Political campaigns are set to spend over $159 million online in 2012, and Facebook is happy to soak up the advertising binge. However, a first-of-its-kind experimental study finds that Facebook banner “ads had no politically consequential effect on knowledge of the candidate, his favorability, or support for his election among voters.” While others in the media have misinterpreted this study to mean that Facebook ads don’t matter at all, the more accurate lesson is that candidates need users to also voluntarily share their content. Recent large scale research proves definitively that widely shared Facebook political ads do have a material impact. In other words, on Facebook, obscure brands can’t buy popularity; they need friends, too. And, this portends good things for democracy, as political ads are increasingly forced into an online arena of conversation.
Berkeley graduate study David E. Broockman teamed up with the academic titan of campaign science, Donald P. Green, to test how popular Facebook ads could make an obscure state legislature candidate. According to Broockman, the local race provided the perfect natural laboratory, since most politicians aren’t running for President, nor are they household names. The experimental study saturated the candidates over-30 target demographic with the maximum allowable Facebook banner ads and measured the impact with a telephone survey of his popularity in the same market. Ultimately, he was a virtual stranger before the ads (85% unaware) and not much changed after the Facebook barrage. Many heavy Facebook users didn’t remember seeing the ad at all.
For a (very) rough comparison to different types of media, Broockman tells us “mail campaigns have been shown to persuade 3 points, TV 3-5 points, radio 6 points, and the candidate knocking on your door and talking to you ~40 points (yes, forty; personal contact is king).”
This doesn’t mean Facebook is ineffective. Just last month in the prestigious journal, Nature, Political scientist James Fowler found that a single get-out-the-vote message could produce a 2.2 percent bump in voter turnout. More importantly, 80% of the impact came from friends sharing the message. In other words, on Facebook, money alone can’t buy popularity. (It should be noted that for Fowler’s study, Facebook took over the newsfeed and placed a vote counter at the top of the timeline, as it does every election year).
In regards to the Berkeley study, as a local politician, it’s likely that few if any voters wanted to share the advertisements or fan page. Additionally, the ads might just not have been very good: using the most advanced mobile targeting data, the Romney campaign tells TechCrunch that they’ve achieved an astonishing 10% click-through-rate (10x higher than Facebook’s average).
While this study will make the Facebook ads team cringe, I suspect it makes Mark Zuckerberg quite happy. Zuckerberg prides Facebook on being a marketplace for the best ideas. He once glowed at the idea that movie studios have difficulty hocking bad films on Facebook, as proof that the wisdom of the users could overshadow clever marketing.
Perhaps then, in an unintentional way, the results of this study are good for democracy. As ads are increasingly forced to rely on social support, it’s difficult for shadowy billionaire partisans to flood radio waves with lies and caricatures. Every ad worth sharing will be attached to a rigorous debate and fact-checking. Facebook could be a very important piece of the puzzle in ending the coercive manipulation of paid political advertising.