Getting involved: the relationship between European
airlines’ Facebook posts and consumer engagement
Kimberley T. P. M. van Nuland
11608994
Master’s Thesis
Graduate School of Communication Master’s programme Communication Science
Supervisor: dr. Sandra M. Zwier 02-02-2018
3
Abstract
The past few years have shown a surge in brands’ social media use. Many brands now use social media to engage with their consumers. Studies on brands and social media aim to provide insight in strategies to strengthen the relationship between brands and consumers. However, social media are constantly changing which causes research to quickly become outdated. So far, limited research has been done to explain the relationship between brands’ social media use and consumer engagement. Previous studies have focused on very specific brands, which leaves room for more research. The present study aims to fill in some of the gap by analyzing European airlines’ Facebook posts. A quantitative content analysis is employed to answer the question to what extent there is a relationship between format characteristics and rational and emotional appeals in European airlines’ posts on Facebook, and consumer engagement. The results show that there is a relationship between the use of engaging format characteristics and the use of emotional appeals in the posts. However, no evidence is found for a relationship between the use of engaging formats, or the use of emotional appeals over rational appeals in European airlines’ Facebook posts and consumer engagement.
Introduction
Many marketers strive for highly engaged consumers as it has many benefits for their brand. Creating consumer engagement has been a part of communication strategies for a while. The rise of web 2.0 has resulted in new opportunities for brands to build relationships with their consumers. Social media make it easier to interact with consumers and are now an important component in stimulating consumer engagement.
Many brands have integrated social media in their communication strategies.
4
social media the past few years. However, social media are constantly developing. New social media arise, algorithms change and the usage changes. This causes research to quickly
become outdated. It is important to keep researching how social media can be used by organizations and consumers to strengthen their relationship and benefit both parties.
Recent research on organizations’ social media posts and consumer engagement has mainly focused on consumers’ motivations to engage with organizations on social media. Research that focused on brands’ social media posts were mostly exploratory. Recent studies have for example, focused on social media posts of brands selling fitness watches (Pinto & Yagnik, 2016) or beauty brands (Shen & Bissel, 2013). These are very specific products which leaves room for more research on other type of organizations’ social media usage, for example on experience service providers like airlines.
Airlines are some of the first organizations to embrace new online technologies to improve their communication strategies and customer service. For example, Delta Airlines launched their first Twitter account only two years after the social media platform was created (Mago, 2017). In addition, consumer engagement is especially important in the airline
industry and has taken a center spot in airlines’ communication strategies. For instance, Southwest Airlines turned social media into a vital piece of their communication strategy in 2015, to engage more and build stronger relationships with their customers (Hickins, 2015). In the Netherlands, airline KLM is considered to be one of the best social media brands. KLM has repeatedly been chosen as ‘Best Social Brand’ (DVJ Insights, 2014) and has received various social media awards (Marketingfacts, 2014; Luchtvaartnieuws.nl, 2015). This makes it interesting to look at how airlines’ social media activities and consumer engagement are related.
To add to the body of knowledge on this topic, the present study will aim to provide more insight in the relationship between brands’ social media use and consumer engagement,
5
by answering the question to what extent there is a relationship between format characteristics and rational and emotional appeals in European airlines’ posts on Facebook, and consumer engagement.
Theoretical Framework
Brands and social media
The past few years have shown a surge in brands’ use of social media. Brands use social media to connect with their consumers and this has become an important part of their promotional mix nowadays. Brands gravitate towards using social media sites because the reach is usually higher than for other platforms (Ashley & Tuten, 2015). A majority of big brands have taken note of the opportunities that social media provide and have been posting messages on their own social media pages during the past few years. In 2013, big
international brands posted on their Facebook pages 24 times a month on average (Parsons, 2013). Considering most consumers follow multiple brands on Facebook, they will receive multiple messages by various brands every day.
Analyses of the content that brands post on Facebook have shown that the emphasis on brands’ Facebook pages often is not on sales promotion, like in traditional advertising. Instead it is mainly focused on creating engagement and building community (Shen & Bissel, 2013). Research by Shen and Bissel (2013) for instance has shown that most of beauty brands’ Facebook posts were entertainment-related, while the same brands had a minimal amount of Facebook posts containing promotions.
Brands’ social media pages appear to have measurable effects on the relationship between customers and brands, which make social media an important marketing tool that should not be overlooked (Jahn & Kunz, 2012). In addition, social media are easy to maintain in comparison to, for example, creating a smartphone app (Ashley & Tuten, 2015).
6
The present study will focus on posts placed by major European airlines on Facebook and how characteristics of these posts relate to consumer engagement. One of the reasons for the focus on the airline industry is that a large part of the revenues in the industry are derived from passengers and this is making customer engagement a top-priority for the industry (Mago, 2017). Further, major European airlines are highly comparable in the customer services they provide. The focus will be on customer engagement through Facebook because it is still the most widely used social medium.
Format characteristics
The present section focuses on the format characteristics Facebook posts can have. A brand can post something on their own timeline, whereby a post can solely consist of text, but it can also include images, videos and links. These characteristics of a Facebook post format can all help to convey the message.
People often say that ‘a picture is worth a thousand words’, meaning that an idea or story can be conveyed with the use of only one picture, and it may do so more effectively than a description in words. It might be that images carry a bigger emotional load than a
description alone does. Therefore, images are often used in charity advertising, in order to gain more donations (Burt & Strongman, 2005). Videos also often contain some form of emotion, since a video most likely conveys a story. Even just the movement in a video can lead to people feeling there is an emotional appeal present (Hiraga & Takahashi, 2011). Brands often use images and videos on social media such as Facebook to try to get the consumers engaged with the content (Parsons, 2013).
7 Emotional and rational appeals
Besides the use of certain formats, brands often try to get consumers to engage with their content, using various kinds of appeals. Johar and Sirgy (1991) distinguished two types of advertising appeals: image appeals and functional appeals. Their research pointed out that which appeals are more effective in persuading consumers was dependent on how the consumers view the product (Johar & Sirgy, 1991).
More recent research on brands’ social media content has also focused on the effects of different types of content appeals. Most often these are nowadays described as emotional appeals and rational appeals. Emotional appeals express a certain emotion. Emotional appeals relate to consumer’s social and/or psychological needs and include, for example, joy, humor, love, sadness, fear, anger, nostalgia, or excitement (Belch & Belch, 2015). An airline can for example, post a story about one of their customers who flew to a different country to meet their grandchild for the first time. This post would contain emotional appeals based on love and family. Rational appeals on the other hand tend to be informational, focusing on the consumer’s practical or functional needs. Rational appeals commonly used are quality, performance, convenience, durability, dependability, efficiency, comfort, convenience, price and health (Belch & Belch, 2015). An example would be if an airline announces a discount on their Facebook page. This would be an example of a rational appeal based on price. Both emotional and rational appeals could be used in the same post.
