Description
Viral marketing, viral advertising, or marketing buzz are buzzwords referring to marketing techniques that use pre-existing social networks and other technologies to produce increases in brand awareness or to achieve other marketing objectives
Stockholm School of Economics Institute of Marketing and Strategy Master Thesis Spring 2009
Content is King
Communication Effects of Viral Video Marketing
The technological paradigm shift has enabled new forms of communication. Consequently, this offers valuable possibilities for marketers in terms of reach, speed and cost-efficiency. Due to similarities with the spread of a virus, the phenomenon of self-replicating content on the web via social networks and word-of-mouth is commonly referred to as viral marketing. Even though this phenomenon is becoming more frequently used by companies to increase brand awareness and improve brand attitude, there is little academic research on the subject. The objective of this study is to identify factors that affect consumer attitude and behaviour, following the exposure to a viral video, as well as studying how these features interact. In contradiction to previous research on adjacent areas of study, the result indicates that contextual factors do not influence neither the attitude towards the message nor the intention to share it with others. Instead, content has proven to be the single most important feature to improve attitudes and consequently viral circulation. Conclusively, a communication effects model for viral video marketing is presented, illustrating the flows of consumer attitude and behaviour in a digital context. Keywords: viral marketing, online videos, digital communication
Authors: Lena Rune, 20263 Hélène Vinberg, 20149
Tutors: Professor Micael Dahlén PhD Student Jonas Colliander Examiner: Associate Professor Patric Andersson
Discussants: Karin Bjerde, 20264 Jakob Ståhl, 20693
Presentation: June 12, 2009, 15.15-17.00 SSE Room C606
Content is King
Rune & Vinberg
Many Thanks to:
Micael Dahlén and Jonas Colliander, Stockholm School of Economics Jan Enander and Annelie Önnerud Åström, YouGov Sweden Marcus Wenner, Désirée Maurd and Caroline Jungsand, United Minds/Prime PR Fredrik Törn and Niclas Öhman, Stockholm School of Economics Hilma Gustavsson, Johan Östberg and Filippa Malmegård Fredrik Egger and Magnus Koch-Schmidt
1
Content is King
Rune & Vinberg
TABLE OF CONTENTS
1. INTRODUCTION.................................................................................................................................. 4 1.1 Background .......................................................................................................................................... 5 1.2 Problem Area ....................................................................................................................................... 6 1.3 Purpose ................................................................................................................................................ 6 1.4 Delimitations ........................................................................................................................................ 6 1.5 Definitions ........................................................................................................................................... 7 1.6 Previous Research and Expected Contribution ..................................................................................... 9 2. THEORETICAL FRAMEWORK ....................................................................................................... 9 2.1 Communication Effects Model ........................................................................................................... 11 2.2 Context .............................................................................................................................................. 12 2.2.1 2.2.2 Social Proof.............................................................................................................................. 12 Sender-Recipient Relationship .................................................................................................. 13
2. 3 Content ............................................................................................................................................. 14 2.3.1 2.3.2 Attitude .................................................................................................................................... 14 Behaviour ................................................................................................................................ 18
3. METHODOLOGY ............................................................................................................................... 24 3.1 Choice of Subject ............................................................................................................................... 24 3.2 Scientific Approach ............................................................................................................................ 24 3.3 Experimental Design .......................................................................................................................... 25 3.3.1 3.3.2 3.3.3 Sample of Respondents ............................................................................................................ 25 Sample of Videos ..................................................................................................................... 26 Construct of Survey.................................................................................................................. 28
3.4 Pre-studies .......................................................................................................................................... 30 3.4.1 3.4.2 Classification of Products ......................................................................................................... 30 Pre-test Survey ......................................................................................................................... 31
3.5 Analytical Tools .................................................................................................................................. 31 3.6 Research Quality ................................................................................................................................. 31 3.7 Sources of Information ....................................................................................................................... 32 4. ANALYSIS AND RESULTS ................................................................................................................ 33 2
Content is King
Rune & Vinberg
4.1 Hypotheses Testing ............................................................................................................................ 33 4.1.1 4.1.2 4.1.3 Contextual Factors ................................................................................................................... 34 Attitude .................................................................................................................................... 36 Behaviour................................................................................................................................. 40
4.2 Conclusion ......................................................................................................................................... 46 5. DISCUSSION AND IMPLICATIONS ............................................................................................... 49 5.1 Motivational Factors Influencing the Intention to Share a Viral Video................................................ 49 5.2 Additional Communication Effects .................................................................................................... 51 5.3 Critique .............................................................................................................................................. 52 5.4 Suggestions for Further Research........................................................................................................ 53 6. REFERENCES ..................................................................................................................................... 55 7. APPENDICES....................................................................................................................................... 59 7.1 Appendix 1: Selected Videos - Summary............................................................................................. 59 7.2 Appendix 2: Selected Videos............................................................................................................... 60 7.3 Appendix 3: Online Survey ................................................................................................................. 62
3
Content is King
Rune & Vinberg
1. INTRODUCTION
”It’s not what you do. It’s what they do with what you do.” 1 - John Grant, The New Marketing Manifesto In June 2006, professional juggler Fritz Grobe and his lawyer friend Stephen Voltz decided to create a video of a prank experiment they had discovered a few months earlier. Dressed in white coats and protective glasses the two men looked like mad scientists when they performed the show, putting mint pastilles from the brand Mentos into bottles of Diet Coke. The reaction caused by the experiment has since then been commonly referred to as a “Mentos Eruption” or “Coke Geyser”, due to the explosion effected by the chemical reaction of the two products. Though only one person was told about the video, once uploaded, views went up at a rapid pace of 1000 hits per hour,2 making Grobe and Voltz world celebrities overnight. However, the massive impact that the online circulation of Grobe and Voltz’ experiment caused was met with scepticism by Coca-Cola. Unlike Mentos who responded positively to the unexpected marketing of their brand, giant Coca-Cola said to the Wall Street Journal "We would hope people want to drink [Diet Coke] more than try experiments with it," referring to the experiment as something that did not fit the “personality” of Diet Coke.3 However, following the upload of a great number of videos from others performing the stunt and a circulation reaching millions of people, the soda manufacturer changed their mind. Eventually, Coca-Cola publically stated their intentions to capitalize on the experiment by announcing competitions on their website on the theme “Poetry in Motion”.4 The happily ending story of Diet Coke and Mentos, which made the odd couple of a juggler and a lawyer world-famous, while reaching millions of potential consumers and providing unexpected exposure of two brands, is one of the many success stories that illustrate the magnificent potential of using the Internet for viral video marketing.
Wipperfürth (2005) p. 46 King (2007-01-16) www.businessweek.com 3 Ibid. 4 www.thecoca-colacompany.com
1 2
4
Content is King
Rune & Vinberg
1.1 Background
The Internet enables consumers to share information cost- and effortless and never before have consumers been able to influence brands to the extent that is now the case. 5 The social networking site Facebook and the video sharing platform YouTube are ranked as the third and fourth most visited sites on the Internet.6 Facebook has over 200 million registered users worldwide7 and even the Vatican has realized the power of YouTube by launching the Pope’s very own channel on the site.8 Recent technology enables not only exposure, but also engagement in a way that has previously not been possible. Brand community fan sites,9 spoof ads,10 uploaded corporate jingles and user generated video commercials11 are all examples of how consumers have been inspired by a brand or advertisement to the degree that they have become part of creating the commercial content. As opposed to traditional marketing where there is an obvious sender–receiver relationship between a company and the consumer, the relationship today has become substantially more complicated. In addition to being the receiver of the message, the consumer is also functioning as the medium and channel of communication. The essence of viral marketing is thus how to motivate consumers to spread a commercial message without any economic incentives or promise of other material gain, hence earning the medium. Without the spread, an online video will simply become a clip among millions of others, lost in cyber space. However, even though the challenges to create a successful viral campaign are numerous, the potential rewards are even greater. Baring this in mind, it is our intention to add knowledge to this topic in order to keep pace with the rapid development of consumer behaviour. As stated by marketing researchers Muñiz and Shau “The revolution won’t be televised. Instead, it will be expertly rendered and edited by dedicated users, and then distributed via e-mail and YouTube.”12
Kincaid & Drummond (2007) www.alexa.com (2009-05-23) 7 www.facebook.com (2009-05-23) 8 Thomas (2009-01-24) www.christiantoday.co.uk 9 Muñiz & Shau (2007) p. 36 10 Björkman & Blomstedt (2008) p. 4 11 Pressler (2008-10-28) www.nymag.com 12 Muñiz & Shau (2007) p. 47
5 6
5
Content is King
Rune & Vinberg
1.2 Problem Area
Even though the most successful viral campaigns have reached millions of people the vast majority have passed unremarked, illustrating the need to understand the forces of viral marketing. The fact that it is the consumers themselves that enable the spread of information is also the challenge. What motivates consumers to share information of commercial substance with fellow human beings? Which attributes of a video influence the perception of its content? Do positive attitudes towards a message equal the intention of forwarding it? Due to the novelty of the subject there is a need for academic research explaining the phenomenon of campaigns that have gone viral. Since viral marketing can be tremendously cost-efficient, but also very unpredictable, it is of great academicand managerial benefit to analyse the reasons behind this phenomenon. This thesis attempts to contribute with an increased understanding of the features that impact a consumer’s intention of forwarding commercial communication in the form of online videos as well as examining the effects that this type of communication will have on brand attitude and purchase intentions.
1.3 Purpose
The purpose of the thesis is to: Identify factors that affect consumer attitude and behaviour following the exposure to a viral video as well as studying how these features interact.
1.4 Delimitations
Due to the scope and the purpose of the thesis, a number of delimitations have been made. Since the purpose of the thesis is to study context- and content related features that impact consumer intentions to forward visual messages online, we have chosen to study videos that have been successful on video sharing sites like YouTube13 and Viral Video Chart14. The fact that they exist on Viral Video Chart, which is a site that compiles videos holding the largest number of views and commercial exposure online at a given time, implies that they are to be considered as relatively successful in terms of reach. With the intention of maximizing the relevance of the thesis we have exclusively used videos that have been uploaded as branded content, i.e. with a commercial purpose.
Website where consumers can stream (watch) and upload videos: www.youtube.com Website that ranks online videos in relation to which extent they have been mentioned on blogs, including data on geographical distribution and reach: www.viralvideochart.com
13 14
6
Content is King
Rune & Vinberg
The sample selection is constituted by 964 (main survey: 913 + control group: 51) respondents representative for the Swedish population in the age of 16-64, who are familiar with the Internet. Even though a cross-cultural sample would increase the validity, the scope of the thesis made it necessary to delimit the respondents to Swedish respondents. However, since the population of Sweden is relatively Internet savvy15 and far developed in terms of social network participation,16 a sample of Swedish citizens is well motivated for the purpose of analyzing online behaviour. The language used in the films is English, with the exception of the Heineken film, which is in Dutch. Since the sample consists of Swedish respondents, a sample of Swedish films would have been preferable in terms of language comprehension. However, as it was not within our power to find a sufficient number of Swedish films that fulfilled the requirements of quality as well as representing various categories of products and brands, the films had to be chosen out of the large number of international films available. Obviously, there is a slight risk of the respondents misinterpreting the messages of the videos since the audio is not in their native language. However, given the visual narrative of the content, the videos should be understood irrelevant of language knowledge. The risk of misinterpretations is also reduced by the fact that Swedes have a high understanding of English.17 Since research in adjacent areas of study, i.e. e-mails, have shown the influence that social proof and the source of the message have on the attitude towards the e-mail as well as the likelihood of sending it forward, these two contextual factors are considered as the most relevant to test. Even though factors as personal relevance and situational mood, as well as personal traits have been stated to affect the perception of commercial content, the limited scope of the thesis has required us to disregard these factors.
1.5 Definitions
Communication: Since this thesis attempts to study communication flows in a web context, we believe it to be of value to explain the term communication. We therefore use Hovland, Janis and Kelley’s definition of
Penetration (Dec 08): 80,2 %, www.internetworldstats.com (2009-03-03) Sweden holds the 6th largest population of Facebook-users in the world: Braw (2008-01-16) www.metro.se 17 www.skolverket.se (2004-03-02)
15 16
7
Content is King
Rune & Vinberg
communication as “the process by which an individual transmits stimuli to influence the behaviour of other individuals.”18 Social Networking sites: These sites are online platforms for people to socially connect and interact with other users. Hagel and Armstrong define them as “computer-mediated spaces where there is a potential for an integration of content and communication with an emphasis on member-generated content.”19 Viral videos are usually shared through these sites, particularly as many of them collaborate with video streaming sites through links in connection to the videos. Popular sites include Facebook, Myspace, Live Spaces, Bebo, Digg, Hi5, Orkut, Mixx, Linkedin, Lunarstorm, Habbo, A Small World, Twitter and Del.ic.ious. Video sharing websites: Due to the novelty of the phenomenon, there is no generally accepted definition of video sharing sites that we have been able to found. In comparison to the social networking sites, the video sharing websites are platforms for users to upload and watch streamed video clips online. The difference between downloading and streaming is that while downloading results in data being stored on your computer, streaming enables you to listen or view data without storing it first. Although social interaction occurs on video sharing websites, especially in terms of commenting on the videos, the purpose of these sites differ from social communities as networking revolves around the theme of videos, rather than the interaction itself. Online Word-of-Mouth: Sundaram defines word-of-mouth as a “form of interpersonal communication among consumers concerning their personal experiences with a firm or a product”.20 The difference between traditional w-o-m and online w-om (e-w-o-m) is that the exchange of information and opinions takes place in an “anonymous asynchronous online environment”.21 This phenomenon is also sometimes referred to as word-of-mouse communication.22 A clear advantage of online w-o-m is generally the potential of rapid reach.23
Chiu (2007) p. 525 Iriberreri & Leroy (2009) p. 2 20 Sundaram (1998) p. 527 21 Davis & Khazanschi (2008) p. 130 22 Blackwell et al. (2006) p. 214
18 19
8
Content is King
Rune & Vinberg
Viral Marketing: Viral marketing is principally the online equivalent to word-of mouth24 and was introduced by Steve Jurvetson and Tim Draper in 1997.25 ”Viral marketing differs from word-of-mouth in that the value of the virus to the original consumer is directly related to the number of other users it attracts”26. Although the expression may imply the information being spread in a virus-like fashion, viral marketing has no connection to spam or other destructive sorts of digital content. Rather, a viral marketing campaign seeks to use the cost-effective forces of spreading information and content exponentially, using social networks online.
1.6 Previous Research and Expected Contribution
Recent studies that focus on consumers’ intentions of interacting with other members in a web context have mostly considered exchange of text messages by e-mail,27 consumer-platforms,28 online reviews29 and chat-rooms.30 Phelps has studied why people forward e-mails, concluding that messages that spark emotions as humour, fear, sadness, or inspiration are more likely to be forwarded.31 He also emphasises the importance of reaching an appropriate target group since people who receive an e-mail from a friend will not consider it as “junk” and will therefore be more likely to forward it than messages received by companies. These findings are confirmed by Chiu, who furthermore states that the extent that a message is perceived as hedonic or utilitarian will affect the probability of the message being forwarded.32 Hennig-Thurau has studied the motivations that are related to affective behaviour and people’s need of social interaction on consumer platforms online.33 The author distinguishes four prime motivations; desire for social interaction, desire for economic incentives, concern for other consumers and the potential to enhance self-worth. An additional study on the incentives of engaging in online communities is Balasubramanian and Mahajan’s34 paper on
Davis & Khazanschi (2008) p. 131 Strauss & Frost (2001) p. 245 25 Phelps (2004) p. 334 26 Modezelewski (2000) p. 334 27 Phelps (2004), Chiu (2007) 28 Hennig-Thurau (2004) 29 Chatterjee (2001) 30 Zinkhan (2003) 31 Phelps (2004) p. 344-345 32 Chiu (2007) p. 530 33 Hennig-Thurau (2004) p. 41-43 34 Balasubramanian & Mahajan (2001) p. 125-126
23 24
9
Content is King
Rune & Vinberg
economic leverage in virtual communities. They identify three sources of utility; focus related utility, which derives from the belief that the contributions of each member will strengthen the community, consumption utility, which refers to the value that is received by consuming the contributions of other community constituents, and approval utility that is the satisfaction that occurs when other members of the network consume and approve of the constituent’s own contributions. Since previous research concerning online w-o-m mainly has focused on messages that contain text of some sort, there is a lack of understanding of how people perceive audiovisual content as well as their motivations and behaviours to spread such content on the web. Our expectation and aspiration is thus to contribute with increased understanding of the implications that exposure to such a message will have on consumer’s attitudes and behaviour. The undeniable future importance of the Internet as a marketing tool and communication channel imply the significant academic- and managerial benefits that an increased understanding of consumer behaviour within this area will have.
10
Content is King
Rune & Vinberg
2. THEORETICAL FRAMEWORK
In this section the theoretical framework upon which we construct the analysed hypotheses is presented.
2.1 Communication Effects Model
When formulating the hypotheses of our study we have used a general communication effects model as reference point, thereby attempting to explain the relationship between exposure of the video and the following effects on attitude and behaviour. This model will thus function as a theoretical framework to increase the understanding on how the various steps are interrelated.35 Figure 2.1 illustrates the causal relationship between consumer attitudes towards a specific advertisement, which will have spill over effects on the brand presented in the advertisement. This will in turn affect the consumer’s intention to purchase the brand’s products as well as his/her intentions to spread the message of the advertisement; generally through w-o-m. The model has been modified to fit the researched subject in terms of viral videos. The grey coloured components of the model are added by the authors to illustrate the significance of content and contextual factors on video attitude as well as the expected direct relationship between video attitude and intentions to share. These consist of three different types of behaviour; send the video forward, show the video to someone or create a similar version of the video.
Figure 2.1 Communication Effects Model36
The model was chosen due to its simplicity and applicability to the subject of studying consumer behaviour. It illustrates the causal relationship of the exposure to advertisement, which has been researched in previous studies, however not in a digital setting. Thus, using this basic model as reference point enables the study to draw on previous research in order to extend its applicability and expand the model further to fit the researched area of viral video marketing. 36 MacKenzie et al (1986), Sundaram & Webster (1999), Notani (1998)
35
11
Content is King
Rune & Vinberg
The following section will focus on the driving factors that influence change in attitude and behaviour, related to viral video marketing. These changes are important marketing objectives as they reflect basic psychological traits of consumer behaviour.37 Hence, understanding the flows and relationships between these motivational factors facilitate value creation.
2.2 Context
The perception of a message is not only affected by the content per se, but also depends on the source of the message as well as contextual factors, as can be seen in the model above (figure 2.1).38 It is therefore of interest to study how contextual factors, such as “social proof” and the relationship the recipient has to the sender will influence the perception of the message content, i.e. the videos.
2.2.1 Social Proof
Emotional gain or mental gratification of some sort is a prerequisite for the sender to engage in online w-o-m. Such value is never material but rather affective and related to the recognition from others.39 In order to be validated by other constituents of the society it is necessary that the consumer adds value to the community, i.e. focus-related utility.40 Adding value has various definitions in this context and can refer to both recommending positive experiences to others41 or preventing them from experiencing bad products or services,42 as well as expressing consideration and care for others.43 Since the desire to get recognition from other members will influence one’s attitude and behaviour44 it is likely that the attitude towards the content of a message would be impacted if the recipient perceives the message to be popular; i.e. if a large number of like-minded people previously have viewed and shared the message. According to online media practitioner Dan Ackerman Greenberg, co-founder of viral video marketing company The Comotion Group,45 driving up the number of views of a video on social networking sites to reach the “most views” page, is essential to any successful viral campaign. The thoughts of Ackerman Greenberg are coherent with the
Blackwell (2006) p. 88 MacKenzie & Lutz (1989) p. 49 39 Dichter (1966) p. 151 40 Hennig-Thurau (2004) p. 42, Chiu (2007) p.526 41 Dichter (1966), Blackwell et al. (2006), Sundaram et al. (1998) 42 Sundaram et al. (1998) 43 Phelps (2004) p. 343 44 Dichter (1966) p. 150 45 Ackerman Greenberg (2007-11-22)
37 38
12
Content is King
Rune & Vinberg
psychological theory on social proof,46 which states that people are drawn to things that others are attracted to. This implies that the number of views will influence the perception of the video both in terms of attitude as well as the likelihood of spreading the message forward. Thus, we formulate the following hypothesis:
Hypothesis 1: The number of views will positively influence the recipients in terms of
a: perceived level of ad creativity b: video attitude c: perceived level of provocation d: perceived level of engagement e: perceived level of authenticity f: perceived level of production quality g: intention to send the video forward h: intention to show the video to someone i: intention to make a similar content video
2.2.2 Sender-Recipient Relationship
Several researchers have shown the impact that the relationship between the sender and the recipient has on the attitude towards the message. If the sender is considered to be a friend, the recipient perceives the message to be of value and that the sender has forwarded it for a good reason.47 Those who receive messages from someone with whom they consider to have an interpersonal relationship are also more willing to forward the message than those who receive messages from unfamiliar acquaintances or commercial sources.48 This implies that there will be a difference between the recipients’ behaviour depending on whether the sender is claimed to be a friend, an acquaintance, or the company that occurs in the video. Thus the closer the relationship between the sender and the recipient, the better the attitude towards the video in terms of ad creativity, video attitude, engagement, authenticity and production quality will be. In similarity with previous reasoning, the closer the relationship the bigger the likelihood of the message being forwarded, hence we formulate the following hypothesis:
Cialdini (2001) Phelps (2004) p. 345 48 Chiu (2007) p. 529
46 47
13
Content is King
Rune & Vinberg
Hypothesis 2: The relationship between the sender and the recipient will positively influence the receiver in terms of
a: perceived level of ad creativity b: video attitude c: perceived level of provocation d: perceived level of engagement e: perceived level of authenticity f: perceived level of production quality g: intention to send the video forward h: intention to show the video to someone i: intention to make a similar content video
2. 3 Content
2.3.1 Attitude
In this section, factors solely related to the content of the video and the effect these might have on the video attitude will be studied. We will also conduct a comparative analysis between the group that has seen the video before and the group that has not. Finally, we will discuss if video attitude have an impact on brand attitude, as well as identifying additional factors that will have an influence on consumer behaviour, which is implied by the communication effects model, figure 2.1. Video Attitude In addition to emotional and cognitive recognition that motivates sender and recipient to engage in word-of-mouth, sharing a message with others also requires the message to have some sort of intrinsic value. Value can be defined as “a person’s relativistic preference with particular things or events”, and thus “both an event's usefulness and an appreciation of activities comprising it can indicate value” .49 Hence, whether the video is perceived to be of some value to the recipient, this will influence his or her intention to forward it, as can be seen in figure 2.1. Value is usually divided into two categories; utilitarian and hedonic50 (informative and transformative51). Utilitarian value is primarily instrumental, functional, and cognitive, providing customer value as the means to an end.52 Hedonic, on the other hand, is an outcome related to spontaneous responses that are more subjective and personal. 53 Examples of hedonic values are entertainment, exploration and self-expression, which are noninstrumental, experiential, and affective.54 In experience- and post-purchase evaluations both terms
Babin et al. (1994) p. 645 Chiu (2007) p. 526 51 Puto & Wells (1984) p. 638 52 Chiu (2007) p. 526 53 Babin et al. (1994) p. 646 54 Chiu (2007) p.526
49 50
14
Content is King
Rune & Vinberg
are relevant. However, since only an insignificant part of the videos that are currently figuring online are of utilitarian value, we have chosen to focus mainly on people’s tendency to share information that they perceive to be of hedonic value. Consumers are likely to provide information to others due to their desire to share joy and positive emotions they have experienced.55 It has also been shown that messages that evoke strong emotions are more likely to be forwarded.56 Features that evoke such emotions include humour,57 novelty and highly engaging content.58 We have therefore used the Ad Creativity model (Novelty, Meaningfulness, Humour, Positiveness and Well-Craftiness), developed by Modig & Lethagen,59 since these factors have been found to influence the attitude towards advertisement. Word-of-mouth will also be stimulated if the consumer perceives to have “inside information” and if the content appears to be authentic.60 Since provocative material can stimulate w-o-m61 it is also of interest to study the impact that such material will have on the attitude towards the video. The general attitude towards advertisement is considered to influence the attitude towards a specific commercial, hence the better the general advertisement attitude the better the attitude towards the specific video.62 We therefore test whether the general attitude towards the video is correlated with the perceived level of ad creativity, provocation, engagement and authenticity. In addition, we also test whether a positive attitude towards advertisement in general will improve the attitude towards the video. Thus we formulate the following hypothesis:
Ibid. p.526 Phelps et al. (2004) p.345 57 Ibid. (2004) p.345 58 Dichter (1966) p. 162 59 Modig & Lethagen (2008)
55 56
Dichter (1966) p. 165 Phelps (2004) p. 345 62 MacKenzie & Lutz (1989)
60 61
15
Content is King
Rune & Vinberg
Hypothesis 3.1: The attitude towards the video will be positively influenced by the following variables:
a: perceived level of ad creativity b: perceived level of provocation c: perceived level of engagement d: perceived level of authenticity e: perceived level of quality f: general advertisement attitude
Repetition According to previous research evaluating people’s attitudes towards TV ads, respondents who were exposed to the ad more than once, compared to respondents that saw the advertisement for the first time, appeared to evaluate the ad more favourable and less dull.63 This can be explained by the fact that the stimulus that consumers are exposed to when watching an advertisement will remain in their mind. Even though the attitude towards the brand is actually related to the repetition of exposure, the consumer does not necessarily understand this relation, and instead relate it solely to the content. Hence, the consumer misattributes the source of the processing simplicity; with a positive attitude towards the advertisement per se.64 To conclude, the attitude towards the video will be better if the respondent has been exposed to the advertisement previously.65 We therefore formulate the following hypothesis:
Hypothesis 3.2: The respondents who have seen the video before will have a more positive attitude towards the
video. Brand Attitude Brand attitude can be defined as a measure of “how much a person likes or dislikes a brand or to the extent to which he or she holds a favourable view of it.” 66 The better the attitude towards the brand, the higher the intention and likelihood of purchasing a product or service will be. Even though attitudes are relatively stable over time they can change and thus one of the main objectives of marketing communication should be to change the attitude in favour of the brand,67 thereby increasing the probability of purchase.68
Mano (1996) De Pelsmecker et al. (2005) p.41 65 Machleit (1988) 66 De Pelsmecker et al. (2005) p. 26 67 De Pelsmecker et al. (2005) p. 41 68 Dahlén & Lange (2007) p. 98
63 64
16
Content is King
Rune & Vinberg
In order to see the impact that viral videos have on brand attitude by examining whether or not the exposure to a video changes attitude, we formulate the following hypothesis:
Hypothesis 4.1: Brand attitude will be more positive for the respondents who have seen a branded video than those
who have not.
