What makes online video ads go viral?
Source of video distribution as moderator of emotional
arousal
Armina Stepan-Cazazian
10494642
Thesis Coordinator: Stephanie Welten
Master Track: Persuasive Communication
1
The current research studies the interaction between emotional arousal and source of distribution with sharing on Facebook as dependent variable. Previous theory on distraction and information diffusion and arousal suggests that when in a state of low arousal people will be likely to pay attention to the context in which the ad is distributed and share what their friends share. An online experiment on a convenience sample of 201 participants has shown a significant effect of arousal (p = .070) and source (p = .011) on sharing, but no interaction between the two independent variables. This suggests that emotionally arousing ads shared by friends are most likely to become viral. Overall the research shows that emotional arousal is not enough by itself to lead to the success of a campaign or increase brand attitude significantly. The research also sheds light on short-term behaviors triggered by emotions and the source of distribution of a message.
Introduction
Previous research on why people share videos online has come up with significant findings related to level or arousal of the emotions conveyed in the video (Berger & Milkman, 2012) as well as the valence of these emotions (Nelson-Field, Riebe & Newstead, 2013) as main triggers for a video ad being forwarded in social media. While the findings clearly indicate that videos which manage to induce strong emotions in the viewer (called “emotional arousal”) of a positive valence (such as admiration and awe) are more likely to be forwarded, no previous research has explored the interaction between emotional arousal and where the ad video is distributed.
This is what will be attempted in the following research: understanding how the source of distribution for social media content moderates the emotional arousal of the receiver with intention to forward as main effect. Whether the user is exposed to an ad shared on an advertiser’s wall or a friend’s wall may trigger the understanding that they are the target of a persuasive message of commercial nature (Friestad & Wright, 1994) and a different level of processing of the message may occur. People may be less likely to accept the message and activate necessary counterarguments in order to resist the persuasive attempt (Knowles & Linn,
2
2004) if their resistance is not distracted in some way, such as bringing them to a high emotional arousal state. However, even if the message is not in itself arousing, the fact that it is being shared by a likable source (such as a friend) can trigger the willingness to share the message.
Intention to forward is important to research on an academic level in order to understand the power of emotions in the age of social media better. What actions are driven by emotions is especially relevant from a communication science perspective. It is also crucial on a societal level because companies invest more and more money in promoting their products and services online (Theonlinecompany.com, 2014) and they measure numerous metrics such as clicks, likes and shares (Gerlitz, Carolin and Helmond, 2011) which are indicators of potential commercial success. In addition, if the content that the brands create spreads organically through users then companies will need to spend less money to reach their audience. Companies could potentially change their approach from one day to another as a result of the response they get in the form of online interactions. Thus, the impact of emotional arousal on performing various actions is important on a scientific and business level.
People get their information and recommendations from word of mouth (WOM), be it online or offline (Allsop & Bassett et al., 2007), as friends are deemed to be more honest and trustworthy than brands broadcasting their commercial messages. This is why now Facebook allows companies to create content in a page structure that mimics the one of human individual user pages. Brands are now able to take part in people’s conversations and to start conversations
themselves. It is, thus, crucial that marketers understand how and why some campaigns that they run are complete successes while others don’t appeal to the audience at all.
3
It appears that people pay more attention to the context in which an ad is distributed especially when they are not fully immersed in the message. The main research question is, thus: How does the source of distributing the video ad interact with its emotional arousal with forwarding intention as outcome?
Theoretical Background
Emotional arousal and intention to forward in social media
Physiological arousal is characterized by activation of the autonomic nervous system (Heilman, 1997), and the mobilization provided by this excitatory state may boost a number of actions related to the arousal trigger. We refer to emotional arousal in advertising as the content that elicits feelings in the viewer, in other words, the physiological arousal that the ad induces. Roseman, Spindle and Jose (1990) describe "hope," "joy," "relief," and "liking" as positive emotions. These emotions are all important from a marketing perspective. Ample research has been done on the kind of emotions that stores should transmit in order to enhance purchasing behavior (Sherman & Mathur, 1997). Pleasure and emotional arousal has been found to influence the time people spend in shops as well as the amount of money spent (Donovan, Robert J et al., 1994). It is expected that this can be extended to the purchasing behavior online as well, especially since much of the customer journey has now shifted to online commerce. Visual cues such as couples in love, beautiful landscapes, touching family moments as well as suggestive background music and the use of color are all elements that can influence the viewer’s emotional arousal caused by visuals (Lane, Chua & Dolan, 1999). The stronger these emotions are generated, the more likely the mobilization in the form of sharing will occur (Berger & Milkman,
4
2011). It is predicted that sharing in social media will occur as direct consequence of this mobilization.
This can be further explained by considering the research done on distraction in advertising (Taylor & Francis, 2001) which has shown that distracting the viewer from the persuasive intent does lead to a better attitude towards the ad, sometimes at the expense of the brand, which will not be well remembered. What distraction does successfully is it reduces counter arguing on the receiver’s part. This is similar in the case of emotional arousal that leaves the viewer more easily convinced to perform immediate, small actions related to the ad such as sharing. Thus, distraction is an important tool in persuasive communication.
Sharing in social media (or forwarding) falls in line with the instant gratification temptation described by Fedorikhin and Patrick (2010). A user sharing content on Facebook receives instant rewards in the form of likes, comments and further sharing from friends. Sharing is, thus, a form of self enhancement (Taylor, Strutton & Thompson, 2012). However, it is a very short term gratification and does not bring with it long term enhancement, as Vincent Miller (2008) shows. Sharing video ads on Facebook falls under the category of phatic communication (Miller & Vettere, 2005): this means that social media has a maintenance role of keeping track with one’s weak ties rather than communicating with close friends and family. As a consequence, it is safe to assume that emotional arousal can act as a trigger for sharing
Emotional arousal triggers the intention to forward the video ad, a quick mobilization and not a part of a more complex communication process. As part of the phatic communication process, the user can share a video with cats on his or her wall solely because of the emotional appeal that the video has generated, and not as part of an ongoing conversation or for the purpose
5
of building reputation. The user can expect instant gratification in the form of a few likes, shares and/or comments. The video will quickly disappear from other people’s Facebook feeds as other news and videos are posted on the platform constantly. The algorithm is constantly improving to make the posts disappear in the archive quite fast. This is part of a larger social media trend of posts that cannot be visualized for a long time after they are published (The Verge, 2014). Thus, it is not necessary to research more complex intentions behind the sharing activity.
Berger & Milkman (2011) have focused their research on the effect of emotional arousal on sharing in social media and they have found that powerful emotions such as awe, admiration and disgust do lead to more sharing than the videos that are relaxing or saddening. This activation in the form of sharing is researched here as well. Jonah Berger has conducted several researches over the past years (2011, 2012, 2013) on emotional arousal as trigger for sharing content online and has shown that among the 7000 most shared New York Times articles the most arousing content was the most shared. More importantly, that positive high arousal was the strongest predictor of sharing. This is because the more relaxing positive content does not have the same power to active the user to perform any promotion-related actions. Moreover, Berger has shown that the more deactivating the emotion presented in a news piece (when manipulating the level of sadness) the less likely that experiment participants were to share the story. The fact that people experience the sharing activation more for positive valence content than negative valence content can in part be explained by the fact that people like to see themselves as being positive persons as well as interact more with other people which they perceive as positive or joyful (Kamins, Folkes & Perner, 1997) but because this field of research is quite new more insight into the topic is needed. Therefore the present research aims firstly to replicate previous findings on the effect of emotional arousal on sharing and focuses solely on positive emotions.
6
H1: Highly emotionally arousing videos of positive valence lead to more intention to forward than low arousing videos of positive valence.
Source of distribution as moderating factor
Distribution channel has a great impact on the receiver’s emotional and cognitive processing of the message. Friestad and Wright (1994) have shown that when the receiver of the message realizes that he or she is being targeted with a persuasive message then the change of meaning occurs. This can occur in any type of communication process, from a simple conversation between people to a call to action from an institution asking people to buy their products on sale. In all cases the agent and the target have specific information about each other and the target manifests coping behaviors based on previous experience with persuasion attempts (Goodstein, 1993). If the persuasion message is not carefully crafted by the agent, then the target may experience reactance, distrust, scrutiny or inertia (Knowles & Linn, 2004). In this phase the receiver is no longer susceptible to emotional arousal because they are consciously resisting the message that is being presented to them. As a consequence the overall effectiveness of the message is compromised.
It is hypothesized that when in a state of emotional arousal the target will be less likely to activate this resistance behavior because they will be under cognitive load. In fact, high emotional arousal can fall under the strategy of distracting resistance (Knowles & Linn, 2004; Petty, Wells and Brock, 1976) which states that distraction will interfere with the counterarguments that the target has built in their mind. A low arousal ad will, however, give way to more counter arguing and the receiver will be more inclined to process the message centrally (Petty & Cacciopo, 1984) rather than peripherally. When this happens they will pay
7
more attention to the context in which the ad is being distributed, therefore the source of distribution can interact with the main effect. The context in which the ad is distributed can be more trustworthy (such as a friend sharing the ad on his/her wall) or less trustworthy when a company shares it. Thus the distraction interrupts the potentially negative thoughts regarding the message and the agent and allows the message to become more effective.
