The Hottest Campaign Ads On Twitter Didn’t Really Work: Study
At various times during the 2020 campaign, an attack ad surfaced on the internet with Twitter watchers gushing about how devastating it would be for Donald Trump. Aside from the fact that the ads were mostly ineffective, at least not for the nominal point of the election: convincing voters on the fence to support Joe Biden.
That was the conclusion reached by the Democratic Party’s top super PAC after analyzing a handful of spots that went viral on Twitter.
The PAC, Priorities USA, spent a good portion of the cycle testing the effectiveness of ads, around 500 total. And along the way, they decided to do an experiment that could potentially have saved them a ton of money. They took five ads produced by a roommate in the Super PAC domain – the Lincoln Project – and tried to measure their persuasiveness with persuasive swing state voters; i.e., the ability of an ad to move Trump voters towards Joe Biden. A control group saw no display at all. Five different treatment groups, each with 683 respondents, saw one of the five ads. They were then asked the same follow-up questions that measured the likelihood that they would vote and whom they would vote.
The idea was not to be petty or hostile to the Lincoln Project, which attracted both fans and critics for the scorched earth who begged Republicans to abandon Trump. Instead, it was about seeing if Twitter virality could be used as a substitute for actual ad testing that was costly and time-consuming. If what the Lincoln Project did turned out to be compelling, it was thought, then Priorities USA could use Twitter as a quasi-barometer to see how strong their own ads were.
But that was not the case. According to Nick Ahamed, Analytics Director of Priorities, the correlation between Twitter metrics – likes and retweets – and persuasiveness was -0.3. The Lincoln Project’s most viral advertisement – a spot called Bounty, which was broadcast 116,000 times and liked more than 210,000 times – turned out to be the least convincing of the priorities tested.
Reed Galen, a senior official on the Lincoln Project, did not deny the conclusions. In fact, he said they made sense. He noted that his group’s advertising strategy shouldn’t be one-dimensional – that in addition to the spots they trumpeted with great impact on social media, they also posted spots with little fanfare in swing states.
“Right from the start, we were pretty clear about the traces of our strategic reach. The first to make the most noise was for the audience by one. That was the stuff that was directed at Trump, the campaign, the White House, and the family. The stuff we knew about would distract her, make her angry, fight her inside, make him fire Brad Parscale, sue us whatever it was, so that her attention would be drawn to others, ”Galen said. “The second is a lot of what we did in the Electoral College states, a lot of times we didn’t even post it on Twitter. But we understood, no one better than us, that Twitter was a megaphone that, from our point of view, drove what we did against Trump sometimes into his head and sometimes into the narrative that the press watched and created and our 2, 7 million people gave Twitter the energy they longed for. “
This energy was no small matter either. Ahamed said the Lincoln Project spots likely helped motivate predisposed Biden voters, even if their persuasiveness to turn Trump voters around didn’t get through. “What we realize is that as political activists or people who are online a lot on Twitter, we don’t necessarily have good judgment about what is convincing,” said Ahamed.
The conclusion that Twitter is indeed not real life seems to underscore much of the 2020 election. Biden’s victory in the Democratic primary came despite the apparent lack of connective tissue between him and his party’s online votes. And when his team ran the general election against Trump, his digital strategy was based on the notion that virality for its own sake was actually a bad idea.
Ahamed looked at the wreckage of the election cycle and expanded the lesson even further. Sometimes, he argued, political entities not only need to protect themselves from the impulse to attract attention on Twitter, they also need to produce content that is understandable to the majority of the population who do not reside on the medium.
To that end, he sent The Daily Beast the most compelling ad that Priorities ran and tested in the same month the group ran their viral ad test. The spot was remarkable because it looked so conventional.
“I think of the underlying persuasion mechanism as a cough suppressant,” said Ahamed. “People don’t enjoy being changed.”