Based on the theory of hierarchies of communication effects,69 brand awareness will influence the attitude towards a brand, which thus is believed to apply to viral video marketing as well. According to the Dual mediation model, the evaluation of the advertisement or commercial has an immediate impact not only on the perception of the content per se, but also an indirect effect on the attitude towards the brand.70 This relation makes it probable to believe that the brand attitude will be influenced by how the respondents have perceived the video. Recent studies have identified an abundance of marketing messages, which has lead to advertisement avoidance and scepticism.71 It is therefore likely to assume that the perception of a commercial message is influenced by the initial advertisement attitude. Due to the immense amount of advertisements that people meet each day and the limited cognitive capability of the human mind, consumers tend to acknowledge only a fraction of the commercial communication they are exposed to, so called “selective scanning”.72 To use inconsistent advertisements73 or so called “shock of difference”74 has therefore been acknowledged as an effective way to get attention and stimulate word-of-mouth. Atypical advertisement requires more effort for the consumer to process and is therefore also given more attention. 75 However, advertising that is inconsistent with brand image might evoke an initial negative attitude towards the ad even though the effect on brand attitude can be positive.76 Inconsistent advertisement also has the power to push through the established
Ibid. p. 97 MacKenzie et al (1986) p. 131 71 Rosengren (2008) 72 Blackwell (2006) 73 Törn (2009) 74 Dichter (1966) p. 162 75 Stafford & Stafford (2002) p. 51 76 Törn (2009)
69 70
17
Content is King
Rune & Vinberg
schemas of well-established brands;77 by creating brand interest as an additional dimension to brand attitude as solely ad-evoked effects will not have a strong influence on mature brands.78 Due to the previously mentioned factors we therefore test how the attitude towards the brand will be affected by the attitude towards the video itself, the degree of brand awareness and the general attitude towards advertisements. It is also of interest to test whether atypical communication will impact the brand attitude. Thus, we formulate the following hypothesis:
Hypothesis 4.2: The attitude towards the brand will be positively influenced by the following variables:
a: video attitude b: brand awareness c: general advertisement attitude d: atypical communication
2.3.2 Behaviour
Intentions are the individual’s subjective estimation of their likelihood to perform a specific action later on. However, intentions should not be considered as certain predictions of the actual behaviour,79 even though studies show that there are correlations between intentions of doing something and then actually doing it.80 We therefore find it interesting to study word-of-mouth- as well as purchase intentions.
Intentions to Share:
Motivational Factors to Send a Video Forward A prerequisite when launching a viral campaign is the willingness of a consumer to forward the received content. A consumer’s tendency to spread information, so called word-of-mouth, has since the 1960’s been recognized as an important tool for marketers81 and is claimed to have a bigger influence on the actual purchase decision than mass media communication.82 Since word-of-mouth communication takes place without the sender expecting economic rewards and because consumers normally are more receptive towards communication that stems from people with whom they have a
Dahlén (2003) Machleit (1993) 79 Notani (1998) p. 248 80 Blackwell et al (2006) p. 411 81 Dichter (1966) 82 Arndt (1967)
77 78
18
Content is King
Rune & Vinberg
relationship,83 word-of-mouth is perceived as more credible than traditional advertising, which consequently increases the attention of the recipient. This implies that the information is processed more thoroughly in the consumer’s mind compared to company communicated commercials, and the final effect will thus be greater.84 The importance of content to motivate consumers to forward messages has been stated by several researchers.85 After being exposed to content that evoke strong emotion, many share their experience as a way of reducing tension created during the processing of the content.86 Messages containing content perceived as humorous, scary, sad and inspirational are therefore more likely to be forwarded. E-mails with “humorous jokes, touchingly sad stories and particularly apt inspirational messages”87 are considered as relevant enough to be sent forward even for those who normally are reluctant to share these types of messages with others. Relevant for the probability of forwarding is also the degree of work required from the sender to be able to pass it on. A prerequisite to reach a large community with one’s message is thus the simplicity of forwarding the message; virus only spread when they are easy to transmit.88 Due to the simplicity of forwarding videos on YouTube, both within the platform as well as to other platforms (Facebook, MySpace, Del.icio.us, Digg etc.), it can be assumed that the context (social network platform) in which the sender evaluates whether s/he should forward the video or not fulfil the requirements of a channel, where it is “easy to transmit”. However, since forwarding the message via Internet, no matter the technological progress, always will require some sort of effort it is also of interest to see whether the probability of showing the video to others depends on the same factors as forwarding it online. The simplicity of forwarding the message is related to the core function of the social network i.e. interacting with others. Due to the relationship between the members of the community it is natural that adding value to the community is frequently occurring.89 However, the belief that each member’s contribution to the community will increase the value or focus (focus related utility) requires that the message in fact contains qualitative content. Thus, whether the recipient will forward the message to other constituents is a direct result of the degree of value s/he perceives the message to have, i.e. the
Dahlén & Lange (2007) p. 110 Dahlén & Lange (2007) p. 112 85 Phelps (2004), Chiu (2007) 86 Dichter (1966) p. 149 87 Phelps (2004) p. 345 88 Chiu (2007) p. 531 89 Balasubramani & Mahajan (2001) p.116
83 84
19
Content is King
Rune & Vinberg
attitude towards the video. The satisfaction that members experience from the recognition of other consumers (approval utility) is also dependent on the quality of the video since the need for approval impedes the sender to forward something that s/he thought would not receive positive approval.90 Since certain research shows that negative or provocative content that cause fear and sadness stimulate a need of word-of-mouth as tension release,91 we have also chosen to test to which degree provocative material will motivate people to forward the message. The majority of the members of video platforms who share content with each other are relatively young.92 Thus, it is also of interest to see whether younger and more Internet savvy people are more likely to forward videos than others. For these reasons we formulate the following hypothesis:
Hypothesis 5: The intention of forwarding the video to others will be positively influenced by the following variables
(except for age which will influence this behaviour negatively): a: perceived level of ad creativity b: video attitude c: perceived level of provocation c: perceived level of engagement d: perceived level of authenticity e: perceived level of production quality f: general advertisement attitude i: video initiator: brand j: video initiator: private person k: age l: Internet usage
Motivational Factors to Show a Video to Others Since the most common objective for a company when up-loading a video online is reach, awareness, positive ad attitude and consequently an improved brand attitude, one can question if it matters whether the consumer decides to forward the message or show it to his/her friends. A significant difference between the two types of actions is the level of processing required by the viewer after the event of watching an online video before deciding to share it with others. Thus, taking the time to show a video to someone else requires more effort than simply forwarding it
Balasubramani & Mahajan (2001) p. 125 Phelps (2004) p. 345 92 52% of 18-34 year old users of YouTube share videos often with friends and colleagues: YouTube Fact Sheet (2009) www.youtube.com/t/fact_sheet
90 91
20
Content is King
Rune & Vinberg
online and accordingly, a real life recommendation imply an even higher level of involvement in the video. It is thus of great interest to study whether any differences between the two exists in terms of motivational features connected to a viral video. Therefore, we formulate the following hypothesis.
Hypothesis 6: The intentions to show the video to others will be positively influenced by the following variables
(except for age which will influence this behaviour negatively): a: perceived level of ad creativity b: video attitude c: perceived level of provocation c: perceived level of engagement d: perceived level of authenticity e: perceived level of production quality f: general advertisement attitude i: video initiator: brand j: video initiator: private person k: age l: Internet usage
Motivational Factors to Create a Similar Video Recent technological progress has enabled consumers to increase their participation in the creation of marketing and brand related content, i.e. user generated content (UGC). The rules of the game are changing and those who learn to adapt - the Crowd Surfers – have much to gain by letting go of their brand in favour of empowering the consumers in order to let them participate in shaping the destiny of the company.93 There is thus a tendency towards consumers making their own versions of commercial advertisements which resembles the originals both in form and intent. Consumers creating such branded content believe themselves to be self-appointed advocates of the brand, with strong opinions of what is right and wrong for it.94 Studies show that this tendency not only is expected to grow stronger but that consumers are increasingly savvy in their understanding of styles, tropes, logic and grammar of advertising and that a great deal of the content is widely spread by being up-loaded on the Internet. This phenomenon is mainly occurring for well established, unique and powerful brands, with which consumers feel connected to.95 In order to examine the tendency of user generated content, we formulate the following hypothesis:
Thomas & Brain (2008) Muñiz & Shau (2007) p. 35 95 Ibid. p. 46
93 94
21
Content is King
Rune & Vinberg
Hypothesis 7: The intention of creating a similar content version of the video will be positively influenced by the
following variables (except for age which will influence this behaviour negatively): a: perceived level of ad creativity b: video attitude c: perceived level of provocation c: perceived level of engagement d: perceived level of authenticity e: perceived level of production quality f: general advertisement attitude i: video initiator: brand j: video initiator: private person k: age l: Internet usage
Brand Attitude Influencing Intentions to Share According to Sundaram & Webster, studies show that brand familiarity and attitude will influence consumer attitude and behaviour in several ways; i.e. in terms of information processing style, ultimate brand choice, recall of advertisements as well as increased purchase intentions. Prior brand exposure is also stated to have a positive effect on the attitude which consequently improves the brand evaluation. Since highly positive evaluations create a need for tension release, this imply that the higher the brand attitude, the stronger the intention to share with others.96 In order to examine the applicability of the previously presented communication effects model, figure 2.1, in relation to the factors mentioned above we formulate the following hypothesis:
Hypothesis 8.1: The intention to send a video forward is positively dependent on brand attitude. Hypothesis 8.2: The intention to show a video to someone is positively dependent on brand attitude. Hypothesis 8.3: The intention to create a similar video is positively dependent on brand attitude.
Purchase Intention The final goal of all marketing communication is to increase likelihood of purchase within the targeted segment; i.e. creating an intention to purchase. A communication strategy must therefore take into account all steps of the purchase process to avoid losing consumers on the way.97 The MAO model (motivation, ability, opportunity) explains the route to the purchase occasion. It indicates that due to obstacles in terms of lack of availability and opportunity to purchase the product, generating motivated consumer is all the
96 97
Sundaram & Webster (1999) p. 666 Dahlén & Lange (2007) p. 98
22
Content is King
Rune & Vinberg
more important.98 Brand awareness has previously been stated to improve brand attitude and consequently purchase intention. Since pre-determined attitudes toward advertisements in general has shown to influence the attitude towards a specific commercial, it is of interest to determine whether this factor also influence purchase intentions.99 Due to advertisement clutter, it is becoming increasingly difficult to gain attention from consumers and therefore we intend to study whether content perceived as atypical compared to previous advertisement from the brand will influence purchase intentions.100 Thus, we formulate the following hypothesis:
Hypothesis 9: Purchase intentions are positively dependent on the following variables:
a: brand awareness b: general advertisement attitude c: atypical communication d: brand attitude
De Pelsmecker et al. (2005) p. 28 MacKenzie & Lutz (1986) p. 131, Törn (2009) p. 16 100 Törn (2009) p. 21
98 99
23
Content is King
Rune & Vinberg
3. METHODOLOGY
In this section we will present the methodology regarding how the thesis has been constructed and carried out.
3.1 Choice of Subject
Following the development of digital media, the term viral communication is increasing in use among academics and managers alike, working in the field of marketing and strategy. When we explored the opportunities of studying digital media, we discussed the subject with a number of people from the PR and marketing field to be able to understand which problem area is of highest interest to the industry. Coherently, we came to the conclusion that the rapid development of technology and the change in consumer behaviour that comes with this, brings new challenges and opportunities for marketers to adapt to. Therefore, we chose to study the phenomenon of spreading commercial video clips virally. As of today, there are a number of success stories of companies that have reached millions of viewers with their message, simply by planting a video clip on YouTube or an equivalent website. However, breaking through the clutter of millions of videos on these sites tends to be difficult. An important step towards creating a viral campaign is to identify the drivers of consumer behaviour in the digital context, why our intention is to contribute with further understanding of the area in question.
3.2 Scientific Approach
Since the objective of the thesis is to make a general conclusion of the features that influence consumers to engage in online w-o-m, a quantitative study was found to be the most appropriate. The reasons for this are that a quantitative approach simplifies establishing the reliability of the thesis and clarifies the analysis. Furthermore, as we have chosen to base the research on theory from adjacent areas of study and testing these using a large sample of respondents, the study called for a conclusive research design. The attempt to create a model that identifies determinates of attitude towards the ad and behavioural implications imply a causal approach. Since general conclusions are made from testing specific hypothesis on a population sample the approach is deductive.101
101
Malhotra & Birks (2007) p. 160
24
Content is King
Rune & Vinberg
3.3 Experimental Design
The quantitative study was conducted in collaboration with the research company YouGov,102 which provides tailor-made market research solutions for its clients, based on high quality internet and information technology with a panel of approximately 2 million members across the world. YouGov’s Swedish Internet Panel consists of 42 000 individuals who receive monetary rewards for their participation. The partnership enabled the experiment to be performed on a large scale as well as in a digital setting. In order to control the variables of the study, an experimental design was constructed based on fifteen video clips that each was preceded by a survey based on 3x2 scenarios. These manipulations were presented referring to who the sender of the video was (close friend/distant acquaintance/the company showing in the video) and how many that had viewed the video before, referring to social proof (many/few viewers). After the respondents had read the background manipulations and watched the short video they were asked to answer questions regarding their attitude and thoughts concerning the video content and the figuring brand, as well as their intentions to pass the clip forward. In order to segment the respondents, questions regarding their attitude towards advertising in general were also included. The name of the videos were not shown in the survey to avoid any kind of priming, especially concerning the film clips where the brand was only subtly evident to the viewer.
Sent from / No. of views Many views - 5 000 000 Few views - 100 Distant Acquaintance 3 4 Company in video 5 6
Best Friend 1 2
Table 3.1 Scenarios; source of message and number of views
3.3.1 Sample of Respondents
The collaboration with YouGov enabled a large sample of 964 respondents representing the Swedish population aged 16-64, based on three criteria: age, gender and geographical region. The participants were rewarded with a small monetary contribution of 5 SEK per individual. The survey contained a main survey of 913 respondents and an additional 51 respondents for the control group, testing only brand attitude for each brand represented in the video clips. Since the target group for a viral video campaign consist of people who are active on the Internet, we have chosen to use this channel to conduct the survey. It is also noteworthy that the design of the experiment allowed for
102
www.yougovsweden.se
25
Content is King
Rune & Vinberg
six different manipulations (3x2) and fifteen video clips, which in turn left only circa ten respondents to each manipulation and film. These sub-samples cannot in other words be perfectly divided among the population. However, the videos were randomly distributed among the recipients and the small sub-groups only serve as dividers between the manipulations where each film responds to a number of approximately 60 respondents.
3.3.2 Sample of Videos
In order to be able to test the variables of the experiment, the video clips were chosen based on a number of criteria. Firstly, as ecological validity103 is preferred, the videos used in the experiment were authentic films featuring a vast selection of different brands known to the Swedish audience. However, the brand did not need to be explicitly evident in the video as this was a factor desired to examine. Furthermore, each of the fifteen videos has been present on the Viral Video Chart,104 proving their success in circulation and popularity. Because of the importance of using fairly new videos, the videos had to be “discovered” no earlier than in September 2006. They should neither be longer than 3 minutes (the longest film is “T-Mobile Dance”, which is 2.41 min) to ensure that the respondents would not get tired before completing the survey, nor should they have been shown on Swedish television to avoid priming. However, there have been a couple of exceptions to the last criteria as a similar version of the “Dove Onslaught” and “Heineken Walk-in Fridge” (see appendix 2) videos have been appearing on Swedish television during the time in which the survey was conducted, which unfortunately is something that could not have been foreseen. The conducted survey shows that 30% of the respondents who watched the Heineken video had seen it before as the commercial shown on television is the same version as the video in the survey, although with English audio instead of Dutch. The Dove Onslaught video had only been seen by 5% as the TVcommercial is a heavily modified version of the viral video used in the survey. Nevertheless, we choose to not see this as a problem as it enabled the findings of additional conclusions regarding the effects of repetition, which are presented in more detail in the analysis. Regarding the time frame of the videos, two of the film clips were shortened due to survey related issues imposed by the collaboration with YouGov. “Diet Coke + Mentos” and ”Bike Hero”, which originally are approximately 3 minutes long, were only shown to the respondents for ca 1 minute. Naturally, the
103 104
Brewer (2000) www.viralvideochart.com
26
Content is King
Rune & Vinberg
shortened videos were made sure to show the core of the content to ensure the same outcome as if the full version had been shown. As the aim of the thesis is to identify characteristics that increase the chances of a video going viral, the sample of film clips were chosen based on the models of Ad Creativity105 and the FCB Grid106. Our intention when choosing the videos was thus to select a sample of film clips that, based on our personal perception as well as comments presented on the video-sharing sites, represented both high and low levels of ad creativity, which has been confirmed when analysing the data. In addition, brands covering the full spectrum of high-/low involvement – think/feel scales, based on the theories of Ratchford, thus the FCB Grid, were also taken into consideration when selecting the videos to ensure a wide range of product categories. (See Fig. 3.1 below and appendix 1.)
Figure 3.1 The FCB Grid, Featuring the Products Presented in the Selected Videos
105 106
Modig & Lethagen (2008) p. 73 Ratchford (1987)
27
Content is King
Rune & Vinberg
3.3.3 Construct of Survey Main survey:
Each respondent group were faced with a scenario and a short video clip before answering the main survey, which is the same independently of video or scenario, except for the questions regarding the specific researched brand. As the survey was conducted in a digital setting, it was possible to ensure that all of the questions were answered and in the desired order. The collaboration with YouGov enabled strict control over the respondents’ background in terms of gender, age and geographical area in which they live, which made it unnecessary to add these questions in the survey. The complete survey is presented in appendix 3. Content The first battery of questions in the survey concerns how the respondents perceive the videos. In order to determine external influence based on the respondent’s previous experience with the video clip, they were asked if they had seen it before. Thereafter, the respondents were asked to state to what extent they agreed with a number of adjectives that would represent the content of the video. When we discussed which adjectives that would achieve the necessary data we used characteristics that Dichter states stimulate traditional word-of-mouth107 as a vantage point. Similar features as those mentioned by Dichter are included in the Ad Creativity model108 (Novelty, Meaningfulness, Humour, Positiveness and Well-Craftiness) and this model was therefore considered as appropriate to measure the perceived level of creativity of the various videos. To measure the attitude towards the video per se we used the well established variables “Good”, “Like it” and “Pleasant”.109 Since the variable “Positive” from the ad creativity model already was included in the survey and was considered as rather similar to “Pleasant”, the latter was removed since it did not add any significant value. The following variables were, during discussion between the authors and the tutors, decided as adequate to evaluate features of the online videos; provocative, engaging and authentic, each represented by three similar questions to ensure reliability in the measurements. Three questions were developed to study the perceived quality of the video in terms of cost, time and level of difficulty that was needed to produce the film.
Dichter (1966) Modig &Lethagen (2008) 109 Dahlén & Lange (2007) p. 100
107 108
28
Content is King
Rune & Vinberg
Perceived Sender To avoid priming the respondents by stating the brand represented by the video too early in the survey (other than manipulations 5 and 6) as well as getting a good view of how the respondents perceive the video in terms of whether or not the sender is evident, they were asked an open-ended question as to which brand(s) they could identify in the video. The respondents were also asked how difficult they perceived it to be to identify the brand as well as how familiar they were with it. In order to test the level of irrationality in the communication,110 the respondents were asked whether they perceived the video to be typical/atypical with the marketing they had been exposed to in the past by the brand. To be able to test the variable of who the respondents perceive as producer of the video (the brand/private individual), these questions were also added. Intentions to Pass the Message Forward Based on empirical studies of video streaming behaviour, we were able to identify three main actions representing the intentions to share an online video; namely passing the film clip forward digitally (mainly through social networking sites, blog posts and e-mail), showing the video to others in a physical setting, and prolonging the life of the video by replicating the content in an amateur production to upload on the Internet. Conventional measurements of intentions were used,111 asking the respondents if they would want to pass the video forward as well as the probability of them doing so. Due to limitation of the length of the survey, only two questions were used for each index; however we believed these to be the most relevant. Brand Attitude In this section of the survey, the respondents were told which brand that was present in the video clip. Attitude toward the brand was calculated through the traditional measurement of “What is your perception of the brand x?". The scale ranged from 1-7, with the options bad/good, dislike/like and negative/positive.112 (See the section on “control group” for more details). The respondents were also asked to answer whether they had changed their attitude toward the brand after watching the video.
Machleit et al (1993) Dahlén & Lange (2007) p. 99 112 Blomstedt & Björkman (2008) p.23
110 111
29
Content is King
Rune & Vinberg
Advertisement Attitude To study whether a predetermined attitude towards advertising in general would influence the attitude towards the video, the respondents were asked to respond to what degree they agreed with the following statements; “I like advertising”, “Advertising is annoying”, and “Advertising is more manipulative than informative”, using a scale of 1-7. Internet Usage To study whether familiarity with the Internet would have an impact on attitudes towards the video and brand as well as w-o-m and purchase intentions, a question regarding Internet usage was added, measured in terms of hours per day. Control Group The 51 respondents in the control group were each asked three questions measuring their attitude towards the full selection of brands featured in the different videos, without being exposed to either manipulations or the researched film clips. The questions measured to what extent the respondents’ experienced the brands to be good/positive/like, which then could be compared to the group of respondents who were exposed to the videos in order to see the differences made by the manipulation of seeing the film clips.
3.4 Pre-studies
In order to ensure a credible outcome of the experiment, two pre-studies have been conducted to (1) categorize the products tested in the survey and (2) to pre-test the survey questions.
3.4.1 Classification of Products
According to Ratchford,113 purchase decisions can be categorized based on two dimensions, namely level of involvement and think/feel. The four categories constitute the FCB Grid, named after Foote, Cone and Belding who developed the model. The pre-study was conducted on March 16th 2009, allowing a random sample of twenty respondents, aged 19-61, to answer eight questions developed by Ratchford for each of the products represented in the sampled videos. Interviews were conducted with 20 respondents by phone or face-to-face.
113
Ratchford (1987)
30
Content is King
Rune & Vinberg
The pre-test confirmed that a range of product categories were represented among the videos (see the results in Figure 3.1).
3.4.2 Pre-test Survey
To ensure comprehension of the survey, including videos and manipulations, a pre-test was conducted, allowing ten respondents to participate in the study and give comments on the questions they had difficulties understanding. The final experiment is the result of a survey fully comprehended by the respondents.
3.5 Analytical Tools
The analysis of the gathered data has been conducted using the software SPSS, including compare of means through independent-sample t-tests and one-way ANOVA as well as linear regression analyses. Each index has been measured using Cronbach’s Alpha or Bivariate Correlation to ensure internal consistency and computed by calculating the mean of the selected variables. When applicable, the data has been re-coded into a numeral language. This has been the case for the variable measuring recognition, which was presented in the survey as an open question. Results have been accepted at a 5% level of significance.
3.6 Research Quality
Research quality is mainly dependent on two factors; validity and reliability. The concept of reliability involves the discrepancy found between the observed and the real value due to random errors in measurement.114 Hence, if the reliability of a research study is high, anyone who attempts to do the same study in the same manner would reach the same results. In order to ensure high reliability, there are a number of alternatives, where the most common is to study the extent to which one reaches the same results from several related measurements, examining the same feature. The study conducted for this thesis uses recognized measurements from marketing research to ensure high reliability and thus high research quality. In addition, multiple questions measurements have been used, signifying a high Cronbach’s Alpha (? > 0.7) and for indexes made by two questions, a high level of Bivariate Correlation (? > 0.7), which concludes that the internal consistency in terms of high correlation between the questions is sufficient.
114
Söderlund (2005) p.134
31
Content is King
Rune & Vinberg
The purpose of validity is to see to what extent a study is liberated both from random and systematic errors,115 which implies that, in order to reach high validity, the study should be designed in such a way that the attempted phenomenon of study is the one that is actually being studied. According to Malhotra & Birks, it is desirable to create an experimental design that has both internal and external validity.116 The first measures whether the manipulations of the independent variables actually cause the effects on the dependent variable, whereas the latter measures the extent to which the cause-and-effects relationships found in the study can be generalized. The research design of our study is experimental, focusing on using manipulations in the form of scenarios and different video clips while keeping the additional design of the survey static. Furthermore, the collaboration with YouGov enabled a greater, heterogeneous sample of respondents who were approached and participated in the study in a digital setting, which diminishes the external factors that could affect the outcome of the study. Thus, both the internal and external validity of the study are strong as the design of the study has been thoroughly examined to measure what was intended, while the large sample of respondents enable the conclusions of the study to be generalized.
