combination of hedonic brands and Facebook is most likely to result in factually informative UGC. Perhaps this is related to higher levels of overall brand involvement towards hedonic brands. As hedonic brands speak more to the emotions of people than utilitarian brands do, one could argue that customers of hedonic brands feel more need to communicate with or about the brand in general. It could also have to do with the fact that hedonic brands are usually more active on social media. As discussed earlier, the communication between a brand and its customers is an interplay. When a brand
communicates more through social media, one can expect a higher level of communication from the customers’ side. Especially on Facebook, which functions as a customer service platform, this can result in more factually informative posts of users. Much research has been conducted on consumer motivations for utilitarian and hedonic products, but little is known about how this affects their online behavior in creating content. It would be interesting to continue this line of research to be able to explain the finding that UGC of hedonic brands is more likely to be factually informative.
Lastly, H2E states that posts about hedonic brands are more likely to feature positive brand
sentiment than posts about utilitarian brands. Even though no significant direct effect on positive sentiment was found, it was found for negative sentiment. UGC about hedonic brands is 2.3 times less likely to feature negative brand sentiment. Additionally, a significant moderation effect was found on positive sentiment. UGC about hedonic brands is 10 times more likely to have a positive brand sentiment on Instagram than on Facebook, which means that the combination of hedonic brands and Instagram leads to the most positive UGC. This is in line with the idea that utilitarian brands are approached in more functional ways, resulting in more neutral sentiment, while hedonic brands are approached in more emotional ways, resulting in expressions of gratification towards the brand. It appears logical that this happens more on Instagram than on Facebook, as Instagram is about sharing and identity (Kietzmann et al., 2011), which makes it a place for users to express themselves and the things they like.
5.4 Limitations
Like any research, there are several limitations to this study. The first relates to the coding of the data. Due to limited time and financial resources, the data was coded by the author. Even though this was done with utmost care and as much objectivity as possible, having multiple independent people code the data could have reduced eventual bias and errors. This method is therefore suggested for future cross-channel content research. Second, the sample consists of the most recent 25 posts of each brand. These posts are already aggregated by the platform. Brands with content that had striking patterns, such as consecutive content about a brand event or contest, were not selected for the sample. However, there is a possibility that time-related events structurally influenced the content. For future research a larger number of posts spread over a wider time span is recommended to reduce the chance of any time-related influences. Third, it is very hard to eliminate influences other than the type of social media platform and brand category, because the setting of this study is not controlled. As the sample draws from two sets of five brands, variance in UGC can be caused by other factors such as the quality of the brand’s products or customer service. A suggestion to tackle this issue in future research is to select a larger number of brands, so that major deviations are balanced out by the rest of the sample.
5.5 Future Research
As this study is the first to consider the effect of utilitarian and hedonic brands on UGC, more research on this relationship is needed to confirm and extend the preliminary findings. In addition to the previously described suggestions for further research on the topic of this study, several suggestions are made for other interesting research directions. To start, it would be valuable to compare brands from multiple industries (e.g. the fashion industry, food industry and car industry) within one cross-channel content analysis to find out how their UGC differs. This can generate more industry specific implications for marketing managers. Second, future research could consider user-generated content created with commercial interests. Instagram is known as a platform where bloggers and online influencers post sponsored content to promote brands. How do consumers respond to this content? How does it change their brand perception? And how does is vary across different product categories or brands? These could be relevant questions for future research. The results could provide valuable insights for marketing managers who want to collaborate with influencers and look for effective strategies. Third, future research could consider new social media platforms. Platforms come and go in a rapid pace. Instagram has become one of the most used platforms over the past years, which is why it was included in this study. A platform that is currently on the rise is Snapchat. Even though analyzing user-generated content on this channel is difficult due to privacy issues and the way the app works, future research could analyze brand-generated content of multiple brands. This would be the visual content that is published and available for 24 hours in the brand’s so called ‘Snapchat story’.
6. Conclusion
This study is the first to conduct an online cross-channel content analysis that examines UGC on Facebook as well as Instagram, and the first to consider the influence of brand category (in terms of utilitarian and hedonic brands). The aim of this study was two answer the two research questions “How does the type of social media platform affect brand-related user-generated content?” and “How is the effect of the type of social media platform on brand-related user-generated content moderated by brand category?”. The findings indicate that when comparing the two platforms, UGC on Facebook is more likely to feature brand centrality, marketer-directed communication, factual information, and negative brand sentiment. On the other hand, UGC on Instagram is more likely to feature promotional self-presentation and positive brand sentiment. With regards to the second research question, the results show that brand category indeed moderates the effect on three of the five UGC dimensions. UGC relating to hedonic brands is more likely to feature promotional self-presentation and positive brand sentiment on Instagram than on Facebook, and is less likely to be factually informative on Instagram than on Facebook.
