Researchers tested different headlines on social media to figure out the best way to get people’s attention. Here’s what they found.

Professionals in direct marketing have spent years testing headlines against each other, trying to find the perfect formulation that will catch consumers’ eyes. Now academics are getting into the hunt. Here’s what they found: a study in the peer-reviewed research journal Social Influence suggests that when marketing online, asking questions is the way to go.

Linda Lai and Audun Farbrot of the Norwegian Business School, Oslo, conducted experiments on both Twitter and a Norwegian shopping site similar to eBay. They posted headlines with links to stories and then analyzed which headlines pulled the best. There tested four types:

• a declarative headline
• a question headline
• a question headline with self-reference cues directed at the reader (such as using second person)
• a rhetorical question headline that suggested its own answer

What Makes People Click

The short of it is the researchers found that question headlines with self-references to the reader were far and away the winner. Next were questions without self-reference cues. Third, declarative headlines. Bringing up the rear were rhetorical questions.

The researchers did try to control for variables. All messages were posted at the same time of day. They also tried to balance the types of topics to ensure that no approach was the beneficiary of subjects that would of themselves attract a greater number of readers.

Now for the caveats. The Twitter account had 6,350 followers, and given the social network’s synchronous nature (you see messages in time), chances are that many of the followers weren’t even aware of most of the tweets. Even if all received the tweets, it would be a self-selecting group because of the way Twitter works; this was no random sample.

Furthermore, there’s no information on nationality of the followers. As the shopping site was Norwegian, chances are good that it was a fairly culturally cohesive group, so there’s no telling whether what worked with them would with people from other cultural backgrounds.

The number of message topics was low. Also, there’s no indication of how effectively any of the headlines were written. Small variations can make for significant changes in readership. It might be that a more effectively constructed declarative headline would out-pull the question.

The Takeaways

Even questions can be badly framed. In general, yes-or-no questions that allow the reader to fill in an answer without reading more can drive down audience interest. Why check the answer to a question that you’ve unconsciously answered?

Also, any gains you make by asking questions could be lost by too much repetition. You’d likely need to switch off the types of headlines (and hopefully do a lot of your own testing) to keep the audience from skipping over your posts because they know what is coming.

In other words, welcome this and other types of research that might get you trying different approaches. But forget about the magic formula that tells you how to do something every time. If it were that easy, everyone would be able to do it.