Social Media in a Controversial Industry: How Different Facebook
Strategies Affect Consumer Engagement
Roland van Lienden
10703055
Master Thesis
Corporate Communication
Supervisor: Lise van Oortmerssen
This research investigates the relation between aspects of social media strategy: a post’s purpose, its vividness, and the ratio of directed communication from the company, and the consumer engagement it generates. As there is no general consensus on what the right approach of a company towards social media is, this research hoped to help clear up these effects. Through a content analysis of four fast food companies’ Facebook pages, it was investigated whether the controversy surrounding this industry grants different results than previous research on this topic. Most of the results showed only very weak relations between the aspects of social media strategy and the consumer engagement. Moreover, directed
communication and post purpose showed no significant relation for most of the measurements of consumer engagement. However, the results were affected by the presence of unrelated negative messages from consumers. Instead, these comments were related to the company’s actions outside of Facebook. So, the main conclusion was that the controversy surrounding the fast food industry does indeed have an impact on the relation between social media strategy and consumer engagement. However, further research will be required to establish how these effects work and what can be done to limit these effects.
Introduction
With the popularisation of social media, new opportunities for communicating with
consumers are created for companies. Although researchers agree that this new type of media gives better opportunities to communicate directly with consumers and establish a dialogue which can benefit both the company and its clients , this potential has not been fully
recognised in a substantial part of the business world (Van Belleghem, 2013; Rybalko & Seltzer, 2010; Goh et al., 2013).
Although research has been done to analyse how companies should use social media (i.e. Goh et al., 2013; Dekay, 2012; Cvijikj & Michahelles 2013) , it is still unclear what factors influence the online attention a company receives. To investigate, the fast food industry is an interesting subject primarily due to the controversy often surrounding it. There have been multiple documentaries concerning the potential link between fast food and obesity among children as well as adults. Perhaps most notably the documentary ‘Super-size me’ by Morgan Spurlock, in which he experiments by solely eating at McDonalds for a 30 day period, showing the health risks involved in fast food (Scott, 2003).
As a result of this controversy, it can be expected that consumers respond more on the social media pages of companies in this particular industry than they do on average. To elaborate, there are several theories concerning the impact of negative messages in proportion to positive messages, most prevalently, the attribution theory. This theory states that
consumers only absorb the key points of the information they are offered and attribute this to the company or product it is related to (Weinberger et al., 1981). The relevance to negative messaging of this theory lies in the fact that negativity stands out more than positivity,
because it is less prevalent in the social environment. As a result, negative information about a company can disproportionally affect a consumer’s opinion regardless of previously positive connotations (Weinberger et al., 1981). Moreover, when this has happened and a majority of
the messaging on the Facebook page of a company is negative in nature, this could have detrimental effects on the reputation of that company (Dekay, 2012). Subsequently, the consumers might become less committed to the company or brand. This commitment is often described as consumer engagement, which will also be the central part of this research. The importance of consumer engagement is not limited to just attention, it also influences the purchase behaviour of consumers, meaning it is valuable to any company with something to sell (Anderson et al., 2014; Jahn & Kunz, 2012).
As social media are still not always being used to its full potential (Van Belleghem, 2013), this research will attempt to establish the links between different Facebook strategies and the consumer engagement towards a company. The choice for Facebook was made as it is still the most used social media, with 71% of all internet users having a Facebook account (Duggan et al., 2015). By combining variables from previous research (i.e. Cvijikj & Michaehelles, 2013; Dekay, 2012) and selecting an industry with substantial reputation challenges due to controversy, this research will attempt to discover what aspects of a Facebook strategy benefit a company the most. In other words:
How do different Facebook strategies influence consumer’s engagement in the fast food industry?
To investigate the link between Facebook strategies and consumer engagement in this industry, this research will look at McDonalds, Burger King, Subway and KFC.
Theoretical Framework
Consumer Engagement
Consumer engagement as a concept has been mostly investigated through psychology. That is, through experiments, researchers have attempted to establish what factors influence the engagement of a consumer towards certain brands. To elaborate, engagement in this case means how actively involved and committed the consumer feels to the brand. For example, Hiller et al. (2013) performed a survey on students about brand engagement to an apparel brand. After planting forms of traditional advertising around the campus, a similar survey was performed to test whether the students’ engagement had increased. Interestingly, the survey showed no effect of traditional advertising on the brand engagement, meaning managers will need to employ different strategies to enable engagement growth. What is more, Higgins & Scholar (2009) compared the results from several studies investigating what factors influence a consumer’s engagement. The most salient conclusion was the disparity between the
commitment of consumers to a brand and the expected positive effect on the company’s reputation or revenue. To clarify, although high engagement led to a more positive experience for the consumer in some cases, when the consumer was confronted with a substantial amount of negative information, the opposite was true. Therefore, while consumer engagement itself can be valuable for a company, the tone of this engagement influences the end result.
As mentioned, consumer engagement itself has been widely researched. However, in relation to social media such as Facebook or Twitter, there is still much to be learned. Due to the vast amount of consumers active on these media, in combination with the opportunity for direct interaction with the company, consumer engagement will have to be considered differently. Cvijikj & Michahelles (2013) analysed the Facebook posting of several retailer companies and the response that each post generated. Consumer engagement was made
tangible by taking the amounts of likes, shares and comments into consideration.
