budgets and the viral ad can be considered the holy grail of digital marketing. ..... 1 http://www.youtube.com/watch?v=o
Marketing content going viral What are the factors that may boost virality?
Master thesis by Melissanthi Mavridou
Supervisor: Jakob Svensson Uppsala University Department of Informatics and Media Digital Media and Society May 2016
Abstract Viral marketing allows companies to promote their products and services using very small budgets and the viral ad can be considered the holy grail of digital marketing. In this light, the aim of this study is to discuss the main factors in the existing literature that are considered to boost virality and articulate a summarized Virality Theoretical Model. The empirical study included in this thesis involves the monitoring of an actual ad and an assessment of whether it went viral or not and if it follows the guidelines of the Virality Theoretical Model. The empirical study showed that the ad did not go viral and did not include and display all the attributes proposed by the Model. This further indicates, in regards to theory as well as for marketing executives and advertising practitioners, that virality is a complex phenomenon that depends on several different factors involving both content characteristics and dissemination. Keywords: viral marketing; going viral; virality; video ad; content characteristics; viral dissemination; factors that boost virality;
2
Table of Contents 1. Introduction .......................................................................................................................... 6 2. Background .......................................................................................................................... 9 2.1 Populate and description of the ad project ........................................................................... 9 3. Theoretical Framework ..................................................................................................... 12 3.1 Definition of Virality ......................................................................................................... 12 3.1.2 Other types of information flow .............................................................................. 18 3.2 Virality in action: Factors that can make virality happen .................................................. 19 3.2.1 Virality and emotion ................................................................................................ 19 3.2.2 Virality and the element of surprise ....................................................................... 20 3.2.3 Virality and other content characteristics ................................................................ 23 3.2.4 Virality and interest networks................................................................................. 24 3.2.5 Virality and influencers: Personal influence and gatekeepers ................................. 24 3.3 A Virality Theoretical Model............................................................................................. 28 4. Methodology ....................................................................................................................... 30 4.1 Operationalization of virality concept and data collection ................................................ 30 4.2 Assessment of ad according to the Virality Theoretical Model ......................................... 31 4.2.1 Qualitative content analysis choice as a research method ....................................... 32 4.2.2 Design and Data....................................................................................................... 32 4.2.3 Reliability and Validity ........................................................................................... 34 5. Results and analysis ........................................................................................................... 35 5.1 Virality assessment results and discussion ........................................................................ 35 5.2 Virality Theoretical Model assessment results and discussion .......................................... 37 5.2.1 “Nya Tunnelbanan” and emotions ........................................................................... 37 5.2.2 “Nya Tunnelbanan” and surprise ............................................................................. 38 5.2.3 “Nya Tunnelbanan” and interesting content ............................................................ 39 5.2.4 “Nya Tunnelbanan” and good presentation ............................................................. 40
3
5.2.5 “Nya Tunnelbanan” and high quality ...................................................................... 40 5.2.6 “Nya Tunnelbanan” and content that resonates ....................................................... 40 5.2.7 “Nya Tunnelbanan” and context.............................................................................. 41 5.2.8 “Nya Tunnelbanan” and branding ........................................................................... 41 5.2.9 “Nya Tunnelbanan” and interest network ............................................................... 42 5.2.10 “Nya Tunnelbanan” and gatekeepers .................................................................... 43 5.2.11 “Nya Tunnelbanan” and affinity groups ................................................................ 44 6. Conclusion .......................................................................................................................... 45 References ............................................................................................................................... 47 Appendix ................................................................................................................................. 50 Appendix 1: Creative Brief............................................................................................... 50 Appendix 2: Swedish transcript of audio of the ad “Nya Tunnelbanan” ......................... 53 Appendix 3: English transcript of audio of the ad “Nya Tunnelbanan” ........................... 54 Appendix 4: Description of the ad “Nya Tunnelbanan”................................................... 55
List of Figures Figure 1: Phelp et al’s (2004) model of passing along emails ................................................ 14 Figure 2: Example of a sigmoid curve pattern ........................................................................ 16 Figure 3: Example of a power law distribution ...................................................................... 17 Figure 4: Cumulative views of “Nya tunnelbanan” ................................................................ 35 Figure 5: Daily views from Google analytics ......................................................................... 36 Figure 6: Views from external sites / 18-28 Aug 2015........................................................... 43
4
Acknowledgements I wish to express my sincere thanks to Populate and especially Fredrik Lundqvist and Anders Land without whose contribution this thesis could not be completed. I would also like to thank my supervisor Jakob Svensson for his valueable input and guidance throughout this journey. Finally, a special thanks goes to my partner in life Nikolaos Simisiroglou for his encouragement, useful advice, love, patience and unconditioned support he offered me throughout this venture. I would like to dedicate this thesis to him and my parents, Marina and Konstantinos, for their encouregement and support.
5
1. Introduction An ad has been viewed 23 million times within 36 hours of original release. This sounds like a fictional marketing success, except this has actually happened like, for example, in the case of the men’s perfume brand Old Spice. In 2010, Procter & Gamble uploaded a video ad on YouTube named “The Man Your Man Could Smell Like”1 to advertise its men’s body wash brand and it was an immediate hit reaching millions of people within a very brief amount of time (Kaplan & Haenlein 2011, 253). This Old Spice ad managed to “go viral”, a buzzphrase widely used by people and media to describe the process of online content of any type – from ads to cat videos or celebrity photos – achieving this kind of success. As Teixeira (2012, 25) claims, the viral ad is the holy grail of digital marketing. One can easily agree, when having in mind the views one ad can get for free when it is voluntarily shared and spread across networks, that viral marketing allows companies to promote their products and services using very small budgets (Kaplan & Haenlein 2011, 254). Moreover, the sharing from consumer to consumer increases knowledge and awareness of products and services (Phelps et al. 2004, 334). In addition, when the message comes from interpersonal communication it is more persuasive and can trigger purchases both quickly and widely as the message spreads and reaches a large number of people (Rogers 1995 in Phelps et al. 2004, 334). As a process, virality is not only relevant for marketing and advertising but also other fields as well as actors, politicians, activists and artists. A typical example of a political video going viral is the American artist will.i.am’s “Yes we can”2 video that promoted Barack Obama’s US presidential candidacy. The video was uploaded on YouTube on the 2nd of February in 2008 and was an immediate success; it reached 150,000 views the first day and was viewed over 5.4 million times within one month (Wallsten 2010, 170). Another very well known video, this one with an activist objective, that went viral is the KONY2012 30-minute documentary3. It was produced by the nonprofit organization Invisible Children and its goal was to raise awareness about and to help in stopping the human rights’ violations of the Lord Resistance Army and its leader Joseph Kony in Central Africa. The video reached 100 million views in six days (Nahon & Hemsley 2013, 132). Virality can, therefore, be perceived as a highly sought after outcome whenever it is achieved, offering a competitive advantage to the actors involved. 1 http://www.youtube.com/watch?v=owGykVbfgUE, accessed 7 March 2015. 2 https://www.youtube.com/watch?v=jjXyqcx-mYY, accessed 7 March 2015. 3 https://www.youtube.com/watch?v=Y4MnpzG5Sqc, accessed 7 March 2015.
6
It should be noted, though, that it can also have detrimental consequences, as the information communicated can be misinterpreted or attract negative attention. One such case is that of a student at the University of Los Angeles. The student, Alexandra Wallace, uploaded a video called “Asians at the Library”4 complaining about Asians’ behavior in the library. The video went viral, it got many responses and remixes, i.e. people manipulated the video making fun of it, and Wallace was accused of racism. As a result, she had to make a public apology and decided to leave university, as she had even received death threats (Nahon & Hemsley 2013, 7). However, this thesis will not delve into the possible negative consequences of virality. It will rather examine the factors that may lead to virality as a positive, welcomed outcome. But what does the concept of virality include? Which factors increase the chances for content to have a potential to go viral? Can virality be controlled or predicted? Stemming from these questions this thesis will (1) present the definition of virality, as it has been described by Nahon & Hemsley (2013) in their book “Going Viral”; a definition that is used as a reference to check the virality5 of an actual ad, and (2) summarize the characteristics of the content that increase chances of going viral into a theoretical model by reviewing the existing literature. Creating a theoretical model out of these characteristics helps towards the operationalization of the concept of virality, so it can be used empiricaly. Part of this empirical study involves the assessment of an actual ad (called “Nya Tunnelbanan”) created by the Communications Agency “Populate”6 which is headquartered in Sweden. The company specializes in creating and producing video ads for different clients. The studied ad’s characteristics were juxtaposed to the theoretical model’s characteristics and the ad was assessed to ascertain if it had indeed achieved virality, according to the definition by Nahon & Hemsley (2013). The research questions in this study are as follows: 1. Can the ad “Nya Tunnelbanan” be considered to have gone viral according to the definition by Nahon & Hemsley (2013)? 2. Which of the factors enumerated in this thesis’ constructed theoretical model does the ad “Nya Tunnelbanan” possess? a) Can the assessment of the ad according to the theoretical model explain why the ad did or did not go viral? b) Which factors of the theoretical model are more important? 4
https://www.youtube.com/watch?v=AQQr3hUepZM, accessed 7 March 2015. for this choice is discussed later in this thesis, in 3.1 Definition of virality. 6 http://populate.se/, accessed 7 March 2015. 5 Argumentation
7
Many scholars (Bakshy et al, 2011; Berger & Milkman, 2012; Dafonte-Gomez, 2014; Dobele et al, 2007; Nahon & Hemsley, 2013; Phelps et al. 2004; Teixeira, 2012) have looked into the topic of virality, acknowledging and analyzing the potential factors that drive virality, through the study of content that has already gone viral. This thesis delves into their findings and gives an overall view of the different scholars’ research on what drives virality and continues to create a more elaborated model of how to have a better chance of achieving it. In this way this thesis tries to contribute to the affirmation or not of what have been considered factors that can lead to virality and to further enlighten not only marketers, but also academics and other social or political actors that are interested in the concept of virality and how it works. Moreover, it proposes a model that can have a practical use for marketing purposes as it acts as a practical guide encompassing factors that when employed may increase the chances of virality. Therefore, the main focus of this thesis is the concept of virality within the field of marketing where “going viral” is generally regarded as a positive, welcome outcome. My interest focuses on marketing especially, as I have worked in the advertising industry as a creative copywriter for over 12 years. Working in the field of advertising during the era when the internet became a primary advertising and marketing tool, I am very much interested in exploring the factors that allow certain ads to go viral. On top of that the ability to create content that will go viral is the dream of perhaps every advertiser, every client and every brand. First the background of the company Populate is given and the ad chosen for the case study is presented. The thesis continues by presenting the definition of virality and providing a discussion of the factors that potentially help in achieving virality, as described in existing literature. A summarized theoretical model is presented and then the empirical study and its findings are analyzed.
8
2. Background This thesis involves an empirical case study by which a video ad, released on YouTube, is evaluated according to this thesis’ theoretical model and its views are monitored to ascertain if it has gone viral or not according to the definition Nahon & Hemsley (2013) provide in their book “Going Viral”. To this end a collaboration with the Swedish advertising and producing agency Populate was established in order to perform the case study. After briefly presenting Populate, the client’s brief about the desired ad is presented and the creative concept of the ad established.