Ashley and Tuten (2015) found that brands use multiple appeals throughout their social media messages, whereas in more traditional marketing messages and campaigns, brands are more likely to focus on one unique selling point. Swani, Brown, and Milne (2014) compared the difference in use of appeals between to-business (B2B) and business-to-consumer (B2C) posts on Twitter. B2B tweets were found to include a higher percentage of rational appeals, whereas B2C tweets included a higher percentage of emotional appeals. That
8
brands’ social media messages targeting consumers contain a higher percentage of emotional appeals was also confirmed by research by Pinto and Yagnik (2016). Their content analysis of Facebook pages from brands selling fitness watches showed that emotional appeals were used far more often than rational appeals, with 43 percent of posts containing emotional appeals versus 23 percent containing rational appeals. Emotional appeals identified included fun/humor, fear, action/adventure, guilt, status, happiness/love and motivation. Rational appeals included product superior quality, price/value, performance, reliability and features/characteristics (Pinto & Yagnik, 2016).
So far, different format characteristics, as well as the use of appeals in brands’ social media messages were discussed. Because images and videos can be assumed to carry a bigger emotional load, just as emotional appeals, it might be that the use of more entertaining format characteristics such as images and videos, tends to go together with the use of more emotional appeals in brands’ social media posts. This relationship has not been studied before, which is why the following hypothesis is proposed for the present study of European airlines’ social media posts:
Hypothesis 1: There is a positive relationship between format characteristics and the use of appeals in European airlines’ posts, such that when engaging formats are used (such as questions, photos and videos), the posts are more likely to include emotional appeals than rational appeals.
Engagement
One of the main advantages of using social media is that it provides brands with options to engage consumers. Consumer engagement is a widespread concept in modern advertising. Consumer engagement is an important driver for consumers’ involvement with advertising
9
messages and is used to measure advertising effectiveness (Wang, 2006). Research on consumer engagement in the past has focused on engagement as consumers having an active role, either emotionally or cognitively engaging or active behavior towards the brand
(Kuvykaité & Taruté, 2015). Before social media emerged, brands tried to engage their consumers through their advertisements in mass media. For TV advertising, engagement can be described as a moment in which the viewer is psychologically immersed in a TV
advertisement (Kim, Ahn, Kwon, & Reid, 2017). This means that a viewer feels like he or she is experiencing what is happening in the advertisement. Literature now defines consumer engagement as an interactive process, based on the context and psychological state of the consumer, and consists of cognitive, emotional and behavioral dimensions (Brodie, Ilic, Juric, & Hollebeek, 2013; Dessart, Veloutsou, & Morgan-Thomas, 2015; Kuvykaité & Taruté, 2015).
Stimulating the behavioral dimension of consumer engagement has become easier by using interactive media like social media platforms. Common aspects stimulating engagement are the use of emotional or rational appeals, photos and videos, and the invitation to interact (Azar, Machado, Vacas-de-Carvalho, & Mendes, 2016). By using their own Facebook pages, brands take the initial responsibility and action to start a dialogue with the customers. Brand managers play an important role in evoking consumer engagement, because they have the tools to create opportunity for engagement on the platform (France, Merrilees, & Miller, 2016).
Consumer engagement on social media has several benefits for the brand. Research shows that consumer engagement on social media is positively related with the perceived reputation of a brand (Dijkmans, Kerkhof, & Beukeboom, 2014) and increases brand trust (Habibi, Laroche, & Richard, 2014). Furthermore, consumer engagement can impact brand value and loyalty (France et al., 2016). This is also confirmed by Gummerus, Liljander,
10
Weman and Pihlström (2012), who found that online consumer engagement has a positive influence on relationship benefits. Specifically, liking and commenting on Facebook posts led to more economic, social and entertainment benefits for the consumer, which then led to higher satisfaction and loyalty to the brand. Entertainment and social benefits seemed to be the most important drivers. These benefits can possibly be provided by using emotional appeals in messages, and by providing entertaining content such as photos and videos.
Companies often use videos and incentives such as coupons or prize competitions to get consumers to interact with the post (Parsons, 2013). Some of these format characteristics have proven to help create more consumer engagement. A study examining restaurants’ Facebook pages for instance, found that posts containing photos or text received more likes and comments than posts containing links or videos (Kwok & Yu, 2013). Though there is no satisfactory explanation for why photos evoked more engagement in this study than videos. An analysis of Australian health organizations’ Facebook posts revealed the opposite, stating that video posts were the most engaging, whereas posts with links or text only were the least engaging (Kite, Foley, Grunseit, & Freeman, 2016).
Since there is no consensus so far on which formats of social media posts actually are related to more consumer-brand engagement on social media, this study proposes the
following hypothesis:
Hypothesis 2: There is a positive relationship between the use of engaging formats (such as questions, photos and videos) in European airlines’ posts on Facebook and consumer engagement.
Pinto and Yagnik (2016) measured indicators of consumer engagement (comments, shares, and hashtags) but did not analyze whether there was a relationship between appeals
11
used in brands’ social media posts and engagement. Various other studies however have shown effects of the use of different appeals on consumer engagement on social media. Jahn and Kunz (2012) proposed that content on social media contains both hedonic and functional values. They found that both hedonic values and functional values led to more intense usage of the brands’ social media fan pages, with intense usage leading to more brand engagement. Ashley and Tuten (2015) found that rational appeals did not lead to better consumer
engagement with social media messages, and this may suggest that people are more motivated by emotional driven appeals to engage with the messages than rational appeals. Kwok and Yu (2013) found that conversational messages received more comments than sales messages. It might be possible that conversational messages contain more emotional appeals, whereas sales messages contain more rational appeals.
Emotional appeals can increase involvement and engagement with a marketing
message, particularly in older markets, as shown by Chandy, Tellis, MacInnis and Thaivanich (2001). In younger markets rational appeals may be more effective, because of the consumer’s need for information. The aviation and travel market in which airlines operate is an older market, which means that based on the previous rationale, emotional appeals are more likely to evoke consumer engagement. The use of emotional appeals in advertising is also more effective for experience services, such as dining at a restaurant or seeing a movie, whereas advertising with rational appeals is more effective for credence services, such as medical treatments or education (Zhang, Sun, Liu, & Knight, 2014). Since flights are considered experience services, it can be expected that emotional appeals in messages will be more effective in creating engagement for airlines than rational appeals:
Hypothesis 3: There is a positive relationship between the use of emotional appeals over rational appeals in European airlines’ posts and consumer engagement.
12
Methods
Sample
Since it was not possible to look at the Facebook posts from all European airlines, a sample of the population was selected. This study examined the Facebook posts from four so-called ‘flag carrier’ airlines. Flag carriers are national airlines that were once fully owned by the state. Initially four Western European flag carriers were selected: Air France, British Airways, KLM and Lufthansa. Upon examining their Facebook pages, however it was noticed that Air France posted a lot less than the three other airlines (only 28 posts total in 2016, compared to for instance KLM with 474 posts). Therefore, the decision was made to replace Air France by Iberia, the Spanish flag carrier airline. In the end, the sample therefore consisted of the
following airlines: British Airways from the United Kingdom, Iberia from Spain, KLM from The Netherlands and Lufthansa from Germany. All four selected airlines are among the biggest airlines of Europe (Pariona, 2016). This study only examined the main Facebook pages of the airlines. This means that for airlines with multiple Facebook pages only the main English Facebook page was used. Other pages that are specific to a certain region (for
example KLM Switzerland) or market (Lufthansa Cargo) were not used in this study. Table 1 shows the total amount of page likes for each airline’s Facebook page on the 10th
of
December 2017, the date the Facebook page was created, the language of the main Facebook page and the number of passengers the airlines carried in 2016.