Additionally, the interaction duration, and the time between the original post and the final comment was looked at. Cvijikj & Michaehelles (2013) found that different types of posts as well as their purpose had varying effects on the consumer engagement. These types and purposes will be discussed later on.
What is more, as social media offer the opportunity for consumers to engage the company themselves more easily, there is more potential for input. That is, listening to consumers’ ideas about products or new promotions can be valuable for the company to increase their reputation (Van Belleghem, 2013; Dekay, 2012) and therefore their revenue (Jahn & Kunz, 2012; Anderson et al., 2014). Moreover, Cheung et al. (2009) showed that the opinions of consumers influence their peers. By analysing several fashion brand communities on forums, Cheung et al. found that when information is shared about a certain brand, other consumers’ opinions as well as their likelihood to purchase the brand’s products are
influenced. As a result, when substantial amounts of negative information is being shared, the company’s reputation and revenue will suffer. Conclusively, not only can a company benefit from interacting with their consumers, by not getting involved there is the potential that negative opinions from consumers will float around unopposed, hurting the company.
As mentioned, the reputation of a company can decrease when faced with a substantial amount of negative communication (Dekay, 2012; Weinberger et al., 1981). Therefore,
investigating what approach to posting on their Facebook page generates the least negative response can benefit companies.
Social media strategy
Purpose of communication
To attempt to map a company’s Facebook strategy so its relation to consumer engagement can be determined, the focus will first be on the purpose behind a company’s communication to the consumers. Firstly, there are several manners in which a company can advertise their product or brand. Most prolifically, information about a product is given or entertainment is supplied through music, jokes or funny videos. For example, Hsu et al. (2007) used a
questionnaire to investigate how consumers perceive different types of advertising on TV and radio. Their findings showed that obviously some information about the product or brand is required, to ensure the link to the company. However, by providing some kind of
entertainment in the advertisement, the consumer felt more positively engaged to the company after seeing or hearing it. In contrast, Petrovici & Marinov (2007), who investigated the effects of different types of advertising in several countries in the EU came to a different conclusion. To elaborate, while advertising with an entertainment purpose was more influential in some Romania, the informative purpose generated the more positive attitude towards the advertisement in Bulgaria.
Besides the before mentioned purposes of a company’s communication, Cvijikj & Michahelles (2013) found the remuneration purpose to be used substantially on a company’s Facebook page. Remuneration in their research implied any posting of the company that contained a contest in which the consumers could enrol to obtain a prize. Cvijikj &
Michaehelles (2013) concluded that different purposes of the posts by a company had varying effects on the consumer engagement that was generated. That is, the remuneration posts enjoyed the least consumer engagement overall, while information posts that contained product information or notifications about events was found to be the 2nd most influential on
engagement. Finally, the entertaining posts which contained funny pictures, videos, or some type of wordplay were found to generate the most consumer engagement.
In previous research it was found that the highest engagement was found in reaction to entertainment-orientated posts, with information and remuneration following respectively. (Cvijikj & Michahelles, 2013). Although these results were obtained through investigation of the retail industry as a whole, rather than the controversial fast food industry, the expectation is to find similar results.
Therefore, the first hypotheses are:
H1a: Posts by the company that contain an entertainment purpose will generate the most consumer engagement
H1b: Posts by the company that have an informative purpose will generate less consumer engagement than entertaining posts, but more than remuneration posts.
H1c: Posts by the company of remunerative purpose will generate the least consumer engagement.
As mentioned, posts with an entertainment purpose generate a more positive connotation towards the company than informative posts (Hsu et al. 2007). Although the effect of remuneration posts has not yet been tested in this sense, it is likely that due to the fact the consumer can win a prize, these posts will generate less negative response than informative or entertaining messages.
Therefore these hypotheses are proposed:
H2a: Posts by the company that have an entertainment purpose will generate more negative consumer engagement than remuneration posts, but less negative than informative posts. H2b: Posts by the company that have an informative purpose will generate the most negative consumer engagement.
H2c: Posts by the company that have a remunerative purpose will generate the least negative consumer engagement
Post vividness
Besides the purpose of a post, the actual content of the post and its relation to consumer engagement is also of value. To investigate this relation, the vividness of the post will first be considered. The vividness of a media type is determined by how it communicates information to the consumer (Steuer, 1992). To clarify, information portrayed only through text is less vivid than a picture or video. That is, due to the fact that the consumer of the information is presented with a richer type of media, they are presented with more vividness, meaning they are more likely to experience the interaction as fulfilling (Coyle & Thorson, 2001). The reasoning behind this according to Steuer (1992) lies in the telepresence felt by the user or consumer. To elaborate, telepresence is explained as the feeling of presence through any medium of communication. For instance, when someone is watching a film they will be more invested and therefore have a greater sense of ‘being there’. In contrast, reading a short letter will grant less telepresence.
What is more, Coyle & Thorson (2001) investigated this feeling of presence in relation to vividness and interactivity of websites. They found that when consumers are faced with media types with a higher level of vividness, they will become more interactive with the website’s brand due to an increased feeling of telepresence. In addition, Griffith et al. (2001) investigated the relationship between different types of interface in online catalogues on consumer engagement. The experiment they created was performed on marketing students, who were shown catalogues with different types of interfaces. Their results showed that when a web interface is more interactive to use, the students felt more engaged to the brand as well as being more likely to purchase the products presented. Cvijikj and Michahells (2013) found
similar results for consumers on Facebook pages. By categorising varying levels of vividness in company’s posts into status updates, pictures, links, and videos, they found that a higher level of vividness had positive effects on consumer engagement. Therefore, the following hypothesis is proposed:
H3a: More vividness in posts results in more consumer engagement.
Besides vividness increasing the amount of interactivity and engagement, researchers also found that the telepresence experienced by the consumers created a more positive feeling towards the company whose website they were shown (Coyle & Thorson 2001). This result was attributed to the more human-like interaction that was established when the message was portrayed more vividly. To clarify, the social needs of the consumer are fulfilled as the interaction felt like a social endeavour (Steuer, 1992). Moreover, Griffith et al. (2001) found that the marketing students felt more connected to a company when their catalogue was more interactive, which also increased positive connotations. As a result, it can be expected that the tone of their comments replying to the company’s posts will be more positive when there is a higher degree of vividness. This leads to the hypothesis:
H3b: Posts with a higher level of vividness will generate more positive consumer engagement.
Directed and undirected communication
In addition to the company posting itself, they also have the opportunity to respond to the consumer’s replies to the original post. However, this interaction implies a different type of communication, namely, directed communication. The difference between directed and undirected communication lies in the range of the message send by the company. To
elaborate, the original posts the company creates on their Facebook page are undirected in the sense that they are broadcasted to all consumers who check the page. In contrast, directed
messages, although they can be read by anyone, are directed at one specific consumer. On the company’s own Facebook page this encompasses any reply to a consumer’s comment.
The added value of directed communication lies in the relationship that is engaged communicating one on one. Goh et al. (2013) analysed Facebook messaging through a combination of qualitative and quantitative content analyses. To elaborate, the connection between the creator of the message and the directedness of this message were evaluated in relation to the purchase behaviour of consumers. The results showed that when the company directs their message to a consumer rather than broadcasting it to everyone, the positive effect on purchase behaviour is increased. In addition, Burke et al. (2011) found that when the consumer is directly communicated with by the company, they are required to be more active, thus increasing their engagement. Burke et al. (2011) came to this conclusion by investigating the different ways in which a consumer can communicate through Facebook. To elaborate, surveys were taken by Facebook users over time and linked to their Facebook activity. The researchers concluded that while passively absorbing undirected information that is being broadcasted had little effect on the engagement of consumers, direct interaction with a company did have a positive effect. Conclusively, by creating this dialogue with the
consumer, the company’s online reputation is positively influenced, as they have shown their commitment to the consumer (Goh et al., 2013; Burke et al., 2011). As a result, the consumer engagement on the Facebook page will increase, leading to the following hypothesis:
H4a: The more often the company replies to its consumers’ comments, the more consumer engagement is generated.
Besides the increase in engagement from the consumers, previous research has also shown that the subjects gained more positive connotations towards the companies (Goh et al., 2013; Burke et al., 2011). Coyle & Thorson (2011) found similar effects in their research on vividness. Their results were explained by a higher degree of media richness in directed
communication. That is, due to the increased sense of telepresence that is created when the consumer is no longer passively absorbing information that is being broadcasted, but engaging in a dialogue with the company, more positive connotations are attributed to the company (Steuer, 1992; Goh et al., 2013). Therefore, the following hypothesis is proposed:
H4b: The more often the company replies to its consumers’ comments, the more positive the consumer engagement will be.
To create a clear model of the factors influencing the amount of consumer engagement as well as its tone, all hypotheses are portrayed in Figure 1.
Methods
Research Design
A manual content analysis was conducted on Facebook posts and comments on the fast food companies’ official Facebook pages. Besides analysing the company’s own messages on the concepts that were mentioned earlier in this proposal, the comments from consumers as well as the interaction between consumers and the company were considered. The aim of this research was to compare the social media strategies of fast food companies and the consumer engagement that resulted from these strategies. Consequently, an advice could be crafted on what strategy creates the largest amount of consumer engagement. What is more, the resulting tone of the consumer engagement was also analysed. The choice for the fast food industry was based on the increased potential for controversy. As explained in the introduction, fast food has often been linked to obesity and other health risks. In addition, recently new issues arose about the wage of the employees in this sector. As negative messages have been shown to have a bigger impact than positive messages (Dekay, 2012; Weinberger et al., 1981), it was interesting to look at how companies in such a potentially controversial industry incorporate this into their social media strategy.
Sample
The sample (N=2460) was comprised of all posts by four fast food companies in the year 2014. Additionally, the first 10 comments and any responses from the company to these comments were included. The fast food companies considered in this research were:
McDonalds, Burger King, Subway and KFC. The choice for these companies lied mostly in their international presence, meaning they are well established brands. As a result, it could be
expected that there was a high degree of activity on a medium such as Facebook, which increased the potential for this research to find enough data to come to an interesting
conclusion. In addition, by selecting companies that were all in the same industry, the results were expected to be more comparable.