3.7 Sources of Information
In addition to gathered quantitative data from the surveys, a number of secondary sources as academic articles and dissertations, as well as books related to the area of study have been used. In addition to the academic material we have also used several articles from online published magazines.
115 116
Ibid. p.149 Malhotra & Birks (2007) p. 308
32
Content is King
Rune & Vinberg
4. ANALYSIS AND RESULTS
In this section we present the analysis and the findings we have made to point out the motivational factors and brand related effects of a viral video campaign. The results are consequently summarized in the conclusion and illustrated in an extended communication effects model.
4.1 Hypotheses Testing
Each hypothesis presented in the theory section is tested based on the main study using an interval scale of 1-7, ranging from negative to positive. The few questions that were stated in an opposing manner, ranging from positive to negative, were inverted. Results are accepted at a 5% significance level, with the exception of one border line case of p=0.051. When using linear regression, each model has been tested for adequate levels of autocorrelation, multicollinearity and heteroscedasticity using Durbin-Watson (close to 2), Condition Index (< 20 or < 30 if supported by theory) and Scatter Plot (satisfactory illustrations). The indexes used in the study have been tested for sufficient levels of correlation (> 0.7) using Cronbach’s Alpha for multiple-question measurements. The content related indexes are as follows: Ad Creativity (?=0.919), Video Attitude (?=0.958), Provocation (?=0.865), Engagement (?=0.938), Authenticity (?=0.945) and Production Quality (?=0.859). Since Cronbach’s Alpha is appropriate using a minimum of three questions,117 Bivariate Correlations at a significance level of 0.01, have been used to study internal consistency for the indexes with only two questions, which are intention to share the videos as well as the intention to purchase from the company: Intention to Forward (?=0.916), Intention to Show (?=0.939), Intention to Create a Similar Version (?=0.766) and Purchase Intentions (?=0.801). Brand Attitude (?=0.981) was measured using Cronbach’s Alpha. Regarding Advertisement Attitude, the result was only ?=0.695, which is somewhat lower than the adequate level of 0.7. However, since the same measurement has been used in previous research we believed it to be supportable to disregard this small difference and index the questions.
117
Söderlund (2005), p.146
33
Content is King
Rune & Vinberg
4.1.1 Contextual Factors
Social Proof Hypothesis H1a-i addresses the potential differences in perception of the videos and intention to share a film clip based on the number of views it has received. This was tested using an independent-samples t-test to compare the groups exposed to the scenarios 1, 3 and 5 versus scenarios 2, 4 and 6 (see the Methodology section for further details), where the manipulations stated that the previous number of views were either 5 000 000 views or 100 views. Our calculations do not indicate any differences between these groups at the 5% significance level (see table 4.1). Table 4.1 Social proof: Independent-samples T-test for H1a-i
Hypothesis Variable a) Ad creativity b) Video attitude c) Provocation d) Engagement e) Authenticity f) Production quality g)Send forward h) Show to others i) Make another version Many views Few views Difference Means (SD) Means (SD) in means Significance 3.70 (1.64) 3.54 (1.68) 0.16 0.16 3.60 (1.87) 3.49 (1.93) 0.11 0.36 2.90 (1.52) 2.88 (1.63) 0.02 0.87 3.12 (1.67) 3.06 (1.73) 0.06 0.63 3.15 (1.70) 3.01 (1.67) 0.14 0.22 3.26 (1.47) 3.29 (1.58) -0.03 0.72 2.43 (1.80) 2.60 (1.95) -0.17 0.19 2.38 (1.80) 2.48 (1.85) -0.10 0.42 1.44 (0.94) 1.46 (1.03) -0.02 0.82 Support for the hypothesis: Yes or No No No No No No No No No No
N=913 (Many views: 457, Few views: 456), Many = 5 000 000 views, Few = 100 views Scale: 1-7 (neg.-pos.)
The results imply that the notion of social proof does not affect the recipients’ perception of a video, nor does it affect their intentions to share the video with others. However, it is important to state that the results do not reject the quest for a viral video campaign to strive after retaining a significant number of views as this does not discard of the fact that a viral video needs to be found and heard of by the consumers to create a spin to its circulation. One potential reason behind the difference in our results compared to the advice given by practitioners like the Comotion Group, which specializes in increasing the number of views of a video in order to enable it to go viral, is the purpose. The aim of many companies is to earn a spot on the “most views” sites and spark other types of media to pick up on the story rather than social proof in itself being the cause of better
34
Content is King
Rune & Vinberg
perception or willingness to share the video. The sole existence of “most views” pages, imply that there is an interest to watch popular videos in terms of number of views. However, our findings show that this phenomenon does not affect the outcome for receivers of a video, opposed to people who actively seek this type of entertainment.118 Sender-Recipient Relationship H2a-i aims to find significant differences in the perception of the measured features of a video and the intention to share its content based on the relationship between the sender of a video and the recipient. The hypothesis was constructed based on the preceding scenarios that the respondents were exposed to before watching the chosen video clips and are thus a comparison between manipulation 1 & 2 (best friend), 3 & 4 (distant acquaintance) and 5 & 6 (company appearing in the video), which is elaborated in further detail in the Methodology section. The three groups were compared in terms of means using ANOVA, however the findings do not illustrate any significant differences at the 5% level between any of the groups (see table 4.2). Table 4.2 Sender-recipient relationship: One-way ANOVA for H2a-i
Hypothesis Variable a) Ad Creativity b) Video attitude c) Provocation d) Engagement e) Authenticity f) Production quality g) Send forward h) Show to others i) Make own version Best Friend Means (SD) 3.50 (1.70) 3.40 (1.94) 2.89 (1.63) 2.98 (1.72) 2.96 (1.69) 3.16 (1.49) 2.48 (1.93) 2.42 (1.89) 1.44 (1.02) Distant Acquaintance Company Means (SD) Means (SD) Significance 3.70 (1.66) 3.65 (1.64) 0.30 3.67 (1.90) 3.57 (1.86) 0.24 2.85 (1.61) 2.93 (1.48) 0.80 3.17 (1.73) 3.10 (1.63) 0.39 3.15 (1.71) 3.14 (1.65) 0.30 3.30 (1.51) 3.37 (1.57) 0.22 2.53 (1.84) 2.52 (1.89) 0.93 2.48 (1.83) 2.39 (1.77) 0.80 1.41 (0.93) 1.50 (1.00) 0.54 Support for the hypothesis: Yes or No No No No No No No No No No
N=913 (Best Friend: 297, Distant Acquaintance: 309, Company: 307), Scale: 1-7 (neg.-pos.) The results contradict the findings of a number of prominent researchers, including Phelps (2004) and Chiu (2007) who state that the closer relationship the recipient of digital content has to the sender, the more likely s/he will be to perceive it as more interesting and the more willing s/he will be to share the message with others. However, as the development of digital media is moving in a
118
The hypothesis was neither supported when examining the 15 videos independently.
35
Content is King
Rune & Vinberg
rapid pace, consumer behaviour changes quickly, which the results of this study demonstrates. There is obviously a difference in behaviour and attitude among the receivers of a viral video in comparison to recipients of e-mails, which is the digital medium mainly researched in the past. Consequently, the comparison between this study and previous research indicates that the change of consumer behaviour in a digital setting must be closely monitored in order to follow the rapid development. At present however, this demonstrates that receivers of a viral video are not sensitive to the relationship they have to the sender, with regards to their perception of the content or willingness to share the video with others.119
4.1.2
Attitude
Content Related Factors:
Dimensions of Video Attitude Hypothesis H3.1a-f was constructed to study the potential dimensions that influence the attitude towards a video. These were tested using linear regression with dependent variable video attitude and independent variables indexed as ad creativity, perceived level of provocation, perceived level of engagement, perceived level of authenticity, perceived level of production quality and general advertisement attitude. Results confirm that ad creativity (?=0.742), engagement (?=0.199) and authenticity (?=0.057) explain video attitude at the 0.1% significance level, however provocation (?=-0.081) affect video attitude negatively. The latter result is notable as it illustrates the negative feelings connected to the notion of provocation and clearly states that provocation should not be a desired goal when aiming to affect the viewer of a viral video in a positive manner. Hence, the more creative, engaging and authentic a viral video is, the better will the viewer think of it, however the less provocative the video is, the higher the attitude. Together the dimensions account for 90.8% of the variation in the measure (R²), which is highly satisfactory. The study does not confirm hypotheses H3.1e and H3.1f, which is somewhat surprising as these factors evidently play as significant part in explaining other types of attitude and behaviour (see the sections on intentions to share and brand attitude). However, when evaluating the actual video, general attitude towards advertising and the quality of the video do not affect the outcome. For details see table 4.3.1.
119
The hypothesis was neither supported when examining the 15 videos independently.
36
Content is King
Rune & Vinberg
Table 4.3.1 Dimensions of video attitude: Linear Regression for H3a-f
Hypothesis Independent variables a) Ad creativity b) Provocation c) Engagement d) Authenticity e) Production quality f) Advertisement attitude Standardized Beta (?) 0.742*** -0.081*** 0.199*** 0.057*** Support for the hypothesis: Yes or No Yes No Yes Yes No No
N=913, R² (adjusted) = 0.908 Significance: *p ? 0.05, **p ? 0.01, ***p ? 0.001 Durbin-Watson: 1.827 Condition Index: 12.544 Repetition According to previously stated theory, people who are exposed to marketing communication repeatedly tend to evaluate the advertisement more favourable and thus, H3.2 aims to see if this reasoning also applies to viral video marketing. The results indicate that there is a significant difference in attitude towards the videos, dependent on whether the respondents watch the film for the first time or not. Those who watch the video for the second time or more will generally evaluate the video as better than those who see it for the first time, illustrating the importance of repetition in digital media as well as conventional counterparts. Table 4.3.2 Repetition: Independent-Sample T-test for H3.1
Hypothesis Variable Video attitude Support for Seen video the before First time view Difference hypothesis: Means (SD) Means (SD) in means Significance Yes or No 4.68 (1.60) 3.48 (1.90) 1.20 0,000 Yes
N= 913 (Seen video before: 54, First time view: 859), Scale: 1-7 (neg.-pos.) Brand Attitude The general change in brand attitude is tested using an independent-sample t-test, comparing the means between the respondents who have been exposed to a viral video and those who have not.
37
Content is King
Rune & Vinberg
The test show no significant differences at the accepted 5% level, however there are indications (p=11%) that the brand attitude among those who have seen a viral video tend to have an improved attitude towards the brand than those who have not seen the video (see table 4.4.1 below). Comparing these groups should naturally acknowledge the consumers’ perception of the researched videos as these are signified by a wide selection of videos generating both positive and negative feelings. However, similar tests using only those videos generating high attitudes towards the video and vice versa show no different results. Hence, the data only indicate that there is no significant difference in brand attitude between those who have been exposed to the video and those that have not. Table 4.4.1 Brand attitude: Independent-sample T-test for H4.1
Support for the Seen video Not seen video Difference hypothesis: Means (SD) Means (SD) in means Significance Yes or No 4.58 (1.58) 4.34 (0.90) 0.24 0.11 No
Hypothesis Variable Brand Attitude
N= 952 (Seen video: 913, Not seen video: 39), Scale: 1-7 (neg.-pos.) Nevertheless, studying each video respectively, a sample of three of the videos indicates a change in brand attitude. “T-Mobile Dance”, which proved to be the most popular of the sampled videos, with a mean of 5.08 in video attitude (Scale: 1-7) among the respondents, indicates a change in brand attitude among those who have seen the video and those who have not. Durex’ video “Get it on” was also able to change the brand attitude among the respondents due to similar reasons as it was the third most popular video (video attitude mean=4.44) and one of the least known brands (T-Mobile was graded the lowest based on brand awareness and Durex fourth to last). What is more interesting is the change in brand attitude among receivers of the “McNugget Rap” video from McDonald’s, which was one of the least popular videos (graded third to last in video attitude: 2.90 on a 1-7 points scale). The change cannot be traced back to low brand awareness as McDonald’s is the most well known brand in the study (mean=6.23). However, it is notable that “McNugget Rap” was graded as the most atypical communication for the brand (mean=4.98) among the sampled videos, indicating the significance that atypical communication may have on brand attitude (see table 4.4.1.1).
38
Content is King
Rune & Vinberg
Table 4.4.1.1 Brand attitude by video: Independent-sample T-test
Not seen Seen video video Difference in Means (SD) Means (SD) means Significance 3.60 (1.23) 3.15 (1.11) 0.45 0.051 4.76 (1.49) 4.07 (1.24) 0.69 0.01 4.35 (1.69) 3.51 (2.01) 0.84 0.02 Support for the hypothesis: Yes or No Yes Yes Yes
Hypothesis Variables T-Mobile Durex McDonald's
T-Mobile: N=110 (Seen video: 61, Not seen video: 49) Durex: N=108 (Seen video: 61, Not seen video: 47) McDonald’s: N=110 (Seen video: 62, Not seen video: 48) Scale: 1-7 (neg.-pos.) As stated in the previous paragraph, the analyses of the data indicate that high video attitude causes respondents to change their attitude towards the brand for the better, especially with regards to less well-known brands. For more well-known brands inconsistency seems to play a more significant part in explaining change in brand attitude even if this does not imply that the respondents are positive towards the ad. Examining these tendencies in closer detail using linear regression demonstrates that the dimensions of brand attitude, with regards to viral video marketing, include video attitude (?=0.236), brand awareness (?=0.204), advertisement attitude (?=0.117) and atypical communication (?=0.082). These variables account for 13.2% of the variance in the measurement (R²), indicating that other factors need to be taken into account as well. The analysis was conducted based on only a limited part of the sample based on the respondents’ recognition of the researched brand. Hence, only 559 of the 913 respondents were used in this analysis as those who were not able to identify the brand in the video would add noise by answering questions regarding their perception of the brand based on the inaccurate idea of which brand that was in the video (See table 4.4.2).
39
Content is King
Rune & Vinberg
Table 4.4.2 Dimensions of brand attitude: Linear Regression for H4.2a-d
Hypothesis Independent variables a) Video Attitude b) Brand Awareness c) Advertisement Attitude d) Atypical Communication Standardized Beta (?) 0.236*** 0.204*** 0.117** 0.082* Support for the hypothesis: Yes or No Yes Yes Yes Yes
N=559, R² (adjusted) = 0.132 Significance: *p ? 0.05, **p ? 0.01, ***p ? 0.001 Durbin-Watson: 1.707 Condition index: 11.310
4.1.3
Behaviour
Intentions to Share:
Intention to Send a Video Forward Hypothesis H5a-k seeks to find the determinant factors of the intention to send a video forward, testing the variables ad creativity, video attitude, perceived level of provocation, perceived level of engagement, perceived level of authenticity, perceived level of production quality, brand awareness, advertisement attitude, brand as video initiator as well as private person as video initiator, age and Internet usage as potential dimensions of a causal relation, using linear regression. The results indicate that the decision to send a viral video forward is dependent on ad creativity (?=0.164), video attitude (?=0.448), perceived level of authenticity (?=0.099), perceived production quality (?=0.076) and advertisement attitude (?=0.106).Together the 5 dimensions account for 52.8% of the variation in the measure (R²). However, no evidence was found that provocation, engagement, video initiator, age or Internet usage are directly correlated with the intention of forwarding. See table 4.5 for further details of the analysis.
40
Content is King
Rune & Vinberg
Table 4.5 Motivational factors that influence the intention to send a video forward: Linear Regression for H5a-k
Hypothesis Independent variables a) Ad creativity b) Video Attitude c) Provocation d) Engagement e) Authenticity f) Production quality g) Advertisement attitude h) Video initiator: Brand i) Video initiator: Private Person j) Age k) Internet Usage Support for the Standardized hypothesis: Beta (?) Yes or No 0.164* Yes 0.448*** Yes No No 0.099** Yes 0.076** Yes 0.106*** Yes No No No No
N=913, R² (adjusted) = 0.528 Significance: *p ? 0.05, **p ? 0.01, ***p ? 0.001 Durbin-Watson: 1.962 Condition index: 22.471
Although the regression analysis does not indicate any dependent relationship between age and the intention to forward at the 5% level of significance, a comparing of means analysis using one-way ANOVA illustrates significant differences between age groups (16-30, 31-45, and 46-64). Young people, aged 16-30, show significantly higher intentions to send a video forward (2.8318 over 2.2570 on a 1-7 scale) than the higher age segment 46-64 (p= 0.001). Respondents 31-45 years old also show significantly higher intentions to forward a video (mean 2.6133, p=0.046) than the group of 46-64 year olds. Intention to Show a Video to Others Similar to the previous section, hypothesis H6a-k addresses the determinant factors that cause the recipient of a video to share it with others; more specifically the intention to show the video to others in a real life setting. The results imply that the motivators of sharing a video differ somewhat between the intentions to pass the video forward in a digital setting compared to showing it to someone in the “real” world. Measuring the same variables as H5 using linear regression provides the outcome of a number of motivators to explain the behaviour of sharing a video by showing it to
41
Content is King
Rune & Vinberg
others. These factors include attitude towards the video (?=0.564), perceived level of engagement (?=0.130), perceived production quality (?=0.078) and advertisement attitude (?=0.100) as motivating to show the video to another constituent of their real life network. The model provides a solid base for explanation, with the 4 independent factors accounting for 52.2% variation in the measure (R²). See further details in table 4.6. Table 4.6 Motivational factors that influence the intention to show a video to a friend: Linear Regression for H6a-k
Hypothesis Independent variables a) Ad creativity b) Video Attitude c) Provocation d) Engagement e) Authenticity f) Production quality g) Advertisement attitude h) Video initiator: Brand i) Video initiator: Private Person j) Age k) Internet Usage Standardized Beta (?) 0.564*** 0.130** 0.078*** 0.100*** Support for the hypothesis: Yes or No No Yes No Yes No Yes Yes No No No No
N=913, R² (adjusted) = 0.522 Significance: *p ? 0.05, **p ? 0.01, ***p ? 0.001 Durbin-Watson: 1.933 Condition index: 11.747 Similar to the intention to forward the video, showing it to others does not indicate to be dependent on age. However, segmenting the sampled population into three age groups illustrates differences at a significance level of less than 0.06%. The results show that “young” and “middle-aged” people aged 16-45 state significant higher intentions to show the video to someone than those aged 46-64 (16-30: 2.7114, 34-45: 2.5933, 45-64: 2.1489, mean based on a 1-7 scale).
42
Content is King
Rune & Vinberg
Intention to Create an Own Version of the Video Hypothesis H7a-k aims to find the motivational factors behind the intention of the respondent making a similar version based on the video s/he has been exposed to. The results indicate that high level of engagement (?=0.334) and production quality (?=0.102) of the content in a video motivates people to make their own version of it, while videos that are perceived as being made by a private person (?=0.082) rather than the brand appearing in the video, as well as a higher level of Internet usage (?=0.095), are additional related factors to consider. However, although the model is satisfactory with regards to measurements of autocorrelation, multicollinearity and heteroscedasticity, it is important to point out that the 4 variables presented in this study only account for 14.5% of the variation in the measurement (R²), implying that these factors alone does not fully explain what motivates the recipient of a video to make a similar-content film clip to upload online. See table 4.7 for further details. Table 4.7 Motivational factors that influence the intention to make an own version: Linear Regression for H7a-k
Hypothesis Independent variables a) Ad creativity b) Video Attitude c) Provocation d) Engagement e) Authenticity f) Production quality g) Advertisement attitude h) Video initiator: Brand i) Video initiator: Private Person j) Age k) Internet Usage Standardized Beta (?) 0.334*** 0.102** 0.082* 0.095** Support for the hypothesis: Yes or No No No No Yes No Yes No No Yes No Yes
N=913, R² (adjusted) = 0.145 Significance: *p ? 0.05, **p ? 0.01, ***p ? 0.001 Durbin-Watson: 1.873 Condition index: 9.677 The summarizing table above, illustrating the motivational factors affecting the intention to create an own version of a viral video, does not indicate that age is a contributing factor in this sense, at the 5% significance level. Nevertheless, comparing the means of these intentions, based on age groups
43
Content is King
Rune & Vinberg
using one-way ANOVA, demonstrates significant differences. Young people aged 16-30 are more willing to create an own version than those aged 46-64 (p=0.007). However, it is important to state that the intention to create an own version of a video is generally very low as the comparison lies between 1.6091 for those aged 16-30 and 1.3499 for the 46-64 age group at a 1-7 points scale. A summarizing illustration of the causal relation for the different kinds of intentions to share a viral video is found below.
Figure 4.1 Motivational Factors for Intentions to Share
The model illustrates the relationship of the features connected to watching a viral video that affects the intention to share the video with others, either by creating a similar version, showing the video to others or sending it forward. The features included in the model are those that have proven to be significant and illustrate both content related factors based on the respondents’ perception of the video as well as anticipatory factors that the respondents possess prior to the exposure of the viral video. Brand Attitude as Motivational Factor for Intentions to Share As indicated by theory,120 word-of-mouth intentions are affected by brand attitude. However, since there is little known research of the phenomenon regarding visual online word-of-mouth, a test was conducted using linear regression to establish the connection between brand attitude and the intentions to respectively send a video forward, show it to a friend and create a similar version for online publishing. Due to the construct of the survey where the questions regarding the intention to share the film clips were stated before the brand presented in the video was revealed, only a limited
120
Sundaram & Webster (1999)
44
Content is King
Rune & Vinberg
selection of respondents (N=559), including those who were able to recognize the brand, was used for this test. The results demonstrate that all of the three variables testing the intentions to share the message of the video are positively affected by brand attitude. However, it is important to state that the intention to send a video forward and to show it to others is more strongly affected by the respondents’ attitude towards the brand than the intention to create a similar content video is. In addition, R² is considerably low for each of the tests, indicating that brand attitude only accounts for a significantly small part of the variation in the measurements, which is expected when using only one independent variable. See table 4.8.1-3 for more details. Table 4.8.1 Brand attitude as motivational factor to send a video forward: Linear Regression for H8.1
Hypothesis Independent variable Brand Attitude Support for the Standardized hypothesis: Beta (?) Yes or No 0.241*** Yes
N=559, R² (adjusted) = 0.057 Significance: *p ? 0.05, **p ? 0.01, ***p ? 0.001 Durbin-Watson: 1.423 Condition index: 6.527 Table 4.8.2 Brand attitude as motivational factor to show a video to a friend: Linear Regression for H8.2
Hypothesis Independent variable Brand Attitude Support for the Standardized hypothesis: Beta (?) Yes or No 0.248*** Yes
N=559, R² (adjusted) = 0.060 Significance: *p ? 0.05, **p ? 0.01, ***p ? 0.001 Durbin-Watson: 1.385 Condition index: 6.527 Table 4.8.3 Brand attitude as motivational factor to make an own version: Linear Regression for H8.3
Hypothesis Independent variable Brand Attitude Support for the Standardized hypothesis: Beta (?) Yes or No 0.083* Yes
N=559, R² (adjusted) = 0.005 Significance: *p ? 0.05, **p ? 0.01, ***p ? 0.001 Durbin-Watson: 1.819 Condition index: 6.527
45
Content is King
Rune & Vinberg
Purchase Intentions The ultimate objective of any successful marketing campaign is to increase sales, which naturally also applies to the use of viral video marketing. To complete the extended communication effects model we test hypothesis H9a-d, which aims to identify factors related to a viral video campaign that determine the intention to purchase. The linear regression implies that brand awareness (?=0.070) and brand attitude (?=0.653) account for 45.3% of the variation in the measurement (R²) and thus provides a solid explanatory model for which viral video-related features that cause the respondents to purchase products or services from the brand in question (see table 4.9 for further details). Table 4.9 Dimensions of purchase intentions: Linear Regression for H9a-d
Hypothesis Independent variable a) Brand Awareness b) Advertisement Attitude c) Atypical Communication d) Brand Attitude Support for the Standardized hypothesis: Beta (?) Yes or No 0.070* Yes No No 0.653*** Yes
N=559, R² (adjusted) = 0.453 Significance: *p ? 0.05, **p ? 0.01, ***p ? 0.001 Durbin-Watson: 1.910 Condition index: 8.164
4.2 Conclusion
The results in the analysis illustrate discrepancies between previously studied digital communication and consumer behaviour connected to audiovisual messages. Our study indicates that the content is the sole most important feature that motivates receivers of an online video to share the message with others. Furthermore, content influence brand attitude positively, which consequently leads to increased purchase intentions. Contextual factors like sender-recipient relationships and number of views of a viral video do not have an effect on consumer attitude and behaviour according to this study. Instead, content related factors including perceived level of creativity (novelty, meaningfulness, humour, positiveness, and well-craftiness), authenticity, provocation and engagement will affect the recipients’ attitude towards the video (video attitude), which in turn encourages a positive attitude towards the brand (brand attitude) and intentions to spread the content of the video (send, show); consequently leading to higher purchase intentions among the receivers of the video. The study also found that repetition increases video attitude as the respondents who had been exposed to the video more than once demonstrated significantly higher evaluations in terms of positive attitudes towards
46
Content is King
Rune & Vinberg
the videos. Additionally, it is illustrated that the intentions to send a video forward, show it to others or create a similar film clip do not merely depend on the consumers’ brand attitude, but a number of factors, which will affect the outcome of the different behaviours respectively. Content is evidently the single most important factor that explains the consumer behaviour; however the dimensions of this feature prove different in terms of explanatory strength. The intentions to show and send forward a video are significantly affected by the respondents’ attitude towards the video, but there are also direct affects concerning the nature of the content influencing the decision. Authentic, creative videos with a high perception of production quality will more likely be forwarded digitally. For the intention to show the video to a friend, engagement is also significantly important as a substitute to authenticity and ad creativity. In addition to content related features, the respondents’ general attitude towards advertisement also proved to be important when deciding to share the message through digital word-of-mouth. In order to encourage the receivers of a viral video to create a similar video and upload it to a video sharing site, production quality and perceived level of engagement are important features of the film clip, while additional factors including high levels of Internet usage and whether the video is made by a private person in comparison to the company in question will affect the outcome positively. The analysis furthermore confirms the belief that segments based upon age will differ in terms of intention to share a viral video, where the younger generations are more likely to act upon this. Regarding brand related effects of a viral video campaign, traditional consumer behaviour theory is confirmed, stating the positive effects on brand attitude from general advertising attitude and brand awareness, in connection to a positive attitude towards the video. Atypical communication in the video also proved to have a positive effect. The analysis confirms the impact that a successful viral video campaign has on purchase intentions, by improving the brand attitude. A summarizing model of the findings is presented in figure 4.2 on the next page. 121
An analysis whether or not the gender of the respondents would illustrate differences in the outcome of the results has been conducted. However, this study does not find any differences in attitude or behaviour at the 5% level of significance beyond obvious differences, which is why this research has not been accounted for within the scope of this thesis.