2.1 Populate and description of the ad project Populate describes itself as a film agency and its slogan is “Creating movies. Creating value.” On their website7 one can find their mission and description of their work split into three distinct pillars: Education, Strategy/Content and Production. What is relevant to this thesis is that Populate provides complete services for clients, i.e. strategy, creative conception and production of one specific type of advertising, which in this case is any type of film (animation or not). This is explicitly stated on their homepage:“...Populate are specialists in communication through the medium of film. We provide strategy and communication that creates measurable results,” and as stated in another section: “We produce everything from commercials, videos for Web TV to animation. With the experience of over 300 tasks we have helped many companies to reach out to their customers, employees and thereby strengthen their brands. Many of our film projects have won awards at international festivals.”8 Populate’s focus on film provides the circumstances for an ideal case study, as “viral video advertising may be the most popular manifestation of viral marketing phenomena” (Dafonte Gomez 2014, 200). Watching videos online was already very popular in 2010 and video as a medium seems to have grown even stronger since then; as a study shows, seven in ten adult internet users (69%) in the U.S. used the internet to watch or download some kind of video in 2010 (Purcell 2010, 2). The types of video watched include short video clips, television shows and movies and they are either watched online on video sharing sites, such as YouTube and Google Video, or they are downloaded for later consumption (ibid., 3). Brands worldwide have come to realize how popular video watching is as one third of them have tried the viral video approach in their marketing strategies 7 http://populate.se/, accessed 8 April 2015. 8 Translation of website text by the author
9
(Lindstrom 2009 in Eckler & Bolls 2011, 2). It is also noteworthy that the online magazine Advertising Age9 features a weekly chart of the 10 most popular viral videos (Eckler & Bolls 2011, 2). Using a video ad for this thesis, therefore, can be considered an ideal type of content to check virality potential and Populate provides a suitable collaborator for this purpose. The specific project chosen for this case study is an animated video ad Populate created and produced for communicating and promoting the metro expansion in the city of Stockholm. The client that hired Populate for this project is the Swedish SLL10, the County Council which is responsible for all publicly-financed healthcare and public transport in Stockholm County, as well as for regional development. As stated on their website, they strive for easily accessible, reliable and environmentally friendly public transport and they are carrying out a number of major projects and investments for the region’s development11. One of their projects is the metro expansion which began in 2016. In the creative brief12 for the project it is stated that the goal of the campaign is to produce two versions of the video - a short one (about 1 minute) and a longer one (max 5 minutes) - with the overall goal of giving a good overview of the forthcoming expansion of the subway. The challenge of the campaign is to place the metro expansion in a larger context and bring into focus the end benefits for citizens, rather than just present technical details. The company ultimately only produced and released one video on YouTube with a duration of three minutes and thirteen seconds. This video was the one that was used for the purposes of this thesis13. According to the creative brief, the purpose of this film is to answer the following questions: “Why build an expansion?”, “What will happen? ”, “When does the project start and when will it be finished?”. The target audiences include the Stockholm residents, as the metro expansion affects their everyday lives, new employees, who can familiarize themselves with the project through the campaign, as well as the media, since the project is of public interest. The main goals of the film are the following: (1) to inform the target audiences that the metro will have three new parts, (2) to explain that Stockholm is expanding with new residential development and (3) to make the target audiences feel happy that the metro is developing. In the creative brief the directive given for the tonality of the ad 9 Advertising Age (or Ad Age) is a magazine, delivering news, analysis, and data on marketing and media. 10 http://www.sll.se/, accessed 8 April 2015. 11 http://www.sll.se/om-landstinget/Information-in-English1/, accessed 8 April 2015. 12 See appendix for translation of original document. 13 https://www.youtube.com/watch?v=n1f3e1Clhus, accessed 19 August 2015.
10
is for it to be easygoing, approachable but factual and serious. Finally, the video ad is suggested to be featured on various sites, including the SLL website as well local government’s sites and the SL site (SL is the organization running all of the land based public transport systems in Stockholm County). Moreover, it is suggested it be featured on YouTube, both in SLL’s and SL’s channels. When the ad was finally produced it was released - to the author’s knowledge – through a special YouTube channel run by SLL, called “Nya tunnelbanan”14 (i.e., “New Metro”) and it was also featured at the SLL’s official site15 in the news section. The style of the video ad is based on graphic design and the goal was for it to be produced in a way that it would still be relevant over a course of two years, so it can be used for the communication during the entire metro expansion. The action of the ad takes place on a map of Stockholm showing the existing metro line as well as the new expansions to the line. The extensions of the metro lines, as well as the residential development of Stockholm are thoroughly presented. With the help of animation, the new metro lines are created on the map and new buildings “pop up” out of the map forming the new neighborhoods of Stockholm. The narrator explains how important the expansion of the metro is for Stockholm, for the decongestion of certain routes and for the development of new residential areas that Stockholm needs16. This ad has a clear informative character and builds the image of SLL and its practices to enhance the quality of life in Stockholm. There is no product to sell, rather the ad tries to make residents of Stockholm feel good about the new developments and to counterbalance any inconvenience the metro expansion construction might bring.
14 https://www.youtube.com/channel/UCzbApDA2RWR-ABkvqhkc8sw, accessed 01 April 2016. 15 http://goo.gl/Neey3F, accessed 25 March 2016. 16 See Appendix for a thorough description of the video and audio of the ad and/or visit https://www.youtube.com/watch?v=n1f3e1Clhus, accessed 30 March 2016.
11
3. Theoretical Framework This section provides (1) the presentation of virality’s definition as it has been accounted for by Nahon and Hemsley (2013) in their book “Going Viral”, as well as the reasoning behind this choice and (2) an elaborate, summarized presentation and discussion of the existing literature regarding the factors that can contribute to content having a greater chance of going viral that leads up to this study’s theoretical model. The first point is covered in 3.1 Definition of virality and the second point in 3.2 Factors that can make virality happen and 3.3 A Virality Theoretical Model.
3.1 Definition of Virality Providing a definition for virality is a key issue for this thesis. It is important to pinpoint what is meant by this commonly used, sometimes as a mere buzzword term. The definition is important because it is used as a reference to determine if the ad in the thesis’ empirical study has gone viral. Virality draws on the concept of viruses in the field of biology. Just like a virus of a disease spreads from person to person and community to community, the internet viral content disseminates from user to user, from one network to another network (Nahon & Hemsley 2013, 17). The main difference is that the biological virus makes no discriminations, while the host of information spreads it selectively (Huberman & Adamic 2004, 379). Viral marketing was coined as a term by Juvertson in 1997, and was used for describing the exponential adoption of Hotmail email services through word-of–mouth networks (Nahon & Hemsley 2013, 17). Scholars, that have engaged in research concerning virality, approach it from different angles. For example, Jean Burgess (2008) refers to viral marketing as “the attempt to exploit the network effects of word-of-mouth and Internet communication in order to induce a massive number of users to pass on marketing messages and brand information voluntarily” (1). Burgess (2008, 3) tries to see viral content and particularly viral videos through the lens of participatory culture in social networks, and understands virality as a condition in which a video, besides being quantitatively popular, is also the initiator that will further induce creative activity and engagement by other people. Bennet & Segerberg (2012) also acknowledge the importance of personal appropriation of a message in the process of social sharing; the ability to individualize and imitate content contributes to its spread (745). Another scholar, Dafonte-Gomez (2014), refers to a viral video ad as a video produced by a brand with direct or indirect commercial goal and with a high number of views, but in his definition he places great importance in the way the views are achieved: virality
12
depends on individual sharing of many people to their networks (200). In addition, DafonteGomez (2014) also takes note of the time frame of the content’s spread; the large number of views should be achieved in a short period of time and we should be able to observe a strong peak and then a dramatic decrease (200). Modzelewski (2000) in an article on Direct Marketing
News site17 an effort is made to distinguish viral marketing from word-of-mouth marketing, claiming that it “differs …. in that the value of the virus to the original customer is directly related to the number of other users it attracts. That is, the originator of each branch of the virus has a unique and vested interest in recruiting people to the network.” Phelps et al. (2004)18 draw on Shirky (2000)19 and Rosen (2000, 6) and view viral marketing “as the process of encouraging honest communication among consumer networks” (334). The concept of honesty here refers to the recipient’s belief that the content he receives from another person in his or her network provides an honest endorsement of the relevant content. Some common characteristics in these definitions of viral marketing are: the concept of network, the act of sharing, the word-of-mouth dissemination, the number of views and the speed of the spread. All of these characteristics can be found in the definition that Nahon & Hemsley suggest in their book “Going Viral” (2013) which is regarded by the author as a more expansive definition of virality and, therefore, more adequate to be used for the purposes of this thesis. A thorough presentation and discussion of their definition follows below. Nahon and Hemsley’s (2013) definition deals with virality as an information flow and very eloquently delineates four components that should characterize it, regardless of if it is marketing content. The definition of virality, according to Nahon & Hemsley (2013) is the following: “Virality is a social information flow process when many people simultaneously forward a specific information item, over a short period of time, within their social networks, and where the message spreads beyond their own [social] networks to different, often distant networks, resulting in a sharp acceleration in the number of people who are exposed to the message” (16). 17 http://www.dmnews.com/dataanalytics/finding-a-cure-for-viral-marketing-ills/article/68355/, accessed 12 April 2015 18 Phelps et al.’s paper is from 2004, but their research is in line with much more recent research, and thus is considered relevant to this thesis. 19 http://www.shirky.com/writings/toughest_virus.html, accessed 12 April 2015
13
The four components can be discerned in this definition as (1) the social sharing between people and the factors behind this sharing, (2) the speed of the spread, (3) the reach in terms of the number of people exposed and (4) the reach in terms of the distance the content covers as a bridge between different networks (ibid., 16). What is also important in their approach is that they distinguish it from other kinds of information flows, such as wordof-mouth marketing, just like Modzelewski (2000) (see above), Memes and Information Cascades. These concepts are discussed in more detail later in this section. It is considered of crucial importance that they are clearly distinguished from virality, because phenomena of this kind proliferate throughout the Internet and should not be misunderstood for virality. Let us now discuss the four components of Nahon & Hemsley’s definition of virality before briefly explaining virality’s difference from the other types of information flow. Regarding the first component of personal and social aspects of sharing information, Nahon & Hemsley (2013) draw attention to two key decisions each person has to make when exposed, for example, to an online video: first they have to decide to watch it, then they have to decide to share it (19). These two decisions are also a part of the model of passing along emails by Phelps et al. (2004). More specifically, according to Phelps et al. (2004), the decision to open an email (the equivalent of watching a video) depends on some factors among which are (1) the receiver’s relationship to the sender, (2) the sender’s identity – for example, if the sender is someone that is perceived as sending low quality content, the email will be deleted, and (3) the personal mood, state of mind and context of the receiver, i.e. if someone is under pressure, stressed or assumes the content is inappropriate, they will not open it. The decision to pass along the email, on the other hand, depends on the content of the email (a factor that will be more thoroughly analyzed later) and the emphasis falls on (1) how fun the content is, and/or (2) how it contributes to social connection (helping others - showing care, thus displaying social connection). A summarized version of their concept is depicted in Figure 1. DECISION TO OPEN AN EMAIL Depends on: 1. Receiver’s relationship to sender. 2. Sender’s identity. 3. Personal mood.
DECISION TO PASS ALONG AN EMAIL Depends on: 1. Content of email: a) is it fun? b) does it contribute to social connection?
Figure 1: Phelp et al’s (2004) model of passing along emails.
14
Nahon & Hemsley (2013) also claim that the decision to watch an online video has to do with the identity and stance of people involved, the context, the content, the form and social forces at stake (19). They continue by explaining that the decision to share the content is a second step that has to do with social factors, meaning that people choose to share or not share information online on the same grounds they make this choice when they meet people offline. As Goffman (1959) suggests, in the presence of others, an individual will act so as to convey an impression to others that is in his interests to convey (3). This is especially the case online, where the online environments give many opportunities for identity construction, and negotiation. As Miller (2011) suggests the identity construction becomes a “reflexive project” (169); the individuals are required to constantly reflect on their identity, think about it, shape and re-shape it, as well as maintain it. So, people will probably avoid sharing content that does not reflect well on their image or content they want to keep for themselves (Nahon & Hemsley 2013, 20). One important difference between online and offline sharing, though, according to Nahon & Hemsley (2013), is that, when the sharing happens in online networks, the content is communicated by many to many (20). As this many-to-many communication occurs, the viral content displays a certain life cycle (Nahon & Hemsley 2013, 21). The increase in number of people who have seen the video presents a slow-fast-slow growth pattern, which is referred to as a sigmoid curve (ibid, 21). The slow-fast-slow growth pattern, an example of which can be seen in Figure 2, means that at first the sharing is slow, then it accelerates until there are only few people that haven’t been exposed to the content, at which point it slows down again (ibid., 21). The acceleration phase begins when the tipping point is reached – the tipping point, according to Gladwell (2000), is the point when any social phenomenon reaches critical mass, a threshold where enough people have been involved in a particular phenomenon, and after which the phenomenon takes off (12). In the case of exposure to a video this translates into enough people having seen the video, and after the tipping point has been reached, the exposure accelerates until almost everyone has seen the video and then the diffusion slows down, resulting in the sigmoid curve growth pattern (Nahon & Hemsley 2013, 21-22).