To ensure the actuality of the research, the time period investigated was the most recent full year, 2014. After thoroughly browsing the Facebook pages of the companies involved in this research, the conclusion was made that it was feasible to include all posts of the companies in 2014 which were shown under that years highlights. Moreover, this choice was also made because Facebook shows this as a default setting, meaning consumers are shown these posts when browsing rather than all posts by the company. However, as the number of comments varied substantially with the maximum being in the thousands, only the first 10 comments were taken into consideration. Due to the way in which Facebook portrays comments, the choice was made to analyse the top comments. To elaborate, the choices Facebook has for portraying comments are: Recent activity, most relevant, and top comments. The obvious issue with ‘recent activity’ is that random spam comments are sometimes added months after, subsequently often being irrelevant to the original post and therefore never responded to. Additionally, ‘most relevant’ is determined in an unknown way and therefore would be difficult to analyse without knowing what Facebook had based this on. Finally, ‘top comments’ are determined by likes and views, which means that these are the comments that had the biggest impact for the company, as they received the most attention.
Finally, The data collection was completed via Facebook, where all messages are publicly accessible.
Procedure
To be able to investigate the varying relationships between social media strategies and consumer engagement, a coding scheme (See Appendix) was created. Besides collecting the numbers of likes, shares, comments, and the interaction duration, this coding scheme was focussed on analysing the companies’ posts by their purpose and vividness. Additionally, the tone of comments from consumers and whether they were replied to by the company was considered.
Firstly, the three possible purposes of posts and their relationship with consumer engagement were analysed through one way ANOVAs. Subsequently, regression analyses were performed to establish the relationship between the vividness of a post and consumer engagement. Finally, to investigate the link between the ratio of directed communication from the companies and consumer engagement, ratios were calculated for each company.
Moreover, these ratios were linked to consumer engagement through more regression analyses.
Operationalisation of Variables
Consumer Engagement
For the purpose of the research, consumer engagement was made tangible for measurement. Firstly, the number of likes, shares, comments, and interaction duration were considered to compare the response to post types and purposes that will be discussed later. To elaborate, the interaction duration was found by calculating the time between the original post and the final comment (Cvijikj & Michahelles, 2013). Additionally, by analysing the tone of comments posted by consumers in response to the company’s posts, the type of consumer engagement generated by different types of posts was investigated. The tone of the comments was compared for several factors to discover what strategy with regard to posting receives more
positive or negative feedback. The coded possibilities in this research were: positive, neutral, and negative. Logically, any posts containing a positive attitude towards the company fell into the positive category, while negative messaging were coded as negative. The neutral category encompassed any posts which did not fit into the other two categories. An example of this would be a question.
Post Purpose
As this research is done from outside of the companies involved, it was not possible to establish the social media strategy and the logic behind it from the companies’ perspective. That is, rather than asking the companies how they approach Facebook, what they attempt to achieve from it and how they do this, the content on their page was evaluated to conceptualise the strategy behind it. The evaluation of the content posted by the company was based on several aspects. Firstly, the purpose of each individual post was established. Moreover, the possible purposes were categorised as: Entertainment, Information, and Remuneration (Cvijikj & Michahelles, 2013). These purposes were then linked to the consumers’ responses to establish what approach generated the greatest consumer engagement. The entertainment category contained all posts that were not directly linked to the products offered by the company, but instead focussed on jokes, word play or other forms of entertaining the consumer without trying to promote the product (Cvijikj & Michahelles, 2013).
The information category contained posts that referred directly to a product or service offered by the company. To elaborate, this category could also be defined as the more
traditional way of promoting the company through broadcasting a new product or discount deal in a manner which would also be used on traditional media such as television or magazines (Cvijikj & Michahelles, 2013).
Finally, the remuneration category encompassed posts that asked participation from the consumer with the possibility of obtaining a prize or discount. In other words, any type of contest which results in a potential reward when participated in fell into this category (Cvijikj & Michahelles, 2013).
Vividness
To explore whether posts in varying levels of vividness generated different amounts and tone of consumer engagement, these levels were first defined. That is, in ranking of least vivid to most vivid, these levels were: status update, picture, link, and video (Cvijikj & Michahelles 2013). What is more, when a post contained several of these categories, the highest level of vividness amongst these was coded.
Directed Communication
Directed communication on Facebook entails the owner of the page directly responding to content posted by consumers (Goh et al., 2013). In the case of the fast food companies in this research, consumers were not permitted to place original posts. As a result, the only consumer content found were the comments to the original posts of the company. What is more, this meant that for the company to engage in directed communication, they would have to reply to these comments. So, to establish the company’s investment in engaging in directed
communication with their consumers, a ratio was calculated for each company that described what percentage of consumer comments they reply to. The new variable was created by dividing the amount of comments replied to by the total amount of comments and the values for each company can be found in Table 1. To clarify, the total amount here meant those coded, rather than the actual total.
Table 1. Company # of replies to comments # of total comments Ratio Burger King 132 870 .15 Subway 280 620 .45 KFC 319 530 .60 McDonalds 313 440 .71 Results
To analyse the relationship between consumer engagement and the aspects of social media strategy as discussed before, multiple t-tests and one way ANOVAs were conducted. These tests focussed on the differences between the categories of post type and purpose as well as the ratio of directed communication. As the amount of consumer engagement is split into likes, comments, shares, and interaction duration, this structure will be used to present the results, in order of hypotheses.