121
47
Content is King
Rune & Vinberg
Figure 4.2 Communication Effects Model for Viral Video Marketing
The model illustrates the overall findings of the study, with the exception of the direct links between the respondents’ perception of the video and the intentions to share, which is found in further detail on page 44. The arrows illustrate a causal relationship between the squares whereas the straight lines connected to a circle illustrate the components of the square.
48
Content is King
Rune & Vinberg
5. DISCUSSION AND IMPLICATIONS
This section will discuss the findings of the thesis as well as present implications based on this knowledge. Furthermore, we will also consider critique towards the study and suggestions for further research. Using theory from traditional marketing communication as reference point, we are able to conclude that traditional communication effect models to a large extent is applicable also in a digital marketing context. Similar spill-over effects that a positive attitude towards a commercial will have on word-ofmouth intentions, brand attitude and consequently purchase intentions, can be applied to digital videos (see Figure 4.2). These findings provide a solid foundation to explore the driving factors and effects of viral video marketing in further detail as well as distinguishing differences. Hence, the traditional communication effects model offers a sustainable basis for the subsequent argumentation.
5.1 Motivational Factors Influencing the Intention to Share a Viral Video
The most imperative motivational factor for the intention of sending a video forward or showing it to other people is the receiver’s attitude and perception of the video content. This is affected by a number of content related features; many of which have a direct influence on the intentions to share a video with others. Hence, we have identified content as the single most important factor in order to make a video go viral. The results of this study imply that low levels of provocation, together with a sense of authenticity as well as high quality content are factors to take into account when constructing a successful viral video campaign. In addition, Ad Creativity (Novelty, Meaningfulness, Humour, Positiveness and Well-Craftiness) strongly affects the attitude towards the video as well as the intention to send the clip forward. The result that provocation affects video attitude negatively is interesting as it may seem to oppose the idea of creating content that is “on the edge” and “sensational” to promote a positive attitude towards a video. This is most likely connected to the respondents’ perception of the word provocation itself, which was found to be associated with sexist content for example. This is illustrated in a quote from the video “Making of the Ryanair Cabin Crew Charity Calendar 2009” (see appendix 1), which was regarded as the most provocative video as well as the least favourite of them all. It is noteworthy that novelty, which is part of the Ad Creativity index and could be seen as the other side of using content that is “on the edge”, is affecting video attitude positively. Thus, the results connected to the content related features examined in this thesis, illustrates the complexity of finding the right balance between creating
49
Content is King
Rune & Vinberg
something new and interesting, while not being too provocative. Furthermore, contradictory to a rather widespread opinion that videos online need to appear amateurish even for branded content in order to be accepted and spread, we can conclude that marketers should regard viral videos the same way as any other audiovisual communication, hence focusing on a creative idea, together with an impeccable execution, i.e. content is king! Studying the influencing factors of video attitude does not merely include independent variables connected to the content. Repetition was also found to be an important feature to take into consideration when constructing a successful viral video campaign as the respondents who had been exposed to the video prior to participating in the experiment showed significantly higher levels of attitude towards the film clip. Implication of this finding is basic, yet noteworthy as it indicates that enabling the consumers to get used to the content of an online video will improve their attitude towards it. Hence, using a marketing mix of different channels such as television, cinema and digital media for a branded video will, according to this study, more likely provide positive attitudes towards the film clip than if the video were only shown a limited number of times online. One important difference between the intentions to show a film clip to someone compared to sending it forward is the importance of engagement. Rather than evaluating the level of creativity and authenticity of the video, respondents who intend to show the video, which evidently requires more effort in terms of recommendations than to simply make a few clicks after watching the film clip, are to a greater extent influenced by the level of engagement in the content. This conclusion has important managerial implications, as it illustrates the significance of the receiver’s perception of being inspired by the content of the video to take the steps required to tell someone about the video that s/he finds to be interesting enough to share even after the event of watching it him-/herself. In order to encourage receivers of a viral video to create a similar content video on their own, additional factors proved vital. Unsurprisingly, people who are more Internet savvy are more likely to make their own version of the original. There are also indications that the intention of making a similar reproduction increases if one perceives the originator of the video not to be connected with the company appearing in the film clip. However, interestingly, additional features affecting this decision are perceived level of engagement in the video as well as production quality. This indicate that the viewer of an online video is more likely to create a similar film clip if the video in question is both high in quality in terms of expenditure, time and effort as well as if it is user generated.
50
Content is King
Rune & Vinberg
Although these features may appear as opposing and not coherent with previous beliefs that user generated content would be of low quality and in an amateur like fashion, the results illustrate that the current technological progress enables user generated content to be as qualitative as that made of professionals. It should however be noted that the intention to make a similar version was generally very low, which may be an outcome due to the lack of incentives provided for the respondents to act in this manner; i.e. asking them to create a video as part of an interactive campaign. Contradictory to our prior beliefs, contextual factors including number of views and senderrecipient relationships did not indicate any effects on the respondents’ perception and behaviour. These results differ from previous research122 on digital communication via e-mails, where contextual factors had a significant impact on the attitude and behaviours of the survey participants. This difference can be explained by the more intense processing that audiovisual content requires,123 which consequently implies a higher degree of attention towards the video content compared to emails with text or picture messages. It should also be noted that the studies on e-mail communication were conducted in 2004 respectively 2007, when the amount of digital communication was less developed than it is at present.124 This information overload has most likely created a degree of jadedness towards advertisement within certain segments, which consequently increases the demands on content quality and creativity, which could be a contributing factor to the insignificance of contextual factors. Hence, earning the medium, which has become a term widely used among practitioners in the field of digital media, implies the importance of hard work in terms of creating a valuable content for the web in order for a video to go viral.
5.2 Additional Communication Effects
Similar to traditional marketing research, viral video communication will positively affect brand attitude if the receiver experience a positive attitude towards the video per se. This will in turn influence both the intentions to share the video, as well as purchase intentions. Brand attitude, based on the exposure to a branded online video, is mainly dependent on two factors in addition to the traditional dimensions of brand awareness and advertising attitude, which are not directly connected to the videos, namely; video attitude and inconsistency.
Chiu (2007), Phelps (2004) De Pelsmacker (2005) p. 142 124 Rosengren (2008)
122 123
51
Content is King
Rune & Vinberg
The results of this study indicate a difference between well-known brands and those the respondents were less familiar with. Firstly we can conclude that brand attitude is significantly dependent on video attitude, which in turn is highly affected by how the content is perceived. Secondly, in order to get attention, the content should be atypical in comparison to previous communication from the brand. The latter mainly applies to strong, well-known brands since a prerequisite for inconsistent communication is already established key associations to the brand, resulting from a long and visible existence on the market.125 In this study McDonalds is the most obvious example of this phenomenon. This implies that large, well-established companies should not be afraid of stepping outside of what they normally want their brands to represent in order to increase the curiosity for the brand. Inconsistent content is thus likely to get more attention, better top-of-mind awareness and greater consumer interest in the brand, which has been stated in previous research. 126 An example of this effect is the initial negative reaction from Coca-Cola’s marketing division when confronted with the video of the Diet Coke and Mentos experiment, which were mentioned in the introduction. This video has proven to be tremendously successful as a marketing campaign, whereupon Coca-Cola now uses the buzz to attract traffic to their website.127 Less known brands have more to gain by focusing on qualitative content and video attitude through creativity, authenticity, engagement and non-provocative content. An example is the “T-Mobile Dance” (see appendix 2) where a large group of people together perform a dance in the middle of rush hour at Liverpool Station in the UK. The clip was ranked highest in terms of video attitude even though T-Mobile does not exist on the Swedish market. This implies that viral campaigns are accessible for all companies independently of the awareness of the brand, although brand awareness still promotes brand attitude and purchase intentions, according to the study.
5.3 Critique
Even though the respondents watched the exact same version that is to be found online, it should be acknowledged that certain aspects differed compared to a real life scenario. Although the experiment was constructed to be as similar to reality as possible, noise related to the use of social networking sites and video sharing sites, i.e. title and comments from other viewers, was removed in the survey. The anonymity of the sender, even though it was stated if the source of the message was
Törn (2009) p. 16 Törn (2009) 127 www.thecoca-colacompany.com
125 126
52
Content is King
Rune & Vinberg
a friend, an acquaintance or a company, could reduce the impact on the sender-recipient relationship results compared to real life. It should also be noted that in real life people make the active choice of watching a video as the ones included in the survey, as oppose to our study where the respondents have taken part in an experiment with a compensation of 5 SEK. Furthermore, the sample of respondents could be questioned as it may not be representative for the users of digital media, which are mainly Internet savvy and relatively younger age groups. This can influence whether or not the conclusions are entirely generalizable.
5.4 Suggestions for Further Research
This study has provided further understanding on the consumer behaviour related to audiovisual communication on the web. There are however several issues that need to be addressed in adjacent areas of study. A prerequisite for this study has been the respondents’ exposure to the video. Hence , the initial behaviour when receiving such a video, i.e. underlying motivations to watch the video, has not been possible to study. It would thus be of interest to determine whether contextual factors, as the number of views or sender-recipient relationship, would influence the motivation to click on the video in the first place, in similarity with findings on the reasons for opening e-mails.128 Instead of segmenting the sample by demographic or psychographic variables, we prioritized to identify general conclusions on the perceptions of viral videos and the motivations for sharing them with others. The reason for this is the lack of theory and previous research that exists within the area of audiovisual web communication. However, since previous studies have shown that the tendency to forward e-mails depends on the personal traits of the sender,129 i.e. if the sender is conscientious or extravert, it would be of interest to see if the same can be said for audiovisual content. To study a cross-cultural sample would also increase the applicability of the results, since the Internet provides a communication channel that goes pass country boarders. User generated content is becoming increasingly frequent and it is thus of the highest interest to further study the underlying reasons for consumers to contribute to the marketing of certain brands, as well as the implications and affects this will have on commercial communication praxis. Studying
128 129
Phelps et al. (2004) Chiu (2007) p. 529
53
Content is King
Rune & Vinberg
the common traits of the most popular videos, they all share a certain element of mysticism. They are frequently commented and the stories of how they were created as well as the originators of the videos are widely discussed, both among consumers as well as in the press (see for example T-Mobile Dance, Taylor Momsen Escapes Paparazzi and Kobe Bryant Jumps over Car). The aspect of storytelling and the connections between guerrilla marketing and digital w-o-m to create buzz and improve the attitude towards the video is something that we have not been able to study more closely but that we believe is of great interest. As previously stated, the findings of this study indicates the importance of engaging the viewer in the content of a viral video in order for that person to show the film clip to others, which is an action required to be performed at a later occasion than immediately after the event of the video being watched. Although we have used a wide set of high- and low involvement products being presented in the videos used in the survey, the limited scope of this thesis has restricted the level of which we were able to study differences between these products as influencers. We look forward to see further research on this subject in the future.
54
Content is King
Rune & Vinberg
6. REFERENCES
Ackermann Greenberg, D. (2007-11-22) “The Secret Strategies Behind Viral Videos”, Tech Crunch,http://www.techcrunch.com/2007/11/22/the-secret-strategies-behind-many-viral-videos/ (retrieved 2009-05-22) Alexa.com (2009-05-23)http://www.alexa.com/topsites (retrieved 2009-05-23) Arndt, J. (1967) “Role of Product-Related Conversations in the Diffusion of a New Product”, Journal of Marketing Research, Aug, 291-295 Babin, B., J., Darden, W. R., Griffin, M., (1994) “Work and/or Fun: Measuring Hedonic and utilitarian Shopping Value“, Journal of Consumer Research, vol 20 Balasubramanian, S., Mahajan, V., (2001), “The Economic leverage of The Virtual Community”, International Journal of Electronic Commerce”, vol 5, pp. 103-138 Blackwell, R.D., Miniard, P.W. & Engel, J.F. (2006), “Consumer Behavior”, 10th edition, Thomson Higher Education, Mason, USA Blomstedt, A., Björkman, J. (2008), ”Är All Reklam Bra Reklam?”, Master Thesis, Department of Marketing & Strategy, Stockholm School of Economics Braw. E. (2008-01-16), “Netanel Sprider Facebook i Världen”, Metro,http://www.metro.se/se/article/2008/01/16/07/1900-42/index.xml (retrieved 2009-03-23) Brewer, M. (2000) “Research Design and Issues of Validity”. In Reis, H. and Judd, C. (eds) Handbook of Research Methods in Social and Personality Psychology. Cambridge: Cambridge University Press., England Chatterjee, P. (2001) “Online Reviews: Do Consumers Use Them?”, Advances in Consumer Research, Vol. 28 Chiu, H., Hsieh, Y., Kao, Y., Lee, M., (2007), “The determinants of Email Receivers’ Disseminating Behaviors on the Internet”, Journal of Advertising Research, Dec 2007 p.524-534 Cialdini, R. B. (2001). “Influence - Science and Practice” 4th edition, Allyn and Bacon, New York, USA The Coca-Cola Company, “Products and Packaging Myths & Rumours”,http://www.thecoca-colacompany.com/contactus/myths_rumors/packaging_mentos.html, (retrieved 2009-05-26) Dahlén, M. (2003) “Marknadsförarens Nya Regelbok”, 1st edition, Liber Ekonomi, Malmö, Sweden
55
Content is King
Rune & Vinberg
Dahlén, M., Lange, F. (2007) ”Optimal Marknadskommunikation”, 2nd edition, Liber Ekonomi, Malmö, Sweden Davis., A., Khazanchi, D., (2008), “An empirical Study of Online Word of Mouth as a predictor for Multi-product category e-commerce Sales”, Electronic Markets, vol 18 (2) p. 130-141 De Pelsmacker, P., Geuens, M., Van den Bergh, J. (2005), “Foundations of Marketing Communications – A European Perspective”, Pearson Education Limited, Essex, England Dichter, E., (1966), “How Word-of Mouth Advertising Works”, Harvard Business Review, Nov-Dec Facebook.com (2009-05-23)http://www.facebook.com/press/info.php?statistics (retrieved 200905-23) Hennig-Thurau, T., Gwinner, K.P., Walsh, G., Gremler D.D. (2004) “Electronic Word-of-Mouth via Consumer-Opinion Platforms: What Motivates Consumers to Articulate Themselves on the Internet”, Journal of Interactive Marketing, Winter, 2004 p. 38-52 Iriberri, A., Leroy, G. (2009), “A Life-Cycle Perspective on Online Community Success”, ACM Computing Surveys, Vol. 41, No. 2, Article 11 Kincaid, D., Drummon, P., S., (2007), “Power Shift”, Marketing Magazine, vol 112, p. 26-26 King, Rachel (2007-01-16), “The Diet Coke and Mentos Explosion”,http://www.businessweek.com/technology/content/jan2007/tc20070116_569213.htm, (retrieved 2009-05-26) Machleit, K., A., Allen, C. T., Madden, T. J. (1993), “The Mature Brand and Brand Interest: An Alternative Consequence of Ad-Evoked Affect”, Journal of Marketing, vol 57, p. 72-82 Machleit, K., A., Wilson, D., R. (1988),"Emotional Feelings and Attitude Toward the Advertisement: The Roles of Brand Familiarity and Repetition", Journal of Advertising, vol 17, Number 3, Pages 27-35. MacKenzie, B., Lutz, R., J., Belch, G., E., (1986), "The role of Attitude Toward the Ad as a Mediator of Advertising Effectiveness: A test of Competing Explanations", Journal of Marketing Research, vol 23. MacKenzie, S. B., & Lutz, R..J. (1989), “An Empirical Examination of the Structural Antecedents of Attitude Toward the Ad in an Advertising Pretesting Context”. Journal of Marketing, April, Vol. 53:2, 48-65. Malhotra, Naresh K, Birks D. (2007) ”Marketing Research - An Applied Approach”, 3rd Updated European Edition. Prentice Hall Mano, H., (1996) “Assessing Emotional reactions to TV Ads: Replication and Extension with a brief Adjective checklist”, Advances in Consumer Research, 23, 63-69
56
Content is King
Rune & Vinberg
Modezelewski, M., F., Niehaus R., W., (2000-09-13),”Finding a cure for Viral Marketing”, Direct Marketing News,http://www.dmnews.com/Finding-a-Cure-for-Viral-Marketing-Ills/article/68355/ (retrieved 2009-05-22) Modig, E., Lethagen, H. (2008), “Unbundling Creativity”, Master Thesis, Department of Marketing & Strategy, Stockholm School of Economics Muñiz, A. M. & Schau, H. J. (2007) “Vigilante Marketing and Consumer-Created Communications”. Journal of Advertising, Vol 36:3, 35-50. Notani, A., S., (1998), "Moderators of Perceived Behavioral Control's Predictiveness in The theory of planned Behaviour: A Meta Analysis", Journal of Consumer Psychology, 7(3), 247-271 Phelps, J.E., Lewis, R., Mobilio, L., Raman, N., (2004), “Viral Marketing or Electronic Word-ofMouth Advertising: Examining Consumer Responses and Motivations to Pass Along E-mail”, Journal of Advertising Research, Dec 04 p 333-348 Pressler, J. (2008-10-28) “Internet Crush: Joel Moss Levinson”, New York Magazine,http://nymag.com/daily/intel/2008/10/internet_crush_joel_moss_levin.html , (retrieved 2009-05-24) Puto, P., C., Wells, D., W., (1984), “Informational and Transformational Advertising: The Differential Effects of Time”, Advances on Consumer Research, Vol.11 Issue 1, p638-643 Ratchford, Brian T. (1987), “New Insights About the FCB Grid”, Journal of Advertising Research; Aug/Sep87, Vol. 27 Issue 4, p24-38, 15p Rosengren S., (2008) “Facing Clutter: On Message Competition in Marketing Communication”, Economic Research Institute, Stockholm School of Economics (EFI) Skolverket, (2004-03-02)http://www.skolverket.se/sb/d/205/a/825;jsessionid=1D88DF692A47C95C0D8942929FE2EF92 (retrieved 2009-05-24) Stafford, T. F. & Stafford, M. R. (2002), “The Advantages of Atypical Advertisements for Stereotypical Product Categories”. Journal of Current Issues and Research in Advertising, Vol 24:1, 25–37. Strauss, J., Frost, A., (2001) “E-marketing”, 2nd edition, Upper Saddle River, N.J: Prentice Hall Sundaram, D., Mitra, K., Webster, C. (1998), “Word-of-Mouth Communication: A Motivational Analysis”, Advances in Consumer Research, 25, p. 527-531 Sundaram, D., S,. Webster, C. (1999) “The Role of Brand Familiarity on the Impact of Word of Mouth Communication on Brand Evaluations”, Advances in Consumer Research, p664- 670 Söderlund, M. (2001) ”Den lojala kunden”, 2nd edition, Liber Ekonomi, Malmö.
57
Content is King
Rune & Vinberg
Söderlund, M., (2005) ”Mätningar och mått- i marknadsundersökarens värld”, Liber, Malmö Thomas, M., Brain, D., (2008),”Crowd Surfing”, A & C Black Publishers Ltd, London, England Thomas, A. (2009-01-24) “Pope hopes new YouTube channel will reach young people”,http://www.christiantoday.co.uk/article/pope.hopes.new.youtube.channel.will.reach.young.people/ 22374.htm, (retrieved 2009-05-23) Törn, F. (2009), “Challenging Consistency – Effects of Brand-Incongruent Communications”, Dissertation, Economic Research Institute, Stockholm School of Economics, Sweden Wipperfürth, A., (2005) “Brand Hijack: Marketing Without Marketing”, Portfolio Trade, New York YouTube Fact Sheet (2009),http://www.youtube.com/t/fact_sheet (retrieved: 2009-05-24) Zinkhan, G., M., Kwak, H., Morrison, M., Peters, C., O., (2003), “Web-Based Chatting: Consumer Communication in Cyberspace”, Journal of Consumer Psychology, 13 (1&2), p. 17-27 YouGov Sweden: www.yougovsweden.se YouTube: www.youtube.com Viral Video Chart: www.viralvideochart.com Worldstats.com, December 2008,http://www.internetworldstats.com/europa.htm#se, (retriever 2009-03-23)
58
Content is King
Rune & Vinberg
7. APPENDICES
7.1 Appendix 1: Selected Videos - Summary
Video name Brand Product No. Of views (16/2-09) Length Discovered Video Ranking Selected quotes from respondents
“I thought that the commercial was festive, but it had absolutely nothing to do with chocolate. That was what I remembered. That it came up a candy bar at the end. But I didn’t think much about the brand unfortunately." ”I guess the brand has something to do with ’low fat’. Maybe Lätta or something similar.” "I have no idea which brand it is. It wasn’t evident in the video, which I think is really bad."
1. Cadbury Eyebrows 2. Taylor Momsen Escapes Paparazzi 3. Tea Partay 4. All I Want for Xmas is a PSP
Cadbury
Milk Chocolate: Low involvement/Feel
Sportswear/sneaker s: High involvement/Think Vodka: Low involvement/Feel Portable gaming console: Low involvement/Think
4,076,069 views
1.00min
23-jan-09
4
Nike Smirnoff Sony Guitar Hero / Activision
588,912 views 5,361,096 views 55,572 views 1,985,285 views 5,531,688 views 4,550,675 views 11,771,013 views 5,721,089 views 520,375 views
59sec
13-jan-09 06-sep-06 15-dec-06
8 12 14
2.10min
5. Bike Hero 6. Durex Get It On 7. Walk-in Fridge 8. Diet Coke + Mentos 9. T-Mobile Dance 10. McNugget Rap 11. Making of the Ryanair Cabin Crew Charity Calendar 2009 12. Nintendo SixtyFOOOOOOOOUR 13. Kobe Bryant Jumps Over Car 14. Lactacyd, "Protect Yourself" 15. Dove Onslaught
Video game: Low involvement/Think
1.22min original: 3.36min, short: 1.00min
"PSP..something. Whatever that is."
18-nov-08
9
Durex Heineken Coca-Cola & Mentos T-Mobile McDonald 's
Condoms: High Involvement/Think Beer: High Involvement/Feel
30sec 32sec original: 2.57min, short: 1.10min
14-jan-09 30-dec-08
3 2
"I didn’t think about which brand it was, I listened to the music
" ”Hehe, I’m sorry. I’ve already forgotten which brand it was. It was a word of approximately 6 letters and it was a condom brand.” ”The brand is Heineken – beer, featuring a very clever allusion to Sex & the City!” “Both Mentos and Coke (which I don’t know of) were in the video. I guess that both companies are marketing their brands at the same time.” ”The brand was T-Mobile, but I didn’t think it had much to do with that really.” "The brand is Chicken McNuggets, McDonald’s. – Anyone who missed that???" “The brand was Ryan Air – but after watching this film clip I never want to fly with this sexist airline!! Totally degrading towards the female cabin-crew-genus and not the least to us as passengers!!” “I don’t know which brand it was. Feels like some kind of cheap crap toys that you trick children into liking, like BR or something.”
Soda & candy: Low Involvement/Feel Mobile phone subscription: High Involvement/Think Fast-food: Low Involvement/Feel
06-sep-06 16-jan-09 11-jul-07
6 1 13
02.41min
32 sec
Ryanair
Airline: High Involvement/Think Gaming console: Low Involvement/Think Sportswear/sneaker s & sports car: High Involvement /Think & Feel Intimate care products: High Involvement/Think Beauty products: High Involvement/Think
242,228 views
1.42min
11-nov-08
15
Nintendo Nike & Aston Martin
10,639,135 9,871,826 views
1.06min
06-sep-06
11
52sec
10-apr-08
7
"The brand was probably Nike but I heard the name Aston Martin more." "It would be interesting to know which brand it was. Is the product as confused as the film clip?"
Lactacyd Dove
106,930 views 2,055,341 views
30sec
12-jan-09 01 Oct 2007
10 5
1.18min
"?????????"