This idea of context as moderator of the effect of arousal is further developed when taking into consideration the importance of source likability (Chaiken, 1980). This research has shown that when a message receiver is immersed in a topic then he or she is less likely to be influenced by the likability of the source while when the receiver is low involved state, then they will be sensitive to the source’s likeability such that more the likable source is more persuasive than the less likeable source. It can be assumed that the high versus low involvement state will affect the receiver in a similar way to high versus low emotional arousal because they both refer to full immersion of the receiver in the message, making the receiver in the low emotional arousal state more willing to accept the message being shared by a likable source (their own Facebook friend).
Later on Campbell and Kirmani (2000) have tested the change of meaning principle within a sales context and have shown that when the persuasion motive of a sales agent’s behavior is immediately accessible, both a busy receiver as well as a not busy receiver will perceive the agent as less sincere. However, when the persuasion motive is revealed later on, the busy target will perceive the agent as more sincere. Facebook News Feed has been improved by its developers in this exact spirit: to provide the persuasive message in a context in which the user is already immersed and between other messages perceived as honest. The stats also confirm the news feed advertisements are more persuasive: Prior to their 2013 update the
8
platform was only offering brands the possibility to advertise in the form of classical banners on the right side of the user’s page. Since the update the number of ad clicks has gone up 49 times and has reduced cost by 45%, according to industry research (AdRoll, 2013) when the sponsored brand messages were moved to the user’s news feed, where usually only their friends’ activity
was visible. This allows users to interact with brands as they would with a person, by commenting, liking and sharing their content. When a user presses the button Share then they are taken to a new screen inside Facebook where they can easily press Post and have that same message broadcasted either publicly or only with their friends (depending on the user’s privacy settings). Non-commercial type of word of mouth is preferred by people when on Facebook (Eisingerich et al., 2014), thus it is likely that people will attend to conversations with friends on Facebook rather than any type of brand related message.
The likelihood of people only attending to messages shared by friends is line with the theory of information diffusion (Valente & Davis, 1999): this states that the most trusted member of the community (the so called “leaders”) will accelerate the diffusion of information as the
other members of his or her network are then more likely to pass that information on as being trustworthy as well. This loyalty to friends makes companies’ mission more challenging. If people are rather immune to messages sent by anyone elsethen viewing a video ad on a brand’s wall will have a similar effect as the sales agent revealing his intention from the very beginning, whereas when the video is presented on a friend’s wall it will induce more positive feelings and a stronger emotional arousal. This will allow for keeping peripheral processing in the receiver rather than the high elaboration state in which the receiver might resist the message. Later it was proved that the receiver experiences further detachment when the manipulation intents are salient inside a message (Wentzel, Tomczak and Herrmann, 2010). A narrative ad where the
9
manipulative intent is made salient was associated with less positive affect, weaker self-brand connections and less positive attitude towards the brand than a non-narrative advertisement, which simply exhibits the product benefits and price. Taking into consideration that people usually attend to the messages shared by friends and tend to counter argue if they are not immersed completely in the message it can be hypothesized that:
H2a: An ad video distributed on a friend’s wall will be more likely to be shared than a video distributed on a brand’s wall, but only when the video induces low emotional arousal.
H2b: When the video is a highly arousing one, then the source of distribution does not have a significant effect on the likelihood of sharing.
Next to intention to forward the ad, other measures were used to better assess the effect of the dependent variable: intention to like the ad on Facebook, attitude toward the ad as well as attitude toward the brand. Together they show what type of brand related activation the viewer experiences. IV: - arousal + arousal DV: Intention to forward message M: Source of distribution a friend’s wall + the brand’s wall -
10
Pre-test method
Seventy-one participants (M age = 25.30, SD = 2.50, 66,7% women) were randomly
distributed into two pre-test conditions: low arousal or high arousal condition. Out of the 71 volunteering participants 8 of the participants’ answers were not completed, as a result their data
was taken out of the analysis and 63 participants were registered as valid.
In the low arousal condition participants were exposed to three video ads from KLM airline, the John Lewis retail store and Innocent Drinks juice. The ads were deemed low arousal by the researchers because they all made use of relaxing images, calm music and generally a slow pace of the story. The 95 Years of KLM video by KLM (see Appendix 1) shows images of various places around the world where people go by with their lives while a low male voice talks about travelling. Images of sunsets, sea, clouds and city skylines give the ad a postcard feel. The Innocent Drinks ad (see Appendix 2) shows how fruit juices are being made and how long it takes for a mango fruit to become ripe. Finally, the John Lewis Christmas advert 2014 (see Appendix 3) builds on the idea of childhood innocence and fantasy and shows a boy going around with his best friend – a fluffy penguin. The ad presents a twist at the end as the boy gets his pet penguin a girlfriend and we find out that the penguins were all stuffed toys.