Purpose of communication
To investigate the link between the purpose of a post and consumer engagement, several one way ANOVAs were conducted between Entertainment (N=1101), Information (N=1289), and Remuneration (N=70) on the amount and tone of consumer engagement.
Likes
The one way ANOVA test showed a significant difference between the three purposes F(2, 2457) =13.82, p < .001. The post-hoc Tukey test indicated that Information posts (M=15148,
SD=25818) generated significantly more likes than Entertainment (M=9954, SD=25761, p <
.001), and Remuneration (M=7447, SD=8948, p=.04). However, no significant difference was found between Entertainment and Remuneration. So, hypotheses H1a, H1b, and H1c were all not proven with regard to likes.
Comments
The one way ANOVA test showed a significant difference between the three purposes F(2, 2457) =6.35, p = .002. The post-hoc Tukey test indicated that Entertainment posts (M=741,
SD=2441) generated significantly more comments than Information (M=506, SD=876, p=.003), However, no significant difference was found between Remuneration (M=310, SD=234) and the other purposes. So, hypotheses H1a was partly proven with regard to
comments, as posts with an entertainment purpose were found to generate the most comments, however the comparison with remuneration did not grant significant results. Additionally, H1b could only be considered partly proven as the posts with an informative purpose did generate less comments than those that were entertaining, but no significant difference could be found with remuneration. Consequently, H1c was not proven with regard to comments.
Shares
The one way ANOVA test showed a significant difference between the three purposes F(2, 2457) =7.97, p < .001. The post-hoc Tukey test indicated that Entertainment posts (M=1358,
SD=8002) generated significantly more shares than Information (M=467, SD=1395, p < .001).
However, no significant difference was found between Remuneration (M=718, SD=777) and the other purposes. As a result, H1a and H1b were only partly proven with regard to shares, while H1c was not proven at all.
The one way ANOVA test showed a significant difference between the three purposes F(2, 2457) =91.55, p < .001. The post-hoc Tukey test indicated that Information posts(M=20.1,
SD=14.8) initiated a significantly longer interaction duration than Entertainment (M=10.9, SD=19.2, p < .001) and Remuneration (M=10.6, SD=7.1 p < .001). However, no significant
difference was found between Entertainment and Remuneration. Consequently, Hypothesis H1b was partly proven as Information generated a longer interaction duration than
Remuneration, but also longer than Entertainment. In addition, H1a and H1c were both not proven.
Tone of consumer engagement
The one way ANOVA test indicated a significant difference between the three purposes F(2, 2457) = 5.30, p=.037. The post-hoc Tukey test showed that Remuneration posts (M=1.81,
SD=.91) generated significantly less negative comments than Entertainment (M=2.15, SD=.91, p=.009) and Information (M=2.08, SD=.93, p=.049). However, no significant
difference was found between Entertainment and Information. So, H2a and H2b were both only partly proven, while H2c was fully confirmed.
Vividness
To establish how the level of vividness relates to the consumer engagement generated, regression analyses were performed on vividness in relation to the amount and tone of consumer engagement.
Likes
The regression analysis indicated that there is a significant positive relationship between the vividness of the company’s post and the amount of likes it generated, F(1, 2458)=7.54,
p=.006. However, this relationship was found to be very weak as only 1% of the variance in
[5442.51, 11609.49]. So, H3a was confirmed in relation to likes. However, this relationship was found to be very weak.
Comments
The regression analysis showed a significant positive relationship between the vividness of the original post and the comments it generated, F(1, 2458)=141.44, p < .001. However, this relationship was found to be rather weak as only 5% of the variance in the amount of
comments (R²=.05) was explained by vividness, b*=522.34, t=11.89, p < .001, CI [436.22, 608.47]. This means H3a was also confirmed in relation comments. However, this
relationship was found to be rather weak.
Shares
The regression analysis showed a significant positive relationship between the vividness of the company’s post and the amount of times it was shared, F(1, 2458)=107.26, p < .001. However, this relationship was indicated as being rather weak, as only 4% of the variance in the amount of shares (R²=.04) was explained through vividness, b*=1424.47, t=10.36, p < .001, CI [5442.51, 11609.49]. Consequently, H3a was confirmed with regard to shares, although this relationship was again rather weak.
Interaction Duration
The regression model for the relationship between the length of interaction duration and vividness was found to be non-significant, F(1, 2458)= 1.26, p=.262. So, H3a was not proven with regard to interaction duration.
Tone of consumer engagement
The regression analysis indicated a significant positive relationship between the vividness of the original post and the tone of the comments it received, F(1, 2458)=12.55, p < .001. However, this relationship was found to be very weak as only 5% of the variance in the tone
of the consumer engagement (R²=.04) was explained by the post’s vividness, b*=.08, t=3.54,
p < .001, CI [.04, .13]. Consequently, H3b was proven, although the relationship found was
very weak.
Directed Communication
To investigate the relationship by the amount of directed communication from the company and consumer engagement, regression analyses were performed. In these analyses a computed variable was assigned to each company to show their ratio of responsiveness, which was then used in regression analysis with relation to the amount and tone of consumer engagement.