59
Content is King
Rune & Vinberg
7.2 Appendix 2: Selected Videos
60
Content is King
Rune & Vinberg
61
Content is King
Rune & Vinberg
7.3 Appendix 3: Online Survey (example from Nike, Scenario 4)
62
Content is King
Rune & Vinberg
63
Content is King
Rune & Vinberg
64
Content is King
Rune & Vinberg
65
Content is King
Rune & Vinberg
66
Content is King
Rune & Vinberg
67
doc_421472117.pdf
Viral marketing, viral advertising, or marketing buzz are buzzwords referring to marketing techniques that use pre-existing social networks and other technologies to produce increases in brand awareness or to achieve other marketing objectives
Stockholm School of Economics Institute of Marketing and Strategy Master Thesis Spring 2009
Content is King
Communication Effects of Viral Video Marketing
The technological paradigm shift has enabled new forms of communication. Consequently, this offers valuable possibilities for marketers in terms of reach, speed and cost-efficiency. Due to similarities with the spread of a virus, the phenomenon of self-replicating content on the web via social networks and word-of-mouth is commonly referred to as viral marketing. Even though this phenomenon is becoming more frequently used by companies to increase brand awareness and improve brand attitude, there is little academic research on the subject. The objective of this study is to identify factors that affect consumer attitude and behaviour, following the exposure to a viral video, as well as studying how these features interact. In contradiction to previous research on adjacent areas of study, the result indicates that contextual factors do not influence neither the attitude towards the message nor the intention to share it with others. Instead, content has proven to be the single most important feature to improve attitudes and consequently viral circulation. Conclusively, a communication effects model for viral video marketing is presented, illustrating the flows of consumer attitude and behaviour in a digital context. Keywords: viral marketing, online videos, digital communication
Authors: Lena Rune, 20263 Hélène Vinberg, 20149
Tutors: Professor Micael Dahlén PhD Student Jonas Colliander Examiner: Associate Professor Patric Andersson
Discussants: Karin Bjerde, 20264 Jakob Ståhl, 20693
Presentation: June 12, 2009, 15.15-17.00 SSE Room C606
Content is King
Rune & Vinberg
Many Thanks to:
Micael Dahlén and Jonas Colliander, Stockholm School of Economics Jan Enander and Annelie Önnerud Åström, YouGov Sweden Marcus Wenner, Désirée Maurd and Caroline Jungsand, United Minds/Prime PR Fredrik Törn and Niclas Öhman, Stockholm School of Economics Hilma Gustavsson, Johan Östberg and Filippa Malmegård Fredrik Egger and Magnus Koch-Schmidt
1
Content is King
Rune & Vinberg
TABLE OF CONTENTS
1. INTRODUCTION.................................................................................................................................. 4 1.1 Background .......................................................................................................................................... 5 1.2 Problem Area ....................................................................................................................................... 6 1.3 Purpose ................................................................................................................................................ 6 1.4 Delimitations ........................................................................................................................................ 6 1.5 Definitions ........................................................................................................................................... 7 1.6 Previous Research and Expected Contribution ..................................................................................... 9 2. THEORETICAL FRAMEWORK ....................................................................................................... 9 2.1 Communication Effects Model ........................................................................................................... 11 2.2 Context .............................................................................................................................................. 12 2.2.1 2.2.2 Social Proof.............................................................................................................................. 12 Sender-Recipient Relationship .................................................................................................. 13
2. 3 Content ............................................................................................................................................. 14 2.3.1 2.3.2 Attitude .................................................................................................................................... 14 Behaviour ................................................................................................................................ 18
3. METHODOLOGY ............................................................................................................................... 24 3.1 Choice of Subject ............................................................................................................................... 24 3.2 Scientific Approach ............................................................................................................................ 24 3.3 Experimental Design .......................................................................................................................... 25 3.3.1 3.3.2 3.3.3 Sample of Respondents ............................................................................................................ 25 Sample of Videos ..................................................................................................................... 26 Construct of Survey.................................................................................................................. 28
3.4 Pre-studies .......................................................................................................................................... 30 3.4.1 3.4.2 Classification of Products ......................................................................................................... 30 Pre-test Survey ......................................................................................................................... 31
3.5 Analytical Tools .................................................................................................................................. 31 3.6 Research Quality ................................................................................................................................. 31 3.7 Sources of Information ....................................................................................................................... 32 4. ANALYSIS AND RESULTS ................................................................................................................ 33 2
Content is King
Rune & Vinberg
4.1 Hypotheses Testing ............................................................................................................................ 33 4.1.1 4.1.2 4.1.3 Contextual Factors ................................................................................................................... 34 Attitude .................................................................................................................................... 36 Behaviour................................................................................................................................. 40
4.2 Conclusion ......................................................................................................................................... 46 5. DISCUSSION AND IMPLICATIONS ............................................................................................... 49 5.1 Motivational Factors Influencing the Intention to Share a Viral Video................................................ 49 5.2 Additional Communication Effects .................................................................................................... 51 5.3 Critique .............................................................................................................................................. 52 5.4 Suggestions for Further Research........................................................................................................ 53 6. REFERENCES ..................................................................................................................................... 55 7. APPENDICES....................................................................................................................................... 59 7.1 Appendix 1: Selected Videos - Summary............................................................................................. 59 7.2 Appendix 2: Selected Videos............................................................................................................... 60 7.3 Appendix 3: Online Survey ................................................................................................................. 62
3
Content is King
Rune & Vinberg
1. INTRODUCTION
”It’s not what you do. It’s what they do with what you do.” 1 - John Grant, The New Marketing Manifesto In June 2006, professional juggler Fritz Grobe and his lawyer friend Stephen Voltz decided to create a video of a prank experiment they had discovered a few months earlier. Dressed in white coats and protective glasses the two men looked like mad scientists when they performed the show, putting mint pastilles from the brand Mentos into bottles of Diet Coke. The reaction caused by the experiment has since then been commonly referred to as a “Mentos Eruption” or “Coke Geyser”, due to the explosion effected by the chemical reaction of the two products. Though only one person was told about the video, once uploaded, views went up at a rapid pace of 1000 hits per hour,2 making Grobe and Voltz world celebrities overnight. However, the massive impact that the online circulation of Grobe and Voltz’ experiment caused was met with scepticism by Coca-Cola. Unlike Mentos who responded positively to the unexpected marketing of their brand, giant Coca-Cola said to the Wall Street Journal "We would hope people want to drink [Diet Coke] more than try experiments with it," referring to the experiment as something that did not fit the “personality” of Diet Coke.3 However, following the upload of a great number of videos from others performing the stunt and a circulation reaching millions of people, the soda manufacturer changed their mind. Eventually, Coca-Cola publically stated their intentions to capitalize on the experiment by announcing competitions on their website on the theme “Poetry in Motion”.4 The happily ending story of Diet Coke and Mentos, which made the odd couple of a juggler and a lawyer world-famous, while reaching millions of potential consumers and providing unexpected exposure of two brands, is one of the many success stories that illustrate the magnificent potential of using the Internet for viral video marketing.
Wipperfürth (2005) p. 46 King (2007-01-16) www.businessweek.com 3 Ibid. 4 www.thecoca-colacompany.com
1 2
4
Content is King
Rune & Vinberg
1.1 Background
The Internet enables consumers to share information cost- and effortless and never before have consumers been able to influence brands to the extent that is now the case. 5 The social networking site Facebook and the video sharing platform YouTube are ranked as the third and fourth most visited sites on the Internet.6 Facebook has over 200 million registered users worldwide7 and even the Vatican has realized the power of YouTube by launching the Pope’s very own channel on the site.8 Recent technology enables not only exposure, but also engagement in a way that has previously not been possible. Brand community fan sites,9 spoof ads,10 uploaded corporate jingles and user generated video commercials11 are all examples of how consumers have been inspired by a brand or advertisement to the degree that they have become part of creating the commercial content. As opposed to traditional marketing where there is an obvious sender–receiver relationship between a company and the consumer, the relationship today has become substantially more complicated. In addition to being the receiver of the message, the consumer is also functioning as the medium and channel of communication. The essence of viral marketing is thus how to motivate consumers to spread a commercial message without any economic incentives or promise of other material gain, hence earning the medium. Without the spread, an online video will simply become a clip among millions of others, lost in cyber space. However, even though the challenges to create a successful viral campaign are numerous, the potential rewards are even greater. Baring this in mind, it is our intention to add knowledge to this topic in order to keep pace with the rapid development of consumer behaviour. As stated by marketing researchers Muñiz and Shau “The revolution won’t be televised. Instead, it will be expertly rendered and edited by dedicated users, and then distributed via e-mail and YouTube.”12
Kincaid & Drummond (2007) www.alexa.com (2009-05-23) 7 www.facebook.com (2009-05-23) 8 Thomas (2009-01-24) www.christiantoday.co.uk 9 Muñiz & Shau (2007) p. 36 10 Björkman & Blomstedt (2008) p. 4 11 Pressler (2008-10-28) www.nymag.com 12 Muñiz & Shau (2007) p. 47
5 6
5
Content is King
Rune & Vinberg
1.2 Problem Area
Even though the most successful viral campaigns have reached millions of people the vast majority have passed unremarked, illustrating the need to understand the forces of viral marketing. The fact that it is the consumers themselves that enable the spread of information is also the challenge. What motivates consumers to share information of commercial substance with fellow human beings? Which attributes of a video influence the perception of its content? Do positive attitudes towards a message equal the intention of forwarding it? Due to the novelty of the subject there is a need for academic research explaining the phenomenon of campaigns that have gone viral. Since viral marketing can be tremendously cost-efficient, but also very unpredictable, it is of great academicand managerial benefit to analyse the reasons behind this phenomenon. This thesis attempts to contribute with an increased understanding of the features that impact a consumer’s intention of forwarding commercial communication in the form of online videos as well as examining the effects that this type of communication will have on brand attitude and purchase intentions.
1.3 Purpose
The purpose of the thesis is to: Identify factors that affect consumer attitude and behaviour following the exposure to a viral video as well as studying how these features interact.
1.4 Delimitations
Due to the scope and the purpose of the thesis, a number of delimitations have been made. Since the purpose of the thesis is to study context- and content related features that impact consumer intentions to forward visual messages online, we have chosen to study videos that have been successful on video sharing sites like YouTube13 and Viral Video Chart14. The fact that they exist on Viral Video Chart, which is a site that compiles videos holding the largest number of views and commercial exposure online at a given time, implies that they are to be considered as relatively successful in terms of reach. With the intention of maximizing the relevance of the thesis we have exclusively used videos that have been uploaded as branded content, i.e. with a commercial purpose.
Website where consumers can stream (watch) and upload videos: www.youtube.com Website that ranks online videos in relation to which extent they have been mentioned on blogs, including data on geographical distribution and reach: www.viralvideochart.com
13 14
6
Content is King
Rune & Vinberg
The sample selection is constituted by 964 (main survey: 913 + control group: 51) respondents representative for the Swedish population in the age of 16-64, who are familiar with the Internet. Even though a cross-cultural sample would increase the validity, the scope of the thesis made it necessary to delimit the respondents to Swedish respondents. However, since the population of Sweden is relatively Internet savvy15 and far developed in terms of social network participation,16 a sample of Swedish citizens is well motivated for the purpose of analyzing online behaviour. The language used in the films is English, with the exception of the Heineken film, which is in Dutch. Since the sample consists of Swedish respondents, a sample of Swedish films would have been preferable in terms of language comprehension. However, as it was not within our power to find a sufficient number of Swedish films that fulfilled the requirements of quality as well as representing various categories of products and brands, the films had to be chosen out of the large number of international films available. Obviously, there is a slight risk of the respondents misinterpreting the messages of the videos since the audio is not in their native language. However, given the visual narrative of the content, the videos should be understood irrelevant of language knowledge. The risk of misinterpretations is also reduced by the fact that Swedes have a high understanding of English.17 Since research in adjacent areas of study, i.e. e-mails, have shown the influence that social proof and the source of the message have on the attitude towards the e-mail as well as the likelihood of sending it forward, these two contextual factors are considered as the most relevant to test. Even though factors as personal relevance and situational mood, as well as personal traits have been stated to affect the perception of commercial content, the limited scope of the thesis has required us to disregard these factors.
1.5 Definitions
Communication: Since this thesis attempts to study communication flows in a web context, we believe it to be of value to explain the term communication. We therefore use Hovland, Janis and Kelley’s definition of
Penetration (Dec 08): 80,2 %, www.internetworldstats.com (2009-03-03) Sweden holds the 6th largest population of Facebook-users in the world: Braw (2008-01-16) www.metro.se 17 www.skolverket.se (2004-03-02)
15 16
7
Content is King
Rune & Vinberg
communication as “the process by which an individual transmits stimuli to influence the behaviour of other individuals.”18 Social Networking sites: These sites are online platforms for people to socially connect and interact with other users. Hagel and Armstrong define them as “computer-mediated spaces where there is a potential for an integration of content and communication with an emphasis on member-generated content.”19 Viral videos are usually shared through these sites, particularly as many of them collaborate with video streaming sites through links in connection to the videos. Popular sites include Facebook, Myspace, Live Spaces, Bebo, Digg, Hi5, Orkut, Mixx, Linkedin, Lunarstorm, Habbo, A Small World, Twitter and Del.ic.ious. Video sharing websites: Due to the novelty of the phenomenon, there is no generally accepted definition of video sharing sites that we have been able to found. In comparison to the social networking sites, the video sharing websites are platforms for users to upload and watch streamed video clips online. The difference between downloading and streaming is that while downloading results in data being stored on your computer, streaming enables you to listen or view data without storing it first. Although social interaction occurs on video sharing websites, especially in terms of commenting on the videos, the purpose of these sites differ from social communities as networking revolves around the theme of videos, rather than the interaction itself. Online Word-of-Mouth: Sundaram defines word-of-mouth as a “form of interpersonal communication among consumers concerning their personal experiences with a firm or a product”.20 The difference between traditional w-o-m and online w-om (e-w-o-m) is that the exchange of information and opinions takes place in an “anonymous asynchronous online environment”.21 This phenomenon is also sometimes referred to as word-of-mouse communication.22 A clear advantage of online w-o-m is generally the potential of rapid reach.23
Chiu (2007) p. 525 Iriberreri & Leroy (2009) p. 2 20 Sundaram (1998) p. 527 21 Davis & Khazanschi (2008) p. 130 22 Blackwell et al. (2006) p. 214
18 19
8
Content is King
Rune & Vinberg
Viral Marketing: Viral marketing is principally the online equivalent to word-of mouth24 and was introduced by Steve Jurvetson and Tim Draper in 1997.25 ”Viral marketing differs from word-of-mouth in that the value of the virus to the original consumer is directly related to the number of other users it attracts”26. Although the expression may imply the information being spread in a virus-like fashion, viral marketing has no connection to spam or other destructive sorts of digital content. Rather, a viral marketing campaign seeks to use the cost-effective forces of spreading information and content exponentially, using social networks online.
1.6 Previous Research and Expected Contribution
Recent studies that focus on consumers’ intentions of interacting with other members in a web context have mostly considered exchange of text messages by e-mail,27 consumer-platforms,28 online reviews29 and chat-rooms.30 Phelps has studied why people forward e-mails, concluding that messages that spark emotions as humour, fear, sadness, or inspiration are more likely to be forwarded.31 He also emphasises the importance of reaching an appropriate target group since people who receive an e-mail from a friend will not consider it as “junk” and will therefore be more likely to forward it than messages received by companies. These findings are confirmed by Chiu, who furthermore states that the extent that a message is perceived as hedonic or utilitarian will affect the probability of the message being forwarded.32 Hennig-Thurau has studied the motivations that are related to affective behaviour and people’s need of social interaction on consumer platforms online.33 The author distinguishes four prime motivations; desire for social interaction, desire for economic incentives, concern for other consumers and the potential to enhance self-worth. An additional study on the incentives of engaging in online communities is Balasubramanian and Mahajan’s34 paper on
Davis & Khazanschi (2008) p. 131 Strauss & Frost (2001) p. 245 25 Phelps (2004) p. 334 26 Modezelewski (2000) p. 334 27 Phelps (2004), Chiu (2007) 28 Hennig-Thurau (2004) 29 Chatterjee (2001) 30 Zinkhan (2003) 31 Phelps (2004) p. 344-345 32 Chiu (2007) p. 530 33 Hennig-Thurau (2004) p. 41-43 34 Balasubramanian & Mahajan (2001) p. 125-126
23 24
9
Content is King
Rune & Vinberg
economic leverage in virtual communities. They identify three sources of utility; focus related utility, which derives from the belief that the contributions of each member will strengthen the community, consumption utility, which refers to the value that is received by consuming the contributions of other community constituents, and approval utility that is the satisfaction that occurs when other members of the network consume and approve of the constituent’s own contributions. Since previous research concerning online w-o-m mainly has focused on messages that contain text of some sort, there is a lack of understanding of how people perceive audiovisual content as well as their motivations and behaviours to spread such content on the web. Our expectation and aspiration is thus to contribute with increased understanding of the implications that exposure to such a message will have on consumer’s attitudes and behaviour. The undeniable future importance of the Internet as a marketing tool and communication channel imply the significant academic- and managerial benefits that an increased understanding of consumer behaviour within this area will have.
10
Content is King
Rune & Vinberg
2. THEORETICAL FRAMEWORK
In this section the theoretical framework upon which we construct the analysed hypotheses is presented.
2.1 Communication Effects Model
When formulating the hypotheses of our study we have used a general communication effects model as reference point, thereby attempting to explain the relationship between exposure of the video and the following effects on attitude and behaviour. This model will thus function as a theoretical framework to increase the understanding on how the various steps are interrelated.35 Figure 2.1 illustrates the causal relationship between consumer attitudes towards a specific advertisement, which will have spill over effects on the brand presented in the advertisement. This will in turn affect the consumer’s intention to purchase the brand’s products as well as his/her intentions to spread the message of the advertisement; generally through w-o-m. The model has been modified to fit the researched subject in terms of viral videos. The grey coloured components of the model are added by the authors to illustrate the significance of content and contextual factors on video attitude as well as the expected direct relationship between video attitude and intentions to share. These consist of three different types of behaviour; send the video forward, show the video to someone or create a similar version of the video.
Figure 2.1 Communication Effects Model36
The model was chosen due to its simplicity and applicability to the subject of studying consumer behaviour. It illustrates the causal relationship of the exposure to advertisement, which has been researched in previous studies, however not in a digital setting. Thus, using this basic model as reference point enables the study to draw on previous research in order to extend its applicability and expand the model further to fit the researched area of viral video marketing. 36 MacKenzie et al (1986), Sundaram & Webster (1999), Notani (1998)
35
11
Content is King
Rune & Vinberg
The following section will focus on the driving factors that influence change in attitude and behaviour, related to viral video marketing. These changes are important marketing objectives as they reflect basic psychological traits of consumer behaviour.37 Hence, understanding the flows and relationships between these motivational factors facilitate value creation.
2.2 Context
The perception of a message is not only affected by the content per se, but also depends on the source of the message as well as contextual factors, as can be seen in the model above (figure 2.1).38 It is therefore of interest to study how contextual factors, such as “social proof” and the relationship the recipient has to the sender will influence the perception of the message content, i.e. the videos.
2.2.1 Social Proof
Emotional gain or mental gratification of some sort is a prerequisite for the sender to engage in online w-o-m. Such value is never material but rather affective and related to the recognition from others.39 In order to be validated by other constituents of the society it is necessary that the consumer adds value to the community, i.e. focus-related utility.40 Adding value has various definitions in this context and can refer to both recommending positive experiences to others41 or preventing them from experiencing bad products or services,42 as well as expressing consideration and care for others.43 Since the desire to get recognition from other members will influence one’s attitude and behaviour44 it is likely that the attitude towards the content of a message would be impacted if the recipient perceives the message to be popular; i.e. if a large number of like-minded people previously have viewed and shared the message. According to online media practitioner Dan Ackerman Greenberg, co-founder of viral video marketing company The Comotion Group,45 driving up the number of views of a video on social networking sites to reach the “most views” page, is essential to any successful viral campaign. The thoughts of Ackerman Greenberg are coherent with the
Blackwell (2006) p. 88 MacKenzie & Lutz (1989) p. 49 39 Dichter (1966) p. 151 40 Hennig-Thurau (2004) p. 42, Chiu (2007) p.526 41 Dichter (1966), Blackwell et al. (2006), Sundaram et al. (1998) 42 Sundaram et al. (1998) 43 Phelps (2004) p. 343 44 Dichter (1966) p. 150 45 Ackerman Greenberg (2007-11-22)
37 38
12
Content is King
Rune & Vinberg
psychological theory on social proof,46 which states that people are drawn to things that others are attracted to. This implies that the number of views will influence the perception of the video both in terms of attitude as well as the likelihood of spreading the message forward. Thus, we formulate the following hypothesis:
Hypothesis 1: The number of views will positively influence the recipients in terms of
a: perceived level of ad creativity b: video attitude c: perceived level of provocation d: perceived level of engagement e: perceived level of authenticity f: perceived level of production quality g: intention to send the video forward h: intention to show the video to someone i: intention to make a similar content video
2.2.2 Sender-Recipient Relationship
Several researchers have shown the impact that the relationship between the sender and the recipient has on the attitude towards the message. If the sender is considered to be a friend, the recipient perceives the message to be of value and that the sender has forwarded it for a good reason.47 Those who receive messages from someone with whom they consider to have an interpersonal relationship are also more willing to forward the message than those who receive messages from unfamiliar acquaintances or commercial sources.48 This implies that there will be a difference between the recipients’ behaviour depending on whether the sender is claimed to be a friend, an acquaintance, or the company that occurs in the video. Thus the closer the relationship between the sender and the recipient, the better the attitude towards the video in terms of ad creativity, video attitude, engagement, authenticity and production quality will be. In similarity with previous reasoning, the closer the relationship the bigger the likelihood of the message being forwarded, hence we formulate the following hypothesis:
Cialdini (2001) Phelps (2004) p. 345 48 Chiu (2007) p. 529
46 47
13
Content is King
Rune & Vinberg
Hypothesis 2: The relationship between the sender and the recipient will positively influence the receiver in terms of
a: perceived level of ad creativity b: video attitude c: perceived level of provocation d: perceived level of engagement e: perceived level of authenticity f: perceived level of production quality g: intention to send the video forward h: intention to show the video to someone i: intention to make a similar content video
2. 3 Content
2.3.1 Attitude
In this section, factors solely related to the content of the video and the effect these might have on the video attitude will be studied. We will also conduct a comparative analysis between the group that has seen the video before and the group that has not. Finally, we will discuss if video attitude have an impact on brand attitude, as well as identifying additional factors that will have an influence on consumer behaviour, which is implied by the communication effects model, figure 2.1. Video Attitude In addition to emotional and cognitive recognition that motivates sender and recipient to engage in word-of-mouth, sharing a message with others also requires the message to have some sort of intrinsic value. Value can be defined as “a person’s relativistic preference with particular things or events”, and thus “both an event's usefulness and an appreciation of activities comprising it can indicate value” .49 Hence, whether the video is perceived to be of some value to the recipient, this will influence his or her intention to forward it, as can be seen in figure 2.1. Value is usually divided into two categories; utilitarian and hedonic50 (informative and transformative51). Utilitarian value is primarily instrumental, functional, and cognitive, providing customer value as the means to an end.52 Hedonic, on the other hand, is an outcome related to spontaneous responses that are more subjective and personal. 53 Examples of hedonic values are entertainment, exploration and self-expression, which are noninstrumental, experiential, and affective.54 In experience- and post-purchase evaluations both terms
Babin et al. (1994) p. 645 Chiu (2007) p. 526 51 Puto & Wells (1984) p. 638 52 Chiu (2007) p. 526 53 Babin et al. (1994) p. 646 54 Chiu (2007) p.526
49 50
14
Content is King
Rune & Vinberg
are relevant. However, since only an insignificant part of the videos that are currently figuring online are of utilitarian value, we have chosen to focus mainly on people’s tendency to share information that they perceive to be of hedonic value. Consumers are likely to provide information to others due to their desire to share joy and positive emotions they have experienced.55 It has also been shown that messages that evoke strong emotions are more likely to be forwarded.56 Features that evoke such emotions include humour,57 novelty and highly engaging content.58 We have therefore used the Ad Creativity model (Novelty, Meaningfulness, Humour, Positiveness and Well-Craftiness), developed by Modig & Lethagen,59 since these factors have been found to influence the attitude towards advertisement. Word-of-mouth will also be stimulated if the consumer perceives to have “inside information” and if the content appears to be authentic.60 Since provocative material can stimulate w-o-m61 it is also of interest to study the impact that such material will have on the attitude towards the video. The general attitude towards advertisement is considered to influence the attitude towards a specific commercial, hence the better the general advertisement attitude the better the attitude towards the specific video.62 We therefore test whether the general attitude towards the video is correlated with the perceived level of ad creativity, provocation, engagement and authenticity. In addition, we also test whether a positive attitude towards advertisement in general will improve the attitude towards the video. Thus we formulate the following hypothesis:
Ibid. p.526 Phelps et al. (2004) p.345 57 Ibid. (2004) p.345 58 Dichter (1966) p. 162 59 Modig & Lethagen (2008)
55 56
Dichter (1966) p. 165 Phelps (2004) p. 345 62 MacKenzie & Lutz (1989)
60 61
15
Content is King
Rune & Vinberg
Hypothesis 3.1: The attitude towards the video will be positively influenced by the following variables:
a: perceived level of ad creativity b: perceived level of provocation c: perceived level of engagement d: perceived level of authenticity e: perceived level of quality f: general advertisement attitude
Repetition According to previous research evaluating people’s attitudes towards TV ads, respondents who were exposed to the ad more than once, compared to respondents that saw the advertisement for the first time, appeared to evaluate the ad more favourable and less dull.63 This can be explained by the fact that the stimulus that consumers are exposed to when watching an advertisement will remain in their mind. Even though the attitude towards the brand is actually related to the repetition of exposure, the consumer does not necessarily understand this relation, and instead relate it solely to the content. Hence, the consumer misattributes the source of the processing simplicity; with a positive attitude towards the advertisement per se.64 To conclude, the attitude towards the video will be better if the respondent has been exposed to the advertisement previously.65 We therefore formulate the following hypothesis:
Hypothesis 3.2: The respondents who have seen the video before will have a more positive attitude towards the
video. Brand Attitude Brand attitude can be defined as a measure of “how much a person likes or dislikes a brand or to the extent to which he or she holds a favourable view of it.” 66 The better the attitude towards the brand, the higher the intention and likelihood of purchasing a product or service will be. Even though attitudes are relatively stable over time they can change and thus one of the main objectives of marketing communication should be to change the attitude in favour of the brand,67 thereby increasing the probability of purchase.68
Mano (1996) De Pelsmecker et al. (2005) p.41 65 Machleit (1988) 66 De Pelsmecker et al. (2005) p. 26 67 De Pelsmecker et al. (2005) p. 41 68 Dahlén & Lange (2007) p. 98
63 64
16
Content is King
Rune & Vinberg
In order to see the impact that viral videos have on brand attitude by examining whether or not the exposure to a video changes attitude, we formulate the following hypothesis:
Hypothesis 4.1: Brand attitude will be more positive for the respondents who have seen a branded video than those
who have not.
