Attitudes towards Brands and Advertisements: Qualitative and Thematic Analysis of Social Media Data

,


Introduction
Social media has been described as a platform for discussing ideas, communicating experiences and exchanging knowledge.It has changed the way individuals interact, providing massive amount of data and rich market insight as customers and brands engage and build relationships.This public declaration is of great concern for any organisation as it transfers the power to shape brand images from the hands of advertisers to the words of consumers' online connections.
This chapter sets an agenda proposing the possibilities of qualitatively analysing usergenerated content on social media platforms to provide insight into attitudes towards advertisements and their brands.Unlike participants being interviewed in a focus group, filling in questionnaires or neuroscience providing insight into how the mind perceives advertisements which typically requires expensive, bulky equipment and lab-type settings that limit and influence the experience, this is readily available public data which can be thematically analysed to add to existing knowledge.
Presenting the idea, publicly declared responses to the advertisements of UK banks on Facebook were analysed in order to gain insight into their perceptions and attitudes towards the advertisements and their brands.An outline of how to perform an analysis of user-generated content was provided to buttress the research method.Challenges and limitations of this research method were also considered.

Social Media and Brand Communication
The power of social media cannot be underestimated, revolutionising personal and organisational communications and interactions worldwide (Ngai et al., 2015).However, as far as the terminology is concerned, Kaplan and Haenlein (2009) noted that the term "social media" is commonly used interchangeably with social computing, social networking, Web 2.0, and virtual social worlds.Ngai et al. (2015) acknowledged the many definitions of these terms, but summarised 'dividing the compound term "social media" into its two components, "social" and "media."The "social" part refers to the activities carried out among people, whereas "media" refers to Internet-enabled tools and technologies used to carry out such activities' (p 771).Malita (2011), however, provided a more suitable definition, by considering social media as the 'tools that facilitate the socialization of content […] social media services encourage collaboration, interaction, and communication through discussion, feedback, voting, comments, and sharing of information from all interested parties' (p.748).This further supports the co-creating of brand meaning enhanced by the interaction on social media which this current chapter intents to explore.Mangold and Faulds (2009) considered social media an obvious choice as a marketing tool and is viewed as an integral part of the integrated marketing communication strategies of firms.Hence, social media can be considered as a new component in marketing promotional material.Ngai et al., (2015) also acknowledged that its use in marketing has increased rapidly in recent years and it has received increasing attention from academia, as well as various industries.Kaplan and Haenlein (2010) provided six classifications according to the level of social presence/media richness (low, medium and high) and self-presentation/self-disclosure (low and high).The six classifications include blogs, social networking sites (e.g.Facebook), virtual social worlds (e.g.Second Life), collaborative projects (e.g.Wikipedia), content communities (e.g.YouTube) and virtual game worlds (e.g.World of Warcraft).
This interaction between brand and social media users is the main focus of this study.Social networking sites (SNS) such as Facebook, Twitter, Google+ and content communities such as YouTube and Instagram which allows allow the creation and exchange of user-generated content are consider suitable for this research method.It is envisaged that not all social media users will be customers of the brand; some users will want to engage with the brand in anticipation of making a purchase.Irrespective of the intention, social media provides central location for online social engagement and a strategic means for building and maintaining a strong brand presence online (Wallace, Wilson, & Miloch, 2011).
As Tsai and Men (2013) noted, academic research struggles to keep pace with the rapid growth of social media network, the possibility of customers' comments on social media providing insights into their attitudes to the brand and advertisements has not been explored.This highlights a gap that this chapter hopes to fill.The study focuses on examining the possibility that self-reports from consumers (through their social media comments) can be used to explore the attitude to advertisement and brand construct.It draws attention to a social function that allows customers to interact with brands and for brands to have a better understanding of how their consumers engage with their advertisement and, importantly, the attitude-toward-the-ad theory which suggests that a positive attitude towards the advertisement, in turn, has a positive effect on purchase intention (PI) (Bruner & Kumar, 2000).
Analysing the User-Generated Qualitative Data This section provided an outline of how to perform an analysis of self-reported comments on social media to gain qualitative insight into the advertisements and the brands.To start with this idea, it is essential to identify the context and idea behind the analysis.Why would you want to carry out analysis on social media content?Perhaps a brand has just launched an advertisement and will like to know what people think of the advertisement, or a Brand Manager observed thumbs down (dislikes) on their YouTube video and wants to know why people disliked it.Analysing the comments provided an insight.Users on social media might find it easier to click a like or dislike button to show their attitude, but it takes effort, and a certain degree of interest to type constructive comments and such feedback should be cherished.This further highlights benefits in this user-generated data which suggests why it is imperative to take note of these comments.