13
Table 1: Overview of Facebook page likes on December 10, 2017, page creation date, page
language and passengers carried per airline
Airline Page likes Creation date Main page language Passengers carried in 2016 (million) British Airways Iberia KLM Lufthansa 3,029,623 1,747,103 12,236,498 3,401,292 July 2009 June 2010 March 2010 December 2009 English Spanish/English English English 44,51 21,82 30,43 109,74
Table 1 shows that there was quite some difference in the number of page likes between the four airlines, therefore this variable was controlled for in the eventual analysis.5
A sample of 100 Facebook posts from each airline in 2016 was selected for analysis. A full year was selected to include various sorts of posts airlines could have posted throughout the year, for example discounts for the summer season or festive wishes for the holidays. Examining posts over a year also decreased the influence of any major events that may have happened, like the terrorist attacks in Brussels in March 2016.
All Facebook posts from each airline from January 1st, to the 31st of December 2016 were collected by using Microsoft Power BI Desktop. A list of all posts in 2016 was
generated, within which all posts that were in any other language than English were deleted. In addition, also posts that solely consisted of updates like “KLM updated their profile
1 British Airways Plc (2017) 2 IB Opco Holding (2017) 3 KLM (2017) 4 Lufthansa Group (2017)
5 Unfortunately, it is not possible to view the amount of page likes on a certain date, thus the current total will be
used. It is not expected that one of the airlines’ total page like count increased largely compared to the other airlines, so the current numbers should be suitable to use as a control variable.
14
picture” or “Lufthansa updated their cover photo” without any other message from the airline were removed. Table 2 shows how many posts each airline had in 2016, and how many were left after removal of previously mentioned posts. Finally, by deleting alternate rows, a sample of 100 posts throughout the year was selected for each airline.
Table 2: Overview of total posts in 2016 for each airline, posts after removal, and final
sample
Airline Total posts Posts after removal Final sample
British Airways 88 88 88
Iberia 221 186 100
KLM 474 384 100
Lufthansa 254 235 100
As can be seen in Table 2, the British Airways Facebook page did not have a total of 100 posts in 2016. It was nevertheless decided to keep the airline in the sample, because any replacing airline would be less similar to the other airlines. None of the British Airways posts had to be removed, so the total of 88 posts was included in the sample.
All posts were saved before the start of the coding process, because Facebook is a live medium. This means that the information can be changed at all times. Facebook posts can be edited, and numbers of likes, shares and comments can change long after they have been posted. For this study, screenshots of all posts were taken and saved under unique numbers.
Research design
In order to test the hypotheses, a quantitative content analysis was conducted. The main asset of a content analysis is that the research materials (here: the Facebook posts) were not
15
purposefully created for research and so the research context had no influence on the Facebook posts (Wester, 2006).
The codebook can be found in Appendix A and contains 15 coding elements per post. The reliability of the coded data was tested through an assessment of intercoder-reliability. The author coded all posts and a second coder analyzed 40 Facebook posts, ten for each airline, which was 10.3 percent of the total sample. The percent agreement, Kappa and
Krippendorff’s Alpha (Hayes & Krippendorff, 2007), were calculated to assess the intercoder-reliability of each variable. The exact results can be found in Appendix B. In conclusion, all variables were reliably coded. Results varied from a minimum agreement of .66 to a
maximum of 1.
Observed variables
The first concept that was coded for each post was format characteristics. This concept was used to describe what the characteristics of the Facebook posts were. The codebook asked whether a post included a) text, b) an external link to a different webpage, c) a picture like an image, photo or .gif image, d) a video, e) emoticons like ‘emoji’ – small images that represent an emotion or other thing, f) a question to the reader, or g) if the post asked the reader to perform a certain social media related action, such as liking, sharing or commenting on the post. Each coding option was coded by stating ‘no’ or ’yes’. No means that the post did not include that specific characteristic, whereas ‘yes’ means it did include said characteristic.
The second coding concept concerned the appeals used in the posts. The appeals represented in the codebook were based on the appeals often used in advertising (Belch & Belch, 2015) and used in recent similar research (Pinto & Yagnik, 2016). The list of appeals was then adjusted to contain only those appeals that were perceived to be relevant for a sample of airline Facebook posts. For example, rational appeals like ‘durability’ or ‘product
16
features’ were not included in the codebook, as they apply more to specific products and not so much to services like flights.
The first coding variable was rational appeals, about informational or functional aspects of the brand. The rational appeals included in the analysis were: price,
quality/performance, convenience, comfort, safety and sustainability. Price included statements about prices and discounts, also if the actual price was not mentioned.
Quality/performance included statements about the quality of the airline’s services, which also included mentions of new products, new airplanes or new destinations. Convenience was about easiness of booking flights or checking in luggage. Comfort included statements about how comfortable flying with said airline is. Safety included any statements about safety of services, but also included mentions of security. Lastly, sustainability described mentions of sustainable development or consumption, like descriptions of being environmentally friendly.
The second variable was emotional appeals, related to the consumer’s social or psychological state. Emotional appeals included in the analysis were: joy, humor,
love/affection, sadness, fear, anger, adventure, and nostalgia. Joy related to expressions of happiness, excitement or celebration. Humor included jokes or mentions of funny situations. Love/affection included any depictions of love, for the airline, but also for family and friends. Sadness included any depictions of sad situations, crying or farewells. Fear included
statements about being afraid of something such as embarking on a flight. Anger included any statements of anger, outrage or aggression. Adventure included any mentions of going on an adventure or exploring new destinations. Finally, nostalgia included any feeling of sentiment for the past, which included longing to places with good memories or old photos throwing the viewer back to old times. All coding instructions and full descriptions can be found in the codebook in Appendix A.
17
The last and dependent variable was consumer engagement. In order to examine how much consumers engaged with the airlines’ Facebook posts, three variables were included: likes, shares and comments. This included the total number of likes, the number of shares, and the number of comments on each post. Any number from zero was possible. In the present sample, the actual scores ranged from a minimum of 11 to a maximum of 697,000 likes, 0 to 554,819 shares, and 0 to 35,000 comments.
In order to rule out the influence of some possible other factors, some control variables were included in the analysis. These included the airline for their total number of page likes, to rule out differences caused by the popularity of the airlines on Facebook. Also included was the month in which a post was placed, to rule out variation due to consumers being more active on Facebook during certain periods in a year.
Results
Descriptive findings
Before testing the hypotheses, the statistics of the variables were inspected. Table 3 shows how often the different format characteristics were included in the examined airlines’ Facebook posts, and Table 4 shows how often rational appeals and emotional appeals were present in the Facebook posts. All format characteristics were scored as ‘not present’ (0) or ‘present’ (1), therefore the maximum score for each variable was 1.