In the high arousal condition the same brands were used but in the context of different campaigns: the KLM Live High Five (see Appendix 4) shows people in the street around the world interacting with each other via a surprising video gadget from KLM. The video is edited to appear as fast paced and the characters (regular people on the street) interact with each other in touching ways, high fiving and making friends through the video messaging installation. In the end they are in great joy because they receive surprise gifts from KLM. The Innocent Drinks Big
11
Knit 2014 (see Appendix 5) is a call to action that people knit little hats for Christmas 2014 and the company will donate to organizations that help senior citizens. The ad goes very fast, a lively voice explains how the old people can be helped and a grandmother is shown being dressed in shawls and hats to keep her warm as she becomes more joyful. Finally, the high arousal John Lewis ad (see Appendix 6) celebrates the 150 years history of the brand by showing people of all ages walk, dance and play through life without focusing on any specific character or story.
All of the ads chosen for the pre-test in both conditions had over 15 000 views on YouTube and more thumbs up than thumbs down. It is, thus, safe to assume that the ads are all well executed and are not controversial in any way. They have a large potential audience being ads for global brands and all have narrative elements to them and as a consequence any of the ads could be deemed relevant by the participants to be shared on their walls. This falls in line with the conditions set by Wentzel, Tomczak and Herrmann (2010): if ads would be chosen in which the persuasive intent would be salient from the very beginning then the viewer would quickly become detached and the experiment manipulation would most likely fail.
The highly arousing video and the low arousal video are chosen by human coders in using the “Self-assessment Manikin” (Morris, 1995). This ensures a high inter-reliability between coders as it eliminates potential cultural barriers from the answers. This is especially important in the current research given the fact that the participants have English as a second language and also they come from various cultural backgrounds within Europe. To control for confounding factors the pair of ads (low versus high arousal) were chosen from the very same brands (retail store John Lewis, KLM airline and juice brand Innocent Drinks), were not controversial in nature and were not within the top 3 most shared ads of the year. This makes sure that people would not be inclined to resist the ad or be more favorable to it because of other
12
underlying factors. During the pretest participants were also asked if they had seen the ad before and where they had remember seeing it to control for potential saturation of the ad online.
The 71 participants were also asked about how professional the ads were perceived (Whitehead, 1968), α = .84 and also about how well they understood them (cognition scale, Petroshius & Crocker, 1989),α = .73. In regards to measurements the self-assessment manikin had two items assessing arousal and happiness on 9 point scale with 1 being “very activated” respectively “very happy” and 9 being “not activated at all” respectively “not happy at all”. The cognition scale was measured with three items: believable, informative and clear on a 7 point scale, where 1 is “very believable”, respectively “very informative” and “very clear” and 7 is “not believable at all”, respectively “not informative at all” and “not clear at all”. The professionalism scale was measured also with three items on a 7 point scale on dimensions: professional, credible and trustworthy with 1 being “very professional”, “very credible” and “very trustworthy” respectively and 7 being “not professional at all”, “not credible at all” and “not trustworthy at all”.
Pretest results
KLM ad. Three independent t-test samples with arousal as independent variable and
happiness, arousal, cognition and professionalism as dependent variables were conducted to assess which pair of ads is the most appropriate to use as stimuli in the research. There was a significant difference in the arousal scores for high (M = 5.45, SD = 2.11) and low arousal (M = 4.28, SD = 2.45) conditions; t (61) = 2.03, p =.047. These results suggest the “High five” KLM ad is indeed significantly more emotionally arousing than the “95 years of KLM” ad. The test for valence also shows a significant difference between them, with high arousal ad (M = 2.22, SD =
13
1.94) being considered more positive than the low arousal (M = 3.58, SD = 1.47), t (61) = 3.13, p = .003, even though both of them are well below the midpoint of the scale, “5”, thus they are both of positive valence. If this pair of ads is chosen for the main research then the valence should be taken into consideration as a potential confound in the research in that the happiness is assessed once more and a mediation check is performed. The test for professionalism shows that the ads scored similarly with no significant difference: the low arousal (M = 1.96, SD = .95) and the high arousal (M = 2.36, SD = 1.14) did not differ, t (61) = -3.28, p = .140. The high arousal ad (M = 3.2, SD = 1.14) faired significantly lower on the cognition dimension than the low arousal (M = 2.27, SD = 1.07), t (61) = -3.28, p = .002. Because of the reversed scale items, this shows that the high arousal ad was perceived less clear, even though both scored well below the scale mid-point of 4, thus they were both rather well understood.
John Lewis ad. An independent t-test was performed for the high and low arousal
conditions of the John Lewis ads. There was no significant difference in arousal for high (M = 4.97, SD = 2.44) and low arousal (M = 5.00, SD = 2.20) conditions; t (61) = .053, p = .958, both being rated equally neutral. When it came to happiness the low arousal condition (M = 3.39, SD = 2.48) did not differ significantly from the high arousal (M = 3.38, SD = 1.51), t (61) = .023, p = .981. On the professionalism and cognition dimensions, the two conditions also faired similarly with no significant differences found: professionalism low arousal (M = 2.23, SD = 1.22) and high (M = 2.63, SD = 1.00), t (61) = -1.40, p = .160 respectively cognition low arousal (M = 3.68, SD = 1.17) and high (M = 3.77, SD = 1.11), t (61) = .053, p = .776. so this pair of ads is not appropriate for using further in the main research.