Based on the theory of hierarchies of communication effects,69 brand awareness will influence the attitude towards a brand, which thus is believed to apply to viral video marketing as well. According to the Dual mediation model, the evaluation of the advertisement or commercial has an immediate impact not only on the perception of the content per se, but also an indirect effect on the attitude towards the brand.70 This relation makes it probable to believe that the brand attitude will be influenced by how the respondents have perceived the video. Recent studies have identified an abundance of marketing messages, which has lead to advertisement avoidance and scepticism.71 It is therefore likely to assume that the perception of a commercial message is influenced by the initial advertisement attitude. Due to the immense amount of advertisements that people meet each day and the limited cognitive capability of the human mind, consumers tend to acknowledge only a fraction of the commercial communication they are exposed to, so called “selective scanning”.72 To use inconsistent advertisements73 or so called “shock of difference”74 has therefore been acknowledged as an effective way to get attention and stimulate word-of-mouth. Atypical advertisement requires more effort for the consumer to process and is therefore also given more attention. 75 However, advertising that is inconsistent with brand image might evoke an initial negative attitude towards the ad even though the effect on brand attitude can be positive.76 Inconsistent advertisement also has the power to push through the established
Ibid. p. 97 MacKenzie et al (1986) p. 131 71 Rosengren (2008) 72 Blackwell (2006) 73 Törn (2009) 74 Dichter (1966) p. 162 75 Stafford & Stafford (2002) p. 51 76 Törn (2009)
69 70
17
Content is King
Rune & Vinberg
schemas of well-established brands;77 by creating brand interest as an additional dimension to brand attitude as solely ad-evoked effects will not have a strong influence on mature brands.78 Due to the previously mentioned factors we therefore test how the attitude towards the brand will be affected by the attitude towards the video itself, the degree of brand awareness and the general attitude towards advertisements. It is also of interest to test whether atypical communication will impact the brand attitude. Thus, we formulate the following hypothesis:
Hypothesis 4.2: The attitude towards the brand will be positively influenced by the following variables:
a: video attitude b: brand awareness c: general advertisement attitude d: atypical communication
2.3.2 Behaviour
Intentions are the individual’s subjective estimation of their likelihood to perform a specific action later on. However, intentions should not be considered as certain predictions of the actual behaviour,79 even though studies show that there are correlations between intentions of doing something and then actually doing it.80 We therefore find it interesting to study word-of-mouth- as well as purchase intentions.
Intentions to Share:
Motivational Factors to Send a Video Forward A prerequisite when launching a viral campaign is the willingness of a consumer to forward the received content. A consumer’s tendency to spread information, so called word-of-mouth, has since the 1960’s been recognized as an important tool for marketers81 and is claimed to have a bigger influence on the actual purchase decision than mass media communication.82 Since word-of-mouth communication takes place without the sender expecting economic rewards and because consumers normally are more receptive towards communication that stems from people with whom they have a
Dahlén (2003) Machleit (1993) 79 Notani (1998) p. 248 80 Blackwell et al (2006) p. 411 81 Dichter (1966) 82 Arndt (1967)
77 78
18
Content is King
Rune & Vinberg
relationship,83 word-of-mouth is perceived as more credible than traditional advertising, which consequently increases the attention of the recipient. This implies that the information is processed more thoroughly in the consumer’s mind compared to company communicated commercials, and the final effect will thus be greater.84 The importance of content to motivate consumers to forward messages has been stated by several researchers.85 After being exposed to content that evoke strong emotion, many share their experience as a way of reducing tension created during the processing of the content.86 Messages containing content perceived as humorous, scary, sad and inspirational are therefore more likely to be forwarded. E-mails with “humorous jokes, touchingly sad stories and particularly apt inspirational messages”87 are considered as relevant enough to be sent forward even for those who normally are reluctant to share these types of messages with others. Relevant for the probability of forwarding is also the degree of work required from the sender to be able to pass it on. A prerequisite to reach a large community with one’s message is thus the simplicity of forwarding the message; virus only spread when they are easy to transmit.88 Due to the simplicity of forwarding videos on YouTube, both within the platform as well as to other platforms (Facebook, MySpace, Del.icio.us, Digg etc.), it can be assumed that the context (social network platform) in which the sender evaluates whether s/he should forward the video or not fulfil the requirements of a channel, where it is “easy to transmit”. However, since forwarding the message via Internet, no matter the technological progress, always will require some sort of effort it is also of interest to see whether the probability of showing the video to others depends on the same factors as forwarding it online. The simplicity of forwarding the message is related to the core function of the social network i.e. interacting with others. Due to the relationship between the members of the community it is natural that adding value to the community is frequently occurring.89 However, the belief that each member’s contribution to the community will increase the value or focus (focus related utility) requires that the message in fact contains qualitative content. Thus, whether the recipient will forward the message to other constituents is a direct result of the degree of value s/he perceives the message to have, i.e. the
Dahlén & Lange (2007) p. 110 Dahlén & Lange (2007) p. 112 85 Phelps (2004), Chiu (2007) 86 Dichter (1966) p. 149 87 Phelps (2004) p. 345 88 Chiu (2007) p. 531 89 Balasubramani & Mahajan (2001) p.116
83 84
19
Content is King
Rune & Vinberg
attitude towards the video. The satisfaction that members experience from the recognition of other consumers (approval utility) is also dependent on the quality of the video since the need for approval impedes the sender to forward something that s/he thought would not receive positive approval.90 Since certain research shows that negative or provocative content that cause fear and sadness stimulate a need of word-of-mouth as tension release,91 we have also chosen to test to which degree provocative material will motivate people to forward the message. The majority of the members of video platforms who share content with each other are relatively young.92 Thus, it is also of interest to see whether younger and more Internet savvy people are more likely to forward videos than others. For these reasons we formulate the following hypothesis:
Hypothesis 5: The intention of forwarding the video to others will be positively influenced by the following variables
(except for age which will influence this behaviour negatively): a: perceived level of ad creativity b: video attitude c: perceived level of provocation c: perceived level of engagement d: perceived level of authenticity e: perceived level of production quality f: general advertisement attitude i: video initiator: brand j: video initiator: private person k: age l: Internet usage
Motivational Factors to Show a Video to Others Since the most common objective for a company when up-loading a video online is reach, awareness, positive ad attitude and consequently an improved brand attitude, one can question if it matters whether the consumer decides to forward the message or show it to his/her friends. A significant difference between the two types of actions is the level of processing required by the viewer after the event of watching an online video before deciding to share it with others. Thus, taking the time to show a video to someone else requires more effort than simply forwarding it
Balasubramani & Mahajan (2001) p. 125 Phelps (2004) p. 345 92 52% of 18-34 year old users of YouTube share videos often with friends and colleagues: YouTube Fact Sheet (2009) www.youtube.com/t/fact_sheet
90 91
20
Content is King
Rune & Vinberg
online and accordingly, a real life recommendation imply an even higher level of involvement in the video. It is thus of great interest to study whether any differences between the two exists in terms of motivational features connected to a viral video. Therefore, we formulate the following hypothesis.
Hypothesis 6: The intentions to show the video to others will be positively influenced by the following variables
(except for age which will influence this behaviour negatively): a: perceived level of ad creativity b: video attitude c: perceived level of provocation c: perceived level of engagement d: perceived level of authenticity e: perceived level of production quality f: general advertisement attitude i: video initiator: brand j: video initiator: private person k: age l: Internet usage
Motivational Factors to Create a Similar Video Recent technological progress has enabled consumers to increase their participation in the creation of marketing and brand related content, i.e. user generated content (UGC). The rules of the game are changing and those who learn to adapt - the Crowd Surfers – have much to gain by letting go of their brand in favour of empowering the consumers in order to let them participate in shaping the destiny of the company.93 There is thus a tendency towards consumers making their own versions of commercial advertisements which resembles the originals both in form and intent. Consumers creating such branded content believe themselves to be self-appointed advocates of the brand, with strong opinions of what is right and wrong for it.94 Studies show that this tendency not only is expected to grow stronger but that consumers are increasingly savvy in their understanding of styles, tropes, logic and grammar of advertising and that a great deal of the content is widely spread by being up-loaded on the Internet. This phenomenon is mainly occurring for well established, unique and powerful brands, with which consumers feel connected to.95 In order to examine the tendency of user generated content, we formulate the following hypothesis:
Thomas & Brain (2008) Muñiz & Shau (2007) p. 35 95 Ibid. p. 46
93 94
21
Content is King
Rune & Vinberg
Hypothesis 7: The intention of creating a similar content version of the video will be positively influenced by the
following variables (except for age which will influence this behaviour negatively): a: perceived level of ad creativity b: video attitude c: perceived level of provocation c: perceived level of engagement d: perceived level of authenticity e: perceived level of production quality f: general advertisement attitude i: video initiator: brand j: video initiator: private person k: age l: Internet usage
Brand Attitude Influencing Intentions to Share According to Sundaram & Webster, studies show that brand familiarity and attitude will influence consumer attitude and behaviour in several ways; i.e. in terms of information processing style, ultimate brand choice, recall of advertisements as well as increased purchase intentions. Prior brand exposure is also stated to have a positive effect on the attitude which consequently improves the brand evaluation. Since highly positive evaluations create a need for tension release, this imply that the higher the brand attitude, the stronger the intention to share with others.96 In order to examine the applicability of the previously presented communication effects model, figure 2.1, in relation to the factors mentioned above we formulate the following hypothesis:
Hypothesis 8.1: The intention to send a video forward is positively dependent on brand attitude. Hypothesis 8.2: The intention to show a video to someone is positively dependent on brand attitude. Hypothesis 8.3: The intention to create a similar video is positively dependent on brand attitude.
Purchase Intention The final goal of all marketing communication is to increase likelihood of purchase within the targeted segment; i.e. creating an intention to purchase. A communication strategy must therefore take into account all steps of the purchase process to avoid losing consumers on the way.97 The MAO model (motivation, ability, opportunity) explains the route to the purchase occasion. It indicates that due to obstacles in terms of lack of availability and opportunity to purchase the product, generating motivated consumer is all the
96 97
Sundaram & Webster (1999) p. 666 Dahlén & Lange (2007) p. 98
22
Content is King
Rune & Vinberg
more important.98 Brand awareness has previously been stated to improve brand attitude and consequently purchase intention. Since pre-determined attitudes toward advertisements in general has shown to influence the attitude towards a specific commercial, it is of interest to determine whether this factor also influence purchase intentions.99 Due to advertisement clutter, it is becoming increasingly difficult to gain attention from consumers and therefore we intend to study whether content perceived as atypical compared to previous advertisement from the brand will influence purchase intentions.100 Thus, we formulate the following hypothesis:
Hypothesis 9: Purchase intentions are positively dependent on the following variables:
a: brand awareness b: general advertisement attitude c: atypical communication d: brand attitude
De Pelsmecker et al. (2005) p. 28 MacKenzie & Lutz (1986) p. 131, Törn (2009) p. 16 100 Törn (2009) p. 21
98 99
23
Content is King
Rune & Vinberg
3. METHODOLOGY
In this section we will present the methodology regarding how the thesis has been constructed and carried out.
3.1 Choice of Subject
Following the development of digital media, the term viral communication is increasing in use among academics and managers alike, working in the field of marketing and strategy. When we explored the opportunities of studying digital media, we discussed the subject with a number of people from the PR and marketing field to be able to understand which problem area is of highest interest to the industry. Coherently, we came to the conclusion that the rapid development of technology and the change in consumer behaviour that comes with this, brings new challenges and opportunities for marketers to adapt to. Therefore, we chose to study the phenomenon of spreading commercial video clips virally. As of today, there are a number of success stories of companies that have reached millions of viewers with their message, simply by planting a video clip on YouTube or an equivalent website. However, breaking through the clutter of millions of videos on these sites tends to be difficult. An important step towards creating a viral campaign is to identify the drivers of consumer behaviour in the digital context, why our intention is to contribute with further understanding of the area in question.
3.2 Scientific Approach
Since the objective of the thesis is to make a general conclusion of the features that influence consumers to engage in online w-o-m, a quantitative study was found to be the most appropriate. The reasons for this are that a quantitative approach simplifies establishing the reliability of the thesis and clarifies the analysis. Furthermore, as we have chosen to base the research on theory from adjacent areas of study and testing these using a large sample of respondents, the study called for a conclusive research design. The attempt to create a model that identifies determinates of attitude towards the ad and behavioural implications imply a causal approach. Since general conclusions are made from testing specific hypothesis on a population sample the approach is deductive.101
101
Malhotra & Birks (2007) p. 160
24
Content is King
Rune & Vinberg
3.3 Experimental Design
The quantitative study was conducted in collaboration with the research company YouGov,102 which provides tailor-made market research solutions for its clients, based on high quality internet and information technology with a panel of approximately 2 million members across the world. YouGov’s Swedish Internet Panel consists of 42 000 individuals who receive monetary rewards for their participation. The partnership enabled the experiment to be performed on a large scale as well as in a digital setting. In order to control the variables of the study, an experimental design was constructed based on fifteen video clips that each was preceded by a survey based on 3x2 scenarios. These manipulations were presented referring to who the sender of the video was (close friend/distant acquaintance/the company showing in the video) and how many that had viewed the video before, referring to social proof (many/few viewers). After the respondents had read the background manipulations and watched the short video they were asked to answer questions regarding their attitude and thoughts concerning the video content and the figuring brand, as well as their intentions to pass the clip forward. In order to segment the respondents, questions regarding their attitude towards advertising in general were also included. The name of the videos were not shown in the survey to avoid any kind of priming, especially concerning the film clips where the brand was only subtly evident to the viewer.
Sent from / No. of views Many views - 5 000 000 Few views - 100 Distant Acquaintance 3 4 Company in video 5 6
Best Friend 1 2
Table 3.1 Scenarios; source of message and number of views
3.3.1 Sample of Respondents
The collaboration with YouGov enabled a large sample of 964 respondents representing the Swedish population aged 16-64, based on three criteria: age, gender and geographical region. The participants were rewarded with a small monetary contribution of 5 SEK per individual. The survey contained a main survey of 913 respondents and an additional 51 respondents for the control group, testing only brand attitude for each brand represented in the video clips. Since the target group for a viral video campaign consist of people who are active on the Internet, we have chosen to use this channel to conduct the survey. It is also noteworthy that the design of the experiment allowed for
102
www.yougovsweden.se
25
Content is King
Rune & Vinberg
six different manipulations (3x2) and fifteen video clips, which in turn left only circa ten respondents to each manipulation and film. These sub-samples cannot in other words be perfectly divided among the population. However, the videos were randomly distributed among the recipients and the small sub-groups only serve as dividers between the manipulations where each film responds to a number of approximately 60 respondents.
3.3.2 Sample of Videos
In order to be able to test the variables of the experiment, the video clips were chosen based on a number of criteria. Firstly, as ecological validity103 is preferred, the videos used in the experiment were authentic films featuring a vast selection of different brands known to the Swedish audience. However, the brand did not need to be explicitly evident in the video as this was a factor desired to examine. Furthermore, each of the fifteen videos has been present on the Viral Video Chart,104 proving their success in circulation and popularity. Because of the importance of using fairly new videos, the videos had to be “discovered” no earlier than in September 2006. They should neither be longer than 3 minutes (the longest film is “T-Mobile Dance”, which is 2.41 min) to ensure that the respondents would not get tired before completing the survey, nor should they have been shown on Swedish television to avoid priming. However, there have been a couple of exceptions to the last criteria as a similar version of the “Dove Onslaught” and “Heineken Walk-in Fridge” (see appendix 2) videos have been appearing on Swedish television during the time in which the survey was conducted, which unfortunately is something that could not have been foreseen. The conducted survey shows that 30% of the respondents who watched the Heineken video had seen it before as the commercial shown on television is the same version as the video in the survey, although with English audio instead of Dutch. The Dove Onslaught video had only been seen by 5% as the TVcommercial is a heavily modified version of the viral video used in the survey. Nevertheless, we choose to not see this as a problem as it enabled the findings of additional conclusions regarding the effects of repetition, which are presented in more detail in the analysis. Regarding the time frame of the videos, two of the film clips were shortened due to survey related issues imposed by the collaboration with YouGov. “Diet Coke + Mentos” and ”Bike Hero”, which originally are approximately 3 minutes long, were only shown to the respondents for ca 1 minute. Naturally, the
103 104
Brewer (2000) www.viralvideochart.com
26
Content is King
Rune & Vinberg
shortened videos were made sure to show the core of the content to ensure the same outcome as if the full version had been shown. As the aim of the thesis is to identify characteristics that increase the chances of a video going viral, the sample of film clips were chosen based on the models of Ad Creativity105 and the FCB Grid106. Our intention when choosing the videos was thus to select a sample of film clips that, based on our personal perception as well as comments presented on the video-sharing sites, represented both high and low levels of ad creativity, which has been confirmed when analysing the data. In addition, brands covering the full spectrum of high-/low involvement – think/feel scales, based on the theories of Ratchford, thus the FCB Grid, were also taken into consideration when selecting the videos to ensure a wide range of product categories. (See Fig. 3.1 below and appendix 1.)
Figure 3.1 The FCB Grid, Featuring the Products Presented in the Selected Videos
105 106
Modig & Lethagen (2008) p. 73 Ratchford (1987)
27
Content is King
Rune & Vinberg
3.3.3 Construct of Survey Main survey:
Each respondent group were faced with a scenario and a short video clip before answering the main survey, which is the same independently of video or scenario, except for the questions regarding the specific researched brand. As the survey was conducted in a digital setting, it was possible to ensure that all of the questions were answered and in the desired order. The collaboration with YouGov enabled strict control over the respondents’ background in terms of gender, age and geographical area in which they live, which made it unnecessary to add these questions in the survey. The complete survey is presented in appendix 3. Content The first battery of questions in the survey concerns how the respondents perceive the videos. In order to determine external influence based on the respondent’s previous experience with the video clip, they were asked if they had seen it before. Thereafter, the respondents were asked to state to what extent they agreed with a number of adjectives that would represent the content of the video. When we discussed which adjectives that would achieve the necessary data we used characteristics that Dichter states stimulate traditional word-of-mouth107 as a vantage point. Similar features as those mentioned by Dichter are included in the Ad Creativity model108 (Novelty, Meaningfulness, Humour, Positiveness and Well-Craftiness) and this model was therefore considered as appropriate to measure the perceived level of creativity of the various videos. To measure the attitude towards the video per se we used the well established variables “Good”, “Like it” and “Pleasant”.109 Since the variable “Positive” from the ad creativity model already was included in the survey and was considered as rather similar to “Pleasant”, the latter was removed since it did not add any significant value. The following variables were, during discussion between the authors and the tutors, decided as adequate to evaluate features of the online videos; provocative, engaging and authentic, each represented by three similar questions to ensure reliability in the measurements. Three questions were developed to study the perceived quality of the video in terms of cost, time and level of difficulty that was needed to produce the film.
Dichter (1966) Modig &Lethagen (2008) 109 Dahlén & Lange (2007) p. 100
107 108
28
Content is King
Rune & Vinberg
Perceived Sender To avoid priming the respondents by stating the brand represented by the video too early in the survey (other than manipulations 5 and 6) as well as getting a good view of how the respondents perceive the video in terms of whether or not the sender is evident, they were asked an open-ended question as to which brand(s) they could identify in the video. The respondents were also asked how difficult they perceived it to be to identify the brand as well as how familiar they were with it. In order to test the level of irrationality in the communication,110 the respondents were asked whether they perceived the video to be typical/atypical with the marketing they had been exposed to in the past by the brand. To be able to test the variable of who the respondents perceive as producer of the video (the brand/private individual), these questions were also added. Intentions to Pass the Message Forward Based on empirical studies of video streaming behaviour, we were able to identify three main actions representing the intentions to share an online video; namely passing the film clip forward digitally (mainly through social networking sites, blog posts and e-mail), showing the video to others in a physical setting, and prolonging the life of the video by replicating the content in an amateur production to upload on the Internet. Conventional measurements of intentions were used,111 asking the respondents if they would want to pass the video forward as well as the probability of them doing so. Due to limitation of the length of the survey, only two questions were used for each index; however we believed these to be the most relevant. Brand Attitude In this section of the survey, the respondents were told which brand that was present in the video clip. Attitude toward the brand was calculated through the traditional measurement of “What is your perception of the brand x?". The scale ranged from 1-7, with the options bad/good, dislike/like and negative/positive.112 (See the section on “control group” for more details). The respondents were also asked to answer whether they had changed their attitude toward the brand after watching the video.
Machleit et al (1993) Dahlén & Lange (2007) p. 99 112 Blomstedt & Björkman (2008) p.23
110 111
29
Content is King
Rune & Vinberg
Advertisement Attitude To study whether a predetermined attitude towards advertising in general would influence the attitude towards the video, the respondents were asked to respond to what degree they agreed with the following statements; “I like advertising”, “Advertising is annoying”, and “Advertising is more manipulative than informative”, using a scale of 1-7. Internet Usage To study whether familiarity with the Internet would have an impact on attitudes towards the video and brand as well as w-o-m and purchase intentions, a question regarding Internet usage was added, measured in terms of hours per day. Control Group The 51 respondents in the control group were each asked three questions measuring their attitude towards the full selection of brands featured in the different videos, without being exposed to either manipulations or the researched film clips. The questions measured to what extent the respondents’ experienced the brands to be good/positive/like, which then could be compared to the group of respondents who were exposed to the videos in order to see the differences made by the manipulation of seeing the film clips.
3.4 Pre-studies
In order to ensure a credible outcome of the experiment, two pre-studies have been conducted to (1) categorize the products tested in the survey and (2) to pre-test the survey questions.
3.4.1 Classification of Products
According to Ratchford,113 purchase decisions can be categorized based on two dimensions, namely level of involvement and think/feel. The four categories constitute the FCB Grid, named after Foote, Cone and Belding who developed the model. The pre-study was conducted on March 16th 2009, allowing a random sample of twenty respondents, aged 19-61, to answer eight questions developed by Ratchford for each of the products represented in the sampled videos. Interviews were conducted with 20 respondents by phone or face-to-face.
113
Ratchford (1987)
30
Content is King
Rune & Vinberg
The pre-test confirmed that a range of product categories were represented among the videos (see the results in Figure 3.1).
3.4.2 Pre-test Survey
To ensure comprehension of the survey, including videos and manipulations, a pre-test was conducted, allowing ten respondents to participate in the study and give comments on the questions they had difficulties understanding. The final experiment is the result of a survey fully comprehended by the respondents.
3.5 Analytical Tools
The analysis of the gathered data has been conducted using the software SPSS, including compare of means through independent-sample t-tests and one-way ANOVA as well as linear regression analyses. Each index has been measured using Cronbach’s Alpha or Bivariate Correlation to ensure internal consistency and computed by calculating the mean of the selected variables. When applicable, the data has been re-coded into a numeral language. This has been the case for the variable measuring recognition, which was presented in the survey as an open question. Results have been accepted at a 5% level of significance.
3.6 Research Quality
Research quality is mainly dependent on two factors; validity and reliability. The concept of reliability involves the discrepancy found between the observed and the real value due to random errors in measurement.114 Hence, if the reliability of a research study is high, anyone who attempts to do the same study in the same manner would reach the same results. In order to ensure high reliability, there are a number of alternatives, where the most common is to study the extent to which one reaches the same results from several related measurements, examining the same feature. The study conducted for this thesis uses recognized measurements from marketing research to ensure high reliability and thus high research quality. In addition, multiple questions measurements have been used, signifying a high Cronbach’s Alpha (? > 0.7) and for indexes made by two questions, a high level of Bivariate Correlation (? > 0.7), which concludes that the internal consistency in terms of high correlation between the questions is sufficient.
114
Söderlund (2005) p.134
31
Content is King
Rune & Vinberg
The purpose of validity is to see to what extent a study is liberated both from random and systematic errors,115 which implies that, in order to reach high validity, the study should be designed in such a way that the attempted phenomenon of study is the one that is actually being studied. According to Malhotra & Birks, it is desirable to create an experimental design that has both internal and external validity.116 The first measures whether the manipulations of the independent variables actually cause the effects on the dependent variable, whereas the latter measures the extent to which the cause-and-effects relationships found in the study can be generalized. The research design of our study is experimental, focusing on using manipulations in the form of scenarios and different video clips while keeping the additional design of the survey static. Furthermore, the collaboration with YouGov enabled a greater, heterogeneous sample of respondents who were approached and participated in the study in a digital setting, which diminishes the external factors that could affect the outcome of the study. Thus, both the internal and external validity of the study are strong as the design of the study has been thoroughly examined to measure what was intended, while the large sample of respondents enable the conclusions of the study to be generalized.
