Testing the Science of Sharing at the Super Bowl: Can Viral Ads Be Manufactured?

After years of trial and error, some believe that viral videos can not only be fostered and quantified but also predicted

  • Share
  • Read Later
Mekanism

Screenshot of the Mekanism created Napster ad that put the creative agency on the map for creating "viral" campaigns.

In 2003, before Facebook and Twitter and YouTube, before Keyboard Cat and Charlie Bit My Finger and Justin Bieber, before viral wasn’t much more than a medical term, start-up ad agency Mekanism was approached by now defunct music-sharing service Napster about creating an advertising campaign centered on its relaunch.

The firm had yet to make an impact in the ad world — Napster was its first big client — so it wanted to try something bold. What they came up with was a series of ambitious, animated ads showing the Napster mascot escaping from jail, being revived in a hospital bed and preparing to rock out to hair metal.

The ads were an overnight success, eventually watched more than 2 million times, making them one of the very first viral hits in the Internet’s social-media infancy — to some extent, admits Mekanism president Jason Harris, because there wasn’t a lot of interesting online content competing with it.

Soon after, Microsoft came calling, hoping Mekanism could do the same for the software giant. “Microsoft said, ‘Well, aren’t you guys the viral guys? Can’t you just make it go viral? Just push a button and do whatever you do to make it happen,’” Harris says. The result: another highly successful campaign featuring a series of videos in which then up-and-coming comedian Demetri Martin searched for spiritual “clearification” — meant to portray Windows Vista’s ability to clear up a user’s desktop.

(MORE: Why Beyoncé Will Make No Bills, Bills, Bills for Her Super Bowl Performance)

Of course, Mekanism didn’t have a viral button — and still doesn’t. But the firm has perhaps gotten as close as any to being able to manufacture viral content on demand. The agency has created dozens of campaigns over the past decade, featuring a claymation Eminem drinking Brisk Iced Tea, a surprisingly hilarious Charles Schwab (yes, Charles Schwab) giving strange and financially misguided folks money advice, and a trailer for Rise of the Planet of the Apes with an all-too-realistic primate firing an AK-47.

In fact, after years of trial and error, many advertising and marketing agencies now claim, like Mekanism, to have developed systematic — or at least reliable — approaches to creating massively shareable advertising content.

While not all firms are convinced that virality can be manufactured, as it were, some believe there’s a science to it. Yesterday’s Super Bowl, for which more ads than ever before were designed to be shared online, was a laboratory for their theories. But now comes the real test: after spending $3.8 million to run 30-second commercials, advertisers are no longer satisfied unless their messages resonate on social networks long after the game is over.

Predicting Virality

The idea that one could anticipate, with any degree of precision, the extent to which a particular video will go viral might seem as far-fetched as predicting which new pop song will become the next megahit or which tech start-up will be the next Facebook. But a company called Unruly Media thinks it has figured it out. In the run-up to the Super Bowl, the advertising world’s biggest night, London-based Unruly began marketing what it calls an algorithmic tool for predicting the “shareability” of video content. “This is the holy grail as far as advertising is concerned,” claims Unruly founder and COO Sarah Wood.

What Unruly does is run sophisticated focus groups in which test-panel participants watch client videos while hooked up to biometric machines that measure psychological and emotional responses through changes in heart rate, eye movements, facial gestures and even skin moisture. It then takes those responses and compares them with benchmarks developed through studies conducted with the Ehrenberg-Bass Institute for Marketing Science. These studies determined, for example, that videos eliciting a strong emotional reaction are twice as likely to be shared — via e-mail, Facebook or Twitter — than those with low emotional responses; and those that trigger “positive” emotions are 30% more likely to be shared than those that elicit other responses.

“If brands want to get people sharing their content, the most important thing they need to do is make an emotional connection,” explains Wood. “They want to make content that gets people laughing out loud or sends shivers down their spine or has their hair standing up on the back of their neck.”

Unruly also found that social motivations play a significant role. We often share a video because it addresses an interest we have in common with a friend, as a form of self-expression, and even for altruistic reasons — think of the success of the viral Kony 2012 video, for example.

For clients, Unruly combines all this information into what it calls a ShareRank, which Wood says can accurately predict the social success of a video. She hopes advertisers will use the system to test advertising content before it’s launched.

(MORE: Why Financial Elites Are Losing Sleep)

“It’s a really exciting time for brands that are genuinely interested in knowing why people are sharing their content,” says Wood, who plans on putting this year’s Super Bowl commercials through the tests to get a sense of what makes this year’s popular ads shareable.

But Unruly isn’t the only outfit applying scientific approach to viral videos. Thales Teixeira, a professor at Harvard Business School, has been doing similar research over the past few years. Using eye- and face-tracking tests, Teixeira has discovered that consumers have an unconscious aversion to forceful brand images; that they seek strong emotional changes that alternate between high and low intensity; and that they tend to share videos that are surprising but not shocking, largely because we don’t know how a shocking video will be received.

He’s also developed a theory that he calls — using a term borrowed from biology — advertising symbiosis. His point, in essence, is that the most successful ads are mutually beneficial for both advertisers and consumers — even though the two groups don’t have the same interests. “The consumer is not interested in helping the company,” says Teixeira. “If the consumer shares an ad, it’s in a very self-interested matter.”

Art, Not Science

Other viral-video creators, however, dismiss the idea that virality can be reduced to a formula.

Jason Norcross is the executive creative director and partner at 72andSunny, an ad agency that’s one of the industry’s leading viral-content makers. The firm created the popular K-Swiss ads featuring Eastbound & Down’s Kenny Powers and last year produced the spot for the Samsung Galaxy S III that gently pokes fun at the legions of Apple fans who stand in line for the company’s heralded product releases.

He says what 72andSunny does isn’t science but art — the result of original thinking, carefully tuned sensibilities and gut instinct.

Norcross says the parameters and metrics cited by Teixeira and Unruly Media don’t even cross his mind when he’s working on an ad campaign. Indeed, many of his firm’s successful viral ads clearly violate the guidelines. Just look at the viral Samsung ad, in which the mobile phone’s brand and name were ubiquitous. “Why that was successful had a lot to do with going after Apple very openly,” says Norcross. “No one had ever done that.”

Norcross’s skepticism about engineering viral videos basically comes down to the irreducibility of true creativity — and the near impossibility of summoning inspiration on demand. “Look, you can do search-engine optimization, p.r. outreach, Facebook, all sorts of algorithms I don’t know anything about,” he says. “But if you want to have Brian Williams do an NBC Nightly News story and feature your ad campaign, I don’t know if you can engineer that. It’s almost like saying, ‘We’re going to make a hit.’ Well, you can’t guarantee that.”

‘Everyone Hates Advertising’

Don’t tell that to Mekanism’s Harris. After the firm’s successful Microsoft campaign, in 2006, Harris says, the firm soon realized that more companies would come looking for a viral hit. “We figured that if people are going to come to us and expect millions of views, we have to back up and figure out a strategy.”

The firm starts with a set of guidelines it calls its Nine Lessons from Epic Wins and Epic Fails. Among them: There’s no such thing as an accidental viral hit. Everyone hates advertising. Start with a simple human truth. Create a memorable story. A great title is everything.

(MORE: Prop Bets, Sports Books and Beyoncé’s Cleavage: Everything You Need to Know About Super Bowl Gambling)

After it uses these principles to develop a piece of content, however, it doesn’t leave things to mere chance. Instead, it relies on network of about 500 online “influencers” — each with a sizable social-media following of his or her own — to tweet, “Like” and discuss the advertising campaigns it creates. A single mention from one of these influencers could earn a video thousands of hits within minutes.

“You have to have the strategy of bringing your friends along, but then you have to have great creative and fundamentally believe that everyone hates advertising, but everyone loves entertainment,” Harris says.

For this year’s Super Bowl, Mekanism was handed one of the biggest responsibilities of the night: creating the Pepsi ad that immediately preceded Beyoncé’s halftime extravaganza, itself sponsored by Pepsi. The assignment, probably the biggest and most complicated Mekanism ever took on, involved narrowing down 100,000 user-generated photos to the 300 that were eventually included in the 30-second spot.

Of course, there’s something ironic about a firm that made its name creating ads that generate their own audience suddenly being handed the biggest audience in the world — 114 million people watched Madonna’s Super Bowl halftime performance last year, which averaged more viewers than the game itself. But that giant stage doesn’t guarantee that Mekanism’s ad will have a cultural shelf life after the game. Some 60 ads will be shown during Sunday’s game, and only a few will get the full digital water-cooler treatment in the days and weeks that follow.

But considering Mekanism’s track record, its Pepsi spot has a sporting chance. “You still never know whether something will go viral,” says Harris. “There’s still the unknown. But over the past six years, we feel like we know the ingredients to put in the stew.”

3 comments
theQuiggler
theQuiggler

Testing after a video's made isn't massively useful.  The core focus should surely be in helping creating better viral ideas, not tweaking bad video ideas.  No amount of post testing is gonna fix a turkey.  This is what TubeRank's designed to help with: http://tuberank.joinvan.com