Table 3: Frequencies of format characteristics
Variable Frequency Percent Mean
Text 388 100.0 1.00 Link Image 168 270 43.3 69.6 .43 .70
18 Video Emoticon Question 110 97 88 28.4 25.0 22.7 .28 .25 .23 Action 14 3.6 .04
Table 4: Frequencies of rational appeals and emotional appeals
Variable Frequency Percent Mean
Price 18 4.6 .05 Quality Convenience Comfort Safety Sustainability Joy Humor Love Sadness Fear Anger Adventure Nostalgia 53 19 1 11 4 72 42 26 4 2 0 86 32 13.7 4.9 .3 2.8 1.0 18.6 10.8 6.7 1.0 .5 .0 22.2 8.2 .14 .05 .00 .03 .01 .19 .11 .07 .01 .01 .00 .22 .08
As can be seen in Table 3, all posts included text, and there were no Facebook posts that did not include any text at all. It can further be seen that, besides text, the airlines’ Facebook posts most often included images, followed by links. As shown in Table 4, appeals regarding
19
adventure and joy were employed most often, while ‘anger’ appeals were never included in any Facebook post. Because both did not show any variation, the variables ‘text’ and ‘anger’ were excluded from the analyses.
Table 5 shows the descriptive statistics of the dependent variables ‘likes’, ‘shares’ and ‘comments’. The table shows that each post received more than 6,700 likes, more than 2,400 shares, and more than 440 comments on average.
Table 5: Descriptive statistics of likes, shares and comments
Variable Mean Standard
Deviation Minimum Maximum Likes 6709.34 37294.59 11.00 697000.00 Shares Comments 2451.40 447.88 31179.68 2902.557 .00 .00 554819.00 35000.00 Hypothesis testing
In order to test the hypotheses, some new variables were computed. This included the creation of variables for engaging format characteristics, rational appeals, emotional appeals and user engagement respectively.
The variable ‘engaging format characteristics' was computed by adding up the variables ‘image’, ‘video’, ‘question’, and ‘action’. The new variable ‘engaging formats’ indicates how many engaging format characteristics a Facebook post had. The variables ‘rational appeals’ and ‘emotional appeals’ were created to indicate whether a Facebook post included one or more rational appeals (0 = no, 1 = yes) or emotional appeals (0 = no, 1 = yes).
20
In addition, to test the first hypothesis, another new variable for appeals was created, which indicates whether the post only included rational appeals (0) or included emotional appeals (1) and is called ‘appeals’.
Lastly, a new variable for ‘user engagement’ was created by adding up the amounts of likes, shares and comments the Facebook posts had to a total number of interactions. Table 6 shows the descriptive statistics of the new variables.
Table 6: Descriptive statistics of engaging formats, rational appeals, emotional appeals,
appeals and user engagement.
Variable Mean Standard
Deviation Minimum Maximum Engaging formats 1.24 .50 .00 3.00 Rational appeals Emotional appeals Appeals User engagement .24 .58 .79 9608.63 .43 .49 .41 68713.45 .00 .00 .00 23.00 1.00 1.00 1.00 1278819.00
Format characteristics and appeals
Hypothesis 1 predicted a positive relationship between format characteristics and the use of appeals in European airlines’ posts, such that when more engaging formats are used, the posts are more likely to include emotional appeals than rational appeals. A multiple regression analysis was performed with ‘engaging formats’ as the independent variable and ‘appeals’ as the dependent variable, while controlling for airline. Posts that did not include any appeals were excluded from the analysis, which means the analysis was performed with a sample of 287 posts. Table 7 shows the regression model of engaging format characteristics on appeals.
21
Table 7: Multiple regression analysis of format characteristics on appeals, controlled for
airline b* t p Constant 8.129 .000* Engaging formats KLM Iberia Lufthansa .139 -.027 -.173 .034 2.328 -.348 -2.168 .437 .021* .728 .031* .662 R2 = .064; R2adj = .050; F = 4.791; p = .001 *p < .05
The regression model shown in Table 7 was significant (F = 4.791, p = .001), although the proportion of variance explained was quite small (R2 = .064). The model showed a significant
relationship between engaging formats and the kind of appeals the Facebook posts included (b* = .139, t = 2.328, p < .05). This indicated that posts that included engaging format
characteristics were more likely to include emotional appeals rather than rational appeals. The first hypothesis could be accepted based on these results.
Because the first model was not very strong, a second analysis was performed to reveal the relationships between individual format characteristics (link, image, video, emotion, question, or action) and appeals. This model is shown in Table 8.
Table 8: Multiple regression analysis of format characteristics on appeals
b* t p
Constant 4.536 .000*
22 Image Video Emoticon Question Action KLM Iberia .301 .198 .062 .115 .002 -.084 -.252 1.929 1.228 .675 1.967 .025 -.084 -.252 .050* .221 .500 .050* .980 .307 .028* Lufthansa -.082 -.082 .342 R2 = .116; R2adj = .087; F = 4.040; p = .000 *p < .05
The regression model shown in Table 8 was significant (F = 4.040, p < .001), and the proportion of the explained variance somewhat improved (R2 = .116) in comparison to the previous model. There was a significant positive relationship between image and appeals (b* = .301, t = 1.929, p < .05), and questions and appeals (b* = .115, t = 1.967, p = .05), which showed that posts that included images or questions were more likely to contain emotional appeals than rational appeals. In addition, there was a significant negative relationship between ‘link’ and appeals (b* = -.227, t = -3.932, p < .05), suggesting that posts containing links to other webpages were less likely to contain emotional appeals and were more likely to contain rational appeals. All in all, while the previous regression analysis suggested that posts that included engaging format characteristics were indeed more likely to contain emotional appeals over rational appeals, the present analysis showed that this did not hold for the use of videos (b* = .198, t = 1.228, p = .221). So, based on these results, the hypothesis could be partly, but not fully accepted.
23 Format characteristics and engagement
Hypothesis 2 predicted a positive relationship between the use of engaging formats in
European airlines’ posts on Facebook and consumer engagement. A multiple linear regression including engaging formats as the predictor and user engagement as the dependent variable was performed to see whether the prediction would hold. The regression model is shown in Table 9. All posts were included in the analysis (n = 388) and the analysis controlled for the airline and month of the posts.
Table 9: Multiple regression analysis of engaging format characteristics on user engagement,
controlled for airline and month of post
b* t p Constant -.266 .791 Engaging formats KLM Iberia Lufthansa December November October September August July June May April .009 .187 .022 .019 .170 -.020 .038 -.016 .009 -.003 .004 -.012 -.004 .160 2.890 .340 .293 2.738 -.330 .588 -.264 .145 -.053 .057 -.190 -.069 .873 .004* .734 .769 .006* .742 .557 .792 .885 .958 .954 .849 .945
24
March -.007 -.110 .913
February -.012 -.194 .847
R2 = .060; R2adj = .022; F = 1.580; p = .077
*p < .05
The regression model in Table 9 was not significant (F = 1.580, p > .05) and showed no significant relationship between engaging format characteristics and user engagement (b* = .009, t = .160, p > 0.5). Therefore, hypothesis 2 could not be accepted.
Because the regression model was not a good fit, another attempt was done to get more insight in the relationship between the individual format characteristics and consumer engagement. Three regression analyses were performed, each with the multiple characteristics of airline’s Facebook posts as the predictors, and by using the separate indicators of consumer engagement: likes, shares and comments respectively. The full models can be found in
Appendix C. Table 10 shows an overview of the results of these analyses.
