Facebook campaign advertisements don’t work, says new study

Facebook advertisements are a popular way for candidates to connect with voters, but a recent academic study suggests they aren't actually that effective.

Donald Green, a political science professor at Columbia University, and David Brookman, a graduate student at the University of California, Berkeley, found that voters exposed to Facebook campaign ads were no more likely than other voters to know the advertised candidate's name. Nor were they more like to have a favorable impression of the candidate or vote for the candidate. 

The authors concluded that online ads are "quite unlikely to play a meaningful role in determining the fate of a political campaign or retailer."

ADVERTISEMENT
Facebook declined to comment on the findings.

To conduct the study, the researchers observed a state legislative candidate who bought one week's worth of ads targeting Facebook users in his district.

The ads appeared in a small box on the right side of a user's page.

There were three types of ads: one trying to build name recognition, one stressing the candidate's experience and one touting his desire to improve farming (an important industry in the unnamed state). 

The price of Facebook ads are determined by an auction process, and the campaign spent the maximum amount to win all of the auctions for the available ad slots.

Facebook allowed the campaign to spend a maximum of about $40 per day, or about $280 for the one week campaign.

After the campaign, the researchers surveyed 2,984 district voters. They found Facebook users who were exposed to the ads were no more likely to support or know of the candidate than people not exposed to the ads.

In fact, the users exposed to the ads were not any more likely to even remember seeing them.

"Indeed, the results allow us to confidently rule out the proposition that this veritable bombardment of online ads increased the candidate's name recognition in his district by more than 1.8 percentage points," the authors concluded.

The study was released on Wednesday and has not yet been accepted for publication in an academic journal. 

The results didn't persuade Keegan Goudiss, a partner at Revolution Messaging, a consulting firm that helps campaigns use online tools.

He said Facebook can be a valuable way to connect with voters, and he said he would continue advising his clients to use the website.

"I don't think it's the right methodology," he said of the study.

Goudiss argued that one week wasn't enough time to observe the impact of the ads. He said researchers should have also studied bigger races, like congressional or statewide campaigns.  

Facebook ads are more useful for building a list of supporters than just raising name recognition, Goudiss said.

"We usually sell Facebook more as a step in the process," he said, explaining that Facebook ads are a "very cheap way to build a list or build a Facebook page."  

He noted that Facebook also allows campaigns to employ other interactive tools like games to engage voters and build enthusiasm.

Also, if the campaign wanted to use Facebook ads to raise name recognition, Goudiss said it should have bought Facebook's "premium" ads. Those are displayed with greater prominence.

"This study, for their goal, they didn't use the right Facebook ad tactics," Goudiss said.

But Brookman, one of the authors, argued that the flaws are with Facebook as an advertising platform — not with the study's methodology.

He said the goal of the study was to determine whether Facebook ads can influence the outcome of elections — not whether they can boost the number of Facebook "likes" achieved by a candidate.

He said the researchers only studied a one-week ad campaign because previous research had shown that ads have the most effect within the first week of being viewed.

He noted that Facebook wouldn't allow the campaign to spend more money on the ads.

"We maxed out the platform," Brookman said. "We consider this the upper bound of the effect this could have politically on someone's race. If there's various limitations ... that just speaks to how difficult it is to use Facebook politically."

Brookman admitted that $280 might not have had a measurable impact if the candidate had spent it on TV ads or yard signs instead. 

"This still could be the most cost-effective campaign tactic out there," he said. "We can't rule that out."

But he argued that Facebook ads are cheap because most groups doubt their value.

"There's a reason TV ads are so expensive," he said.