Likes
The regression analysis showed a significant negative relationship between the ratio of directed communication and the amount of likes generated, F(1, 2137)=13.15, p < .001. However this negative relationship was very weak, as only 1% of the total variance in the amount of likes (R²=.01) was explained by the ratio of directed communication, b*=-9186.71,
t=3.63, p < .001, CI [-14155.60, -4217.82]. As the relationship between directed
communication and the amount of likes that was generated had a negative nature, H4a was not proven in this case.
Comments
The regression analysis indicated a significant positive relationship between the ratio of directed communication and the amount of comments that were posted by consumers in total,
F(1, 2137)=4.31, p=.038 However this positive relationship was very weak, as only 1% of
the total variance in the amount of total comments (R²=.01) was explained by the ratio of directed communication, b*=-409.57, t=2.08, p=.038, CI [22.50, 796.63]. Subsequently, H4a in the case of comments was proven, but the relationship was very weak.
The regression analysis showed a significant positive relationship between the ratio of directed communication and the amount of times posts were shared, F(1, 2137)=13.27, p < .001. However this relationship was very weak, as only 1% of the total variance in the amount times the posts were shared (R²=.01) was explained by the ratio of directed communication,
b*=2239.82, t=3.64, p < .001, CI [1033.84, 3445.80]. So, H4a in relation to shares was
proved. However, this relationship was very weak.
Interaction Duration
The regression analysis showed a significant negative relationship between the ratio of directed communication and the length of the interaction duration, F(1, 2137)=86.17, p < .001. However this negative relationship was very weak, as only 4% of the total variance in the interaction duration’s length (R²=.04) was explained by the ratio of directed
communication, b*=-9186.71, t=3.63, p < .001, CI [-14155.60, -4217.82]. H4a is not proven with regard to interaction duration, as the relationship found was negative rather than positive.
Tone of consumer engagement
The regression analysis showed a significant positive relationship between the ratio of directed communication and tone of the consumer’s responses, F(1, 2137)=24.67, p < .001. However this negative relationship was very weak, as only 1% of the total variance in the tone of the comments (R²=.01) was explained by the ratio of directed communication, b*=.48,
t=4.97, p < .001, CI [.29, .67]. So, H4b was proven, although this relationship was again very
Discussion
Engaging consumers via social media is important as more engagement has been shown to positively influence a company’s reputation and their consumers’ purchase behaviour (Jan & Kunz, 2012; Anderson et al., 2014). What is more, to avoid negative effects on their
reputation, companies will need to ensure that this engagement has a positive tone (Dekay, 2012). This research analysed the content on fast food companies’ Facebook pages to
investigate how the online communication of these companies was related to the engagement of the consumers.
Post Purpose
The first factor that was considered was the purpose of posts by the company.
Previous research indicated that the most consumer engagement was generated through posts of an entertainment nature, followed by informative posts, and finally remunerative posts. However, in the case of the fast food companies, the relation between a post’s purpose and the amount of consumer engagement was either not significant or very weak. With regard to comments and shares, entertainment posts did generate the highest amount, but no significant difference was found with remunerative posts. On the other hand, informative posts rather than entertaining ones generated the longest interaction duration.
As a result, most of the hypotheses regarding post purpose were not proven. What is more, when there was a significant relation, it was found to be very weak. However, these results might not be conclusive. To elaborate, the remuneration category contained around 4 percent of all posts, resulting in very skewed data. In addition, many of the comments from consumers were unrelated spam messages. As a result, the interaction duration was often much longer than it would have been without spam. Additionally, the number of comments was obviously boosted in this way, which causes the original post to be rebroadcasted on
people’s own Facebook page, possibly stimulating shares and likes as well. So, the analysis of post purpose in relation to the amount consumer engagement was inconclusive.
Furthermore, remunerative posts were found to generate the least negative responses, while information and entertainment showed no difference amongst each other. While the spam messaging does not affect these results as only the top comments were coded, the small portion of remunerative posts makes these results somewhat unreliable. Finally, many of the comments by consumers were found to be unrelated to the original posts, instead commenting on controversial predicaments the company was in during that time. For example, Burger King moved their corporate headquarters to Canada to avoid taxes, which infuriated many American consumers. These conclusions suggest that due to the controversial aspect of the industry, managers will have to approach social media differently to obtain positive results. Vividness
Besides the purpose of the post, the vividness level was also analysed with regard to its effects on consumer engagement. Previous research suggested that more vividness in communication created a greater telepresence for the consumer, which through making the interaction more human-like positively influences their opinion of the company (Cvijikj & Michahelles, 2013; Coyle & Thorson, 2001; Steuer, 1992). Although the regression analyses showed that the level of vividness did positively correlate with the amount of likes,
comments, and shares, these relations were very weak. Again, similar issues with spam messaging may have affected these results. Additionally, over 60 percent of all posts
contained only a picture, potentially biasing the results as one group was so much larger than the others. Furthermore, the interaction duration was not significantly influenced by the vividness of the company’s post.
