Social Video Lab Launch: Brand Prevalence Has No Negative Effect On Video Sharing

To celebrate the launch of our ground-breaking Unruly Social Video Lab – the first of its kind in the world – we are writing a number of articles that looks under the bonnet of social video advertising. The aim is to provide insight into why people share certain ads and not others, what the current trends to look out for are, and what role social video has in the future of advertising.To find out more about the Social Video Lab, click here.

Brand prevalence in social video content has no negative effect on sharing activity.

That’s according to fascinating new research carried out by a university in Australia, which found that the number of times a brand appears visually or verbally in a commercial hadlittle or no impact on its popularity online.

The study is the second led by Dr Karen Nelson-Field and her team at the Ehrenberg-Bass Institute for Marketing Science into the emotions that are the most likely to trigger video sharing across the social web.

The first report, published back in November 2011, focused on non-branded videos. The second, which used data from Unruly’s Viral Video Chart, compared the original results with 400 of the most shared commercial videos of all time.

The result? People share commercial videos for the same reasons they share non-commercial clips.

The popularity of ads, just like YouTube videos of funny fails or cute kittens, is based on the strength or valence of the emotions they elicit.

In other words, content that elicits a marked physiological response is the key to sharing success.

Advertisers therefore should aim for material that makes us laugh, not just smile, makes us shocked, not just irritated, and makes us exhilarated, not just happy.

Here are some of the key findings from the scientific study:

• Brand presence does not affect the degree to which the video will be shared. This is perhaps counter to industry wisdom, which suggests that removing branding will facilitate sharing (Red Bull Case Study Harbison 2011);
• Videos (branded and non-branded) that evoke marked physiological responses (laughter, anger, crying, shock) are the most likely to be shared – so it's content that makes uslaugh or cry, not just smile or frown, that triggers sharing; 
• Videos (branded and non-branded) that evoke positive emotions (exhilaration, hilarity, astonishment, happiness, inspiration) are more likely to be shared than those that evoke negative emotions (anger, disgust, sadness, shock, frustration). Sadness and disgust (in the commercial data) did punch above their weight relative to other negative emotions, however, it is perhaps a brave brand manager who would venture into this highly provocative negative space;
• Designing high arousal content is difficult and in many cases material that was developed to be emotionally evocative fell well short of expectations. In particular, commercial material seems not to be as funny as non-commercial material. This may be due to agencies focusing on the creative appeal (i.e. the use of animals, babies), as opposed to the degree of arousal it evokes.

So how did they come up with the data? Well, like the first study, Karen and her team took 400 of the top branded videos and worked out the average number of daily shares each clip generated on Facebook, Twitter and the blogosphere (source: Viral Video Chart).

They then asked 14 independent people to watch a sub-set of the sample and indicate the emotions they personally felt from a list of 16 potential emotional responses.

These were: astonishment, exhilaration, inspiration, hilarity, surprise, happiness, calmness, amusement, shock, anger, frustration, disgust, discomfort, sadness, boredom and irritation. 

Each video was ‘coded’ twice to lessen the impact of subjectivity.

They then allocated each emotion an average number of daily shares. Here are the results comparing non-commercial (N-C) with commercial (C):

Table 1: Total average shares per day by individual emotions (000’s) 

* Important caveat – we can’t compare the sharing numbers of two data sets directly i.e. that inspiring non-commercial videos share nearly 4 times more than inspiring commercial videos due to variation in the mean of the dependant variable (daily shares). The strength of these findings lies in the patterns found to hold at the aggregate level.

While the results are consistent across both data sets at the arousal/valence group level, a few differences in the rank order of the individual emotions was apparent. In particular, in the non-commercial data hilarity is the top sharer, while in the commercial data it ranks the lowest.

To understand these differences, Karen and her team sub-coded the data in two different ways.

First by brand prevalence i.e. how many times a brand appeared visually or was spoken verbally in the video.

Dr Nelson-Field said: “The presence of a brand’s logo in a social video is a regularly debated subject in industry. We wanted to understand whether an overuse of branding might explain why hilarious videos don't fare as well in the commercial data.”

Here are the results:

Table 2: Total average brand frequency

Table 2 shows the total average brand frequency per emotion. Karen and her team found only a weak positive correlation between the video sharing of individual emotions and the total amount of times a brand is present.

This suggests that the difference in the ranking order of the individual emotions in the commercial data is not related to the amount of branding present. This table also shows a bigger implication: that brand prevalence does not affect the degree of sharing overall.

Secondly, both sets of data were sub-coded by creative characteristics (Table 3) to find out whether the creative execution of commercial and non-commercial material vary.

Table 3: Prevalence creative appeals (%)

The results found that percentages are largely similar for either commercial or non-commercial videos in relation to the low and high arousal emotions that are elicited by videos that use similar creative appeals.

This means that no one creative appeal will always elicit a particular emotional response, nor will it guarantee sharing success.  So babies and animals can be both successful and unsuccessful in terms of sharing.

However, the table does also show that commercial and non-commercial videos use different creative appeals to generate the same emotional response (perhaps not by design).

Even though creative agencies would seem to be designing more of these (more parodies/comedy skits/pranks in the commercial data than the non-commercial data) fewer of them hit the hilarious mark.

It would seem that staged hilarity may be harder to achieve.

This is a likely explanation for why hilarity in the commercial data does not do as well as non-commercial data.

To read the full presentation, click here.