We consider A/B headline testing as a powerful tool for optimising our content for users, so we recently doubled down on putting it to good use.
every author is instructed to develop alternative russian business mailing list headlines for his or her articles and forward ideas for developing headlines to the editorial team.
For the main topics, we collate alternative headlines in our central planning tools for everyone to see and use when needed. More specifically, we have increased the use of A/B testing even at off-peak hours, always looking for the best headlines – or at least identifying better options – not afraid of re-doing tests later, sometimes with the same variants, often with newly developed ideas.
For example, the late shift runs A/B tests for new content even at night-time, which might not have the biggest impact right away, but does provide instant options for optimising content for the early shift – it gives them a head start and they can push on from there. This also includes A/B testing different images, not just headlines, and for videos or galleries, too, not just articles. Should we show Messi holding a cup or Messi in tears etc…
How did you get the team on board with your new strategy?
Most editors, authors and reporters are keen on tweaking content, to begin with, so there has always been a passion for optimising headlines.
What we have done via meetings, memos and setting examples is raising awareness for a more regular usage of A/B testing, deliberately making room (and time) for that as well as showing the impact of data-based optimisation. Headlines are even more debated in meetings and chats now.
How have you communicated those results across the editorial team?
A/B test results are regularly shared and talked about by the daily team or presented in meetings, if relevant, for current workflows.
What we’ve implemented on top of that is a “best of headline A/B testing” initiative. We gather data from Taboola Newsroom, as well as corresponding examples, and analyse and summarise the findings to closely monitor our usage and its impact, making sure we don’t overdo a headline and we stay true to the message of the article. This creates key learnings for the team.