15
Figure 2: Example of a sigmoid curve pattern.
The other factor we should also have in mind is the speed in which this life cycle occurs, and this brings us to the second component of Nahon & Hemsley’s definition. Citing different studies (Kwak et al. 2010, Bakshy et al. 2012, Wu & Huberman, 2007) Nahon & Hemsley argue that most of the re-sharing process (sharing beyond the original source) takes place within a day but it also depends on the platform: most of Facebook resharing might take place on average within 6 hours, while a tweet might be mostly retweeted within an hour (23-24). According to Nahon & Hemsley (2013), promoted messages usually attract and lose people’s attention more rapidly than socially driven messages, but they also note that viral events are rarely strictly promoted or strictly social (25). Messages considered to be promoted are those that are professionally produced and promoted by organizations, like for example a promotional movie trailer, whereas a socially driven video is one that receives views from social sharing, i.e. it circulates from person to person (ibid., 24). The line between these categories, though, may be blurred as in the example of Susan Boyle’s video20. 20 https://www.youtube.com/watch?v=RxPZh4AnWyk, accessed 15 April 2015.
16
47-year-old Susan Boyle took part in “Britain’s got talent” show and performed “I dreamed a dream” song from “Les Miserables”. The video of her performance was viewed by almost 100 million people in less than 10 days (Nahon & Hemsley 2013, 4). The content of this video was not amateur, as it involved professional production, some news services discussed the video and, thus, promoted its sharing, but still the sharing occurred on a one to many individuals basis (ibid., 5). In any case, Nahon & Hemsley (2013) determine virality speed through the rate of decay: if the decline in views follows a power law distribution (an example of which can be seen in Figure 3), where the graph visualization is characterized by a tall peak to the left, a sharp downhill and then a long right tail, then the content most probably can be claimed to have gone viral (26). The important conclusion regarding the discussion about speed is that the number of views does not necessarily define virality- it is rather the way the content spreads that tells us if it is viral. In this context, for example, if some particular content would gradually gain a million views over a period of one year, it could be characterized popular, but not viral (ibid., 28).
Figure 3: Example of a power law distribution.
The third and fourth component of virality, according to Nahon & Hemsley (2013), are the reach by numbers and the reach by networks. When talking about reach by numbers, they take into account not only how many people re-shared the viral content, but also how many people were exposed to it (ibid., 31). The other aspect of reach has to do with how deep into networks the content spreads. Nahon & Hemsley (2013) suggest that a viral event, moving from network to network, creates a short-lived interest network, i.e. a network of people that show interest in the specific viral content (32). This network is created thanks to the weak ties’ connections that act as bridges between different networks (ibid., 32).
17
According to Granovetter (1983) the weak ties are the acquaintances of an individual, as opposed to their close friends or family that constitute their strong ties (201), and it is from the weak ties that people receive new information, new ideas and resources (209). The new information received is then shared in one’s network of strong ties, until the network is saturated and then, again through weak ties, the content spreads farther (Nahon & Hemsley 2013, 32-33). As the viral content involves a network of interest, it is concluded that virality is ephemeral and contextual, which means that when, for example, a country is in election period, a politician’s video might go viral but as years pass by it will not have this spread or be viewed anymore (ibid., 35). 3.1.2 Other types of information flow Besides defining what we mean by virality, it is equally important to distinguish it from other types of information flow to avoid confusion and misunderstandings. Nahon & Hemsley (2013) differentiate between virality and word-of-mouth, as also differentiated by others aforementioned, memes and information cascades (35). Word-of-mouth, in the marketing context, is considered the dissemination of information about a product or service from mouth to mouth without a commercial goal (ibid., 36). Memes refer to content that spreads from person to person by copying or imitation, like for example in the case of the Pepper Spray meme21, when a police officer was caught on camera pepper-spraying peaceful protesters. This image was copied and manipulated with Photoshop in different ways and then shared extensively, creating a buzz around the incident (ibid., 37). Information cascades are situations whereby people imitate other people’s behaviors; they follow the herd instead of their own individual thinking (ibid., 38). In this sense information cascades in relation to information flows relate the sharing process to people’s impulse to watch and share content because other people have done so (ibid., 38). According to Nahon & Hemsley (2013), though, and as already discussed and will be further analyzed in this thesis, people share information and content for a variety of reasons (ibid., 39). In a nutshell, word-of-mouth and memes encompass social sharing as so does virality - while information cascades do not - but the other three components of virality’s definition that have been discussed (sharp acceleration, reach by numbers and by networks) are not necessarily required for any of these other concepts to occur (ibid., 39).
21 http://knowyourmeme.com/memes/casually-pepper-spray-everything-cop, accessed 15 April 2015
18
Identifying virality is one thing, but knowing how virality is to be achieved is another story. Nahon & Hemsely (2013) suggest that viral processes cannot be totally controlled and that there is not a standard recipe for virality, nevertheless, using the right elements, content may have a higher possibility of going viral (96-97). In the following section an overall view of the different scholars’ research on what drives virality is given and the factors that may boost virality are discussed before being summarized into a theoretical model. 3.2 Virality in action: Factors that can make virality happen Two of the most discussed factors of virality in the existing literature are that of the content characteristics and that of the role of influencers. In this section different studies on virality are presented and both of these factors are analyzed in an attempt to map out the factors that can make virality happen. These factors in the relevant literature involve not only marketing content, but also other forms of content, like newspaper articles or political videos; nevertheless, this literature is considered useful and relevant to this thesis’ purpose to explain the dynamics of the concept of virality in an all-embracing and accurate way. 3.2.1 Virality and Emotion The factor of emotions is one of the key issues that Nahon and Hemsley (2013) focus on regarding the sharing of content, along with information characteristics and context which will be analyzed later (61). They claim that content that has emotional impact has more potential to become viral and they refer to Bakshy et al.’s (2011) research in which they argue that content that invokes positive emotions is more likely to be shared (Nahon and Hemsley 2013, 62). By eliciting emotion content has more chances to be considered remarkable, i.e. “worth remarking on with the people we are connected to”, and, therefore, to attract people’s attention and get shared (ibid., 61- 62). Phelps et al. (2004) also conclude that the messages that invoke strong emotion (humor, fear, sadness, inspiration) have more chances to be shared (345). In their research on the motivations for passing along emails, they found that most emails that involved content with naked pictures, jokes about gender, jokes about work or computers, crime warnings, games, and luck-oriented chain letters were shared, while items that were about old content or were characterized as uninteresting, stupid or inappropriate were not passed along (ibid., 342). Phelps et al. (2004) acknowledge two main motives for passing along emails and those are enjoyment / entertainment and the display of social connection, by offering help or
19
communicating care (343). What they recommend for advertisers is to appeal with their messages to desires for fun, entertainment and social connections, while nonprofit organizations should give attention to the social motivation role (ibid., 345-346). Their main suggestion is that, regardless whether they are positive or negative, strong emotions can be considered as playing a significant role in the decision of sharing specific content. Berger and Milkman (2012) also make the connection between emotion – negative or positive – and virality, in their study of the diffusion of New York Times articles over a three-month period. Their main findings suggest that content that evokes emotion, regardless of if it is negative or positive, has more potential to become viral than content not evoking emotions (ibid., 201). Nevertheless, within these findings, they argue that positive emotions drive virality more than negative emotions (ibid.). They further discern between high and low arousal emotions, meaning there are emotions that induce more physiological arousal and lead to activity (Smith & Ellsworth 1985 in Berger & Milkman 2012, 193); for example anger and sadness are both negative emotions, but anger is characterized by high arousal, while sadness by low arousal. In this context, Berger and Milkman (2012) suggest that content that evokes high arousal emotions, i.e. awe, anger or anxiety, leads to more virality than content that evokes low arousal emotions, such as sadness, because high arousal emotions induce action which translates into sharing behavior (193). They manage to generalize their findings to other content besides newspaper articles and conclude that emotion and arousal determine which content will be shared (ibid., 199-200). Moreover, they claim surprising and interesting or useful and positive content leads to virality, as people share to entertain others, but also to inform or to boost their mood (ibid., 201). In regards on how to design successful viral marketing campaigns they suggest that marketers should opt for content with high-arousal emotions, for example they should strive to amuse the audience rather than make people feel content or relaxed, and they sometimes do not need to avoid negative emotions, like anger or anxiety, as these emotions might boost sharing behavior (ibid., 202). To sum up, this study also acknowledges the power of emotions, and claims positive and high arousal emotions more probably lead to virality than the negative or low arousal ones. 3.2.2 Virality and the Element of Surprise Another element of viral content discussed in literature is that of surprise. Dafonte-Gomez (2014) in his paper about the key elements of viral advertising, claims that content-wise the emotion of joy along with the element of surprise are the most common features that the 25
20
most shared viral video ads between 2006 and 2013 he had in his research (204). Humorous content is a component of 58% of the videos in his research, while 76% of the video ads make use of the surprise element (ibid., 205), always in combination with some other basic emotion as described by Ekman (1972, 29). Apart from surprise, basic emotions are fear, sadness, happiness, disgust or rage (ibid.). Dafonte-Gomez (2014) distinguishes between 4 different types of surprise: “real stunt” (surprise caused by dangerous scenes, generally of a sporty nature, with stunt actors or experts), “fictional stunt” (surprise caused by activities impossible to perform by the person doing them, usually with the aid of digital effects), “surprise event” (surprise caused by a thrilling action of “street marketing” developed in public spaces and recorded with a hidden camera) and “narrative surprise” (meaning a narrative turn was used to achieve an unexpected ending) (202-203). The most common types of surprise in the viral videos Dafonte-Gomez researched was pinpointed as the “real stunt” and the “narrative stunt” (ibid., 204). An example of a “real stunt” is Volvo Trucks’ ad featuring Jean Claude Van Damme22 performing a split, with two legs supported on two moving Volvo trucks. A “narrative stunt” can be considered Volkswagen’s ad of the Passat model23, where a child dressed as Darth Vader tries unsuccessfully to use the Force on several objects, until his father remotely starts the Passat to make his son think he achieved his goal. Dafonte-Gomez (2014) also notes that happiness is an emotion found in 92% of the videos (204). Furthermore, he searched for the presence of celebrities and confirmed their use in 32% of the videos, associated primarily with the “real stunt” type of surprise (sportspeople or actors that are connected with sporty, dangerous activities or risky situations) (ibid., 204). His overall conclusion characterizes the emotional tone of the videos as “agreeable”; even when fear or sadness are involved, they end up achieving positive feelings through twists or happy endings (ibid., 204). The main findings, therefore, of Dafonte-Gomez (2014) indicate that virality is connected with surprise used along with some other emotion, preferably to ultimately create positive feelings. Moreover, it seems that celebrities are used in content that has gone viral, but mostly in the context of a surprise, which leads to the notion that celebrity use alone will not necessarily lead to virality if the content is uninteresting. Surprise as a factor that drives forwarding behavior is also supported by Dobele et al. (2007). Apart from agreeing with Dafonte –Gomez (2014) in that surprise works well for virality when combined with another of the primary emotions which include fear, sadness, happiness, disgust or rage, they moreover argue that emotion in itself is not enough to trigger 22 https://www.youtube.com/watch?v=M7FIvfx5J10, accessed 16 April 2015. 23 https://www.youtube.com/watch?v=FCDKQaH2-_s, accessed 17 March 2016.
