On Facebook ads, users may dislike ‘likes’

AUSTIN, Texas — Scroll through your Facebook feed, and you’ll get pelted by advertisements begging for a click. Like any other type of post, these ads allow you to react. Often, you’ll notice that one or more of your friends has already “liked” them.

Advertisers hope that a high number of such endorsements, especially from familiar faces, might make users more likely to click. But new research from Texas McCombs finds it depends on the type of ad — and the type of friend.

The wrong ads and friends could have the opposite effect, making a viewer less likely to click. So say Ashish Agarwal, associate professor of information, risk, and operations management (IROM), and Andrew Whinston, professor of IROM. Whinston is also the Hugh Roy Cullen Centennial Chair in Business Administration and director of the Center for Research in Electronic Commerce at The University of Texas at Austin.

Agarwal, Whinston, and Shun-Yang Lee of Northeastern University focused on call-to-action (CTA) ads. Such ads use assertive wording to urge users to do something specific, such as purchase a product or download a mobile app. They’re different from the passive wording of informational ads, which politely invite users to click to “learn more.”

Advertisers tend to prefer CTA ads, Agarwal says, because they put social media users “directly into purchase mode.” But past research had shown a downside to CTA ads: They often rubbed users the wrong way, especially when people felt manipulated.

The researchers wondered whether an accumulation of “likes” could overcome that resistance. Says Agarwal, “Given that these are assertive ads, how would these social cues help or hurt?”

They conducted two rounds of studies.

• In a field experiment, they teamed up with a mobile app developer to place a CTA ad on Facebook, asking users to download an app. It appeared 710,445 times, resulting in 799 “likes” and 4,052 clicks.
• For a lab test, they evaluated different combinations of ads and cues: informational vs. CTA and generic “likes” vs. “likes” from friends. Each of the 982 study participants provided the names of five friends.

The studies found that users had different responses, depending on the ad and the cue. For informational ads, more “likes” led to more clicks. The odds of a click rose 3% for every 100 generic likes and even more — 21% — for each “like” by a friend.

For CTA ads, the opposite was true. The overall number of “likes” had no meaningful impact on clicks.

But “likes” from friends did have effects — both ways. They were positive or negative, depending on whether a user believed a friend had similar or dissimilar interests.

• Having similar interests increased odds of a click 180%.
• Having dissimilar interests decreased odds 66%.

Why the difference? In a follow-up lab study, the team found that users responded negatively to CTA ads, because they felt advertisers were trying to manipulate them. They saw the highlighting of “likes” as part of that strategy.

They set aside that resistance, though, when they saw that friends with similar interests “liked” an ad. They saw the ad as having higher credibility.

By contrast, they found informational ads less intrusive than CTAs. They felt less resistance and were more open to being swayed by “likes.”

The team’s findings have implications for advertisers, Agarwal says, as well as for social media companies that rely on advertising revenue. Displaying “likes” may be effective for informational ads but not for CTAs.

“You have to be a bit careful about the value of these endorsements,” Agarwal says. “Maybe social media companies can make their presence optional. Maybe advertisers should have a choice: Do I want my content to be promoted with these endorsements or not?”

“The Effect of Popularity Cues and Peer Endorsements on Assertive Social Media Ads” is published online in Information Systems Research.