Innocent Drinks ad. Finally the last pair of ads for Innocent Drinks was also not fit to
14
low (M = 5.65, SD = 2.33) and high arousal conditions (M = 5.13, SD = 2.52), t (61) = .85, p = .399. The ads were thought to be almost equally joyful by the participants with low (M = 3.45,
SD = 1.74) and high arousal (M = 3.44, SD = 2.00) showing a non-significant difference, t(61) =
.03, p = .976. The Innocent Drinks scored similarly on cognition and professionalism with cognition low (M = 2.72, SD = 1.31) and high arousal (M = 2.51, SD = 1.4), t (61) = .611, p = .544, respectively professionalism low (M = 2.76, SD = 1.26) and high arousal ( M = 2.33, SD = 1.07) , t (61) = 1.45, p = .152.
As a consequence the pair of KLM ads was used further as stimuli in the main research with the “happiness” ratings being a potential confound, thus being assessed and entered as
covariate in the analyses.
Method main research
A convenience sample was used through publishing the invitation in English to take part in the research on the researcher’s Facebook wall as well as several Facebook groups of students in Amsterdam. This makes sure that all of the participants are actively using the Facebook platform, have an advanced understanding of the English language and agree to voluntarily participate in the research without any other added incentives.
Two hundred and forty participants were approached on Facebook and randomly distributed to a 2 (arousal: low or high) x 2 (source: friend’s or brand’s Facebook wall) design. Out of the 240 participants to the research 39 did not finalize the survey so their data was taken out, leaving 201 valid participants in the experiment. The average age was 25.50, (SD = 2.45), females 55,7%. The sample was highly educated with 95% either Bachelor or Master’s
15
graduates. The majority of the participants’ origin was Romanian (71,7%) with the rest of the
participants having European origins.
Participants were exposed to the ad in a neutral visual environment. They were asked to envision either the friend condition scenario or the {brand condition} scenario: while they are scrolling through their news feed either a good friend of theirs {the KLM airline page} shared this video ad on Facebook and the video ad appeared in their news feed. This ensures that participants imagine that a friend has shared the ad on his or her wall rather than the researcher artificially trying to recreate the Facebook wall for them in a potentially unrealistic way. The participants were required the watch the video for at least 40 seconds before moving on to the next question. This ensures that viewers viewed at least half of the ad before making up their mind about it. Then the participants were asked to rate the video according to how likely it is that they would “like” it followed by a question in which they were asked to think how likely it was that they would share it on their own Facebook profile. Both questions were assessed on a 7 point scale where 1 is “highly unlikely” and 7 is “very likely”. Then the “happiness” item from
the Self-Assessment Manikin was used again on a 9 point scale with 1 being “very happy” and 9 – “not happy at all”.
Two scales assessed the effect of the independent variables further: attitude toward ad (De Pelsmacker, Decock & Geuens 1998), α = .88 and attitude toward brand (Spears & Singh,
2004), α = .94.The attitude toward ad scale has 6 items; all measured on a 7 point scale where 1 is “Strongly agree” and 7 is “Strongly disagree”. The dimensions measured are “positive impression”, “really something for me”, “interesting”, “credible”, “exaggerated” (reversed item), “attractive”. The attitude toward brand scale is measured through 5 items on a 7 point scale with
16
1 being the most negative and 7 being the most positive on the following dimensions: appealing, good, pleasant, favorable, and likeable.
A number of other control variables are researched, such as gender, age, country of origin, education level as well as previous knowledge of the ad.
Results
In order to analyze the participants’ attitude toward the ad and the KLM brand, the scales were computed into two new variables. The item 5 (“exaggerated”) was reversed after data was collected so that all items are expressed on the same scale. The two new scales, attitude toward ad and attitude toward brand were computed into new variables so that the data could be further analyzed.