Table 10: Regression models of individual engaging format characteristics on likes, shares
and comments, controlled for airline and month of post
Likes Shares Comments
b* b* b* Constant Link -.110 -.070 -.051 Image .090 .040 .091 Video .196 .171 .120 Emoticon .021 .085 .047 Question -.038 -.019 .010
25 Action -.026 -.040 .246* R2 .097 .067 .112 R2adj .048 .016 .064 F 1.975 1.311 2.315 p .008* .168 .001* *p < .05
The first model showed the relationship between format characteristics and likes, which was significant (F = 1.975, p < .05), there were however no significant relationships between the various format characteristics of the posts and the number of likes of a post. The second model showed the regression model of the relationship between posts’ format characteristics and the number of shares. Similar to the previous model, there were no significant predictors of the relationship between the format characteristics and the number of shares of a post, and the regression model was not significant either (F = 1.311, p > .05). The third model showed the regression model of the relationship between posts’ format characteristics and number of comments. The model was significant (F = 2.315, p < .05) and the explained portion of variance was 11 percent (R2 = .112). The most notable relationship was between the presence
of ‘action’ in a post and the number of comments a post received (b* = .246. t = 4.505, p < .05). Specifically, the model showed that posts that ask the reader to perform a certain social media-related action (such as liking, sharing or commenting), were more likely to receive more comments. However, this was the only significant prediction, which means that the second hypothesis could not be supported based on the findings: the presence of engaging format characteristics in European airlines’ Facebook post does not predict signals of consumer engagement among users of the Facebook pages.
26 Appeals and engagement
Hypothesis 3 predicted a positive relationship between the use of emotional appeals over rational appeals in European airlines’ posts and consumer engagement. To test the third hypothesis rational and emotional appeals were added to the previous regression model. Table 11 shows the regression model with engaging format characteristics and rational and
emotional appeals as predictors and user engagement as the dependent variable, controlling for airline and the month the post was placed.
Table 11: Multiple regression analysis of engaging format characteristics, rational appeals
and emotional appeals on user engagement, controlled for airline and month of post
b* t p Constant -.353 .725 Engaging formats Rational appeals Emotional appeals KLM Iberia Lufthansa December November October September August July June .006 -.036 .050 .189 .030 .015 .162 -.029 .030 -.027 .004 -.005 .005 .116 -.647 .912 2.885 .441 .229 2.578 -.478 .453 -.448 .067 -.082 .084 .908 .518 .362 .004* .660 .819 .010* .633 .651 .654 .947 .935 .933
27 May April March -.011 -.005 -.007 -.188 -.077 -.120 .851 .939 .905 February -.013 -.013 .833 R2 = .064; R2adj = .021; F = 1.494; p = .094 *p < .05
The regression model was not significant (F = 1.494, p > .05), and showed no significant relationship between rational appeals (b* = -.036, t = -.647, p > .05) or emotional appeals (b* = .050, t = .912, p > .05) and user engagement. This means that hypothesis 3 could not be supported by this model.
To get more insight into the underlying relationships, three more analyses were performed with all individual indicators of format characteristics, rational appeals and emotional appeals as predictors, and likes, shares and comments as dependent variables. The full models can again be found in Appendix C. Table 12 shows an overview of the most relevant results.
Table 12: Regression models of individual engaging format characteristics, rational appeals
and emotional appeals on likes, shares and comments, controlled for airline and month of
post
Likes Shares Comments
b* b* b*
Constant
Price .030 .020 .145*
28 Convenience -.046 -.032 -.045 Comfort -.046 -.035 -.017 Safety -.018 -.005 -.015 Sustainability -.006 .004 .058 Joy .107 .104 -.001 Humor .009 .007 -.025 Love -.016 -.045 .085 Sadness .003 .001 -.010 Fear -.060 -.049 -.030 Adventure -.007 -.011 .037 Nostalgia .018 -.011 .077 R2 .117 .085 .151 R2adj .035 -.001 .072 F 1.421 .992 1.902 p .067 .484 .003* *p < .05
The first regression model in Table 12 was not quite significant (F = 1.421, p > .05) and showed no significant relationships. The same was the case for the second regression model, showing the relationships between rational and emotional appeals and shares. The model showed no significant predictors and the model itself was not a good fit (F = .992, p > .05). Whereas the models with the numbers of likes and shares as dependent variables were not significant, the regression model with comments as dependent variable was significant (F = 1.902, p < .05) and explains 15 percent of the proportion of variance (R2 = .151). Adding the appeals somewhat decreased the significance level, but slightly increased the explained
29
proportion of variance in comparison with the regression model shown in the ‘comments’ column in Table 10.
The only predictor that showed a significant relationship with comments was price (b* = .145, t = 2.610, p < .05), meaning that posts that included statements about the price or discounts were more likely to obtain comments as opposed to posts that did not include this appeal. However, since price is a rational appeal, this relationship contradicts the prediction in Hypothesis 3, namely that there is a positive relationship between the use of emotional
appeals over rational appeals in European airlines’ posts and consumer engagement. Thus, the third hypothesis could not be accepted.
Conclusion & Discussion
Conclusion
This study aimed to provide more insight in the relationships between format characteristics and appeals used in the Facebook posts of European airlines, and signals of consumer engagement. While results showed that there was a positive relationship between the use of engaging format characteristics and the use of emotional appeals in the airlines’ posts, no significant relationships between the use of format characteristics or emotional appeals and consumer engagement were found.
Some specific engaging format characteristics such as the use of images and questions had a positive relationship with the use of emotional appeals in the European airlines’
Facebook posts, which was predicted in the first hypothesis. Posts that included images or a question were more likely to include emotional appeals rather than rational appeals.
The present study could not confirm the second hypothesis, predicting a positive relationship between the use of engaging format characteristics in the airlines’ Facebook posts and consumer engagement. There were no significant relationships between various format
30
characteristics and the number of likes, shares and comments the posts received. While European airlines did use various formats to interact with consumers, their Facebook pages did not show that posts that contained various engaging format such as photos, received more likes than posts that did not. This finding contradicted earlier research by Kwok and Yu (2013), who found that restaurants’ Facebook posts containing photos received more likes than posts containing links or videos. The only notable finding related to this hypothesis was that posts that asked the consumer to perform a social media related action, received more comments than posts that did not ask this.
The third hypothesis predicting that European airlines’ Facebook posts including emotional appeals would receive more signals of consumer engagement than posts including rational appeals could not be confirmed either. While most relationships were in the expected direction, they were not strong enough to conclude such a relationship. Further, the study contradicted the prediction based on research by Ashley and Tuten (2015) who found that rational appeals did not lead to more consumer engagement on social media, while the current study found that Facebook posts that included the rational appeal ‘price’ received more comments than posts that included emotional appeals or no appeals at all. The results also contradicted the prediction that emotional appeals are more effective for the aviation industry on social media, based on earlier research by Chandy et al. (2001) and Zhang et al. (2014).
In conclusion, the results could not confirm a significant relationship between engaging format characteristics or emotional appeals in European airlines’ Facebook posts and consumer engagement in the form of likes, comments or shares.