In addition, the tone of consumer engagement only showed a weak positive correlation with the vividness of the original posts, but these results may not be usable either. To
elaborate, in spite of the attempts to filter out most of the unrelated spam messaging from consumers by not including comments that were over a week apart from previous comments, it is highly likely that the spam content still influenced the coding, resulting in unreliable results. So, the controversy surrounding these companies may cause a different method of research to be required. Potentially, a more in depth content analysis may shed more light on how much negative messaging is actually on the pages. Additionally, interviews with
consumers as well as corporate employees may give more insight on the company’s strategy with regard to social media. Moreover, the consumer’s response to these strategies rather than their general opinion of the company could be measured.
Directed Communication
Finally, the degree in which the companies engage their consumers in directed communication, in other words, how often they replied to comments, was calculated and analysed in relation to consumer engagement. Goh et al. (2013) found that increasing directed communication, rather than the broadcasting of messages to a wider public, positively
affected the consumer engagement as well as their opinions of the company. While this research also found a positive correlation in the case of the amount of comments and shares, these were again very weak. Moreover, the relation between directed communication and likes as well as interaction duration was actually found to be mildly negative. Besides the before mentioned issues with spam messaging, there may be another issue with regard to this analysis. That is, the ratio was calculated for the company over the entire span of 2014. However, by doing this, the influence of time was not taken into consideration. To clarify, if a company started replying an increasing amount to the consumers’ comments at the start of the year, the effects on consumer engagement may not be clear immediately. Moreover, by then investigating the entire year as a whole, the potential increase in engagement that was caused by the increase of directed communication would be nullified by the results from earlier in the
year. Finally, if a company changed their approach with regard to directed communication halfway through the year, this change would also remain unnoticed.
Besides that, the tone of consumer engagement was shown to have a positive relationship with more directed communication. However, again the relation that was found was very weak.
Conclusion
While previous research showed the positive relation between the purpose of a post, its vividness, and the directed type of communication, this research shows that these results may not be valid for all industries. Although there are some issues with this research, such as having only a single coder, it does seem clear that the controversy surrounding a particular industry has an effect on the consumer engagement. That is, a substantial amount of negative messaging from consumers appears in the comments to a company’s post, regardless of that post’s purpose or vividness. As a result, most of the conclusions about these relations cannot be reliable drawn from this research. However, this research does underpin the
interconnectedness of all media. That is, the Social Media Crisis Communication (SMCC) model explains how through word of mouth, both offline and online, the reach of an
interaction between a consumer and a company is extended (Fisher et al., 2015). For example, if someone sees a negative news segment on a company, they may share this with other consumers, who could then post on that company’s Facebook page. In the case of the fast food industry this model explains how a negative message broadcast on TV will negatively impact the consumer engagement generated on Facebook, suggesting that a media strategy needs to cover all types of media. To summarise, the main conclusion that can be drawn is that the controversy surrounding the fast food industry impacts the interaction between these companies and their consumers. Consequently, these companies may require a different
strategic approach. However, this will need to be confirmed through additional research, potentially of a more qualitative nature.
References
Anderson, Kelley C., Knight, Dee K., Pookulangara, S. & Josiam, B. (2014). Influence of hedonic and utilitarian motivations on retailer loyalty and purchase intention: a Facebook perspective, Journal of Retailing and Consumer Services, 21(5), 773-9 Bortree, D. S., & Seltzer, T. (2009). Dialogic strategies and outcomes: An analysis of
environmental advocacy groups' Facebook profiles. Public Relations Review, 35(3), 317‐319.
Burke, M., Kraut, R. & Marlow, C. (2011). Social capital on Facebook: differentiating uses and users. Human Factors in Computing Systems: Proceedings of the SIGCHI
Conference. 571-580.
Cheung, C.M.K, Xiao, B.S. & Liu, I.L.B. (2014). Do actions speak louder than voices? The signalling role of social information cues in influencing consumer purchase decisions.
Decision Support Systems, 65, 50-58.
Coyle, J. R. & Thorson E. (2001). The effects of progressive levels of interactivity and vividness in web marketing sites. Journal of Advertising, 30(3), 65-77.
Cvijikj, I.P. & Michahelles, F. (2013). Online engagement factors on Facebook brand pages.
Social Network Analysis and Mining, 3, 843-861.
Dekay, S. (2012). How large companies react to negative Facebook comments. Corporate
Communications: An International Journal, 17(3), 289-299.
Duggan, M., Ellison, M. B., Lampe, C., Lenhart A., & Madden, M. (2015). Social media update 2014. Pew Research Center. Retrieved From:
Fisher Liu, B., Fraustino, J. & Jin. Y. (2015). Social media use during disasters: How
information form and source influence intended behavioral responses. Communication
Research, published online 13 January 2015.
Griffith, D.A., Krampf, R. F. & Palmer, J. W. (2001). The role of interface in electronic commerce: Consumer involvement with print versus on-line catalogs. International
Journal of Electronic Commerce, 5(4), 135-153.
Johnson, L. (2014). See how the top fast food companies are performing on Facebook, and if fans are liking it. Adweek. Retrieved from:
http://www.adweek.com/news/technology/see-how-top-fast-food-chains-are-performing-facebook-and-if-fans-are-liking-it-161236
Hiller, C., Kim, Y. & Kozar, J.M. (2012). Social normative influence: An exploratory study investigating its effectiveness in increasing engagement in sustainable apparel-purchasing behaviors. Journal of Global Fashion Marketing, 3(4). 172-179. Higgins, T.E. & Scholar A.A. (2009). Engaging the consumer: The science and art of the
value creation process. Journal of Consumer Psychology, 19(2), 100-114 Hsu, J.L., Yang, S. & Su L.C. (2007). Who is watching TV? Who is listening to radio?
consumer perception of TV and radio advertising information. Social Behaviour and
Personality: an international journal, 35(2). 157-168.