A multivariate analysis was performed to assess the likelihood of liking, sharing, attitude toward ad and toward brand for the two arousal level and two sources. The analysis has shown a significant main effect of emotional arousal on liking intent on Facebook,F (1, 196) = 0.44, p =
.030 for high arousal (M = 4.65 , SD = 2.11) and low arousal conditions (M = 3.63, SD = 2.05) such that low arousing videos are less liked on Facebook than high arousing videos. There was also a significant effect on sharing (F (1, 196) = 3.30, p = .070) such that the high arousal video (M = 2.90, SD = 1.78) is more likely to be shared than the low arousal video (M = 2.13, SD = 1.66). Finally a significant effect on attitude toward the ad (F (1,196) = 8.21, p = .005) was found such that the high arousal video (M = 2.42, SD = 1.02) gained a better attitude from viewers than the low arousal video (M = 3.25, SD = 1.35). No significant effect of arousal on brand attitude was found, however ( F (1, 196) = 1.50, p = .224, M = 5.74, SD = 1.00) with the high arousal video accounting for a non-significant increase in brand attitude compared to the
17
low arousal video. This shows that in the high arousal conditions participants were significantly more likely to like, share and appreciate the ad more than in the low arousal, however the brand attitude was rather high (well above the mid-point of the scale) for all conditions, though in the high-arousal condition the attitude toward brand was slightly more positive. Therefore, H1 is supported and the null hypothesis is rejected.
Table 1: Sharing intent Table 2: Liking intent
Table 3: Attitude toward ad (1= strongly positive, 7= not positive at all)
2,57
2,13
3,41
2,9
Source:
friend
Source:
brand
High arousal
Low arousal
3,59
3,63
4,63
2,9
Source:
friend
Source:
brand
High arousal
Low arousal
3,31
3,25
2,73
2,42
Source:
friend
Source:
brand
High arousal
Low arousal
18
The analysis has also shown a significant effect of source on sharing (F (1, 196) = 6.57, p = .011), with the video being shared by a friend (M = 3.00, SD = 2.00) being more likely to be forwarded than a video shared by a brand (M = 2.50, SD = 1.75). However, no significant effect of source was found on dependent variables liking (F (1, 196) = .23, p = .632), attitude toward ad (F (1, 196) = .44, p = .505) and attitude toward brand (F (1, 196) = 1.47, p = .226).
No interaction was found between the independent variables arousal and source of distribution on liking (F (1, 196) = .01, p = .908), sharing (F (1, 196) = .00, p = .980), attitude toward ad (F (1, 196) = 1.33, p = .250) and attitude toward brand (F (1, 196) = 1.42, p = .235). As such, both high and low arousal condition participants were influenced by the source of distribution in their intention to forward. Thus, H2a and H2b are rejected.
Additionally, happiness was a significant covariate,strengthening the effect of source and arousal on liking (F (1, 196) = 91.44, p < .001), sharing (F (1, 196) = 60.31, p < .001), attitude toward ad (F (1, 196) = 157.51, p < .001) and attitude toward brand (F (1, 196) = 125, 39, p < .001). Thus, the perceived happiness that came across through the video strengthened the effect of the independent variables on all of the dependent variables in the research.
Conclusions and discussion
Berger’s research has been successfully replicated in this research with emotional arousal of a video having a significant positive effect on people’s willingness to press “like”, share the ad, as well as improving the overall attitude toward the ad. However, the sharing intent was not correlated with more profound changes in opinions, having no significant effect on brand attitude. Therefore, people are temporarily excited about the images and the story that they are
19
exposed to, making them more likely to perform small actions in favor of the brand, especially if they discover the video through a friend and not a brand.
Source of distribution did not prove to interact with emotional arousal in the current research. While users who were told that their friend had shared the video were significantly more likely to like and share the video as well than when they were told the KLM brand was posting the video ad, this happened both in low and high arousal condition. The ad was indeed the most shared when the video was both high arousal and posted by a friend. There are several explanations for this. One is that, as shown in the pretest, while the high and low arousal videos did score significantly different on arousal, the high arousal video was barely above the midpoint of the scale, still not expressing a very intense change of emotion. This hints to the fact that people were not overwhelmed by the emotional arousal they experienced, as a consequence they may have been attentive to the context in which the ad was shared in both conditions, not just in the low arousal one, giving way to counter arguing and a more acute observation of the context (Petty & Cacciopo, 1984). Another possible (complementary) explanation is that the sample used in the research is a highly educated one with 95% of them being either Bachelor’s or Master’s graduates. It is possible, therefore, that the results cannot be generalized to the overall population. Some of the respondents even mentioned in the comments section that they hardly ever share ads on their Facebook walls because they do not want to promote brands, however entertaining they may be. It is possible that for a population with lower education and/or persuasion knowledge, the understanding that they could be associating their own image with a brand by sharing would not occur. An entertaining video would, thus, be considered nothing more than a short movie made by a company rather than paid manipulation intent. So it is
20
possible that the effect of source would be weakened in the high arousal condition, contributing to validating the second hypothesis of the research.
Another interesting finding in the study is the confound “happiness” that mediated the effect of source of distribution and arousal on liking, sharing as well as attitudes toward ad and brand. The highly arousing video was also perceived as more joyful, a dimension which was separately evaluated within the Self-Assessment Manikin. Thus, people not only were likely to like and share the video ad because they were emotionally aroused, but also because they perceived the “High Five” video as more joyful than the “95 years of KLM” one. People share content not only because they are aroused but also because they take joy out of the video. This is line with Nelson-Field, Riebe & Newstead’s research that has shown that videos with a positive valence are more often shared even if they are rather relaxing (low arousal), even though less so than positive highly arousing videos. So valence should always be taken into consideration when building a communication campaign.