21
the sharing process: the content has to capture the viewer’s imagination in a unique or unforgettable way, in the sense that it is original, different, inventive or novel (Dobele et al. 2007, 301). Finally, they suggest marketers should use combinations of surprise with specific emotions for better chances of achieving virality, depending on their brand or campaign (ibid., 301): (1) surprise combined with joy works for fun brands, like Apple or Virgin, for invigorating interest in mature categories, like that of cars (an example is the Ford Ka Evil Car ad24), for revitalizing brands’ image or when the target group is younger customers, like for example a fashion brand for young people (2) surprise with sadness, assuming it does not create guilt, can be effective for immediate response to disasters, e.g. for an ad campaign about raising money after an earthquake (3) anger can work well for cases that involve injustice, e.g. a motivational campaign about the environmental damage a company causes (4) fear should be used carefully, for example in campaigns that try to change behaviors such as drug usage or speeding, and the relevant ads should always provide a solution, talk about punishment or give links to further information that consumers can refer to, (5) disgust resonates more with young males and rebel-style brands that may find disgusting situations humorous. To sum up, Dobele et al. (2007) confirm that the presence of surprise along with some other primary emotions helps to achieve virality, and they note the importance of content being interesting and unique. They also provide a guide for using primary emotions depending on the brand or campaign type. Besides using surprise in an ad, it seems that the timing of when the element of surprise is introduced in the course of the narration, also plays a role in achieving virality. Teixeira (2012), suggests that the emotions of surprise or joy – both of which he found play an important role in keeping the viewer’s constant attention - should be introduced early in the narration, so they can work as a hook for the viewer; the online viewer needs to see something that will catch his interest and attention from the opening seconds (26). Furthermore, he stresses that the narration of the ad should be built with emotional ups and downs. This means that the joy and surprise should not be continuous, but rather create an emotional roller coaster (ibid.). Alternating between tension and relief, a technique implemented in suspense movies, will give the maximum of attention. Teixeira (2012) refers to the excellent use of this technique in Bud Light’s video “Swear Jar”25, where, after establishing the storyline, the ad continues with many different scenes, each with its own humor and surprise (26). He also suggests that the surprise should be moderate and not too 24 https://www.youtube.com/watch?v=2eoPyrgBllU, accessed 16 April 2015. 25 https://www.youtube.com/watch?v=JI3Y1auTFpU, accessed 18 April 2015.
22
extreme or shocking, as shocking material, e.g. extreme nudity, will most probably put off the audience’s intention to share the content (ibid., 26). Surprise along with other information characteristics (humor, novelty, resonance and quality) is also considered important for achieving virality according to Nahon and Hemsley (2013, 63). A prominent example of using surprise in a successful way is the Susan Boyle video26, where a charismatic singer takes part in a TV talent show. The singer is an older woman with none of the stereotypical characteristics one would expect from someone who hopes to start a singing career. This simplicity and lack of glamour, though, combined with an astonishing performance is what eventually surprises everyone and makes the content of the considered video worth sharing (ibid., 63). 3.2.3 Virality and Other Content Characteristics Apart from surprise, Nahon & Hemsley (2013) also claim, referring to various studies (Petrovic et al. 2011, Shifman & Blondheim 2010, Kirby & Marsden 2012, Suh et al 2010), that good content, with fun elements or novelty along with a good presentation, including, for example, a short commentary, an accompanying url or hashtag, are factors that promote virality (ibid., 65). In terms of presentation of the brand in a specific content, Teixeira (2012) also mentions that branding in an ad should not be too prominent. As he puts it “People seem to have an unconscious aversion to being persuaded, so when they see a logo, they resist” (25). The solution he suggests is to discreetly weave the brand throughout the ad, like in the case of Coca Cola’s ad “Happiness Factory”27; the bottle of Coca Cola is seen in the ad, but even if you remove it, the plot remains interesting and worth watching (ibid., 26). Moreover, the high quality of the content is considered important for virality and may refer to the production quality and how the narrative is built, manipulating the emotion elicitation, and attracting people’s attention (Nahon & Hemsley 2013, 64). For example, in the video previously mentioned of Susan Boyle, the narrative seems to be designed and produced in a clever way to elicit certain emotions: the judges explicitly express their enthusiasm, a reaction that dictates that the viewers should feel the same way too. Then another judge makes a comment “I knew you would be great”, at which the audience laughs, further stressing the surprise everyone is feeling or supposed to be feeling (ibid.,63). Finally, regarding information characteristics, Nahon & Hemsley (2013) consider content that resonates with people an important factor of virality as well (64). It 26 https://www.youtube.com/watch?v=RxPZh4AnWyk, accessed 15 April 2015. 27 https://vimeo.com/76368217, accessed 19 April 2015.
23
comes as no surprise that people will more likely share content that they can relate to. As an example, Nahon & Hemsley (2013) refer to the viral video of “United breaks guitars”28, which resonates with people’s general frustration towards big companies that sometimes fail to provide the expected high quality of customer service (64). All in all, when content combines humor, surprise, novelty, quality and resonance, the chances of virality are bigger. In some cases, the lack of some of these factors can be made up for by the presence of context. When particular content is salient, which means it is important in the moment, in the context in which it takes place, it is more likely to go viral (ibid., 66). A political video, for example, has more chances to go viral when shared in the context of a country’s elections period. 3.2.4 Virality and Interest Networks Another very important information characteristic of virality, according to Nahon & Hemsley (2013), is the interest around a particular topic of content. Bakshy et al. (2011) argue that if some content is interesting for people, they will share it (71). So, when people find remarkable content (i.e. content worth remarking on, as aforementioned) in a specific area of their own interest, a viral event may emerge as the content spreads from network to network – what Nahon & Hemsley (2013) refer to as an interest network (71). An interest network is created around topics of interest (ibid., 68). This means that the interest network will exist as long as the content flows within the network; once it stops flowing, the interest network will not exist anymore (ibid., 71). So, content can go viral within a certain interest network without having to reach millions of people; as we have already discussed when defining virality, it is the way it spreads, not necessarily how many people it reaches, that makes some particular content viral (ibid., 70). The way the interest network works, also, means that an event can go viral without necessarily having all the content characteristics discussed so far, i.e. positive emotional impact, quality, surprise, resonance, humor etc. (ibid., 68). For example, in a hypothetical case where an individual shares a video about how to catch the perfect fish could go viral within an interest network of people that are interested in fishing. 3.2.5 Virality and Influencers: Personal Influence and Gatekeepers What is also worth discussing regarding the phenomenon of virality is the factor of influencers. The concept of influence can refer to the role of gatekeepers in information flow 28 https://www.youtube.com/watch?v=5YGc4zOqozo, accessed 19 April 2015.
24
(Nahon & Hemsley 2013, 42). Network gatekeepers, that can be people, collectives, or institutions, can exercise great control over the flow of information in three main ways: by choosing which information can or cannot pass, by connecting networks or by regulating the movement of information (ibid., 43). Castells (2011) refers to this as networking power; the power over who and what is included in the network (774). This power is exercised by the programmers and the switchers, social actors that have the ability to control others via two basic mechanisms respectively: “(a) the ability to constitute network(s) and to program/reprogram the network(s) in terms of the goals assigned to the network; and (b) the ability to connect and ensure the cooperation of different networks by sharing common goals and combining resources while fending off competition from other networks by setting up strategic cooperation.” (ibid., 776). Programmers have the ability to let a person, a medium or a message enter the network, or not, through gatekeeping practices and switchers have the power to control the connecting points between various networks (ibid., 776-777). The reliability of a
gatekeeper and their ability to act as a bridge between different networks (through many connections) are qualities of successful gatekeeping that can boost virality. (Nahon & Hemsley 2013, 49-50). Nevertheless, Nahon & Hemsley (2013) suggest that, as virality is a social process of sharing, traditional gatekeepers sharing content is not enough; people must share the content between themselves, and in doing so they also become gatekeepers of information (59). Additionally, viral events that spread in social media can circumvent traditional gatekeepers, and can therefore reduce some of the forms of power of traditional gatekeepers (ibid., 58). This occurs due to the nature of the social infrastructure of the platforms we share content on, which support distributed networks, i.e. networks where many paths connect any two people (Barabasi 2003 cited in Nahon & Hemsely 2013, 93). As Castells (2011) claims, “anything that reaches the Internet, may reach the world at large” (780). Nevertheless, he acknowledges that traditional gatekeepers still hold power; no one can deny that the mainstream media websites rank high in information retrieval, or that most communication still goes through mainstream media (ibid., 789). To sum up, though the role of gatekeepers is never guaranteed, as their power is dynamic and changeable (Nahon & Hemsley 2013, 43), getting the right gatekeepers to share content might boost virality of specific content. Apart from gatekeepers, the concept of influence can also refer to personal influence, or how an individual plays a role in the spread of content (Nahon & Hemsley 2013, 72), a concept which also relates to the two step flow theory (Katz, 1957). The two step flow theory suggests that “ideas often flow from radio and print to opinion leaders and from these
25
to the less active sections of the population” (ibid., 61). This means that the opinion leaders play a crucial role in influencing other people, and that personal contacts appear to influence behavior more successfully than mass media (ibid., 63). Thus, for example, messages that come from other consumers, and not from a company can be more influential (Nahon & Hemsley 2013, 72). According to Bakshy et al. (2012) people are more likely to be influenced by those close to them, people they know and have a personal connection with, the so called strong ties, but they are more likely to come in contact with new information and ideas through their weak ties, the people they are less familiar with (519). Phelps et al.’s (2004) study also concludes that it is important to carefully select the initial target of a message, as people will most probably not delete a commercial message/content if it comes from someone they know, as it could be considered valuable and worth of consuming and maybe of sufficient worth to share further (345). Nahon & Hemsley (2013) claim that strong ties help distribute information faster, as people will more likely decide to pay attention to the content their close friends share, but weak ties are the ones that help the content spread farther, as it is the weak ties that act as bridges between networks (74-75). Given the way personal influence works, for content to have a greater chance to go viral it should be shared by opinion leaders of some sort, so it can “travel” to clusters of strongly linked ties through the weak ties of the opinion leaders (ibid., 78). The selection of the appropriate opinion leaders to achieve the best outcome can be more complicated. Opinion leaders are usually divided in to different categories, as for example, the connectors, the mavens and the persuaders (Gladwell 2002 cited by Nahon & Hemsley 2013, 77); the connectors are people that act as bridges between groups of people, the mavens are considered experts in some field and as a result people pay attention to them and the persuaders have the talent to make people agree with them. What Nahon & Hemsley (2013) suggest is that for content to go viral one needs to choose to share it with the right kind of influencers. For example, if the content is a light/fun video and its goal is to be shared with different kinds of groups, a connector is the ideal influencer, while if the content is a technical video, a maven would be more appropriate (78). Once the strong ties and a few opinion leaders share the content and it manages to reach a tipping point, it can go viral, as people enter into ‘cascade mode’ and start paying attention to the content because it is trending or “because everyone else has seen it” (ibid., 78). What is considered essential at all times is to reach the right consumers with interesting, relevant content. Towards this purpose companies should try to find people that are interested in what the company has to say, for example through affinity groups, i.e.