In a time where consumers have more power than ever, it is highly relevant for managers to understand how they engage in different types of social media. Social media is much more than a one-way advertising tool. By getting actively involved it can be leveraged to express the brand identity, take care of customer service issues, build personal relationships with customers, and even realize a co-creation environment in which customers share their ideas (Smith et al., 2012). The findings of this study can help managers decide which platforms have most potential for the brand, and how they should allocate resources accordingly. Depending on what the desired outcome is, an effective online presence strategy can be developed.
In conclusion, Facebook is a highly valuable platform for both utilitarian and hedonic brands that aim to realize a strong and functional customer service hub. Marketers can bring their brands alive by posting relevant content and providing quick and qualitative responses to UGC. Additionally, creating a space where customers can have conversation with the brand as well as with each other offers opportunities for marketers to gather consumer opinions and ideas (Muñiz & Schau, 2007).
Instagram is more suitable for brands that want to express their identity and create more emotional brand commitment amongst consumers. Especially for hedonic brand, this platform can be leveraged to spread positive brand sentiment and to encourage users to promote the brand. Creative initiatives can be set up to realize this, such as a contest in which consumers share their own brand-related images, or share their ideas for new products and services. This strategy can be valuable when a brand is going through an identity transformation, or when a brand is trying to generate more positive brand visibility online.
7. References
Batra, R., & Ahtola, O. T. (1991). Measuring the hedonic and utilitarian sources of consumer attitudes. Marketing letters, 2(2), 159-170.
Burmann, C. & Arnhold, U. (2009). User generated branding: State of the art of research (Vol. 8). LIT Verlag Münster.
De Swert, K. (2012). Calculating inter-coder reliability in media content analysis using Krippendorff’s Alpha. Center for Politics and Communication, 1-15.
Eggert, A., & Ulaga, W. (2002). Customer perceived value: a substitute for satisfaction in business markets?. Journal of Business & industrial marketing, 17(2/3), 107-118.
Hayes, A. F., & Krippendorff, K. (2007). Answering the call for a standard reliability measure for coding data. Communication methods and measures, 1(1), 77-89.
Kaplan, A. M., & Haenlein, M. (2010). Users of the world, unite! The challenges and opportunities of Social Media. Business Horizons, 53(1), 59-68.
Kietzmann, J. H., Hermkens, K., McCarthy, I. P., & Silvestre, B. S. (2011). Social media? Get serious! Understanding the functional building blocks of social media. Business
Horizons, 54(3), 241-251.
Kolbe, R. H., & Burnett, M. S. (1991). Content-analysis research: An examination of applications with directives for improving research reliability and objectivity. Journal of consumer research, 18(2), 243-250.
Lin, C. A., & Rauschnabel, P. A. (2016, March). Social media platforms as marketing channels. Retrieved May 18, 2017, from
https://www.researchgate.net/publication/301325194_Social_media_platforms_as_marketing _channels
Muñiz, Jr, A. M., & Schau, H. J. (2007). Vigilante marketing and consumer-created communications. Journal of Advertising, 36(3), 35-50.
OECD. (2007). Participative web and user-created content: Web 2.0, wikis, and social networking. Paris: Organisation for Economic Co-operation and Development.
Pan, Y., & Zhang, J. Q. (2011). Born unequal: a study of the helpfulness of user-generated product reviews. Journal of Retailing, 87(4), 598-612.
Smith, A. N., Fischer, E., & Yongjian, C. (2012). How does brand-related user-generated content differ across YouTube, Facebook, and Twitter?. Journal of Interactive Marketing, 26(2), 102-113.
Statista. (2016). Facebook users worldwide 2016. Retrieved March 01, 2017, from
https://www.statista.com/statistics/264810/number-of-monthly-active-facebook-users-worldwide/
Statista. (2017). Instagram: active users 2016. Retrieved April 15, 2017, from
https://www.statista.com/statistics/253577/number-of-monthly-active-instagram-users/ Templeman, M. (2015, December 22). How Facebook Is Becoming A Customer Service Hub.
Retrieved June 18, 2017, from
https://www.forbes.com/sites/miketempleman/2015/12/22/how-facebook-is-becoming-a-customer-service-hub/#6ddf8a143c2b