It is interesting that source had a significant effect by itself on the willingness to share. Users that were asked to imagine that the “KLM High Five” ad was shared by a friend actually then perceived the same ad to be more likely to be forwarded further on their own Facebook walls than the ones which were asked to imagine that it was shared by the KLM brand directly, revealing a willingness of participants to find it acceptable to promote the things that are recommended by their friends, even if they of commercial nature and clearly meant as persuasive message created by a brand.
Unsurprisingly, the distraction caused by the highly arousing ad did not lead people to change their more profound relationship with the KLM as no significant differences in brand
21
opinion were found between the two conditions. Thus, while the ad helped to raise more views and awareness organically through all of the users that would voluntarily share the ad, such ad would not build a successful brand by itself.
As mentioned previously, the research has a few limitations – the limited resources to create the stimuli meant that the participants had to imagine that the ad was shared by a friend or by a brand without seeing the ad in a normal, seamless Facebook page context. It is possible that their attention fluctuated more to the context in which the ad was distributed solely because they were asked to imagine it, as a consequence their counter arguing capacities were activated. If software with a Facebook plugin would allow the research to extract a name and a photo from the participants’ friends list then the experience could be matched closer to reality. Secondly the convenience sample of highly educated people possibly does not reflect the general population’s behavior and thirdly the difference between the high and low arousal videos in arousal was too small to immerse the viewer fully. Furthermore the level of happiness that the participants experienced was different for the two ads, making “happiness” a confound. For further research it is recommended that these findings are replicated with improved material and sample selection. For instance, a set of videos created especially for the study could be used and a sample which contains a balanced educational level can be recruited through Facebook groups by offering incentives in the form of shopping vouchers.
Also important to note is that the users were exposed to only one Facebook “story” at a time in the Qualtrics software and did not have the option to scroll up and down as they would on a normal Facebook feed. They were also forcefully exposed to the story for 40 seconds before they could move to the next screen, which is very different from what their normal experience would be inside the social network. Normally, users would decide from the video preview photo
22
if they want to press play or not and could potentially move to the next story in less than a second. For this reason the text with which the video is introduced as well as a carefully chosen preview photo would likely increase the conversion to liking and sharing if the text and preview are already emotionally arousing and evoke happiness. So in a communication campaign the preview photo and the text that introduces it should be carefully chosen.
For a more complete understanding of the potential customer’s behavior it would be useful to provide the viewer with a more complete experience and the option to actually purchase a product when in state of arousal to understand if the effect of emotional arousal goes beyond the mere “liking” and “sharing” actions and also what is the exact moment in time when a video
reaches tipping point and starts to gain exponential success. It could be that an ad has the right qualities: is positive and highly arousing but does not reach, in other words, the right friends in order to be exposed to a sufficient amount of people. Network nodes, important leaders of opinion (Valente & Davis, 1999) should be reached first to ensure that the diffusion of the video will then occur smoothly, rather than exposing the video to random or peripheral members of a social network. However, such a research would also be limited by the ever changing algorithm of Facebook that is only selecting some friends and brands at a time posts to broadcast with the rest of their network to prevent information clutter.
From an academic perspective this is the first research that has attempted to understand the potential interaction between emotional arousal and the source of distribution in triggering people to perform various actions. It is clear now that emotional arousal alone does not explain people’s liking and sharing of a video, even though it has a strong effect. Viewers are not easily immersed in a video and their potential for counter arguing is strong even when the arousal occurs. Even if they are persuaded to perform a brand related action, this does not mean that they
23
will convert into long term changes of opinion. This is an encouraging finding as it proves that viewers are not easily manipulated and that they tend to keep their opinions intact. From the perspective of the findings, however, it would be interesting to find out what other small actions does emotional arousal have the capacity to trigger – for instance donating a small amount of money to charity: the same kind of arousal that makes people hit the “share” or “like” button could just as likely trigger them to press a “donate” or a “join the cause” button. However, because the effect of arousal appears to disappear quite quickly, it should be made very easy and straightforward for people to actually donate or join the cause after they press the buttons.
From a professional perspective this does not, however, discourage companies to invest a large amount of resources into viral marketing. If their goal is to increase awareness regarding their product or service, brands will more likely invest less in distributing the video if they invest in producing a surprising, awe-inducing, highly entertaining video. However, this by itself will not be enough – it is recommended that brands combine this effort with an excellent customer journey where the video, for instance, leads to a very positive, highly arousing website with content written in a friend-like voice rather than a more impersonal corporate voice.