3.7 Sources of Information
In addition to gathered quantitative data from the surveys, a number of secondary sources as academic articles and dissertations, as well as books related to the area of study have been used. In addition to the academic material we have also used several articles from online published magazines.
115 116
Ibid. p.149 Malhotra & Birks (2007) p. 308
32
Content is King
Rune & Vinberg
4. ANALYSIS AND RESULTS
In this section we present the analysis and the findings we have made to point out the motivational factors and brand related effects of a viral video campaign. The results are consequently summarized in the conclusion and illustrated in an extended communication effects model.
4.1 Hypotheses Testing
Each hypothesis presented in the theory section is tested based on the main study using an interval scale of 1-7, ranging from negative to positive. The few questions that were stated in an opposing manner, ranging from positive to negative, were inverted. Results are accepted at a 5% significance level, with the exception of one border line case of p=0.051. When using linear regression, each model has been tested for adequate levels of autocorrelation, multicollinearity and heteroscedasticity using Durbin-Watson (close to 2), Condition Index (< 20 or < 30 if supported by theory) and Scatter Plot (satisfactory illustrations). The indexes used in the study have been tested for sufficient levels of correlation (> 0.7) using Cronbach’s Alpha for multiple-question measurements. The content related indexes are as follows: Ad Creativity (?=0.919), Video Attitude (?=0.958), Provocation (?=0.865), Engagement (?=0.938), Authenticity (?=0.945) and Production Quality (?=0.859). Since Cronbach’s Alpha is appropriate using a minimum of three questions,117 Bivariate Correlations at a significance level of 0.01, have been used to study internal consistency for the indexes with only two questions, which are intention to share the videos as well as the intention to purchase from the company: Intention to Forward (?=0.916), Intention to Show (?=0.939), Intention to Create a Similar Version (?=0.766) and Purchase Intentions (?=0.801). Brand Attitude (?=0.981) was measured using Cronbach’s Alpha. Regarding Advertisement Attitude, the result was only ?=0.695, which is somewhat lower than the adequate level of 0.7. However, since the same measurement has been used in previous research we believed it to be supportable to disregard this small difference and index the questions.
117
Söderlund (2005), p.146
33
Content is King
Rune & Vinberg
4.1.1 Contextual Factors
Social Proof Hypothesis H1a-i addresses the potential differences in perception of the videos and intention to share a film clip based on the number of views it has received. This was tested using an independent-samples t-test to compare the groups exposed to the scenarios 1, 3 and 5 versus scenarios 2, 4 and 6 (see the Methodology section for further details), where the manipulations stated that the previous number of views were either 5 000 000 views or 100 views. Our calculations do not indicate any differences between these groups at the 5% significance level (see table 4.1). Table 4.1 Social proof: Independent-samples T-test for H1a-i
Hypothesis Variable a) Ad creativity b) Video attitude c) Provocation d) Engagement e) Authenticity f) Production quality g)Send forward h) Show to others i) Make another version Many views Few views Difference Means (SD) Means (SD) in means Significance 3.70 (1.64) 3.54 (1.68) 0.16 0.16 3.60 (1.87) 3.49 (1.93) 0.11 0.36 2.90 (1.52) 2.88 (1.63) 0.02 0.87 3.12 (1.67) 3.06 (1.73) 0.06 0.63 3.15 (1.70) 3.01 (1.67) 0.14 0.22 3.26 (1.47) 3.29 (1.58) -0.03 0.72 2.43 (1.80) 2.60 (1.95) -0.17 0.19 2.38 (1.80) 2.48 (1.85) -0.10 0.42 1.44 (0.94) 1.46 (1.03) -0.02 0.82 Support for the hypothesis: Yes or No No No No No No No No No No
N=913 (Many views: 457, Few views: 456), Many = 5 000 000 views, Few = 100 views Scale: 1-7 (neg.-pos.)
The results imply that the notion of social proof does not affect the recipients’ perception of a video, nor does it affect their intentions to share the video with others. However, it is important to state that the results do not reject the quest for a viral video campaign to strive after retaining a significant number of views as this does not discard of the fact that a viral video needs to be found and heard of by the consumers to create a spin to its circulation. One potential reason behind the difference in our results compared to the advice given by practitioners like the Comotion Group, which specializes in increasing the number of views of a video in order to enable it to go viral, is the purpose. The aim of many companies is to earn a spot on the “most views” sites and spark other types of media to pick up on the story rather than social proof in itself being the cause of better
34
Content is King
Rune & Vinberg
perception or willingness to share the video. The sole existence of “most views” pages, imply that there is an interest to watch popular videos in terms of number of views. However, our findings show that this phenomenon does not affect the outcome for receivers of a video, opposed to people who actively seek this type of entertainment.118 Sender-Recipient Relationship H2a-i aims to find significant differences in the perception of the measured features of a video and the intention to share its content based on the relationship between the sender of a video and the recipient. The hypothesis was constructed based on the preceding scenarios that the respondents were exposed to before watching the chosen video clips and are thus a comparison between manipulation 1 & 2 (best friend), 3 & 4 (distant acquaintance) and 5 & 6 (company appearing in the video), which is elaborated in further detail in the Methodology section. The three groups were compared in terms of means using ANOVA, however the findings do not illustrate any significant differences at the 5% level between any of the groups (see table 4.2). Table 4.2 Sender-recipient relationship: One-way ANOVA for H2a-i
Hypothesis Variable a) Ad Creativity b) Video attitude c) Provocation d) Engagement e) Authenticity f) Production quality g) Send forward h) Show to others i) Make own version Best Friend Means (SD) 3.50 (1.70) 3.40 (1.94) 2.89 (1.63) 2.98 (1.72) 2.96 (1.69) 3.16 (1.49) 2.48 (1.93) 2.42 (1.89) 1.44 (1.02) Distant Acquaintance Company Means (SD) Means (SD) Significance 3.70 (1.66) 3.65 (1.64) 0.30 3.67 (1.90) 3.57 (1.86) 0.24 2.85 (1.61) 2.93 (1.48) 0.80 3.17 (1.73) 3.10 (1.63) 0.39 3.15 (1.71) 3.14 (1.65) 0.30 3.30 (1.51) 3.37 (1.57) 0.22 2.53 (1.84) 2.52 (1.89) 0.93 2.48 (1.83) 2.39 (1.77) 0.80 1.41 (0.93) 1.50 (1.00) 0.54 Support for the hypothesis: Yes or No No No No No No No No No No
N=913 (Best Friend: 297, Distant Acquaintance: 309, Company: 307), Scale: 1-7 (neg.-pos.) The results contradict the findings of a number of prominent researchers, including Phelps (2004) and Chiu (2007) who state that the closer relationship the recipient of digital content has to the sender, the more likely s/he will be to perceive it as more interesting and the more willing s/he will be to share the message with others. However, as the development of digital media is moving in a
118
The hypothesis was neither supported when examining the 15 videos independently.
35
Content is King
Rune & Vinberg
rapid pace, consumer behaviour changes quickly, which the results of this study demonstrates. There is obviously a difference in behaviour and attitude among the receivers of a viral video in comparison to recipients of e-mails, which is the digital medium mainly researched in the past. Consequently, the comparison between this study and previous research indicates that the change of consumer behaviour in a digital setting must be closely monitored in order to follow the rapid development. At present however, this demonstrates that receivers of a viral video are not sensitive to the relationship they have to the sender, with regards to their perception of the content or willingness to share the video with others.119
4.1.2
Attitude
Content Related Factors:
Dimensions of Video Attitude Hypothesis H3.1a-f was constructed to study the potential dimensions that influence the attitude towards a video. These were tested using linear regression with dependent variable video attitude and independent variables indexed as ad creativity, perceived level of provocation, perceived level of engagement, perceived level of authenticity, perceived level of production quality and general advertisement attitude. Results confirm that ad creativity (?=0.742), engagement (?=0.199) and authenticity (?=0.057) explain video attitude at the 0.1% significance level, however provocation (?=-0.081) affect video attitude negatively. The latter result is notable as it illustrates the negative feelings connected to the notion of provocation and clearly states that provocation should not be a desired goal when aiming to affect the viewer of a viral video in a positive manner. Hence, the more creative, engaging and authentic a viral video is, the better will the viewer think of it, however the less provocative the video is, the higher the attitude. Together the dimensions account for 90.8% of the variation in the measure (R²), which is highly satisfactory. The study does not confirm hypotheses H3.1e and H3.1f, which is somewhat surprising as these factors evidently play as significant part in explaining other types of attitude and behaviour (see the sections on intentions to share and brand attitude). However, when evaluating the actual video, general attitude towards advertising and the quality of the video do not affect the outcome. For details see table 4.3.1.
119
The hypothesis was neither supported when examining the 15 videos independently.
36
Content is King
Rune & Vinberg
Table 4.3.1 Dimensions of video attitude: Linear Regression for H3a-f
Hypothesis Independent variables a) Ad creativity b) Provocation c) Engagement d) Authenticity e) Production quality f) Advertisement attitude Standardized Beta (?) 0.742*** -0.081*** 0.199*** 0.057*** Support for the hypothesis: Yes or No Yes No Yes Yes No No
N=913, R² (adjusted) = 0.908 Significance: *p ? 0.05, **p ? 0.01, ***p ? 0.001 Durbin-Watson: 1.827 Condition Index: 12.544 Repetition According to previously stated theory, people who are exposed to marketing communication repeatedly tend to evaluate the advertisement more favourable and thus, H3.2 aims to see if this reasoning also applies to viral video marketing. The results indicate that there is a significant difference in attitude towards the videos, dependent on whether the respondents watch the film for the first time or not. Those who watch the video for the second time or more will generally evaluate the video as better than those who see it for the first time, illustrating the importance of repetition in digital media as well as conventional counterparts. Table 4.3.2 Repetition: Independent-Sample T-test for H3.1
Hypothesis Variable Video attitude Support for Seen video the before First time view Difference hypothesis: Means (SD) Means (SD) in means Significance Yes or No 4.68 (1.60) 3.48 (1.90) 1.20 0,000 Yes
N= 913 (Seen video before: 54, First time view: 859), Scale: 1-7 (neg.-pos.) Brand Attitude The general change in brand attitude is tested using an independent-sample t-test, comparing the means between the respondents who have been exposed to a viral video and those who have not.
37
Content is King
Rune & Vinberg
The test show no significant differences at the accepted 5% level, however there are indications (p=11%) that the brand attitude among those who have seen a viral video tend to have an improved attitude towards the brand than those who have not seen the video (see table 4.4.1 below). Comparing these groups should naturally acknowledge the consumers’ perception of the researched videos as these are signified by a wide selection of videos generating both positive and negative feelings. However, similar tests using only those videos generating high attitudes towards the video and vice versa show no different results. Hence, the data only indicate that there is no significant difference in brand attitude between those who have been exposed to the video and those that have not. Table 4.4.1 Brand attitude: Independent-sample T-test for H4.1
Support for the Seen video Not seen video Difference hypothesis: Means (SD) Means (SD) in means Significance Yes or No 4.58 (1.58) 4.34 (0.90) 0.24 0.11 No
Hypothesis Variable Brand Attitude
N= 952 (Seen video: 913, Not seen video: 39), Scale: 1-7 (neg.-pos.) Nevertheless, studying each video respectively, a sample of three of the videos indicates a change in brand attitude. “T-Mobile Dance”, which proved to be the most popular of the sampled videos, with a mean of 5.08 in video attitude (Scale: 1-7) among the respondents, indicates a change in brand attitude among those who have seen the video and those who have not. Durex’ video “Get it on” was also able to change the brand attitude among the respondents due to similar reasons as it was the third most popular video (video attitude mean=4.44) and one of the least known brands (T-Mobile was graded the lowest based on brand awareness and Durex fourth to last). What is more interesting is the change in brand attitude among receivers of the “McNugget Rap” video from McDonald’s, which was one of the least popular videos (graded third to last in video attitude: 2.90 on a 1-7 points scale). The change cannot be traced back to low brand awareness as McDonald’s is the most well known brand in the study (mean=6.23). However, it is notable that “McNugget Rap” was graded as the most atypical communication for the brand (mean=4.98) among the sampled videos, indicating the significance that atypical communication may have on brand attitude (see table 4.4.1.1).
38
Content is King
Rune & Vinberg
Table 4.4.1.1 Brand attitude by video: Independent-sample T-test
Not seen Seen video video Difference in Means (SD) Means (SD) means Significance 3.60 (1.23) 3.15 (1.11) 0.45 0.051 4.76 (1.49) 4.07 (1.24) 0.69 0.01 4.35 (1.69) 3.51 (2.01) 0.84 0.02 Support for the hypothesis: Yes or No Yes Yes Yes
Hypothesis Variables T-Mobile Durex McDonald's
T-Mobile: N=110 (Seen video: 61, Not seen video: 49) Durex: N=108 (Seen video: 61, Not seen video: 47) McDonald’s: N=110 (Seen video: 62, Not seen video: 48) Scale: 1-7 (neg.-pos.) As stated in the previous paragraph, the analyses of the data indicate that high video attitude causes respondents to change their attitude towards the brand for the better, especially with regards to less well-known brands. For more well-known brands inconsistency seems to play a more significant part in explaining change in brand attitude even if this does not imply that the respondents are positive towards the ad. Examining these tendencies in closer detail using linear regression demonstrates that the dimensions of brand attitude, with regards to viral video marketing, include video attitude (?=0.236), brand awareness (?=0.204), advertisement attitude (?=0.117) and atypical communication (?=0.082). These variables account for 13.2% of the variance in the measurement (R²), indicating that other factors need to be taken into account as well. The analysis was conducted based on only a limited part of the sample based on the respondents’ recognition of the researched brand. Hence, only 559 of the 913 respondents were used in this analysis as those who were not able to identify the brand in the video would add noise by answering questions regarding their perception of the brand based on the inaccurate idea of which brand that was in the video (See table 4.4.2).
39
Content is King
Rune & Vinberg
Table 4.4.2 Dimensions of brand attitude: Linear Regression for H4.2a-d
Hypothesis Independent variables a) Video Attitude b) Brand Awareness c) Advertisement Attitude d) Atypical Communication Standardized Beta (?) 0.236*** 0.204*** 0.117** 0.082* Support for the hypothesis: Yes or No Yes Yes Yes Yes
N=559, R² (adjusted) = 0.132 Significance: *p ? 0.05, **p ? 0.01, ***p ? 0.001 Durbin-Watson: 1.707 Condition index: 11.310
4.1.3
Behaviour
Intentions to Share:
Intention to Send a Video Forward Hypothesis H5a-k seeks to find the determinant factors of the intention to send a video forward, testing the variables ad creativity, video attitude, perceived level of provocation, perceived level of engagement, perceived level of authenticity, perceived level of production quality, brand awareness, advertisement attitude, brand as video initiator as well as private person as video initiator, age and Internet usage as potential dimensions of a causal relation, using linear regression. The results indicate that the decision to send a viral video forward is dependent on ad creativity (?=0.164), video attitude (?=0.448), perceived level of authenticity (?=0.099), perceived production quality (?=0.076) and advertisement attitude (?=0.106).Together the 5 dimensions account for 52.8% of the variation in the measure (R²). However, no evidence was found that provocation, engagement, video initiator, age or Internet usage are directly correlated with the intention of forwarding. See table 4.5 for further details of the analysis.
40
Content is King
Rune & Vinberg
Table 4.5 Motivational factors that influence the intention to send a video forward: Linear Regression for H5a-k
Hypothesis Independent variables a) Ad creativity b) Video Attitude c) Provocation d) Engagement e) Authenticity f) Production quality g) Advertisement attitude h) Video initiator: Brand i) Video initiator: Private Person j) Age k) Internet Usage Support for the Standardized hypothesis: Beta (?) Yes or No 0.164* Yes 0.448*** Yes No No 0.099** Yes 0.076** Yes 0.106*** Yes No No No No
N=913, R² (adjusted) = 0.528 Significance: *p ? 0.05, **p ? 0.01, ***p ? 0.001 Durbin-Watson: 1.962 Condition index: 22.471
Although the regression analysis does not indicate any dependent relationship between age and the intention to forward at the 5% level of significance, a comparing of means analysis using one-way ANOVA illustrates significant differences between age groups (16-30, 31-45, and 46-64). Young people, aged 16-30, show significantly higher intentions to send a video forward (2.8318 over 2.2570 on a 1-7 scale) than the higher age segment 46-64 (p= 0.001). Respondents 31-45 years old also show significantly higher intentions to forward a video (mean 2.6133, p=0.046) than the group of 46-64 year olds. Intention to Show a Video to Others Similar to the previous section, hypothesis H6a-k addresses the determinant factors that cause the recipient of a video to share it with others; more specifically the intention to show the video to others in a real life setting. The results imply that the motivators of sharing a video differ somewhat between the intentions to pass the video forward in a digital setting compared to showing it to someone in the “real” world. Measuring the same variables as H5 using linear regression provides the outcome of a number of motivators to explain the behaviour of sharing a video by showing it to
41
Content is King
Rune & Vinberg
others. These factors include attitude towards the video (?=0.564), perceived level of engagement (?=0.130), perceived production quality (?=0.078) and advertisement attitude (?=0.100) as motivating to show the video to another constituent of their real life network. The model provides a solid base for explanation, with the 4 independent factors accounting for 52.2% variation in the measure (R²). See further details in table 4.6. Table 4.6 Motivational factors that influence the intention to show a video to a friend: Linear Regression for H6a-k
Hypothesis Independent variables a) Ad creativity b) Video Attitude c) Provocation d) Engagement e) Authenticity f) Production quality g) Advertisement attitude h) Video initiator: Brand i) Video initiator: Private Person j) Age k) Internet Usage Standardized Beta (?) 0.564*** 0.130** 0.078*** 0.100*** Support for the hypothesis: Yes or No No Yes No Yes No Yes Yes No No No No
N=913, R² (adjusted) = 0.522 Significance: *p ? 0.05, **p ? 0.01, ***p ? 0.001 Durbin-Watson: 1.933 Condition index: 11.747 Similar to the intention to forward the video, showing it to others does not indicate to be dependent on age. However, segmenting the sampled population into three age groups illustrates differences at a significance level of less than 0.06%. The results show that “young” and “middle-aged” people aged 16-45 state significant higher intentions to show the video to someone than those aged 46-64 (16-30: 2.7114, 34-45: 2.5933, 45-64: 2.1489, mean based on a 1-7 scale).
42
Content is King
Rune & Vinberg
Intention to Create an Own Version of the Video Hypothesis H7a-k aims to find the motivational factors behind the intention of the respondent making a similar version based on the video s/he has been exposed to. The results indicate that high level of engagement (?=0.334) and production quality (?=0.102) of the content in a video motivates people to make their own version of it, while videos that are perceived as being made by a private person (?=0.082) rather than the brand appearing in the video, as well as a higher level of Internet usage (?=0.095), are additional related factors to consider. However, although the model is satisfactory with regards to measurements of autocorrelation, multicollinearity and heteroscedasticity, it is important to point out that the 4 variables presented in this study only account for 14.5% of the variation in the measurement (R²), implying that these factors alone does not fully explain what motivates the recipient of a video to make a similar-content film clip to upload online. See table 4.7 for further details. Table 4.7 Motivational factors that influence the intention to make an own version: Linear Regression for H7a-k
Hypothesis Independent variables a) Ad creativity b) Video Attitude c) Provocation d) Engagement e) Authenticity f) Production quality g) Advertisement attitude h) Video initiator: Brand i) Video initiator: Private Person j) Age k) Internet Usage Standardized Beta (?) 0.334*** 0.102** 0.082* 0.095** Support for the hypothesis: Yes or No No No No Yes No Yes No No Yes No Yes
N=913, R² (adjusted) = 0.145 Significance: *p ? 0.05, **p ? 0.01, ***p ? 0.001 Durbin-Watson: 1.873 Condition index: 9.677 The summarizing table above, illustrating the motivational factors affecting the intention to create an own version of a viral video, does not indicate that age is a contributing factor in this sense, at the 5% significance level. Nevertheless, comparing the means of these intentions, based on age groups
43
Content is King
Rune & Vinberg
using one-way ANOVA, demonstrates significant differences. Young people aged 16-30 are more willing to create an own version than those aged 46-64 (p=0.007). However, it is important to state that the intention to create an own version of a video is generally very low as the comparison lies between 1.6091 for those aged 16-30 and 1.3499 for the 46-64 age group at a 1-7 points scale. A summarizing illustration of the causal relation for the different kinds of intentions to share a viral video is found below.
Figure 4.1 Motivational Factors for Intentions to Share
The model illustrates the relationship of the features connected to watching a viral video that affects the intention to share the video with others, either by creating a similar version, showing the video to others or sending it forward. The features included in the model are those that have proven to be significant and illustrate both content related factors based on the respondents’ perception of the video as well as anticipatory factors that the respondents possess prior to the exposure of the viral video. Brand Attitude as Motivational Factor for Intentions to Share As indicated by theory,120 word-of-mouth intentions are affected by brand attitude. However, since there is little known research of the phenomenon regarding visual online word-of-mouth, a test was conducted using linear regression to establish the connection between brand attitude and the intentions to respectively send a video forward, show it to a friend and create a similar version for online publishing. Due to the construct of the survey where the questions regarding the intention to share the film clips were stated before the brand presented in the video was revealed, only a limited
120
Sundaram & Webster (1999)
44
Content is King
Rune & Vinberg
selection of respondents (N=559), including those who were able to recognize the brand, was used for this test. The results demonstrate that all of the three variables testing the intentions to share the message of the video are positively affected by brand attitude. However, it is important to state that the intention to send a video forward and to show it to others is more strongly affected by the respondents’ attitude towards the brand than the intention to create a similar content video is. In addition, R² is considerably low for each of the tests, indicating that brand attitude only accounts for a significantly small part of the variation in the measurements, which is expected when using only one independent variable. See table 4.8.1-3 for more details. Table 4.8.1 Brand attitude as motivational factor to send a video forward: Linear Regression for H8.1
Hypothesis Independent variable Brand Attitude Support for the Standardized hypothesis: Beta (?) Yes or No 0.241*** Yes
N=559, R² (adjusted) = 0.057 Significance: *p ? 0.05, **p ? 0.01, ***p ? 0.001 Durbin-Watson: 1.423 Condition index: 6.527 Table 4.8.2 Brand attitude as motivational factor to show a video to a friend: Linear Regression for H8.2
Hypothesis Independent variable Brand Attitude Support for the Standardized hypothesis: Beta (?) Yes or No 0.248*** Yes
N=559, R² (adjusted) = 0.060 Significance: *p ? 0.05, **p ? 0.01, ***p ? 0.001 Durbin-Watson: 1.385 Condition index: 6.527 Table 4.8.3 Brand attitude as motivational factor to make an own version: Linear Regression for H8.3
Hypothesis Independent variable Brand Attitude Support for the Standardized hypothesis: Beta (?) Yes or No 0.083* Yes
N=559, R² (adjusted) = 0.005 Significance: *p ? 0.05, **p ? 0.01, ***p ? 0.001 Durbin-Watson: 1.819 Condition index: 6.527
45
Content is King
Rune & Vinberg
Purchase Intentions The ultimate objective of any successful marketing campaign is to increase sales, which naturally also applies to the use of viral video marketing. To complete the extended communication effects model we test hypothesis H9a-d, which aims to identify factors related to a viral video campaign that determine the intention to purchase. The linear regression implies that brand awareness (?=0.070) and brand attitude (?=0.653) account for 45.3% of the variation in the measurement (R²) and thus provides a solid explanatory model for which viral video-related features that cause the respondents to purchase products or services from the brand in question (see table 4.9 for further details). Table 4.9 Dimensions of purchase intentions: Linear Regression for H9a-d
Hypothesis Independent variable a) Brand Awareness b) Advertisement Attitude c) Atypical Communication d) Brand Attitude Support for the Standardized hypothesis: Beta (?) Yes or No 0.070* Yes No No 0.653*** Yes
N=559, R² (adjusted) = 0.453 Significance: *p ? 0.05, **p ? 0.01, ***p ? 0.001 Durbin-Watson: 1.910 Condition index: 8.164
4.2 Conclusion
The results in the analysis illustrate discrepancies between previously studied digital communication and consumer behaviour connected to audiovisual messages. Our study indicates that the content is the sole most important feature that motivates receivers of an online video to share the message with others. Furthermore, content influence brand attitude positively, which consequently leads to increased purchase intentions. Contextual factors like sender-recipient relationships and number of views of a viral video do not have an effect on consumer attitude and behaviour according to this study. Instead, content related factors including perceived level of creativity (novelty, meaningfulness, humour, positiveness, and well-craftiness), authenticity, provocation and engagement will affect the recipients’ attitude towards the video (video attitude), which in turn encourages a positive attitude towards the brand (brand attitude) and intentions to spread the content of the video (send, show); consequently leading to higher purchase intentions among the receivers of the video. The study also found that repetition increases video attitude as the respondents who had been exposed to the video more than once demonstrated significantly higher evaluations in terms of positive attitudes towards
46
Content is King
Rune & Vinberg
the videos. Additionally, it is illustrated that the intentions to send a video forward, show it to others or create a similar film clip do not merely depend on the consumers’ brand attitude, but a number of factors, which will affect the outcome of the different behaviours respectively. Content is evidently the single most important factor that explains the consumer behaviour; however the dimensions of this feature prove different in terms of explanatory strength. The intentions to show and send forward a video are significantly affected by the respondents’ attitude towards the video, but there are also direct affects concerning the nature of the content influencing the decision. Authentic, creative videos with a high perception of production quality will more likely be forwarded digitally. For the intention to show the video to a friend, engagement is also significantly important as a substitute to authenticity and ad creativity. In addition to content related features, the respondents’ general attitude towards advertisement also proved to be important when deciding to share the message through digital word-of-mouth. In order to encourage the receivers of a viral video to create a similar video and upload it to a video sharing site, production quality and perceived level of engagement are important features of the film clip, while additional factors including high levels of Internet usage and whether the video is made by a private person in comparison to the company in question will affect the outcome positively. The analysis furthermore confirms the belief that segments based upon age will differ in terms of intention to share a viral video, where the younger generations are more likely to act upon this. Regarding brand related effects of a viral video campaign, traditional consumer behaviour theory is confirmed, stating the positive effects on brand attitude from general advertising attitude and brand awareness, in connection to a positive attitude towards the video. Atypical communication in the video also proved to have a positive effect. The analysis confirms the impact that a successful viral video campaign has on purchase intentions, by improving the brand attitude. A summarizing model of the findings is presented in figure 4.2 on the next page. 121
An analysis whether or not the gender of the respondents would illustrate differences in the outcome of the results has been conducted. However, this study does not find any differences in attitude or behaviour at the 5% level of significance beyond obvious differences, which is why this research has not been accounted for within the scope of this thesis.