26
creating groups that are willing to hear what the company has to say (Phelps el al. 2004, 345). What Phelps et al. (2004) suggest is that companies should identify opinion leaders who would be interested in the information and initially send it only to them, so the content has better chances to be further forwarded (345), as the two step flow theory suggests. Carefully choosing who to share the content with is a point that other researchers also make. Dobele et al. (2007) also claim that companies should target their audience as shrewdly as possible, and send out messages to the people that are receptive to a brand, product or service for better chances of forwarding (303). Teixeira (2012) as well suggests that the act of sharing an ad depends not only on the ad but also on the personality type of the viewer; his research suggests that extrovert and egocentric people will share content more (27). Extroverted people tend to be more social and egocentric who seek to increase their social status, by displaying their own taste, media savvy and connectedness (ibid., 27). Knowing the personality of the audience to target accordingly is difficult, but Teixeira (2012) suggests that the ability to find this kind of person will become easier as social media evolves and gives opportunities to better understand people’s personalities and will eventually substitute traditional demographic groups (27). Before finishing the discussion on network gatekeeping and personal influence, it is worth mentioning two characteristics of network structure that play a role in virality: (1) the power law distribution of networks, which means that a few people have many links, while most people have comparatively few links, (2) the fact that some links are close to the core of a network (Nahon & Hemsley 2013, 84). Regarding the power law distribution, Nahon & Hemsley (2013) claim that each link, whether it is one that links to many or one that links to a few, plays its role in virality: the elite links, like for example the Huffington Post blog, can ignite viral events by posting links to content early, but then links with fewer connections, like for example individuals, share it to their own networks and contribute to the phenomenon of virality (87). In addition, when a link is situated at the core of a network, and by core it is meant a set of people who are highly interconnected but as a group are also well connected throughout the network (ibid., 88), it is more influential in spreading messages than people that present the same amount of links but lie further out in the network (Kitsak et al. cited in Nahon & Hemsley 2013, 88). In this light, Huffington Post blog, for example, is a blog at the core of its network, and as such, when it features certain content, this particular content has more chances to go viral. Nahon & Hemsley (2013) note, though, that viral events can be scalable, which means that content can go viral in a certain interest network, as it has previously been discussed. For example, a video featured in Huffington Post appeals to
27
a broad audience, but a video that has a theme interesting for e.g. pet lovers could go viral in a pet loving related network. The main point to remember is that “the closer content gets to the core of a network, the more likely it is to be picked up and re-shared by media elites” (ibid., 91). To sum up the discussion about gatekeepers and personal influence, sharing the content with the right people, i.e. people that influence others, will create better circumstances for virality to take place. It is also important to initially choose to share the content with people that are perceptive to the original sender; for example, a good idea for a company would be to share the content with consumers in an affinity group they have created, so these people- already perceptive to the company- can further share it into their networks. Finally, though content is more likely to go viral if links at the core of a network share it, content can also go viral within an interest network. 3.3 A Virality Theoretical Model Having discussed all these concepts in existing literature, a theoretical model
can be
articulated including factors which, when applied, can create better chances for content to go viral. The two most important categories of virality boosting factors seem to be (A) Content Characteristics and (B) The Role of Gatekeepers and Influencers. Therefore, to have a better chance of going viral content should have the following characteristics: (1) It should elicit emotions, preferably positive, though negative emotions, that are characterized by high arousal (like anger or anxiety), can also boost virality. (2) The content should be surprising; the surprise should be combined with some other primary emotion (fear, sadness, happiness, disgust or rage), but always ultimately end with a positive feeling. Moreover, surprise should be introduced early in the narration to attract attention, and then follow a roller coaster pattern with waves of tension and relief. “Real stunts” (surprise caused by dangerous scenes, generally of a sporty nature, with stunt actors or experts) and “Narrative stunts” (content with narrative turn used to achieve an unexpected ending) seem to be the more common types of surprise used in videos that have gone viral. (3) The content should be above all interesting and unique or novel. (4) A good presentation of the content might help virality; a good presentation might include a short commentary or the use of a proper hashtag (#). (5) High quality of content, in the sense of the production quality and how the narrative is built, manipulating the emotion elicitation, and attracting people’s attention might boost
28
virality. (6) Content that resonates with people’s feelings about a situation has better chances of going viral, as people can relate to it. (7) Context is also significant. When particular content is important in the moment, it is more likely to go viral. (8) Branding should be discreet. The perfect situation is when the brand is discreetly weaved through the storyline and the content remains interesting even if the brand was to be omitted. (9) There is a possibility of content going viral within an interest network. This means that content can go viral if it is interesting to a specific group of people. In this way, it can go viral without exhibiting all the content characteristics discussed above. Regarding the role of personal influence and gatekeepers the following should be taken into account: (10) Getting gatekeepers or influencers to share content might boost virality, e.g. online/offline media, popular blogs or people with many contacts and influence. Getting someone at the core of a network to share the content increases the chances of virality; individuals, institutions or groups who are highly interconnected, but alltogether as a group also well connected throughout the network, e.g. Huffington Post blog, can be considered to be at the core of a network. Nevertheless, there is the possibility of content going viral within an interest network, in which case gatekeepers sharing the content as a step towards achieving virality may be omitted. (11) Initially sharing the content with people that are perceptive to the sender, e.g. with an affinity group of a company, might help promote virality.
29
4. Methodology This thesis, as explicitly aforementioned, involves the monitoring and assessment of an actual ad29 created by the Communications Agency Populate. The ad in question is about the expansion and development of the metro in the region of Stockholm, which also entails Stockholm’s housing development (see section 2. Background or Appendix for more specific information and description of the ad). The purpose is to determine whether it has gone viral or not and which of the factors, clearly specified in the Virality Theoretical Model, characterize the ad. The methodology section provides information on the method selection and operationalization towards this goal. First, the operationalization of the concept of virality as defined by Nahon & Hemsley (2013) and how the data for determining if the ad went viral or not was collected is described and second, the qualitative content analysis on how it was determined which of the factors of the Virality Theoretical Model characterize the specific ad is discussed. Furthermore, limitations, validity and reliability and potential sources of error or interpretative ambiguities are discussed. 4.1 Operationalization of virality concept and data collection The video ad “Nya tunnelbanan” (i.e. “New Metro”) was launched on YouTube on the 18th of August 2015. The video has a duration of three minutes and thirteen seconds and it was launched from a special YouTube channel named “Nya tunnelbanan”, which one can identify from the explicit presence of SLL logo is an official channel of the SLL30 governmental council. In order to determine if the video ad “Nya tunnelbanna” by Populate went viral after its launch, its views were monitored on YouTube. A practical guide was established to determine what would be looked for in the pattern of the views. This guide follows the virality definition by Nahon & Hemsley (2013), as thoroughly discussed in section 3.1 Definition of virality of this thesis, and includes three points: (1) The life cycle of the ad is monitored to determine if the growth of the spread of the views follows a sigmoid curve pattern (slow-fast-slow growth pattern); this means, as aforementioned, that at first the sharing is slow, then it accelerates, until there are a few people that haven’t been exposed to the content, which is the point when it slows down again (ibid., 21). Moreover, through the 29 https://www.youtube.com/watch?v=n1f3e1Clhus, accessed 19 August 2015. 30 SLL is, as aforementioned, the County Council which is responsible for all publicly-financed healthcare and public transport in Stockholm County.
30
sigmoid curve pattern it can be determined whether the content has reached deep enough in the network to be characterized as viral, as the tipping point must be reached for the sigmoid curve pattern to occur (ibid., 21-22 ; 33). (2) The rate of decline in views is tracked to establish if the views follow a power law distribution; this indicates that in the beginning a lot of people were exposed to the content, but then the decay started and the shares dropped drastically as the content was less shared (ibid., 26). (3) Regarding the time frame, the spread of the content is to be monitored for ten days. Most sharing may usually happen within only one day depending on the platform used (ibid. 23-24), and, moreover, the specific content is considered promoted content, which tends to gain and lose attention more quickly than socially driven content (ibid., 25). For these reasons ten days of views monitoring is considered a fair amount of time for this empirical study’s purposes. Nevertheless, as the rate of decline in views needs to be also assessed, the views are to be further tracked over a longer period of time to specify when the decline starts and determine if virality was indeed achieved. The data to determine all this was collected through Google analytics, a web analytics service offered by Google that tracks and reports website traffic and also provides statistics and analytical tools for marketing purposes. The information one can retrieve from Google analytics includes, for example, views, average duration of viewing time, by which sites the views were generated etc. Access to the data was provided by Populate. The data was retrieved in raw files and plotted manually using Microsoft Excel, an easy to use spreadsheet program for data management and graphics. The plotted data was then statistically analyzed and compared to the equivalent reference graphics for the sigmoid curve pattern and the power law distribution to conclude if the related ad had gone viral (always according to the definition by Nahon & Hemsley, 2013). The results are presented and discussed later in this thesis. 4.2 Assessment of ad according to the Virality Theoretical Model This thesis further involves the assessment of the empirical study’s ad regarding the factors that may boost virality and that are summarized in the Virality Theoretical Model which can be found in section 3.3 A Virality Theoretical Model. In order to assess which of the factors the ad includes, qualitative content analysis was used.
31
4.2.1 Qualitative content analysis choice as a research method An advertisement, as “Nya Tunnelbanan”, can be considered a form of communication which conveys messages to all potential receivers (general public/individuals, organisations, institutions etc.). As such, the message or messages it conveys, through sound (audio) and moving image (video), may be different from observable events, things or properties in the sense that they inform about something other than themselves (Krippendorff 1989, 403). The audio and video of the ad together form a “document” or “text”, consisting of “words and images in multi-media and digital form” that can be considered “an artefact of social communication” created by the company Populate for public consumption (Daymon & Holloway 2011, 277). Content analysis, “goes outside the immediately observable physical vehicles of communication and relies on their symbolic qualities to trace the antecedents, correlates or consequences of communications, thus rendering the (unobserved) context of data analyzable” (Krippendorff 1989, 403). This means that content analysis’ purpose is to support inferences rather than literally describe the communications content (ibid., 404). As Daymon & Holloway (2010) claim, content analysis “offers a means of revealing features that are hidden or latent in the content” (277-278). Moreover, content analysis as a research tool is widely used in the social sciences, as all social processes, in this thesis’ case the messages conveyed by the advertisement in hand, can be seen as having meaning to the participants (Krippendorff 2004, 44). It is, therefore, considered that content analysis is a proper method to analyze the data in hand, as part of the aim of this study is the assessment of the ad “Nya Tunnelbanan” according to the criteria explicitly defined in the Virality Theoretical Model of this thesis. Using content analysis it can be assessed what meaning the ad has for the recipients and more specifically which of the factors enumerated in the model can be traced in it. 4.2.2 Design and data This thesis tries to assess the advertisement “Nya Tunnelbanan” according to a specific model that is summarized in 3.3 A Virality Theoretical Model. This model’s characteristics serve as the context of the content analysis, in the sense that they define what the analysis tries to explore and is not directly observable (Krippendorff 1989, 406). Moreover, they help towards the articulation of well-defined criteria that formalize the knowledge available about the data-context relationship (Krippendorff 1989, 406). The analytical constructions, in short,
32
model the context chosen and the theoretical model used makes the context explicit so that the results of the analysis are clear (Krippendorff 2004, 34-35). More specifically, the framework for the research is defined through the Virality Theoretical Model in the sense that the well described Content Characteristics and the Role of Influencers and Gatekeepers are sought after in the ad. The Content Characteristics include 9 separate points which are emotion elicitation, element of surprise, interesting content, good presentation, high quality, resonance, context, branding and interest network (see 3.3 A Virality Theoretical Model). The Role of of Influencers and Gatekeepers includes 2 separate points which are gatekeepers/influencers sharing the content and sharing with perceptive audience (see 3.3 A Virality Theoretical Model). The data chosen for the analysis is the ad “Nya Tunnelabanan”. Time constraints led to the establishment of a collaboration on a sole company and ad, which were Populate and the certain ad. It was important to have access to specific information on the ad to conduct the study. Establishing a collaboration with an advertising agency was, therefore, considered a crucial prerequisite in order to have access to all the necessary material, not publicly available for a feasible study. A collaboration was, thus, established with Populate, which is an appropriate partner and which functioned as an intermediate to aquire the consent of an advertised client (in this case SLL) in order to use the specific ad in the study. Moreover, Populate had access to and provided the creative brief (see appendix 1) and data from Google analytics that were used complementary in the assessment of the ad. As already explained, data from Google analytics helped determine whether the ad went viral or not (see 4.1 Operationalization of virality concept and data collection) but they were also used to argument and support the content analysis of the ad by making proper inferences. For example, some of the necessary information for a complete assessment was not retrievable from Populate as the company was responsible solely for the creation of the ad campaign. This means that the relevant public relations activities or contacts were not Populate’s responsibility and Populate could not have access to detailed information on how and if the ad was sent out to influencers and gatekeepers to further share. Nevertheless, the information gathered from Google analytics about external sites traffic (i.e. views coming from sources other than the YouTube channel itself) was used to draw conclusions about the role of gatekeepers and influencers in the specific ad. In another example, statistics from Google analytics showing the average time of viewing helped to support inferences about how interesting the ad was. The results of the complete assessment are presented and more thoroughly discussed in the analysis section of this thesis.