Discussion
With increasing marketing competition, consumer engagement is critical to the airline industry and has taken a central spot in airlines’ communication strategies. Social media
31
provide an opportunity to interact and engage with consumers. Many organizations have taken note of this and specifically airlines were among the first organizations to fully integrate social media into their communication strategies (Mago, 2017). However, social media are rapidly evolving and changing, which causes research on this topic to quickly become outdated. That is why continuing research on organizations’ social media strategies and their relationship with consumer engagement is so important.
The present study attempted to provide more insight in this relationship for the European airline industry. Like many other studies into this topic, this study focused on one specific type of organization. This study however was among the first to research the direct relationship between the use of appeals in social media posts and consumer engagement. The focus on appeals in relation to consumer engagement is what makes this research interesting for other experience service organizations as well. Most likely, other airlines or other
experience service organizations will use similar types of appeals in their social media posts. Organizations that sell physical products or search goods might, on the other hand, use other appeals to market their products on social media and the results of the present study might not hold for those brands. The results regarding engaging format characteristics however, may be of relevance for anyone using social media as a strategy to gain consumer engagement, since those are not specific to any type of good or market.
Since the results of this study are not in line with previous studies, more research is needed to come to a better insight in the relationship between organizations’ Facebook posts such as those of airlines, and consumer engagement on social media. While the present study tried to accurately examine the relationships between format characteristics, appeals and consumer engagement, some limitations can be identified.
First, this study applied a manual quantitative content analysis to gauge the number of likes, comments and shares each post had received. However, when the number of likes of a
32
post increases, Facebook starts to round up the numbers that are shown and posts that received over 1,000 likes do not show the exact numbers anymore. For example, if a post received over 7,300 likes, the Facebook page states it has “7.3K” likes instead of showing the exact number. The same goes for high numbers of comments, although the issue did not present itself for number of shares, which were always shown in exact numbers. Because this study utilized the rounded numbers, some nuances might be lost in the results This issue can possibly be fixed in future research by using scraping software to register the exact number of likes and comments of the posts.
The second issue concerns the comments on the posts. This study solely looked at the number of comments the posts had received and related this to characteristics of the posts. However, it was not taken into account whether the comments were actual responses to the post or included a different topic. On observation, a lot of consumers seem to post comments to ask questions or complain to the airline. These comments were included, even though they might be related to the content of the airline’s post. Since this happens for all posts, it might not influence the relationship with certain appeals used in the posts. Nevertheless, future research could analyze the comments in more detail, to see whether a comment is an actual response to the airline’s post or not and control for the influence of these comments.
The last issue to be discussed concerns the sample of the appeals. As can be seen in the results section, some of the analyzed appeals were only present in a small amount of posts (< 5%). In addition, only 287 out 388 posts (73.9%) included any of the coded appeals. Some posts possibly included appeals that were not analyzed in this study. To contribute to the knowledge on the relationship between appeals and consumer engagement for European airlines’ Facebook posts, future research could re-evaluate which rational and emotional appeals are relevant in these Facebook posts.
33
In relation to consumer engagement the question arises how involved consumers really are with airlines on social media. Consumers that ‘like’ a Facebook posts might not be as engaged as one would like to think, since liking a post is as easy as one click. Leaving
comments however are stronger signals of involvement. It takes effort to type a comment and consumers need to be rather involved with the content of the post to leave a response. Sharing a post means the organization’s post now also shows up on the consumers Facebook page, which means the consumer is now actively showing their involvement with the brand.
Because this study could not confirm all hypotheses, it is not possible to advice organizations to use certain post characteristics or appeals in order to gain more consumer engagement based on the present findings. Yet, posts that ask the reader a to perform a certain action, are likely to receive more comments. This might be relevant when an organization seeks to increase level of interaction with their consumers. Further research is needed to see whether this actually leads to more meaningful interaction with consumers as well.
References
Ashley, C., & Tuten, T. (2014). Creative strategies in social media marketing: An exploratory study of branded social content and consumer engagement. Psychology & Marketing,
32, 15-27. doi:10.1002/mar.20761
Azar, S. L., Machado, J. C., Vacas-de-Carvalho, L., & Mendes, A. (2016). Motivations to interact with brands of Facebook: Towards a typology of consumer-brand interactions.
Journal of Brand Management, 23, 153-178. doi:10.1057/bm.2016.3
Belch, G. E., & Belch, M. A. (2015). Advertising and Promotion. An Integrated Marketing
Communications Perspective. Singapore: McGraw-Hill Education.
34
brand community: An exploratory analysis. Journal of Business Research, 66, 105-114. doi:10.1016/j.jbusres.2011.07.029
Burt, C. D. B., & Strongman, K. (2005). Use of images in charity advertising: Improving donations and compliance rates. International Journal of Organisational Behaviour, 8, 571-580. Retrieved from: http://citeseerx.ist.psu.edu/viewdoc/summary?doi=
10.1.1.512.2664
British Airways Plc (2017). British Airways Report And Accounts Year Ended 31 December
2016. Retrieved from:
http://www.iagshares.com/phoenix.zhtml?c=240949&p=irol-reportsannual
Chandy, R. K., Tellis, G., MacInnis, D., & Thaivanich, P. (2001). What to say when: Advertising appeals in evolving markets. Journal of Marketing Research, 38, 399– 414. doi:10.1509/jmkr.38.4.399.18908
Dessart, L., Veloutsou, C., & Morgan-Thomas, A. (2015). Consumer engagement in online brand communities: A social media perspective. Journal of Product and Brand
Management, 24, 28-42. doi:10.1108/JPBM-06-2014-0635
Dijkmans, C., Kerkhof, P., & Beukeboom, C. J. (2015). A stage to engage: Social media use and corporate reputation. Tourism Management, 47, 58-67. doi:10.1016/
j.tourman.2014.09.005
DVJ Insights (2014). KLM Voor 3e Keer Op Rij Best Social Brand van Nederland. Retrieved
from: https://dvj-insights.nl/klm-voor-3e-keer-op-rij-best-social-brand-van-nederland France, C., Merrilees, B., & Miller, D. (2016). An integrated model of customer-brand
engagement: Drivers and consequences. Journal of Brand Management, 23, 119-136. doi: 10.1057/bm.2016.4
35
Facebook brand community. Management Research Review, 35(9), 857-877. doi:10.1108/01409171211256578
Habibi, M. R., Laroche, M., & Richard, M. O. (2014). The roles of brand community and community engagement in building brand trust on social media. Computers in Human
Behavior, 37, 152-161. doi:10.1016/j.chb.2014.04.016
Hayes, A. F., & Krippendorff, K. (2007). Answering the call for a standard reliability measure for coding data. Communication Methods and Measures, 1, 77-89. doi:10.1080/19312450709336664
Hickins, M. (2015). Southwest Airlines Adds Social To Its Customer Engagement Strategy. Retrieved from: https://www.forbes.com/sites/oracle/2015/09/21/southwest-airlines-adds-social-to-its-customer-engagement-strategy/#2cb6263a58ae
Hiraga, R., & Takahashi, K. (2011). Conveying emotion with moving images: Relationship between movement and emotion. In International Conference on Affective Computing
and Intelligent Interaction (pp. 558-567). Berlin, Heidelberg: Springer. Retrieved
from: https://link.springer.com/chapter/10.1007/978-3-642-24600-5_59
IB Opco Holding (2017). Consolidated Financial Statements And Management Report For
2016. Retrieved from:
http://www.iagshares.com/phoenix.zhtml?c=240949&p=irol-reportsannual
Jahn, B., & Kunz, W. (2012). How to transform consumers into fans of your brand. Journal of
Service Management, 23, 344–361. doi:10.1108/09564231211248444
Johar, J. S., & Sirgy, M. J. (1991). Value-expressive versus utilitarian advertising appeals: When and why to use which appeal. Journal of Advertising, 20, 23-33.