Identifying the Content
The content for analysis could be a brand logo or a slogan but, for this study, UK bank television advertisements have been selected to illustrate the possibilities of analysing user-generated contents on social media to gain insight into attitudes towards the advertisements and their brands.Specifically, in the light of the global financial crisis and the negative attitude towards bank brands (Kottasz & Bennett, 2014), the study explores user-generated comments on UK bank advertisements (See Mogaji (2016) and Mogaji, Farinloye and Aririguzoh (2016) for a full study).

Selecting the Media
Bank advertisements can be distributed in different media -newspapers, billboards or social media to name but a few.However, for this study, Facebook was selected to illustrate the possibilities of analysing user-generated content on social media to gain insight into attitudes towards the advertisement and the brand.It is essential also to note that possibilities abound on alternative social media -tweets from Twitter, comments on YouTube and Instagram, as well can be thematically analysed.What is more, with Facebook being the most prominent social media network, this research process can be transferable.

Extracting the Data
Since UK bank advertisements and Facebook were selected for content and media respectively for this analysis, the next stage was to extract the data.The user-generated qualitative data in this research were the comments and replies made by Facebook users relating to bank advertisements.The data collection process involved viewing the comments, as well as clicking "see more" and replies.These comments could either be extracted manually or just copied, extracted using NCapture, a web browser extension on NVivo (Version 10) -see Bazeley and Jackson, 2013 for Qualitative Data Analysis with NVivo -or using an Application Programming Interface (API) developed especially for that social media site.The extraction method, however, depends on the volume of the data.For example, hundreds of comments for a student project on Facebook can be copied and pasted in a word processing software for further analysis, whereas a more extensive project for a business enterprise might need an API.

Analysing the Data
Analysing the extracted data requires various steps, to deduce a meaningful conclusion.It is crucial to repeatedly read over the data to get a feel for it, remove inappropriate text and have a better understanding of the data, in order to identify labels and elements that would be of theoretical significance and of particular importance (Bryman, 2015), as this focuses on the identification of patterns which are interpreted in terms of themes (Aronson, 1995).Braun and Clarke (2006, p. 81) noted: "Thematic analysis can be an essentialist or realist method, which reports experiences, meaning and the reality of participants, or it can be a constructionist method, which examines the way in which events, realities, meaning, experiences and so on are the effects of a range of discourse operating within society".
Different methods abound for thematic analysis -Interpretative Phenomenological Analysis explores personal meaning and lived experiences, exploring in detail how individuals make sense of their personal and social world (Smith and Osborn, 2015).Template analysis involves analysing data from a textual format, creating codes and categories of interrelating importance which can be interpreted by the researcher to "uncover the real beliefs, attitudes, values and so on of the participants in their research" (King, 1998, p. 119).Another method is the Ground Theory which follows a three-part process coding of data: (1) open coding -breaking down, comparing, conceptualising, and categorising data; (2) axial coding -reassembling data into groupings or families; and (3) selective coding -developing core themes and relating them to other identified themes (Strauss andCorbin 1994, 1998).
This chapter adopted the Grounded Theory for the data analysis (see Smith, 2015, for a practical guide to qualitative research methods).The extracted comments were coded, following a three-part process in order to understand consumers' attitudes towards the bank's brand image and the advertisement.As part of the extensive coding, all extracted comments were read several times, with everything carefully examined, no comments were ignored, and tentative labels were coded.This was followed by axial coding where the relationship between the open codes were established, making connections and researching these codes in more detail.The analysis was completed with selective coding where the core variables were integrated and refined which included all the central activities of consumer attitudes towards the brand and the advertisement.

Reporting the Results
In reporting the thematic analysis, one or two key themes emerged which have subthemes as well in order to provide empirical details into the phenomenon being observed.As with the study on attitude towards UK bank advertisements (Mogaji 2016), two key findings emerged: firstly, consumers are able to express their attitudes towards the brand and advertisement publicly; and, secondly, creative features of the advertisement, such as the background music and the black horse evoked a response towards the advertisement from the viewers (see https://bit.ly/2HflAwb).Additionally, the study on attitudes towards UK bank brands highlights the critical factor of reliability, which is supported by seven other subfactors.Shenton (2004) suggested that steps must be taken in order to help ensure as far as possible that the findings from the research are the result of the experiences and ideas of the informants, rather than the characteristics and preferences of the researcher.
Detailed methodological descriptions to enable the readers to determine how far the data and constructs emerging from it may be accepted, should also been provided.This includes guiding the reader through the inquiry process, providing a detailed account of the methods, procedures and decision points in carrying out the study (Merriam and Tisdell, 2015), which allows any observer to trace the course of the research step-bystep via the decisions made and procedures described and also using quotes from the comments to buttress the emerged themes.