33
Apart from the ad “Nya Tunnelabanan” and the data from Google analytics, the data chosen for the content analysis also include the creative brief of the ad (see appendix 1). A creative brief explicitly gives information on the ad’s intented goals, target groups, role of communication, tonality, challenges, objectives, effects etc. This choice was, therefore, considered not only relevant but also necessary in order to assess the data according to the creators’ own explanations and intentions (Daymon & Holloway 2010, 292). As Krippendorff (2004) claims, “a text means something to someone, it is produced by someone to have meanings for someone else, and these meanings therefore must not be ignored and must not violate why the text exists in the first place” (19). The use of the creative brief (see appendix 1) can also be claimed to help in compensating for biases or preconceptions in the choice of the specific data, as the author constantly compared the established assessment with the ad’s prerequisites and strategic goals, taking into account the intentions that brought the data into being (Krippendorff 2004, 121). 4.2.3 Reliability and validity Reliability and validity are important components of content analysis. As Krippendorff (1989) claims “content analysis is a research technique for making replicable and valid inferences from data to their context” (403). This means that techniques are expected to be reliable in the sense of replicability and research should produce valid results (Krippendorff 2004, 18). In regards to this study, explicit context and framework for the content analysis are provided through the discussion of literature and the allocated Virality Theoretical Model. A limitation to validation in content analysis is inherent in the sense that content analysis attempts to make inferences from what is not directly observable (Krippendorff 1989, 407). In terms of this study, a certain degree of validity can be claimed to be ensured by using well defined criteria, by cross-referencing the analyzable data with the intentions of the creator (using the creative brief) and by supporting the analysis through statistical facts retrieved through Google analytics.
34
5. Results and analysis This section presents the results of this study and discusses them in context with the theoretical framework and the overall aim of the study, i.e. to assess which factors of this study’s theoretical model can be traced in a specific actual ad and whether this ad went viral or not. First, the issue of whether the ad went viral or not is addressed and then the assessment of the ad according to the Virality Theoretical Model follows. Once the ad has been assessed a summary of the thesis’s results along with some additional comments is presented. 5.1 Virality assessment results and discussion The data about the ad’s views were collected and compiled into graphs following the practical guide thoroughly presented in 4.1 Operationalization of virality concept and data collection in the Methodology section. The first pattern that was sought in the ad’s views was the sigmoid curve pattern. This was sought in the views that the ad achieved the first ten days following its launch. As one can clearly see in Figure 4, the views the ad “Nya Tunnelbanan” achieved during this period do not demonstrate the sigmoid curve pattern (see Figure 2 for a reference).
Figure 4: Cumulative views of “Nya tunnelbanan”
35
The second pattern that was sought in the ad’s views was the power law distribution; this distribution was specifically sought in the rate of decline of the views. The power law distribution cannot be seen in the pattern of the graph in Figure 4 depicting the first ten days of views (see Figure 3 for reference). Nevertheless, the views were monitored, as mentioned in the methodology section, for a longer period of time, in order to establish when the decay period of the ad was and draw possible conclusions that can help understand the life cycle of this specific ad. As one can clearly see in Figure 5, the decay in views for the ad “Nya Tunnelbanan” happened more than two months after its launch, around the 29th of September 2015.
Figure 5: Daily views from Google analytics Looking at both of the graphs (Figure 4 and 5) one can easily conclude that the ad “Nya tunnelbanan” did not go viral according to the criteria of this study that follow the definition by Nahon and Hemsley (2013), and are thoroughly discussed in the theoretical framework. The sigmoid curve pattern was not achieved, which means that the ad did not start with a slow growth in views, then switch to a fast spread among people and further, when almost everyone had already seen it, start fading, resulting in slow growth once again. Moreover, a power law distribution in the decay rate would suggest that the acceleration in views was fast enough for the ad to be considered viral (Nahon & Hemsley 2013, 26). In the case of “Nya tunnelbanan” a power law distribution could not be pinpointed during the first ten days. In fact, the decay happened more than two months after the initial launch of the ad,
36
when the views dropped drastically. This can be explained by the fact that the ad was promoted, i.e. there was a YouTube campaign for it. According to Populate, the ad campaign started with the launch of “Nya tunnelbanan” on YouTube on the 18th of August and finished on the 1st of October 2015. This means that during these dates internet users watched the ad on YouTube, because it was played as promoted content before some other content. Having this piece of information one can assume that most of the views may have been a result of having to watch the ad before moving on to other content and not from sharing between people. So it may be concluded that the power law distribution picture we get in Figure 5 after the two months of the launch is because the promotion ended and not because the spread had reached its peak, accelerated fast enough and then slowed down again as one would expect in the life cycle of a viral phenomenon. All in all the data collected for “Nya Tunnelbanan” do not indicate that the ad went viral, as the views do not indicate a sigmoid curve distribution. Moreover, the decay in views started long after the designated period of monitoring (i.e., ten days) and, as it seems, for other reasons than the natural course of a viral phenomenon. Why this ad did not reach virality may be explained, if further assessed according to this study’s theoretical model. In the next section the factors of the Virality Theoretical Model, as pinpointed in this study, are thoroughly discussed in relation to the ad “Nya Tunnelbanan”. 5.2 Virality Theoretical Model assessment results and discussion The factors of the Virality Theoretical Model are analyzed in relation to the ad “Nya Tunnelbanan” in the order they are presented in section 3.3 A Virality Theoretical Model. 5.2.1 “Nya Tunnelbanan” and emotions According to this study’s Virality Theoretical Model, content should elicit emotions in order to have more chances to go viral. Various scholars refer to positivity being a driver for virality, while negative emotions, that are characterized by high arousal (like anger or anxiety), are claimed to also boost virality (Phelps et al. 2004; Berger & Milkman 2012; Nahon & Hemsley 2013). The emotions that are sought in the ad “Nya tunnelbanna” and are studied in this thesis, are the basic emotions, as defined by Ekman (1972, 29), i.e. surprise, fear, sadness, happiness, disgust and rage. The creative brief (see appendix 1) of the ad clearly states that the goal is for it to make viewers feel happiness. More specifically the answer to the question “what do we want the target audience to feel?” is “Happy that the metro is developing.” It can be claimed
37
that the metro expansion and the residential development is something that can create feelings of happiness and joy, but the way this information is given in the ad by an enumeration of facts about the process and results, does not elicit emotions of happiness in itself; the ad refers to an event that may cause positive emotions, but the ad per se is not a joyful event. This means that if a viewer is not going to benefit from the metro expansion and the subsequent residential development, they will probably not feel happiness watching this ad. To sum up, it can be claimed that the ad “Nya tunnelbanan” elicits the emotion of happiness for the viewers that are in some way positively affected by or interested in the metro expansion and residential development. Fear, sadness, disgust or rage are not evidently elicited in “Nya Tunnelbanan”; both audio and video follow an agreeable, easygoing, factual and serious tone as stated in the creative brief (see appendix 1). No negative, sad, fearful or disgusting pictures or words are used throughout the ad (see audio transcript and description of the ad in appendix 3 and 4). Surprise will be assessed as an individual factor in the Virality Theoretical Model of this study and its analysis and discussion follows in the next section as a separate point. 5.2.2 “Nya Tunnelbanan” and surprise One of the characteristics content should display to have more chances to go viral is that of surprise, combined with some other primary emotion (fear, sadness, happiness, disgust or rage), and even if the emotion is negative at the beginning, it should always end up in a positive feeling. Surprise can be considered as the feeling resulting from something unexpected happening; something that gives the storyline of the narration a turn that is not expected by the viewer. When it concerns the ad “Nya tunnelbanan” it can be claimed that there is no surprise element in the narration. As one can read in the creative brief (see appendix 1), the primal goal of the ad is to inform the public about the metro expansion and the tonality should be easygoing and approachable but factual and serious. The result in the actual ad seems to follow this guideline well; the narration has a clear informative character, both in video and audio presentation with the narrator explaining facts and numbers. The overall feeling can be described as agreeable and positive, but one can not detect any kind of surprise. More specifically, there is no “narrative stunt” or “real stunt” that seem to be the more common types of surprise used in videos that have gone viral (Dafonte-Gomez 2014, 204). Moreover, there cannot be detected any emotional roller coaster pattern with waves of tension and relief, a feeling similar to how suspense movies work (Teixeira 2012, 26).
38
Finally, as already discussed in the previous section, while the ad may be claimed to elicit happiness for those affected by or interested in the metro expansion and residential development, no other emotion can be detected. Nevertheless, according to this study’s Virality Theoretical Model the content should display surprise, combined with some other primary emotion (fear, sadness, happiness, disgust or rage), but since the surprise factor can not be detected, the second factor of this model cannot be verified. 5.2.3 “Nya Tunnelbanan” and interesting content According to the Virality Theoretical Model of this thesis the content should be interesting and unique or novel to have more chances to go viral. Uniqueness or novelty can be considered as not cliché. It is a storyline that is not common or seen before. Interesting can be considered a storyline that intrigues the viewer enough that they want to know more and can maintain the viewer’s attention. The ad “Nya Tunnelbanan” is an animated film and the storyline, as already discussed, is of clear informative character, following the guidelines of the creative brief (see appendix 1). Therefore, the storyline or the way it is presented cannot be considered novel or unique; the information flows as a simple narration of facts and the animation can be considered of high quality, but not something one has not seen before. When it concerns whether the ad is interesting or not, one can assume that it is interesting for the viewers that are affected by the metro expansion and subsequent residential development. Nevertheless, to have a clear picture of whether it is interesting or not, the average viewing time of the ad can be assessed, with a longer viewing time meaning that the content is interesting enough to maintain the viewer’s attention. The relevant data was obtained through Google Analytics for the period that the ad campaign lasted, i.e. from the 18th of August until the 1st of October 2015. Most of the views, specifically 97.3% of them, during this period came from the ad promotion campaign on YouTube (94%) and from external websites (3.4%). The data show that the average view length coming from the ad promotion campaign on YouTube is 1:50 minutes and the average view length coming from external websites is 2:30 minutes. Considering the video ad lasts 3:13 minutes, it can be claimed that the viewers lost interest before the storyline was completed and all information communicated. This indicates content of low interest level.