Kaplan, A.M. & Haenlein, M. (2010). Users of the world, unite! The challenges and opportunities of social media. Business Horizons, 53, 59-68.
Kelleher, T. (2009). Conversational voice, communicated commitment, and public
relations outcomes in interactive online communication. Journal of Communication, 59(1), 172-188.
Jahn, B. & Kunz, W. (2012). How to transform consumers into fans of your brand. Journal of
Petrovici, D. & Marinov, M. (2007). Determinants and antecedents of general attitudes towards advertising; A study of two EU accession countries. European Journal of
Marketing, 41(3/4), 307-326
Scott, A.O. (2003). Film review: When all those big macs bite back. NYtimes. Retrieved from:
http://www.nytimes.com/movie/review?res=9C0CE5DC163CF934A35756C0A9629C 8B63
Steuer, J. (1992). Defining virtual reality. Dimensions determining telepresenee, Journal of
Communication, 42 (4), 73-93.
Van Belleghem, S. (2013). De conversation company. Van Duuren Management.
Weinberger, M.G. Allen, C.T. & Dillon, W.R. (1981). Negative information: Perspectives and research directions. Advances in Consumer Research, 8. Association for Consumer Research, 398-404.
Worstall, T. (2015). Fast food could handle a 15$ minimum wage without firing anyone, hurrah! Forbes. Retrieved from:
http://www.forbes.com/sites/timworstall/2015/01/21/fast-food-could-handle-a-15-minimum-wage-without-firing-anyone-hurrah/
Appendix Codebook
Research Question: How do different Facebook strategies of fast food companies influence
their consumer’s engagement?
The first section of this codebook will be used to analyse the original post from the company, the second section will focus on the comments from consumers in response to this post, and the third section analyses any replies from the company to these comments in turn.
Section 1
Part 1: general
1. What company created the post? 1) Burger King
2) Subway 3) KFC 4) McDonalds
2. In what month was the post created?
3. How long was the interaction duration in days? (See Explanation 1)
Part 2: Type of post
4. What purpose did the post have? (See Explanation 2) 1) Entertainment
2) Information 3) Remuneration
5. In what category of vividness does the post belong? (See Explanation 3) 1) Status update
2) Picture 3) Link 4) Video
6. How many likes did the post receive?
7. How many comments did the post generate? (See Explanation 4)
8. How many times was the post shared?
Section 2
This section of the codebook only applies to the comments made by consumers. This means that for every post, 10 comments will be coded.
9. How many replies does the comment have? 1) 0-50
2) 51-100 3) 101-150 4) 151-200 5) 201+
10. Is the comment replied to by the company? [if no Skip Section 3&4] 1) Yes
2) No
11. What is the tone of the comment? (See Explanation 5) 1) Positive
2) Neutral 3) Negative
Section 3 [Skip if Q10=no]
This section of the codebook only applies to the replies of the company itself to a comment on one of their original posts.
12. How quick was the response rate in days? (See Explanation 6) 13. What type of response does the company give?
1) Apology
2) Thanking / Positive enforcement 3) Information / Explanation
Section 4 [Skip if Q10=no]
This section of the codebook applies to the consumer’s reaction to the interaction shared with the company
14. Does the consumer respond to the company’s reply? [If no skip 15] 1) Yes
2) No
15. Does the consumer seem content with the company’s reply? [Skip if 14=No] 1) Yes
Explanation 1
The interaction duration is the time between the original post and the final comment.
However, spam like comments which are added longer than a week after any other comment will not be included in this variable.
Explanation 2
For this question, the categories are defined in the following way:
1) Entertainment covers any posts that do not directly refer to the company or one of its products.
2) Information covers the posts that inform the consumers about events, products or the company itself.
3) Remuneration covers posts which contain a possibility for compensation for the consumer. For example, posts about a competition where the consumer has the opportunity to win a prize.
Explanation 3
The categories for this questions are defined in the following way: 1) A status update contains only text.
2) The picture category contains all posts which contain 1 or more pictures. 3) The link category contains all posts which link to a different section or website. 4) The video category contains all posts which have a video in the body of the post, so
not just a link
When coding this question the highest category which applies to the post should always be chosen. For example, a post containing a picture and a link, falls in the link category.
Explanation 4
Due to the unavailability of the number of comments per post, an estimate of this number was more practical. As these estimates can be time consuming, the answers to this question are divided up into steps of 50, with 201+ being the maximum.
Any comments containing criticism towards the company, its employees or its products are considered negative. For example, a call to boycott a company.
Any comments containing any type of positive attitude will be considered positive. For example, a comment saying “I love this food”.
Finally, any comments not containing either positive or negative notes will be coded as neutral.
Explanation 6
The response rate is calculated as the time it takes the company to reply to the comment directly in days. This amount is rounded up, meaning 1 day stand for within 24 hours, 2 days stands for within 48 hours, etc.