doi:10.1080/00913367.1991.10673345
36
of immersion and presence. Journal of Business Research, 76, 67-76. doi:10.1016/ j.jbusres.2017.03.001
Kite, J., Foley, B. C., Grunseit, A. C., & Freeman, B. (2016). Please like me: Facebook and public health communication. PLoS One, 11, e0162765. doi:10.1371/
journal.pone.0162765
KLM (2017). KLM Carried More Than 30 Million Passengers In 2016. Retrieved from: https://news.klm.com/klm-carried-more-than-30-million-passengers-in-2016/ Kuvykaité, R., & Taruté, A. (2015). A critical analysis of consumer engagement
dimensionality. Procedia – Social and Behavioral Sciences, 213, 654-658. doi:10.1016/j.sbspro.2015.11.468
Kwok, L., & Yu, B. (2013). Spreading social media messages on Facebook: An analysis of restaurant business-to-consumer communications. Cornell Hospitality Quarterly, 54, 84-94. doi:10.1177/193896551245836
Luchtvaarnieuws.nl (2015). KLM Drie Keer In Prijzen Bij Social Media Awards. Retrieved from: https://www.luchtvaartnieuws.nl/nieuws/categorie/2/airlines/ klm-drie-keer-in-prijzen-bij-social-media-awards
Lufthansa Group (2017). Annual Report 2016. Retrieved from: https://investor- relations.lufthansagroup.com/en/finanzberichte/annual-report/2016.html
Mago, V. (2017). Three Customer Engagement Lessons From The Airline Industry. Retrieved from: http://customerthink.com/three-customer-engagement-lessons-from-the-airline-industry/
Marketingfacts (2014). KLM En Trijntje Oosterhuis Winnen De Beste Social Media Awards. Retrieved from: https://www.marketingfacts.nl/berichten/klm-en-trijntje-oosterhuis-winnen-de-beste-social-media-awards
37
http://www.worldatlas.com/articles/which-are-the-largest-airlines-in-europe.html Parsons, A. (2013). Using social media to reach consumers: A content analysis of official
Facebook pages. Academy of Marketing Studies Journal, 17(2), 27–36. Retrieved from: https://www.abacademies.org/articles/amsjvol17no2-2013.pdf#page=35
Pinto, M. B., & Yagnik, A. (2016). Fit for life: A content analysis of fitness tracker brands use of Facebook in social media marketing. Journal of Brand Management, 24, 49-67. doi:10.1057/s41262-016-0014-4
Shen, B., & Bissell, K. (2013). Social media, social me: A content analysis of beauty companies’ use of Facebook in marketing and branding. Journal of Promotion
Management, 19, 629-651. doi:10.1080/10496491.2013.829160
Swani, K., Brown, B. P., & Milnec, G. R. (2014). Should tweets differ for B2B and B2C?: An analysis of Fortune 500 companies' Twitter communications. Industrial Marketing
Management, 43, 873-881. doi:10.1016/j.indmarman.2014.04.012
Wang, A. (2006). Advertising engagement: A driver of message involvement on message effects. Journal of Advertising Research, 46, 355-368. doi:10.2501/
S0021849906060429
Wester, F. (2006). Inhoudsanalyse als onderzoeksontwerp. In F. Wester (Ed.)
Inhoudsanalyse: Theorie en Praktijk (pp. 11-44). Nederland: Kluwer.
Zhang, H., Jin S., Fang L., & Knight, J. G. (2014). Be rational or be emotional: Advertising appeals, service types, and consumer responses. European Journal of Marketing, 48, 2105-2126. doi:10.1108/EJM-10-2012-0613
38
Appendix A
Codebook
Use the code sheet to code the posts. Fill in when asked for answers or check the appropriate box(es) on the code sheet. This codebook will give a short explanation for each element that needs to be coded.
1. Number of post
The number of the Facebook post, starting with 1 for the very first post coded, up to 400 for the last post coded.
Descriptives 2. Airline
The abbreviation of the airline which the post belongs to: BA = British Airways, IB = Iberia, KL = KLM, LU = Lufthansa
3. Date
The date the Facebook post is published, written dd-mm-yy.
Form characteristics 4. Text
Does the post include any text? Code only text that is in the actual post, do not consider text that is included in a link, picture, or video. Choose NO or YES.
39
5. External link
Does the post include an external link? This includes the post having a clickable link that leads to an external URL. For example, a link to the airline’s website. Choose NO or YES.
6. Image
Does the post include an image like a photo or picture? Do not include emoticons or ‘emoji’. GIF images are considered to be an image. Choose NO or YES.
7. Video
Does the post include a video? This includes any video or animation but does not include GIF images. Choose NO or YES.
8. Emoticons
Are emoticons included in the text? This includes ‘emoji’- small images that express emotions or resemble other things. Choose NO or YES.
9. Question
Does the text in the post pose a question to the consumer? Only answer YES for questions that are directed to the reader: questions that can be responded to by answering in the comments or by ‘liking’ the post. Do not include rhetorical questions: questions the brand does not actually expect an answer to, for example “Nice day, isn’t it?” is answered with NO, “What day of the week is your favorite day to fly?” is answered with YES. Choose NO or YES.
40
10. Action
Does the post ask the reader to perform a certain action? Actions like ‘liking’ the post or page, sharing the post, or commenting on the post. Only consider Facebook-related actions,
statements like “go to our website to book your flight” do not count. This can be in
combination with a question. For example, if a post says: “What will be you next destination? Comment your answer below!” both 9. Question and 10. Action will be answered with YES. If a post does not ask a specific question but urges readers to ‘like’ the post for example to get a chance to win a gift, only question 10. Action will be answered with YES. Choose NO or YES.
Appeals
For these variables check the boxes on the code sheet of the appeals that are included in the post. Take into consideration the complete post, this includes the text but also images and videos if included. If there is an external link, only consider the information already available in the post, do not click the link. If a post contains text in multiple languages, only take English text into account. Multiple appeals, both rational and emotional, can be present. Only check the boxes for those appeals that are clearly present in the post, when in doubt do not check the box for that appeal.
11. Rational appeals
Price: any statements about the price of a product or service, including statements about discount. For example, mentions of a certain percentage off the original price or for statements that the product has an attractive price (also if actual price is not
41
Quality/performance: Any statements about the quality or performance of the product or service. Statements about the quality of the brand, for example when a brand states they are the ‘best’. Also includes statements about flying to new destinations for example.