In conclusion if the impact of emotional arousal on liking and sharing is corroborated with more significant actions such as donating or joining a social cause, then it is possible that the Facebook post, as quickly disappearing as it is, can make a long lasting impact. If leaders within a community see it and post it first then the transmission of information will happen all the more successfully.
24
References
Allsop, D. T., Bassett, B. R. & Hoskins, J. A. (2007). Word-of-mouth research: principles and applications.Journal Of Advertising Research, 47 (4), p. 388--411.
Berger, J. (2011). Arousal increases social transmission of information. Psychological Science,
22 (7), pp. 891--893.
Berger, J. & Milkman, K. L. (2012). What makes online content viral?. Journal Of Marketing
Research, 49 (2), pp. 192--205.
Cacioppo, J. T. & Petty, R. E. (1982). The need for cognition. Journal Of Personality And Social
Psychology, 42 (1), p. 116.
Cacioppo, J., & Petty, R. (1984). The elaboration likelihood model of persuasion. Advances In
Consumer Research, 11(1).
Campbell, M. C., & Kirmani, A. (2000). Consumers' use of persuasion knowledge: The effects of accessibility and cognitive capacity on perceptions of an influence agent. Journal of
Consumer Research, 27(1), 69-83.
Chaiken, S., & Eagly, A. H. (1983). Communication modality as a determinant of persuasion: The role of communicator salience. Journal of personality and social psychology, 45(2), 241.
Donovan, R. J., Rossiter, J. R., Marcoolyn, G., & Nesdale, A. (1994). Store atmosphere and purchasing behavior. Journal of retailing, 70(3), 283-294.
Eisingerich, A. B., Chun, H. H., Liu, Y., He, J. M., & Bell, S. J. (2014). Why recommend a brand face-to-face but not on Facebook? How word-of-mouth on online social sites differs from traditional word-of-mouth. Journal of Consumer Psychology.
Fedorikhin, A. & Patrick, V. M. (2010). Positive mood and resistance to temptation: the interfering influence of elevated arousal. Journal Of Consumer Research, 37 (4), pp. 698--711.
Friestad, M., & Wright, P. (1994). The persuasion knowledge model: How people cope with persuasion attempts. Journal of consumer research, 1-31.
Gerlitz, C., & Helmond, A. (2011, January). Hit, Link, Like and Share. Organizing the social and the fabric of the web in a Like economy. In Paper presented at the DMI
mini-conference (Vol. 24, p. 25).
Goodstein, R. C. (1993). Category-based applications and extensions in advertising: motivating more extensive ad processing. Journal of Consumer Research, 87-99.
Kamins, M. A., Folkes, V. S., & Perner, L. (1997). Consumer responses to rumors: Good news, bad news. Journal of Consumer Psychology, 6(2), 165-187.
25
Lane, R. D., Chua, P. M., & Dolan, R. J. (1999). Common effects of emotional valence, arousal and attention on neural activation during visual processing of pictures. Neuropsychologia, 37(9), 989-997.
Miller, V. (2008). New media, networking and phatic culture. Convergence: The International
Journal Of Research Into New Media Technologies, 14(4), 387--400.
Morris, J. D. (1995). Observations: SAM: The self-assessment manikin An efficient cross-cultural measurement of emotional response. Journal of advertising research, 35(6), 63-68. Nelson-Field, K., Riebe, E., & Newstead, K. (2013). The emotions that drive viral
video.Australasian Marketing Journal (AMJ), 21(4), 205-211.
doi:10.1016/j.ausmj.2013.07.003
Petty, R. E., Wells, G. L., & Brock, T. C. (1976). Distraction can enhance or reduce yielding to propaganda: Thought disruption versus effort justification.Journal of Personality and Social
Psychology, 34(5), 874.
Roseman, I. J., Spindle, M. S., & Jose, P. E. (1990). Appraisals of emotion-eliciting events. Testing a theory of discrete emotions. Journal of Personality and Social Psychology, 59, 899–915
Sherman, E., Mathur, A., & Smith, R. B. (1997). Store environment and consumer purchase behavior: mediating role of consumer emotions.Psychology & Marketing, 14(4), 361-378. Taylor, D. G., Strutton, D., & Thompson, K. (2012). Self-enhancement as a motivation for
sharing online advertising. Journal of Interactive Advertising, 12(2), 13-28.
Valente, T. W. (2005). Network models and methods for studying the diffusion of innovations. Models and methods in social network analysis, 28, 98.
Wentzel, D., Tomczak, T., & Herrmann, A. (2010). The moderating effect of manipulative intent and cognitive resources on the evaluation of narrative ads.Psychology & Marketing, 27(5), 510-530.
26 Appendix 1: 95 years of KLM video ad (low arousal)
27 Appendix 3: John Lewis Christmas video ad (low arousal)
28 Appendix 5: Innocent Big Knit video ad (high arousal)