39
5.2.4 “Nya Tunnelbanan” and good presentation A factor that can help boost virality is good presentation of the content. According to the Virality Theoretical Model a good presentation might include a short commentary, a url or the use of a proper hashtag (#). The ad “Nya Tunnelbanan” on the YouTube channel is indeed accompanied by a short commentary that says: “In little more than three minutes the entire major metro expansion in the county is summarized. Do you want to know where the metro will go, when it will be built and why? Here are the answers. More information about the building of the metro at http://nyatunnelbanan.sll.se”31. This commentary is short and communicates well the video content. Moreover, it is prompting and tries to raise interest and curiosity regarding the content of the video by posing questions. It also gives a url, where one can find out more about the metro expansion. It can, therefore, be considered as an appropriate presentation for the specific video. 5.2.5 “Nya Tunnelbanan” and high quality When talking about “high quality”, the Virality Theoretical Model in this study refers to how the narrative is built, manipulating the emotion elicitation and attracting people’s attention. In this light, it can be claimed that the narrative of the ad “Nya Tunnelbanan” is not of high quality. The narration is of informative character, which services the purpose of the creative brief (see appendix 1) to communicate facts and benefits from the metro expansion. Emotions, as already discussed, are not considered to be elicited, except happiness and that, only for the viewers that are in some way affected by or interested in the metro expansion and subsequent residential development. To sum up, the narration follows a linear storyline that enumerates facts and discusses benefits for the target audience; neither the audio nor the video of this ad may be claimed to have been skillfully produced in a way to manipulate emotion elicitation (Nahon & Hemsley 2013, 64). 5.2.6 “Nya Tunnelbanan” and content that resonates According to this study’s Virality Theoretical Model, content that resonates with people’s feelings about a situation has better chances of going viral, as people can relate to it. This 31 Text translated by the author. Original text: “På lite drygt tre minuter sammanfattas hela den stora tunnelbaneutbyggnaden i Stockholms län. Vill du veta vart tunnelbanan ska gå, när den ska byggas och varför? Här får du svaren. Mer information tunnelbanans utbyggnad hittar du på http://nyatunnelbanan.sll.se”
40
specific ad is about the metro expansion and the residential development that starts in 2016 and is estimated to finish, at least according to the ad’s audio, in 2025 (see transcript of audio in appendix 3). It can be claimed that this ad resonates with the target audience’s feelings in the sense that Stockholm suffers from a housing shortage; one can read in various newspaper articles32 about the need to address Stockholm’s housing problem. It can, therefore, be claimed that the expansion of the metro that further involves residential development is a kind of content that resonates with the target group’s (i.e. Stockholm residents, new SLL employees and the Media- see creative brief in appendix 1) feelings. 5.2.7 “Nya Tunnelbanan” and context Context is also significant in increasing the chances for content to go viral. As already discussed, when particular content is important in the moment, it is more likely to go viral (Nahon & Hemsley 2013, 66). In the previous section about the ad’s resonance it has been pointed out that the content of the ad, meaning the information about the expansion and the subsequent residential development, is of high importance for the target group. Nevertheless, the ad refers to the building process as one that starts in 2016 with a planned delivery date in 2025. This is too far in the future to be considered important in the moment. In this light, it could be predicted that the video would be considered important in the moment in 2025, a fact that could possibly boost viewing and sharing at that point in time. 5.2.8 “Nya Tunnelbanan” and branding Branding can be considered the presence of the brand in the form of a logo of an advertised product or service (Teixeira 2012, 25). As already discussed about branding and virality, the ideal situation is when branding is discreet, because the more viewers feel that someone tries to persuade them to buy or use a service through an intrusive logo, the more likely it is they will resist watching the ad (ibid., 25). As one can clearly observe in the “Nya Tunnelbanan” ad33, the logo of the SLL (Stockholms Läns Landsting, i.e. Stockholm’s County Council) is discreetly displayed for a few seconds at the beginning of the ad in one corner of the animated map. It then reappears at the closing scene for a few seconds again, only this time it occupies the whole screen as a 32 http://qz.com/264418/why-its-nearly-impossible-to-rent-an-apartment-in-stockholm/ , http://
www.thelocal.se/20150324/more-homes-are-needed-to-solve-housing-shortage, both accessed 01 May 2016 33 https://www.youtube.com/watch?v=n1f3e1Clhus, accessed 30 August 2015
41
signature, a very common way to close an ad, the so called “Pack Shot” (usually displaying product, logo and slogan). So, in the case of “Nya Tunnelbanan” it can be claimed that branding is discreet. Nevertheless, it is clear that the specific ad does not try to sell any product. Its purpose is to inform the public and any other interested actors about the metro expansion. It can, yet, be claimed that it also tries to build a good image for SLL that drives the metro expansion; one can read in the creative brief (see appendix 1) that the target audience should feel appreciative that the metro is developing, and the ad should put the expansion in perspective in order to show the end benefits for the public. This can be considered as a persuasive aspect, so in this light, it can be said that discreet branding is of importance in the assessment of this ad and is indeed implemented. Moreover, Teixeira (2012) clearly states that a perfect situation is when the brand is discreetly weaved through the storyline and the content remains interesting even if the brand is to be omitted (26). The question of whether the content is interesting has already been previously addressed, but, nevertheless, it is worth remarking that the removal of the branding seems not to affect the storyline of the specific ad. This means that even if it was omitted, the storyline would stay the same. 5.2.9 “Nya Tunnelbanan” and interest network As explained in the Theoretical Framework section if content is interesting for a specific group of people, there is a possibility that it can go viral in a specific interest network, i.e. a network created around a specific topic (Nahon & Hemsley 2013, 71). The ad “Nya Tunnelbanan” is about the metro expansion and the subsequent residential development of Stockholm, a topic that concerns a specific target group; the ones that are affected by these developments and are specified as (a) the Stockholm residents, (b) the SLL employees and (c) the Media in the creative brief (see appendix 1). Nevertheless, as already discussed, the content cannot be characterised as being of a high interest level, due to the short viewing time the data from Google Analytics suggest (see 5.2.3 “Nya Tunnelbanan” and interesting content for elaborate explanation). Though the target group was exposed to the ad through a promotional campaign, the interest it raised was not high enough to create an interest network in which the ad could have gone viral, without necessarily exhibiting the other characteristics (positive emotional impact, quality, surprise, resonance, context etc.) so far dicussed (Nahon & Hemsley 2013, 68) .
42
5.2.10 “Nya Tunnelbanan” and gatekeepers According to the Virality Theoretical Model getting gatekeepers or influencers, e.g. online/offline media, popular blogs or people with many contacts and influence, to share content might boost virality. Populate itself has not access to this specific information about “Nya Tunnelbanan”, nevertheless, conclusions can be derived from relevant data from Google Analytics. Through Google Analytics it can be determined how many views were generated through external sites (i.e. not the YouTube campaign) and which external sites these were. The data on views for the period between 18th of August until the 28th of August 2015 show that 92% of the total views were generated from the video ad campaign, while external sites represented only 4.6% of the views. This 4.6% - among others - further breaks down to 30% of views coming from stockholmdirekt.se, 17% coming from sll.se, 18% from Facebook, 3.8% from blogspot.se and 3.3% form fastighetsvarlden.se (see Figure 6). Among these sites stockholmdirekt.se can be considered a gatekeeper; as it is stated in its web page “Stockholm Direct is a news site with local news from all over Stockholm.”34 and “Every Saturday throughout the year, we reach almost 850,000 readers.”35 This news site shared a story about the metro expansion on the 23rd of August 201536 and as it can be seen in Figure 6 it generated 3,023 views. The sharing of the “Nya Tunnelbanan” by stockholmdirekt.se can also possibly explain the raise in views that can be seen in Figure 5 on the 23rd of August.
Figure 6: Views from external sites / 18-28 Aug 201537
34 http://www.stockholmdirekt.se/om-oss/, translation by the author, accessed 01 May 2016 35 http://www.stockholmdirekt.se/foretagsannonser/, translation by the author, accessed 01 May 2016 36 http://www.stockholmdirekt.se/nyheter/exklusivt-se-filmen-som-visar-nyatunnelbanan/aRKohr!ijVYIjDAn@HjgFLOebjUUw/, accessed 01 May 2016 37 Source: Google Analytics
43
The overall percentage 4.6% of views coming from the external sites, though, is not considered high enough to conclude that the sharing of the video by external sources, including possible gatekeepers like stockholmdirekt.se, boosted views considerably enough in order for the content to reach virality. 5.2.11 “Nya Tunnelbanan” and affinity groups According to the Virality Theoretical Model the content should initially be shared with people that are receptive to the sender, e.g. with an affinity group, an act that might help promote virality. In order to assess if the ad “Nya Tunnelbanan” was shared by sources which the target group is receptive to, data retrieved from Google Analytics were used. A source that can be characterised as being receptive to is the company’s SLL website; it can be considered a website that the target group willingly visits to get information about the metro and about housing. Looking again at Figure 6 it can be concluded that the views generated by this external source (1,661 in total) - representing only a small percentage of total views to that point of time and are negligible. To sum up, though the ad “Nya Tunnelbanan” was shared by sll.se38, which can be considered as a source that the target group is receptive to, the amount of generated views cannot be claimed to have increased views to the point that the content could reach virality.
38 http://www.sll.se/verksamhet/kollektivtrafik/Aktuella-projekt/Nya-tunnelbanan/Nyheter/2015/08/Allt-omnya-tunnelbanan-pa-tre-minuter/
44
6. Conclusion The aim of this study was to contribute to and understand the topic of virality in marketing through the articulation of a Virality Theoretical Model, a discussion of the model in relation to an actual ad and an assessment of whether the model can indeed lead to virality. The Virality Theoretical Model summarizes the most prominent existing literature on the topic of virality, identifying the main factors that different scholars have already recognized as possible drivers of virality phenomena. For the purposes of this thesis, an elaborate presentation of the definition of virality as described by Nahon & Hemsley (2013) was also given, in order to establish if the ad used in this thesis’ empirical study reached virality. Qualitative content analysis, together with statistical data from Google analytics and relevant information from the company that produced the ad were used to draw conclusions about whether the ad went viral and if it fitted the Virality Theoretical Model of this thesis. Results indicated that the ad “Nya Tunnelbanan” that was used in the empirical study did not go viral (RQ 1). It was also concluded that the ad did not include all the factors enumerated in the Virality Theoretical Model. More specifically the ad was not found to elicit emotions or surprise, other maybe than joy; nevertheless, the ad per se was not considered as a joyful event. The ad was assessed as not being interesting enough, novel or unique and was not of high quality in the sense that it lacked skillfull production in the direction of emotion elicitation. It was, furthermore, assessed as not important in the moment. Moreover, though the ad resonated with a specific target group it was assessed as not interesting enough to create an interest network and become viral within it. On the contrary, the ad was assessed to be well presented, it displayed resonance to a specific target group and had discreet branding. In terms of dissemination, it was found that the ad was indeed shared with what can be considered an affinity group through the SLL site and at least one gatekeeper shared the ad. To sum up, the ad “Nya Tunnelbanan” was found to only encompass five (good presentation, resonance, discreet branding, shared with an affinity group and by at least one gatekeeper) out of the eleven factors that together form the Virality Theoretical Model of this thesis (RQ 2). Theoretically, as well as for marketing executives and advertising practitioners, it can be said that this thesis’ contribution to the field of virality is that it makes it explicit that virality is a complex phenomenon, depending on several different factors involving both content characteristics and dissemination. Moreover, this thesis offers an easy to use practical guide for practitioners in the form of the Virality Theoretical Model, which when followed may increase the chances of virality. Though the ad encompassed almost half of the factors of
45
the Model, it did not reach virality. This may lead to the conclusion that the factors that were not included in the ad have greater importance in increasing the chances of virality. The fact that the ad included all of the dissemination factors from the second category of the Model (the role of Personal Influence and Gatekeepers) may lead to the same conclusion. Practically, this may mean that Content Characteristics, and more specifically emotion elicitation, surprise, interesting, unique content that is of high quality and that is important in the moment, are a sine qua non for virality to happen (RQ 2a and 2b). All in all it must be noted that a major limitation for this study is the fact that the Virality Theoretical Model of this thesis is not fully implemented in the empirical study’s ad. This could not happen due to temporal and technical restrictions, such as the author not having been allowed to participate in the making of the ad, so as to make sure the Model was fully implemented. This brings on an opportunity for further research, i.e. an empirical study by which this study’s Virality Theoretical Model will be fully implemented in an actual ad in order to explore its full potential. Moreover, to more accurately answer RQ 2a and 2b further research can also take place to more expicitly assess which of the factors articulated in the Virality Theoretical Model are more important in achieving virality. This could entail empirical research using separate ads that include certain factors of the Model at a time and yielding proper inferences by making comparisons.
46
References Bakshy, E., J. M. Hofman, W. A. Mason, and D. J. Watts (2011). Everyone’s an influencer: quantifying influence on twitter. In Proceedings of the fourth ACM international conference on Web search and data mining, pp. 65–74. ACM. Bakshy, E., I. Rosenn, C. Marlow, and L. Adamic (2012). The role of social networks in information diffusion. In Proceedings of the 21st international conference on World Wide Web, pp. 519–528. ACM. Barabasi, A.-L. (2003). Linked. New York: Plume. Bennett, W. L. and A. Segerberg (2012). The logic of connective action: Digital media and the personalization of contentious politics. Information, Communication & Society 15 (5), 739–768. Berger, J. and K. L. Milkman (2012). What makes online content viral? Journal of marketing research 49 (2), 192–205. Burgess, J. (2008). All your chocolate rain are belong to us? Viral videos, YouTube and dynamics of participatory culture,(ed) Video Vortex Reader: Responses to YouTube. Institue of Network Cultures, Amsterdam, 101–109. Castells, M. (2011). Network theory— a network theory of power. International Journal of Communication 5, 15. Dafonte-Gomez, A. (2014). The key elements of viral advertising. from motivation to emotion in the most shared videos. arXiv preprint arXiv:1505.02002 . Daymon, C. and I. Holloway (2010). Qualitative research methods in public relations and marketing communications. Routledge. Dobele, A., A. Lindgreen, M. Beverland, J. Vanhamme, and R. van Wijk (2007). Why pass on viral messages? because they connect emotionally. Business Horizons 50 (4), 291–304. Eckler, P. and P. Bolls (2011). Spreading the virus: Emotional tone of viral advertising and its effect on forwarding intentions and attitudes. Journal of Interactive Advertising 11 (2), 1–11. Ekman, P. and W. V. Friesen (2003). Unmasking the face: A guide to recognizing emotions from facial clues. Ishk. Erving, G. (1959). The presentation of self in everyday life. Doubleday Anchor Books. Gladwell, M. (2000). The tipping point: How little things can make a big difference. Little, Brown.