Convenience: any statements about the convenience of the product of service. Statements about easiness, for example how easy it is to book a flight or convenience to take luggage on a flight.
Comfort: any statements about comfort, for example about how comfortable flying with said airline is. Also include statements about the comfort of airport lounges etc. Safety: any statements about safety of the product or service. For example, when there
is a statement that proves the actual safeness of the product or service or when extra safety measures are mentioned. For example, statements about airport security. Sustainability: any statements about sustainable development or consumption. For
example, statements about being environmentally friendly.
12. Emotional appeals
Joy: any statements that express joy, happiness or celebration. Also includes being excited about something.
Humor: any statements that include humor, like jokes or mentions of funny situations. Love/affection: any statements that express or include love or affection. Either
statements of the brand loving the customer, or stories about customers that are about love. For example, a statement about how a customer flew to a different country to meet their lover or mentions of Valentine’s Day or finally meeting a grandchild or old friend.
42
Sadness: any statements that express sadness. For example, statements about sad situations, expressing sadness that summer is over or other farewells, and stories about sad customers. Includes any mention of the act of crying.
Fear: any statements that express fear. Can be related to any topic: fear of flying, fear of losing something or being afraid about something.
Anger: any statements that express anger or outrage. Includes statements about the brand expressing anger as well as customers being angry. Also includes statements about aggression.
Adventure: any statements about being adventurous. For example, the brand mentioning adventure in combination with certain destinations. Includes questions about which destination you want to fly to or mentions of exploring a destination. Nostalgia: any statements about feeling sentimental for the past. Can be described as a
longing to a better time or place. For example, statements about longing to fly to a place with good memories attached. Also includes ‘throwbacks’ to other times.
Consumer engagement
For the next elements, write down the complete number. Do only use numbers, do not use dots, commas or other punctuation marks or letters. For example, write 2400 (not 2.400 or 2,4k).
13. Likes
Write down the amount of likes of the post in numbers
14. Shares
43
15. Comments
Write down the amount of comments on the post in numbers
Code sheet
Write down:
1. Number
2. Airline
3. Date
Check one option:
4. Text o NO o YES 5. External link o NO o YES 6. Image o NO o YES
44 7. Video o NO o YES 8. Emoticons o NO o YES 9. Question o NO o YES 10. Action o NO o YES
Check none, one or multiple options:
11. Rational appeals o Price o Quality/performance o Convenience o Comfort o Safety o Sustainability
45 Check none, one or multiple options:
12. Emotional appeals o Joy o Humor o Love/affection o Sadness o Fear o Anger o Adventure o Nostalgia
Write down in numbers:
13. Likes
14. Shares
46
Appendix B
Table 13: Overview of percent agreement, Kappa and Krippendorff’s Alpha for all variables
Variable Percent agreement Kappa Krippendorff’s
Alpha Airline 1 1 1 Date 1 1 1 Text 1 -* -* Link .88 .75 .75 Image .93 .85 .85 Video .95 .90 .90 Emoticon .98 .94 .94 Question .93 .68 .69 Action 1 1 1 Price 1 1 1 Quality .95 .72 .73 Convenience 1 1 1 Comfort 1 1 1 Safety 1 1 1 Sustainability 1 -* -* Joy .90 .75 .75 Humor 1 1 1 Love .92 .75 .75 Sadness 1 -* -* Fear 1 -* -* Anger 1 -* -*
47 Adventure .95 .87 .87 Nostalgia .98 .66 .66 Likes - .95 .92 Shares - .93 .94 Comments - .89 .90 Appeals User engagement Engaging formats Rational appeals Emotional appeals .90 - .90 .95 .90 .71 .87 .67 .86 .71 .72 .87 .67 .86 .72 * no variation Appendix C
Table 14: Multiple regression analysis of format characteristics on user engagement,
controlled for airline and month of post
b* t p Constant -.284 .776 Link Image Video Emoticon Question Action KLM Iberia -.110 .090 .196 .021 -.038 -.026 .228 .041 -1.900 .602 1.286 .253 -.716 -.464 3.354 .434 .058 .547 .199 .801 .474 .643 .001* .664
48 Lufthansa December November October September August July June May April March -.008 .151 -.042 -.002 -.038 -.009 -.039 -.006 -.016 -.015 -.012 -.105 2.419 -.683 -.035 -.635 -.147 -.636 -.102 -.268 -.251 -.188 .916 .016* .495 .972 .526 .883 .525 .919 .789 .802 .851 February -.024 -.386 .700 R2 = .097; R2adj = .048; F = 1.975; p = .008 *p < .05
Table 15: Multiple regression analysis of format characteristics on shares, controlled for
airline and month of post
b* t p Constant -.323 .747 Link Image Video Emoticon Question Action -.070 .040 .171 .085 -.019 -.040 -1.192 .265 1.106 1.026 -.363 -.711 .234 .791 .270 .306 .717 .478
49 KLM Iberia Lufthansa December November October September August July June May April March .156 .008 -.007 .126 -.043 .035 -.031 -.020 -.042 -.021 -.009 -.016 -.017 2.255 -.078 -.098 1.983 -.682 .531 -.541 -.303 -.669 -.331 -.144 -.264 -.264 .025* .938 .922 .048* .496 .596 .607 .762 .504 .741 .885 .792 .792 February -.024 -.383 .702 R2 = .067; R2adj = .016; F = 1.311; p = .168 *p < .05
Table 16: Multiple regression analysis of format characteristics on comments, controlled for
airline and month of post
b* t p Constant -.169 .866 Link Image Video Emoticon -.051 .091 .120 .047 -.892 .612 .795 .581 .373 .541 .427 .562
50 Question Action KLM Iberia Lufthansa December November October September August July June May April March .010 .246 .018 -.015 -.137 .079 -.056 .006 -.003 -.156 -.010 .036 -.015 .005 -.025 .198 4.505 .260 -.158 -1.902 1.278 -.916 .095 -.043 2.440 -.168 .597 -.251 .083 -.413 .843 .000* .795 .874 .058 .202 .360 .924 .966 .015* .867 .551 .802 .934 .680 February -.021 -.350 .727 R2 = .112; R2adj = .064; F = 2.315; p = .001 *p < .05
Table 17: Multiple regression analysis of format characteristics and appeals on likes,
controlled for airline and month of post
b* t p Constant -.225 .822 Link Image -.081 .050 -1.284 .320 .200 .749
51 Video Emoticon Question Action KLM Iberia Lufthansa December November October September August July June May April March February Price Quality Convenience Comfort Safety Sustainability Joy .158 -.013 -.039 -.036 .241 .054 .006 .160 -.040 -.001 -.050 -.012 -.024 .006 -.017 -.003 -.013 -.021 .030 -.024 -.046 -.046 -.018 -.006 .107 .999 -.153 -.711 -.603 3.348 .559 .080 2.404 -.632 -.011 -.806 -.186 -.375 .099 -.268 -.041 -.197 -.340 .538 -.422 -.846 -.894 -.336 -.108 1.888 .318 .879 .478 .547 .001* .577 .936 .017* .534 .991 .421 .853 .708 .921 .789 .967 .844 .734 .591 .673 .398 .372 .737 .914 .060