Challenges
This section of the chapter builds on the works of Zimmer (2010) and Convery and Cox (2012), highlighting a set of ethical challenges that be addressed before embarking on future research in qualitative analysis of social media, including the nature of consent, privacy and anonymity on social network sites, providing strategies, as well as overcoming these challenges.

Demography
To better understand the results and interpret the findings of research, information about the demography of the participants is essential.However, in the context of social media research for attitudes towards a brand and its advertisement, this is considered a challenge, as personally identifiable data is not being collected.For example, it might not be possible to know the age, gender or race of those who have left comments on a social media post as this data was not collected.However, the general demography of the social media can be used to get an insight into the demography of the participants.Also, the brand location can provide further insight into the demography of the participants.For example, Mogaji (2016) explored Lloyds Bank advertisements in the UK.The demography of monthly active users (MAU) of Facebook and the context of the UK were considered.This is because the majority of the respondents are more likely to be in the UK, Therefore, for a brand based in a South American country, most of the comments will be from users within that country who use the brand and are willing to engage with the brand.

Whose Data?
LeBoeuf (2016) noted that the amount of data generated is increasing every day, every minute and every second, as users are continuously utilising the Internet more and more each day; this leaves many opportunities for communication and media providers to experience its benefits.The challenge here, however, is who owns this data -the users, the brands or the social media network?Facebook acknowledges that its users own all of the content and information they post on its network.According to Twitter (2016), users retain their rights to any Content they submit, post or display on or through their services.Also, Instagram does not claim ownership of any content that is posted on or through their service (Instagram, 2017).
Instead, the users grant these social media networks a non-exclusive, fully paid and royalty-free, transferable, sub-licensable, worldwide license to use the content that they post on or through the service.
Extracting these data can also get complicated in the knowledge that the brands also have the right to use the comments generated by their customers on their pages.According to social media terms of use of Lloyds Bank and Halifax Banks in the UK, for example, the customers grant the brands a worldwide, non-exclusive, royalty-free licence to freely use, copy, edit, alter, reproduce, publish, display and/or distribute such material for any commercial and non-commercial purposes in any media or through any distribution method.
Paragraph 4a of Lloyds Bank Social media terms of use noted that their social media channels are public and any contribution users make will be available for anyone to see (Lloyds Bank, 2016).Paragraph 9 of Halifax Bank Social media terms of use also states, 'We will assume that any material you transmit or post to our social media channels is not confidential.We will have the right to use freely, edit, alter, reproduce, publish and distribute such material for any commercial and non-commercial purposes, including the publication of such content on our website and third party websites' (Halifax, 2016).
In the understanding of the ownership of online comments, especially when posted on brand pages, it can be considered to be publicly available for public use -more like the word of the mouth and other comments or reviews online.These can no longer be treated as a private conversation between the brand and the users, and that is why in most cases brand will encourage users to use direct/private mail to send confidential information.So, in overcoming this challenge about ownership, it can be concluded that the transfer of exclusive rights to the social media networks and the brands makes this data readily available for analysis, albeit after permission from the brands.

Accessing the Data
It is expected that students, academic researchers and brand managers will be able to analyse qualitative social media data to gain insight into brand knowledge and attitudes towards advertisements.However, accessing this data is also considered a challenge.As noted by the incident reported by Zimmer (2010), Facebook data was extracted through profile pages of research assistants, highlighting a peculiarity inherent with using innetwork RAs to access the Facebook profile data, as each of the RAs may have had a different level of default access based on individual students' privacy settings.
Additionally, one RA might have different access to a student's profile than a different RA.
If a researcher is collecting data from a brand's page, privacy settings may be of concern.However, based on the focus of this idea -analysing users -generated comments on brand's social media pages, privacy settings is of no real concern, primarily because the data is readily available and can be extracted as the researcher doesn't have to create a profile to join in the conversation, but instead can still have access to the comments.It would, however, be a different case for the brand manager as they can have access to the comments.