47
Granovetter, M. (1983). The strength of weak ties: A network theory revisited. Sociological Theory 1, 201–233. Huberman, B. A. and L. A. Adamic (2004). Information dynamics in the networked world. In Complex networks, pp. 371–398. Springer. Kaplan, A. M. and M. Haenlein (2011). Two hearts in three-quarter time: How to waltz the social media/viral marketing dance. Business Horizons 54 (3), 253–263. Katz, E. (1957). The two-step flow of communication: An up-to-date report on an hypothesis. Public opinion quarterly 21 (1), 61–78. Kirby, J. and P. Marsden (2012). Connected marketing. London: Routledge. Kitsak, M., L. Gallos, S. Havlin, F. Litjeros, L. Muchnik, H. Stanley, and H. Makse (2010). Identifying spreaders in complex networks. Nature Physics 6, 888–893. Krippendorff, K. (1989). Content analysis. Krippendorff, K. (2004). Content analysis: an introduction to its methodology 2nd edition. Kwak, H., C. Lee, H. Park, and S. Moon (2010). What is twitter, a social network or a news media? In Proceedings of the 19th international conference on World wide web, pp. 591–600. ACM. Lindstrom, M. (2009). Essential elements of viral video success. Advertising Age 80. Miller, V. (2011). Understanding digital culture. Sage Publications. Nahon, K. and J. Hemsley (2013). Going viral. Polity. Petrovic, S., M. Osborne, and V. Lavrenko (2011). Rt to win! predicting message propagation in twitter. In ICWSM. Phelps, J. E., R. Lewis, L. Mobilio, D. Perry, and N. Raman (2004). Viral marketing or electronic word-of-mouth advertising: Examining consumer responses and motivations to pass along email. Journal of advertising research 44 (04), 333–348. Purcell, K. (2010). The state of online video. Pew Internet & American Life Project Washington, DC. Rogers Everett, M. (1995). Diffusion of innovations. New York , 12. Rosen, E. (2000). The anatomy of buzz: How to create word-of-mouth marketing.
48
Shifman, L. and M. Blondheim (2010). The medium is the joke: Online humor about and by networked computers. New Media & Society, 1461444810365311. Shirky, C. (2000). The toughest virus of all. Business 2.0 . Smith, C. A. and P. C. Ellsworth (1985). Patterns of cognitive appraisal in emotion. Journal of personality and social psychology 48 (4), 813. Suh, B., L. Hong, P. Pirolli, and E. H. Chi (2010). Want to be retweeted? large scale analytics on factors impacting retweet in twitter network. In Social computing (socialcom), 2010 ieee second international conference on, pp. 177– 184. IEEE. Teixeira, T. (2012). Five techniques can help companies make commercials that people will watch and share. Harvard Business Review March 2012 (March 2012), 25–27. Wallsten, K. (2010). yes we can: How online viewership, blog discussion, campaign statements, and mainstream media coverage produced a viral video phenomenon. Journal of Information Technology & Politics 7 (2-3), 163–181. Wu, F. and B. A. Huberman (2007). Novelty and collective attention. Proceedings of the National Academy of Sciences 104 (45), 17599–17601.
49
APPENDIX Appendix 1: Creative Brief39 Customer: Administration for expansion of metro, SLL Project: Expansion of the metro Revised: 5/3 2015 1. Background – Please describe the project. Producing a basic film in two versions - a short (about 1 minute) and a longer one (max 5 min.). The end result should be films that withstand through time and can easily give a good overview of the forthcoming expansion of the metro. 2. Challenge –what is it we want to achieve? - Films that are used and are seen. - Coherent message in an easygoing way. - Place the metro expansion in a larger perspective. - Do not immerse into the technical details and construction plans, but focus on customer benefits. 3. The role of communication – How do films contribute to solving our challenge? The short film will answer the questions “Why build?”, “What will happen?”, “When does it start and when will it be finished?”. The long film will answer the same questions as the short, but will go somewhat deeper in detail and talk briefly about each project (4 parts) and about the routes. Long film may optionally contain elements that point out: -
Existing Metro traffic will not be affected.
-
Work with the additions will be somewhat noticeable.
-
Almost all the work will be done underground.
-
Any information about the metro profile.
4. Target audience – who we are talking to and what we know about them that is
39 Translation
by the author.
50
relevant? -
The general public, i.e. Stockholm residents – it affects their everyday lives.
-
New employees – to be able to quickly get an overview and familiarize themselves with the project.
-
Media – we want them to follow the trend because of the project’s public interest.
5. Change – what do we want the target audience to know, feel and do? Short film Know: The metro will have three new parts. Feel: Happy that the metro is developing! Do: Watch the long movie if they want more information and visit [URL]. Long film Know: Stockholm is growing and the metro as well, 3 new parts will be clear by 2025. Feel: The metro is a “main artery” that simplifies life and contributes to development. Do: Subscribe to the channel for future videos/information. 6. Main messages – what do we want the target audience to primarily remember? Short film: the metro will have three new parts. Long film: Stockholm metro is growing, which contributes to the simplification of everyday life and development. 7. Personality / tonality – how do we want to be perceived as the sender? Easygoing and approachable but factual and serious. Narration from the public’s perspective. 8. Distribution – how should films be used and complement to other communication? Online conferences and lectures, field trips, new employees, the media. 9. Objectives and effect – how can we measure the impact of the project? -
The films should be watched, be brief and be easily understood.
-
The films will help explain complicated things (e.g. tracks’ profile and depth).
-
Movies that will withstand through time, at least a couple of years.
-
Other online goals and metrics should be discussed with the administration’s web coordinator.
51
10. Spread -On which sites will the films be featured? SLL sites: [URL] -Collaboration opportunities: Local government sites SL.se SL’s YouTube Channel -YouTube YouTube Channel (administration’s) or SLL’s YouTube? Decision is to continue the dialogue on online optimization, advertising etc. along with the administration webmaster. 11. Other The style will be based on graphic design animation.
52
Appendix 2: Swedish transcript of audio of the ad “Nya Tunnelbanan”40 “Stockholms län är en av de snabbast växande storstadsregionerna i Europa. Vi blir 35 000 40 000 fler varje år. Det motsvarar två fullsatta SL bussar om dagen vilket ställer krav på fler bostäder och bättre resmöjligheter. Tunnelbanan har idag runt en miljon påstigande varje dag och ska byggas ut för att fler ska kunna resa kollektiv och för att 78 000 nya bostäder ska kunna byggas. Satsningen är så viktig att staten, landstinget och kommunerna som får ny tunnelbana bekostar de drygt 25 miljarder kronor som utbyggnaden kostar. I nordväst förlängs Blå linje från Akalla via Barkarbystaden till Barkarby station i Järfälla, som blir en knutpunkt där man kan byta mellan t-bana, buss och pendeltåg. I Barkarbystaden byggs en ny stadsdel med bostäder och verksamheter. Den nya tunnelbanan är en förutsättning för den utvecklingen. Från Odenplan dras en ny Gul linje via Hagastaden och Nya Karolinska Solna till Arenastaden med bostäder, arbetsplatser, handel och evenemang. Odenplan blir en viktig knutpunkt för buss, tunnelbana och pendeltåg vilket avlastar T-Centralen. Nya stadsdelen Hagastaden får också en tunnelbanestation med en uppgång rakt in i Nya Karolinska Solnas huvudentré. I sydost förlängs Blå linje från Kungsträdgården till Nacka Centrum via en ny station vid Sofia kyrka på Södermalm samt en station under Hammarby kanal med uppgångar på båda sidor. I Nacka byggs stad längs hela den nya tunnelbanesträckningen. Sickla blir en knut för tunnelbanan, Saltsjöbanan, Tvärbanan och bussar. I Järla når man Saltsjöbanan. Sista stopp är Nacka Centrum med bostäder, arbetsplatser, handel, kultur och samhällsservice. Blå linje får en sydlig gren via Sofia och Gullmarsplan till Sockenplan där den knyts samman med Gröna linjen till Hagsätra. Det möjliggör tätare trafik från hela Söderort och därmed möjlighet att bygga många nya bostäder. Den nya Blå linjen kommer att avlasta tunnelbanenätet som ett alternativ till den idag hårt belastade sträckan mellan Slussen och TCentralen. Nya tunnelbanan byggs under jord och konkurrerar inte med andra trafikslag. Den är snabb, smidig och miljövänlig. Ovanpå finns det plats att bygga både bostäder och arbetsplatser. 2016 börjar de första delarna byggas och 2025 ska allt vara klart. Nya tunnelbanan knyter ihop Stockholmsregionen och förenklar vardagen för alla oss som lever här - nu och i framtiden.”
40 Transcript
taken from Youtube page featuring the video: https://www.youtube.com/watch?v=n1f3e1Clhus, accessed 30 March 2016.
53
Appendix 3: English transcript of audio of the ad “Nya Tunnelbanan”41 “Stockholm is one of the fastest growing metropolitan regions in Europe. The population is increasing by 35 to 40,000 people each year, the equivalent of two full buses every day. That’s why we need more housing and improved public transport. The metro currently carries more than one million travellers a day and expansion is underway to accommodate even more passengers as well as to allow the construction of 78,000 new homes. This investment is of such importance to the future of the region that local and county councils have joined with the national government to fund the project at a cost of 25 billion Swedish crowns. In the northwest the Blue Line will be extended from Akalla through Barkarbystaden to Barkarby station in Järfälla, which will become a key intersection for passengers transferring between the metro, bus and commuter train services. At Barkarbystaden a new residential and business neighborhood will be built. The new subway will be integral to the development. A new Yellow Line will run from Odenplan via Hagastaden and via Nya Karolinska Solna to Arenastaden, with its housing, commercial, retail developments and events center. Odenplan will become a key intersection for changing between buses, the metro and commuter trains, which will relieve the Central Station. The new Hagastaden district will also have a metro station with an escalator leading directly into the main entrance of Nya Karolinska Solna. In the southeast, the Blue Line will be extended from Kungsträdgården to Nacka Centrum via a new station at Sofia Kyrka in Södermalm, and an additional station will be built under the Hammarby canal with escalators exiting on both sides of the water. In Nacka, new developments will be built along the length of the new metro line. Sickla will become an important intersection between the metro, the commuter train, the light rail line and the buses. At Järla one can reach the commuter train. The last stop is Nacka Centrum with housing, commercial, retail and community services. The Blue Line will also have a southern branch via Sofia and Gullmarsplan to Sockenplan, where it will connect with the Green Line to Hagsätra. This will allow for more frequent services from all over the southern suburbs and for the construction of a lot of new housing. The new Blue Line will relieve the metro network as an alternative to the currently overburdened section between Slussen and TCentralen. The new subway will be built underground and will not compete for space with other types of transportation. Fast, convenient and environmentally friendly, it will open up space for the construction of new housing and workplaces. Building will start in 2016 and 41 Translation by the author
54
completion is planned for 2025. The new metro will bring the Stockholm region closer together and it will make everyday life easier for everyone who lives here – now and in the future.”
Appendix 4: Description of video of the ad “Nya Tunnelbanan”42 The ad starts with the camera moving over an engineer's table where a map of Stockholm lies. We can see buildings gradually being built up with animation technique in different parts of Stockholm. Then the buildings disappear and the map turns into Stockholm’s subway map as it looks today in white background. The next scenes show closeups of Stockholm neighbourhoods, where we can see regular traffic of cars, buses and people (all animated) and new buildings being built up at empty spaces. At the same time, through an incision to the ground, we can see the subway train passing by underneath the ground. Then we can see the subway map once again but now the extensions of the existing subway lines appear on the map to show the new stations. We can also see new neighborhoods “growing out” of the map, i.e. being built up with animation technique, as separate pieces of ground. We can see the subway train going from one to another connecting the different neighborhoods through tunnels. At a separate part of the engineer’s table we can see the present connection between Slussen and T-Centralen and with an incision to the ground we can also see the future subway train that will connect the two stations. We can then see a vertical incision of the ground; under the ground and in a tunnel a train is passing by, while over the ground new buildings are being built. The camera then zooms in the train and the film becomes a real life video of a train stopping at a station. There we can see real people getting on and off the train. The ad closes with a frame of Stockholms Läns Landsting’s logo.
42 For a better understanding, please watch the ad at https://www.youtube.com/watch?v=n1f3e1Clhus, accessed 30 March 2016
55