121
47
Content is King
Rune & Vinberg
Figure 4.2 Communication Effects Model for Viral Video Marketing
The model illustrates the overall findings of the study, with the exception of the direct links between the respondents’ perception of the video and the intentions to share, which is found in further detail on page 44. The arrows illustrate a causal relationship between the squares whereas the straight lines connected to a circle illustrate the components of the square.
48
Content is King
Rune & Vinberg
5. DISCUSSION AND IMPLICATIONS
This section will discuss the findings of the thesis as well as present implications based on this knowledge. Furthermore, we will also consider critique towards the study and suggestions for further research. Using theory from traditional marketing communication as reference point, we are able to conclude that traditional communication effect models to a large extent is applicable also in a digital marketing context. Similar spill-over effects that a positive attitude towards a commercial will have on word-ofmouth intentions, brand attitude and consequently purchase intentions, can be applied to digital videos (see Figure 4.2). These findings provide a solid foundation to explore the driving factors and effects of viral video marketing in further detail as well as distinguishing differences. Hence, the traditional communication effects model offers a sustainable basis for the subsequent argumentation.
5.1 Motivational Factors Influencing the Intention to Share a Viral Video
The most imperative motivational factor for the intention of sending a video forward or showing it to other people is the receiver’s attitude and perception of the video content. This is affected by a number of content related features; many of which have a direct influence on the intentions to share a video with others. Hence, we have identified content as the single most important factor in order to make a video go viral. The results of this study imply that low levels of provocation, together with a sense of authenticity as well as high quality content are factors to take into account when constructing a successful viral video campaign. In addition, Ad Creativity (Novelty, Meaningfulness, Humour, Positiveness and Well-Craftiness) strongly affects the attitude towards the video as well as the intention to send the clip forward. The result that provocation affects video attitude negatively is interesting as it may seem to oppose the idea of creating content that is “on the edge” and “sensational” to promote a positive attitude towards a video. This is most likely connected to the respondents’ perception of the word provocation itself, which was found to be associated with sexist content for example. This is illustrated in a quote from the video “Making of the Ryanair Cabin Crew Charity Calendar 2009” (see appendix 1), which was regarded as the most provocative video as well as the least favourite of them all. It is noteworthy that novelty, which is part of the Ad Creativity index and could be seen as the other side of using content that is “on the edge”, is affecting video attitude positively. Thus, the results connected to the content related features examined in this thesis, illustrates the complexity of finding the right balance between creating
49
Content is King
Rune & Vinberg
something new and interesting, while not being too provocative. Furthermore, contradictory to a rather widespread opinion that videos online need to appear amateurish even for branded content in order to be accepted and spread, we can conclude that marketers should regard viral videos the same way as any other audiovisual communication, hence focusing on a creative idea, together with an impeccable execution, i.e. content is king! Studying the influencing factors of video attitude does not merely include independent variables connected to the content. Repetition was also found to be an important feature to take into consideration when constructing a successful viral video campaign as the respondents who had been exposed to the video prior to participating in the experiment showed significantly higher levels of attitude towards the film clip. Implication of this finding is basic, yet noteworthy as it indicates that enabling the consumers to get used to the content of an online video will improve their attitude towards it. Hence, using a marketing mix of different channels such as television, cinema and digital media for a branded video will, according to this study, more likely provide positive attitudes towards the film clip than if the video were only shown a limited number of times online. One important difference between the intentions to show a film clip to someone compared to sending it forward is the importance of engagement. Rather than evaluating the level of creativity and authenticity of the video, respondents who intend to show the video, which evidently requires more effort in terms of recommendations than to simply make a few clicks after watching the film clip, are to a greater extent influenced by the level of engagement in the content. This conclusion has important managerial implications, as it illustrates the significance of the receiver’s perception of being inspired by the content of the video to take the steps required to tell someone about the video that s/he finds to be interesting enough to share even after the event of watching it him-/herself. In order to encourage receivers of a viral video to create a similar content video on their own, additional factors proved vital. Unsurprisingly, people who are more Internet savvy are more likely to make their own version of the original. There are also indications that the intention of making a similar reproduction increases if one perceives the originator of the video not to be connected with the company appearing in the film clip. However, interestingly, additional features affecting this decision are perceived level of engagement in the video as well as production quality. This indicate that the viewer of an online video is more likely to create a similar film clip if the video in question is both high in quality in terms of expenditure, time and effort as well as if it is user generated.
50
Content is King
Rune & Vinberg
Although these features may appear as opposing and not coherent with previous beliefs that user generated content would be of low quality and in an amateur like fashion, the results illustrate that the current technological progress enables user generated content to be as qualitative as that made of professionals. It should however be noted that the intention to make a similar version was generally very low, which may be an outcome due to the lack of incentives provided for the respondents to act in this manner; i.e. asking them to create a video as part of an interactive campaign. Contradictory to our prior beliefs, contextual factors including number of views and senderrecipient relationships did not indicate any effects on the respondents’ perception and behaviour. These results differ from previous research122 on digital communication via e-mails, where contextual factors had a significant impact on the attitude and behaviours of the survey participants. This difference can be explained by the more intense processing that audiovisual content requires,123 which consequently implies a higher degree of attention towards the video content compared to emails with text or picture messages. It should also be noted that the studies on e-mail communication were conducted in 2004 respectively 2007, when the amount of digital communication was less developed than it is at present.124 This information overload has most likely created a degree of jadedness towards advertisement within certain segments, which consequently increases the demands on content quality and creativity, which could be a contributing factor to the insignificance of contextual factors. Hence, earning the medium, which has become a term widely used among practitioners in the field of digital media, implies the importance of hard work in terms of creating a valuable content for the web in order for a video to go viral.
5.2 Additional Communication Effects
Similar to traditional marketing research, viral video communication will positively affect brand attitude if the receiver experience a positive attitude towards the video per se. This will in turn influence both the intentions to share the video, as well as purchase intentions. Brand attitude, based on the exposure to a branded online video, is mainly dependent on two factors in addition to the traditional dimensions of brand awareness and advertising attitude, which are not directly connected to the videos, namely; video attitude and inconsistency.
Chiu (2007), Phelps (2004) De Pelsmacker (2005) p. 142 124 Rosengren (2008)
122 123
51
Content is King
Rune & Vinberg
The results of this study indicate a difference between well-known brands and those the respondents were less familiar with. Firstly we can conclude that brand attitude is significantly dependent on video attitude, which in turn is highly affected by how the content is perceived. Secondly, in order to get attention, the content should be atypical in comparison to previous communication from the brand. The latter mainly applies to strong, well-known brands since a prerequisite for inconsistent communication is already established key associations to the brand, resulting from a long and visible existence on the market.125 In this study McDonalds is the most obvious example of this phenomenon. This implies that large, well-established companies should not be afraid of stepping outside of what they normally want their brands to represent in order to increase the curiosity for the brand. Inconsistent content is thus likely to get more attention, better top-of-mind awareness and greater consumer interest in the brand, which has been stated in previous research. 126 An example of this effect is the initial negative reaction from Coca-Cola’s marketing division when confronted with the video of the Diet Coke and Mentos experiment, which were mentioned in the introduction. This video has proven to be tremendously successful as a marketing campaign, whereupon Coca-Cola now uses the buzz to attract traffic to their website.127 Less known brands have more to gain by focusing on qualitative content and video attitude through creativity, authenticity, engagement and non-provocative content. An example is the “T-Mobile Dance” (see appendix 2) where a large group of people together perform a dance in the middle of rush hour at Liverpool Station in the UK. The clip was ranked highest in terms of video attitude even though T-Mobile does not exist on the Swedish market. This implies that viral campaigns are accessible for all companies independently of the awareness of the brand, although brand awareness still promotes brand attitude and purchase intentions, according to the study.
5.3 Critique
Even though the respondents watched the exact same version that is to be found online, it should be acknowledged that certain aspects differed compared to a real life scenario. Although the experiment was constructed to be as similar to reality as possible, noise related to the use of social networking sites and video sharing sites, i.e. title and comments from other viewers, was removed in the survey. The anonymity of the sender, even though it was stated if the source of the message was
Törn (2009) p. 16 Törn (2009) 127 www.thecoca-colacompany.com
125 126
52
Content is King
Rune & Vinberg
a friend, an acquaintance or a company, could reduce the impact on the sender-recipient relationship results compared to real life. It should also be noted that in real life people make the active choice of watching a video as the ones included in the survey, as oppose to our study where the respondents have taken part in an experiment with a compensation of 5 SEK. Furthermore, the sample of respondents could be questioned as it may not be representative for the users of digital media, which are mainly Internet savvy and relatively younger age groups. This can influence whether or not the conclusions are entirely generalizable.
5.4 Suggestions for Further Research
This study has provided further understanding on the consumer behaviour related to audiovisual communication on the web. There are however several issues that need to be addressed in adjacent areas of study. A prerequisite for this study has been the respondents’ exposure to the video. Hence , the initial behaviour when receiving such a video, i.e. underlying motivations to watch the video, has not been possible to study. It would thus be of interest to determine whether contextual factors, as the number of views or sender-recipient relationship, would influence the motivation to click on the video in the first place, in similarity with findings on the reasons for opening e-mails.128 Instead of segmenting the sample by demographic or psychographic variables, we prioritized to identify general conclusions on the perceptions of viral videos and the motivations for sharing them with others. The reason for this is the lack of theory and previous research that exists within the area of audiovisual web communication. However, since previous studies have shown that the tendency to forward e-mails depends on the personal traits of the sender,129 i.e. if the sender is conscientious or extravert, it would be of interest to see if the same can be said for audiovisual content. To study a cross-cultural sample would also increase the applicability of the results, since the Internet provides a communication channel that goes pass country boarders. User generated content is becoming increasingly frequent and it is thus of the highest interest to further study the underlying reasons for consumers to contribute to the marketing of certain brands, as well as the implications and affects this will have on commercial communication praxis. Studying
128 129
Phelps et al. (2004) Chiu (2007) p. 529
53
Content is King
Rune & Vinberg
the common traits of the most popular videos, they all share a certain element of mysticism. They are frequently commented and the stories of how they were created as well as the originators of the videos are widely discussed, both among consumers as well as in the press (see for example T-Mobile Dance, Taylor Momsen Escapes Paparazzi and Kobe Bryant Jumps over Car). The aspect of storytelling and the connections between guerrilla marketing and digital w-o-m to create buzz and improve the attitude towards the video is something that we have not been able to study more closely but that we believe is of great interest. As previously stated, the findings of this study indicates the importance of engaging the viewer in the content of a viral video in order for that person to show the film clip to others, which is an action required to be performed at a later occasion than immediately after the event of the video being watched. Although we have used a wide set of high- and low involvement products being presented in the videos used in the survey, the limited scope of this thesis has restricted the level of which we were able to study differences between these products as influencers. We look forward to see further research on this subject in the future.
54
Content is King
Rune & Vinberg
6. REFERENCES
Ackermann Greenberg, D. (2007-11-22) “The Secret Strategies Behind Viral Videos”, Tech Crunch,http://www.techcrunch.com/2007/11/22/the-secret-strategies-behind-many-viral-videos/ (retrieved 2009-05-22) Alexa.com (2009-05-23)http://www.alexa.com/topsites (retrieved 2009-05-23) Arndt, J. (1967) “Role of Product-Related Conversations in the Diffusion of a New Product”, Journal of Marketing Research, Aug, 291-295 Babin, B., J., Darden, W. R., Griffin, M., (1994) “Work and/or Fun: Measuring Hedonic and utilitarian Shopping Value“, Journal of Consumer Research, vol 20 Balasubramanian, S., Mahajan, V., (2001), “The Economic leverage of The Virtual Community”, International Journal of Electronic Commerce”, vol 5, pp. 103-138 Blackwell, R.D., Miniard, P.W. & Engel, J.F. (2006), “Consumer Behavior”, 10th edition, Thomson Higher Education, Mason, USA Blomstedt, A., Björkman, J. (2008), ”Är All Reklam Bra Reklam?”, Master Thesis, Department of Marketing & Strategy, Stockholm School of Economics Braw. E. (2008-01-16), “Netanel Sprider Facebook i Världen”, Metro,http://www.metro.se/se/article/2008/01/16/07/1900-42/index.xml (retrieved 2009-03-23) Brewer, M. (2000) “Research Design and Issues of Validity”. In Reis, H. and Judd, C. (eds) Handbook of Research Methods in Social and Personality Psychology. Cambridge: Cambridge University Press., England Chatterjee, P. (2001) “Online Reviews: Do Consumers Use Them?”, Advances in Consumer Research, Vol. 28 Chiu, H., Hsieh, Y., Kao, Y., Lee, M., (2007), “The determinants of Email Receivers’ Disseminating Behaviors on the Internet”, Journal of Advertising Research, Dec 2007 p.524-534 Cialdini, R. B. (2001). “Influence - Science and Practice” 4th edition, Allyn and Bacon, New York, USA The Coca-Cola Company, “Products and Packaging Myths & Rumours”,http://www.thecoca-colacompany.com/contactus/myths_rumors/packaging_mentos.html, (retrieved 2009-05-26) Dahlén, M. (2003) “Marknadsförarens Nya Regelbok”, 1st edition, Liber Ekonomi, Malmö, Sweden
55
Content is King
Rune & Vinberg
Dahlén, M., Lange, F. (2007) ”Optimal Marknadskommunikation”, 2nd edition, Liber Ekonomi, Malmö, Sweden Davis., A., Khazanchi, D., (2008), “An empirical Study of Online Word of Mouth as a predictor for Multi-product category e-commerce Sales”, Electronic Markets, vol 18 (2) p. 130-141 De Pelsmacker, P., Geuens, M., Van den Bergh, J. (2005), “Foundations of Marketing Communications – A European Perspective”, Pearson Education Limited, Essex, England Dichter, E., (1966), “How Word-of Mouth Advertising Works”, Harvard Business Review, Nov-Dec Facebook.com (2009-05-23)http://www.facebook.com/press/info.php?statistics (retrieved 200905-23) Hennig-Thurau, T., Gwinner, K.P., Walsh, G., Gremler D.D. (2004) “Electronic Word-of-Mouth via Consumer-Opinion Platforms: What Motivates Consumers to Articulate Themselves on the Internet”, Journal of Interactive Marketing, Winter, 2004 p. 38-52 Iriberri, A., Leroy, G. (2009), “A Life-Cycle Perspective on Online Community Success”, ACM Computing Surveys, Vol. 41, No. 2, Article 11 Kincaid, D., Drummon, P., S., (2007), “Power Shift”, Marketing Magazine, vol 112, p. 26-26 King, Rachel (2007-01-16), “The Diet Coke and Mentos Explosion”,http://www.businessweek.com/technology/content/jan2007/tc20070116_569213.htm, (retrieved 2009-05-26) Machleit, K., A., Allen, C. T., Madden, T. J. (1993), “The Mature Brand and Brand Interest: An Alternative Consequence of Ad-Evoked Affect”, Journal of Marketing, vol 57, p. 72-82 Machleit, K., A., Wilson, D., R. (1988),"Emotional Feelings and Attitude Toward the Advertisement: The Roles of Brand Familiarity and Repetition", Journal of Advertising, vol 17, Number 3, Pages 27-35. MacKenzie, B., Lutz, R., J., Belch, G., E., (1986), "The role of Attitude Toward the Ad as a Mediator of Advertising Effectiveness: A test of Competing Explanations", Journal of Marketing Research, vol 23. MacKenzie, S. B., & Lutz, R..J. (1989), “An Empirical Examination of the Structural Antecedents of Attitude Toward the Ad in an Advertising Pretesting Context”. Journal of Marketing, April, Vol. 53:2, 48-65. Malhotra, Naresh K, Birks D. (2007) ”Marketing Research - An Applied Approach”, 3rd Updated European Edition. Prentice Hall Mano, H., (1996) “Assessing Emotional reactions to TV Ads: Replication and Extension with a brief Adjective checklist”, Advances in Consumer Research, 23, 63-69
56
Content is King
Rune & Vinberg
Modezelewski, M., F., Niehaus R., W., (2000-09-13),”Finding a cure for Viral Marketing”, Direct Marketing News,http://www.dmnews.com/Finding-a-Cure-for-Viral-Marketing-Ills/article/68355/ (retrieved 2009-05-22) Modig, E., Lethagen, H. (2008), “Unbundling Creativity”, Master Thesis, Department of Marketing & Strategy, Stockholm School of Economics Muñiz, A. M. & Schau, H. J. (2007) “Vigilante Marketing and Consumer-Created Communications”. Journal of Advertising, Vol 36:3, 35-50. Notani, A., S., (1998), "Moderators of Perceived Behavioral Control's Predictiveness in The theory of planned Behaviour: A Meta Analysis", Journal of Consumer Psychology, 7(3), 247-271 Phelps, J.E., Lewis, R., Mobilio, L., Raman, N., (2004), “Viral Marketing or Electronic Word-ofMouth Advertising: Examining Consumer Responses and Motivations to Pass Along E-mail”, Journal of Advertising Research, Dec 04 p 333-348 Pressler, J. (2008-10-28) “Internet Crush: Joel Moss Levinson”, New York Magazine,http://nymag.com/daily/intel/2008/10/internet_crush_joel_moss_levin.html , (retrieved 2009-05-24) Puto, P., C., Wells, D., W., (1984), “Informational and Transformational Advertising: The Differential Effects of Time”, Advances on Consumer Research, Vol.11 Issue 1, p638-643 Ratchford, Brian T. (1987), “New Insights About the FCB Grid”, Journal of Advertising Research; Aug/Sep87, Vol. 27 Issue 4, p24-38, 15p Rosengren S., (2008) “Facing Clutter: On Message Competition in Marketing Communication”, Economic Research Institute, Stockholm School of Economics (EFI) Skolverket, (2004-03-02)http://www.skolverket.se/sb/d/205/a/825;jsessionid=1D88DF692A47C95C0D8942929FE2EF92 (retrieved 2009-05-24) Stafford, T. F. & Stafford, M. R. (2002), “The Advantages of Atypical Advertisements for Stereotypical Product Categories”. Journal of Current Issues and Research in Advertising, Vol 24:1, 25–37. Strauss, J., Frost, A., (2001) “E-marketing”, 2nd edition, Upper Saddle River, N.J: Prentice Hall Sundaram, D., Mitra, K., Webster, C. (1998), “Word-of-Mouth Communication: A Motivational Analysis”, Advances in Consumer Research, 25, p. 527-531 Sundaram, D., S,. Webster, C. (1999) “The Role of Brand Familiarity on the Impact of Word of Mouth Communication on Brand Evaluations”, Advances in Consumer Research, p664- 670 Söderlund, M. (2001) ”Den lojala kunden”, 2nd edition, Liber Ekonomi, Malmö.
57
Content is King
Rune & Vinberg
Söderlund, M., (2005) ”Mätningar och mått- i marknadsundersökarens värld”, Liber, Malmö Thomas, M., Brain, D., (2008),”Crowd Surfing”, A & C Black Publishers Ltd, London, England Thomas, A. (2009-01-24) “Pope hopes new YouTube channel will reach young people”,http://www.christiantoday.co.uk/article/pope.hopes.new.youtube.channel.will.reach.young.people/ 22374.htm, (retrieved 2009-05-23) Törn, F. (2009), “Challenging Consistency – Effects of Brand-Incongruent Communications”, Dissertation, Economic Research Institute, Stockholm School of Economics, Sweden Wipperfürth, A., (2005) “Brand Hijack: Marketing Without Marketing”, Portfolio Trade, New York YouTube Fact Sheet (2009),http://www.youtube.com/t/fact_sheet (retrieved: 2009-05-24) Zinkhan, G., M., Kwak, H., Morrison, M., Peters, C., O., (2003), “Web-Based Chatting: Consumer Communication in Cyberspace”, Journal of Consumer Psychology, 13 (1&2), p. 17-27 YouGov Sweden: www.yougovsweden.se YouTube: www.youtube.com Viral Video Chart: www.viralvideochart.com Worldstats.com, December 2008,http://www.internetworldstats.com/europa.htm#se, (retriever 2009-03-23)
58
Content is King
Rune & Vinberg
7. APPENDICES
7.1 Appendix 1: Selected Videos - Summary
Video name Brand Product No. Of views (16/2-09) Length Discovered Video Ranking Selected quotes from respondents
“I thought that the commercial was festive, but it had absolutely nothing to do with chocolate. That was what I remembered. That it came up a candy bar at the end. But I didn’t think much about the brand unfortunately." ”I guess the brand has something to do with ’low fat’. Maybe Lätta or something similar.” "I have no idea which brand it is. It wasn’t evident in the video, which I think is really bad."
1. Cadbury Eyebrows 2. Taylor Momsen Escapes Paparazzi 3. Tea Partay 4. All I Want for Xmas is a PSP
Cadbury
Milk Chocolate: Low involvement/Feel
Sportswear/sneaker s: High involvement/Think Vodka: Low involvement/Feel Portable gaming console: Low involvement/Think
4,076,069 views
1.00min
23-jan-09
4
Nike Smirnoff Sony Guitar Hero / Activision
588,912 views 5,361,096 views 55,572 views 1,985,285 views 5,531,688 views 4,550,675 views 11,771,013 views 5,721,089 views 520,375 views
59sec
13-jan-09 06-sep-06 15-dec-06
8 12 14
2.10min
5. Bike Hero 6. Durex Get It On 7. Walk-in Fridge 8. Diet Coke + Mentos 9. T-Mobile Dance 10. McNugget Rap 11. Making of the Ryanair Cabin Crew Charity Calendar 2009 12. Nintendo SixtyFOOOOOOOOUR 13. Kobe Bryant Jumps Over Car 14. Lactacyd, "Protect Yourself" 15. Dove Onslaught
Video game: Low involvement/Think
1.22min original: 3.36min, short: 1.00min
"PSP..something. Whatever that is."
18-nov-08
9
Durex Heineken Coca-Cola & Mentos T-Mobile McDonald 's
Condoms: High Involvement/Think Beer: High Involvement/Feel
30sec 32sec original: 2.57min, short: 1.10min
14-jan-09 30-dec-08
3 2
"I didn’t think about which brand it was, I listened to the music
" ”Hehe, I’m sorry. I’ve already forgotten which brand it was. It was a word of approximately 6 letters and it was a condom brand.” ”The brand is Heineken – beer, featuring a very clever allusion to Sex & the City!” “Both Mentos and Coke (which I don’t know of) were in the video. I guess that both companies are marketing their brands at the same time.” ”The brand was T-Mobile, but I didn’t think it had much to do with that really.” "The brand is Chicken McNuggets, McDonald’s. – Anyone who missed that???" “The brand was Ryan Air – but after watching this film clip I never want to fly with this sexist airline!! Totally degrading towards the female cabin-crew-genus and not the least to us as passengers!!” “I don’t know which brand it was. Feels like some kind of cheap crap toys that you trick children into liking, like BR or something.”Soda & candy: Low Involvement/Feel Mobile phone subscription: High Involvement/Think Fast-food: Low Involvement/Feel
06-sep-06 16-jan-09 11-jul-07
6 1 13
02.41min
32 sec
Ryanair
Airline: High Involvement/Think Gaming console: Low Involvement/Think Sportswear/sneaker s & sports car: High Involvement /Think & Feel Intimate care products: High Involvement/Think Beauty products: High Involvement/Think
242,228 views
1.42min
11-nov-08
15
Nintendo Nike & Aston Martin
10,639,135 9,871,826 views
1.06min
06-sep-06
11
52sec
10-apr-08
7
"The brand was probably Nike but I heard the name Aston Martin more." "It would be interesting to know which brand it was. Is the product as confused as the film clip?"
Lactacyd Dove
106,930 views 2,055,341 views
30sec
12-jan-09 01 Oct 2007
10 5
1.18min
"?????????"
59
Content is King
Rune & Vinberg
7.2 Appendix 2: Selected Videos
60
Content is King
Rune & Vinberg
61
Content is King
Rune & Vinberg
7.3 Appendix 3: Online Survey (example from Nike, Scenario 4)
62
Content is King
Rune & Vinberg
63
Content is King
Rune & Vinberg
64
Content is King
Rune & Vinberg
65
Content is King
Rune & Vinberg
66
Content is King
Rune & Vinberg
67
doc_421472117.pdf