Consent
The idea of what constitutes ''consent'' within the context of social media research needs to be addressed, especially in light of this contextual understanding of norms of information flow within specific spheres (Zimmer, 2010).Eysenbach andTill (2001:1104) argued that to determine whether informed consent is required, it is first necessary to decide whether or not communication is private or public.Convery and Cox (2012) also noted that what constitutes 'public' and 'private' spaces, with corresponding implications for whether or not informed consent is required, is one of the central issues with Internet-mediated research and literature concerning regarding the use of informed consent in private spaces is less clear.
The British Psychological Society (2009) Code of Ethics and Conduct, suggested that, unless informed consent has been obtained, restrict research based upon observations of public behaviour to those situations in which persons being studied would reasonably expect to be observed by strangers ' (p.25).This shares a close resemblance with comments posted on the brand's page on social media networks which can be seen and read by strangers, including researchers suggesting there are practical limitations and implications for procedures around valid consent.Though it might be theoretically possible to reach out to everyone for consent, the technicalities do not make this encouraging (e.g. based on the considerable amount of comments generated by around hundreds of users).This was also corroborated by Convery and Cox (2012), who suggested that it would be time-consuming and cumbersome to reach out to each person whose postings have been used.Also, not every participant would be reachable through messages on social media, simply because the researcher might not be in their network, meaning they may not be able to send them a message to request their permission.However, as it has been previously discussed about whose data it is, the brands now own the right to these comments, and the best option will be to contact the brand in order to get informed consent to use comments for analysis.Convery and Cox (2012) also concluded that the Internet is usually considered a public place and therefore informed consent may not always be required as public behaviour does not necessarily require informed consent.

Ethics
The final challenge, however, summarises all the previous challenges into the ethics of social media research: How ethical can it be?What other considerations need to be taken?Can this be treated as research with human subjects?Different research bodies have tried to provide answers to these questions, but the ever-changing face of social media research has always posed a challenge.One of which is the Ethical DecisionMaking and Internet Research provided by The Association of Internet Researchers, which is an academic association dedicated to the advancement of the cross-disciplinary field of Internet studies (Markham and Buchanan, 2012).The British Psychological Society has also provided Ethics Guidelines for Internet-mediated Research (British Psychological Society, 2013) acknowledging that advancements in Internet-mediated research (IMR) may have introduced 'additional, and sometimes non-obvious, complexities around adherence to ethical principles.' (p3) as the boundaries between social media research and other designs can be blurred where research includes elements of both face-to-face observation/interaction and remote data collection.
As expected from qualitative research, the responders' anonymity is considered crucial.However, on social media, this often poses a challenge as it is possible for the brand, the researcher and even the general public know who made those comments.Therefore, remaining anonymous online has become increasingly tricky (Davisson and Booth, 2016).Dan Kaminsky, chief scientist at New York-based security vendor, White Ops, has also argued that absolute anonymity online is not a particularly achievable thing.Comments can still be anonymised as much as possible by using initials in the reporting of the result and in case the actual comments needs to be embedded in the report.It is advisable to blur out the profile picture and also the names, probably leaving out the first letter as illustrated in Image 3.These ethical challenges in social media research are acknowledged, and more reasons at this point are essential to further iterate on the expectations from these ideas proposed in this chapter, as well as other ethical considerations through Internet-mediated research can be in various forms.This chapter focuses on qualitative analysis of social media to gain insight into the brand and its advertisements and how they are perceived.Therefore, ethical concerns will be interpreted in that light.Smith et al. (1996, p. 172) noted that privacy violations can occur when ''extensive amounts of personally identifiable data are being collected and stored in databases''.In this case, data gathered is user-generated comments on brand's social media postings, and this data does not give the researcher the ability to discern gender, age, ethnicity, and other physically-observable characteristics.Therefore, as long as the objectives of the research are strictly adhered to, privacy of the participants has not been violated.

Conclusion
This chapter acknowledges that social media is changing how brands and consumers relate.Customers are co-creating values and meanings with the brands, generating data which can be qualitatively analysed to further enhance our understanding about attitudes towards brands and their advertisements.
It is important to note, that this user-generated data is not only restricted to Facebook but also comments on YouTube videos, Instagram posts and tweets on Twitter -in fact, all possible avenues for brands to engage with consumers.
The possibilities of analysing user-generated comments towards brand content on social media to understand consumers' responses and attitudes towards the brand and advertisement has been presented.With case studies of a UK bank Facebook post, it is believed that this proposed technique will allow researchers and practitioners to have a greater insight into consumers' attitudes towards their advertisements.Also; to capture the public display of attitude towards their brand and advertisements, and to provide managers with valuable insights that can guide the development and execution of their advertising campaigns.
Significant theoretical and marketing practice implications for students, academics, advertisers, brand managers and social media marketing practitioners are presented within this chapter.However, ethical issues surrounding this research method discussed must be considered -limitations in understanding the demography, getting informed consent and developing sufficient strategies for data anonymisation.It has, however, argued that these ethical considerations may not pose much of a challenge, provided that the main objective of the research is to analyse data already available in the public domain and not to search further, invading the privacy of the users and digging deep into their personal information.
The thematic analysis of a user-generated brand does not have to do with profiling the individuals.It is anticipated that this approach will solely be based on what the user has expressed about the brand and its advertisement at that particular time, with no need to intrude into their profile to further gain insight into who they are.