Wednesday, February 03, 2010

Big is beautiful when it comes to viral media too

Last week I published a post titled "Big is Beautiful for Social Media." What I did not say in that post  was that the same principle - the advantage of being big - applies to viral video as well.  If you want to hit critical mass with a video (generally considered to be 1 million views or more) you need both great creative and wide dissemination.

I also posted on this topic back in 2007 (click here to read), but since then we have uncovered a lot more evidence that mass marketing is necessary for viral success.

In a paper titled "Viral Marketing for the Real World," written by Duncan Watts in conjunction with Jonah Peretti and Michael Frumin, Watts finds that both the propagation rate (the degree to which people are willing to pass an ad on to others) and the scale of initial seeding determine the size of the viral audience.

In Millward Brown terms, the propagation rate represents the creative power of the ad, that is, how engaging people find it and how willing they are to pass it on. And it turns out that the viral power of an ad is eminently predictable.

Millward Brown recently completed a study to assess the creative attributes that drive online viral viewing of TV advertising.  The analysis is based on 102 video ads from the U.K. and the U.S. that were shown both on TV and online. The results, which we will be previewing at MAP2010 and eventually hope to publish in the International Journal of Advertising, demonstrate that the same advertising pre-test measures that predict an ad's ability to generate offline TV advertising awareness (enjoyment, involvement, and branding) can also be used, with modifications, to predict its ability to generate viral viewings. Until the paper is accepted to be published, I can share only this topline observation: Using existing measures from the Link pretest, we can explain well over half the variation we see in viral performance on YouTube. The analysis also confirms the hypothesis that viral ads should have LEGS (that is, be Laugh out loud funny, Edgy, Gripping or Sexy). Ads that have these properties tend to perform better than the general relationship would predict. (Click here for more on this topic.)

However, this finding should also be kept in context.  Many other factors can influence a campaign‘s success at "going viral," not the least of which is how well the viral campaign is supported by other media.

Some campaigns may be supported by paid media.  The video ad itself may have appeared on TV, in cinemas, or in paid online video slots.  Of the UK's Campaign's  "Top 10 Virals of the Decade," several are TV ads: John West Salmon "Bear," Honda "Cog" and Budweiser "Whassup."

One other ad in the Top 10, Dove "Evolution," relied not on TV but PR for its success. As reported by Jack Neff in AdAge (Chicago: Oct 30, 2006. Vol. 77, Iss. 44; pg. 1, 2 pgs), Unilever credits the success to a major PR blitz from Edelman New York. This helps explain the vast difference in views between "Evolution" and the follow-up ad titled "Daughters." The former received over 10 million views and the latter only 500,000. The difference was not due to the response to the creative. When tested, people were equally likely to say that they would pass the ad on to others (34% for "Evolution" and 39% for "Daughters"). What was different was how many people were exposed to the two campaigns by PR and news coverage.

My overall conclusion is that in the burgeoning world of social media, there is no free lunch for marketers. Word of mouth applies to both mass media campaigns and viral campaigns, though a key difference exists: When viewers can share links to ads, they can ensure that their friends have the opportunity to see the ad they are talking about rather than leave it to chance. So what do you think? What exceptions are there to this finding?
(average: 4 out of 5)


Email this post to a friend



This entry was posted on Wednesday, February 03, 2010 and is filed under Creative, Media. You can leave a response.

5 Responses

  1. Tuesday, February 02, 2010

    miro

    True enough Nigel
    I'ld be interested in reading the paper connecting the viralness of things to the ubiquitous TV ads
    I think some of its power might be lost in mass ads because there will be less of a "discovery" of it by the viewer (leading to Push vs Pull)...but your point about sharing links to ads is interesting.
    I had floated a different idea (see TV Commercials 2.0  http://miroslodki.wordpress.com/2009/05/29/tv-commercials-2-0-revisited/) that viewers should be able to vote for which ads they get to see..that could achieve the same effect in a different manner.
    Setting aside for the moment the larger issue of the declining 'quality' of programming on TV...and the negative feedback loop that creates.
    Your readers might be interested in this post should they wish more background on what I call The Measurement of Whispers.(http://miroslodki.wordpress.com/2009/08/26/the-measurement-of-whispers/ )
     
    cheers
    Miro

  2. Wednesday, February 03, 2010

    Tom Kasperski

    I don't see any holes in your conclusion. The next logical step would be to estimate a reproduction rate based on creative test results. That way you could determine seeding budgets for maximum ROI. In other words, viral campaigns could actually be predictable.
     

  3. Wednesday, February 03, 2010

    Nigel

    Thanks for the comments guys. Helps pass the time waiting in Terminal 5 at Heathrow.


    Miro, could you use the votes in the same way that Netflix does? If you liked this ad you might like...?


    Tom, I believe Duncan Southgate in our Innovations team is working on a pilot as we speak.


  4. Thursday, February 04, 2010

    miro

    Tom, wouldn`t it be nice if we could predict viral rates.  However one only needs to look at the most recent global effort to predict/manage the H1N1flu virus..where exposure, contageon  and propagation are relatively simple factors...and how a concerted global effort failed to predict the outcome.
     
    Imagine the difficulty of predicting more complex, willful, more trivial and  temporally displaced behavioral outcomes. The formulas themselves are simple, .. it's the data modelling which is more problematic because it needs to have a dynamic feedback loop...although I am sure Nigel`s team will give it a good try.
     
    Nigel, the prompted suggestion could be readily accomodated. The primary thrust of the idea is that mass media can adapt/adopt new technologies on various scales to make themselves more responsive to their viewers..hence deciding what commercials get screened could be implemented right now and applied on a mass scale..then as we devlop more sophisticated internet viewing/interactive delivery models this becomes trivial.
    But it also raises the bar for advertisers to begin thinking about developing more engaging commercials that consumers will want to watch ...will be compelled to pick and will be compelled to act ...rather than the passive assault by the krazy hakkim house of carpets commercials (with all due apologies to krazy hakkim, his family and all carpet sellers reading this)
     
    cheers

  5. Thursday, February 04, 2010

    trevor attridge

    A viral AD that most definitely has ' LEGS '
    http://en.tackfilm.se/?id=1265288915051RA49
    with an addition to the "L" - longevity.

Leave a Reply