Reaching and engaging audiences on social media has become increasingly difficult for media companies over the past year. Meta’s decision to deprioritise news content has had a deeply damaging impact on engagement on both Facebook and Instagram, even in those countries where political content hasn’t been banned altogether.

This is clearly not a tenable position for news companies whose business models are still deeply intertwined with mass distribution through social media. What we’ve seen at Echobox, however, is that necessity is the mother of invention. The seemingly intractable position news publishers find themselves in has led to new waves of experimentation and innovation.

One such insight from this experimentation is a practice that seemingly exploits a loophole in Facebook’s algorithm and can radically improve impressions and engagement on the platform: incorporating photo posts into your strategy.

A short history of photo posts

Back in 2017, Facebook released advice to media companies that mixing link posts (that is, standard posts with clickable images) with photo posts (posts containing an image with a link inserted in the share message) would produce the best results due to recent changes to its algorithm.

We studied 35,000 posts to understand whether this was true and quickly determined that it was not. Our study found that, not only did mixing link and photo posts have a deleterious effect on everything from impressions to engagement, the higher the proportion of photo posts, the worse the performance.

Fast forward to 2024, and we now have some evidence that the reverse seems to hold true: Photo posts are producing significantly higher levels of engagement than link posts.

Photo posts should be a part of your sharing strategy

We work with thousands of publishers worldwide and analyse hundreds of millions of data points every day. That allows our algorithms to detect and adapt to the smallest changes in Facebook’s algorithms.

From analysing the shares of thousands of publishers, we began to note they were achieving positive results by sharing photo posts with the link in either the comments or the share message.

In response, we ran some experiments to evaluate the effect of photo posts on page performance. We took 40,000 link shares published on 33 Facebook pages, randomly converted 25% into photo posts, and measured the results.

Our findings were comprehensive. In each case, we found that every single one experienced an increase in impressions and engagement after incorporating photo posts into the sharing strategy.

Frankly, these increases were massive. On average, pages posting photo posts saw an increase in impressions of around 114%, while engagement increased by around 100%. While our study wasn’t large enough to exclude the possibility of regional variations, the size of the increase would lead us to believe these increases are happening across the board.

The effect on traffic was a little more mixed, although it was broadly positive for the majority of pages. There were some declines in traffic, but, on the whole, we saw a general increase in pageviews.

Why are photo posts outperforming link posts?

Why the sudden over-performance of photo posts relative to link posts?

Our hypothesis is that by throttling link shares, Facebook can keep users on its site for longer and maximise advertising income. Including the link elsewhere in the post seems to exploit a loophole in Facebook’s algorithm and avoid deprioritisation.

Whatever the reason, for now this seems to be a very effective tool to increase news publishers’ performance on Facebook, which remains their largest driver of referral traffic from social media.

We’re still early in our testing of photo posts, so these findings come with the caveat that we don’t know things like whether there is an optimal ratio of photo to link posts or if there is a best place to include the article’s URL. We would therefore advise news publishers to test this for themselves and carefully assess the results before committing to reshaping their sharing strategy.

Caveats aside, though, there is no doubt our own tests indicate there is real scope for news publishers to improve Facebook performance with only a simple change.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .