“Headline testing” refers to the process of developing multiple title variations for an article or piece of online media, which can then be tested on multiple audience segments to determine which one performs the best.
Headline testing can be used to optimize the title of an article or blog post for metrics such as click-throughs, social shares, and other social media and user engagement metrics. The poor performing variations can then be eliminated, so that all future site visitors will be exposed to the best-performing headline choice to increase their likelihood of engagement.
The process for conducting headline testing is to set up the original headline as a control, and choose at least one or two variants. Viral media sites such as Upworthy test as many as 25 headlines per article, but many sites do not have large enough audiences to justify testing that many variants, since it would take a significant amount of traffic to get a statistically significant sample size.
In order to display the headline variants simultaneously to website visitors, an A/B testing tool is used to split the traffic that reaches a website and display the different variations of the headline. The A/B testing platform will gather data from the test, and declare a winner once a sufficient number of people have viewed the test for the results to be statistically significant.
Once the headline test has determined a winner, 100% of the traffic can then be directed at the winning variation. Since the headline of an article is one of the key things that readers look at when reading or sharing an article, optimizing the headline can result in significant boosts in readership and other metrics.
Studies have shown that certain types of words and phrasing in a headline can lead to higher rates of social sharing and higher click-through rates from a website, newsletter or social media platform.
For instance, an analysis of more than 100 headlines by StartupMoon found that articles with more violent terminology were more likely to have social shares: Technology-focused posts that included words like “kill” and “bleeding” in their headlines, such as TechCrunch’s “Oracle makes more moves to kill open source mySQL,” were top-shared posts within those subject areas.
Articles that use numbers in their headlines also perform well. A study by Moz found that the headline variant “30 Ways to Make Drinking Tea More Delightful” performed 15% better than the nearly-identical “Ways to Make Drinking Tea More Delightful.”
For example, if you are writing an article about the proposed new overtime regulations aimed at business owners, you could consider several headline options such as:
All of these headlines address the same issue—the new proposed legislation that will provide overtime pay to salaried workers making less than $50,440. Yet each of them takes a slightly different approach to driving interest in the topic, and it’s likely that you’ll see more people clicking on #2 and #3 than on the first choice, based on results from these studies.
Headline testing is a key practice that can increase readership, social shares and engagement. Start testing your headlines today with Optimizely!