Meta Says That its On Track to Double the Amount of AI Recommendations in User Feeds

Along with its Q1 earnings report, which showed that Facebook usage has continued to grow in all markets, Meta has also shared some new insights into is evolving use of AI-based content recommendations, and how it’s increasingly relying on machine learning to help populate user feeds, as opposed to relying solely on your social graph.

Meta has been leaning into AI over the past year, with Meta chief Mark Zuckerberg noting last July that it planned to double the amount of AI recommendations in user feeds. At that stage, around 15% of the content users were shown on Facebook came via AI recommendations, and ‘a little more than that’ on IG.

Now, it’s pretty much on track to meet its expanded AI usage aims.

According to Meta, as of today, more than 20% of the content displayed in people’s Facebook feeds has been recommended by its AI system, while on Instagram AI recommendations now account for some 40% of the content shown.

Which is a huge shift, especially given that Meta’s market advantage thus far has been your social graph, and the fact that pretty much everyone uses Facebook and/or Instagram, which gives it a big leg up over the competition.

But with the arrival of TikTok, the approach to recommendations has shifted, with TikTok’s algorithm focusing on each individual post, not on who you’re connected to, necessarily. That’s a step away from ‘social’ media, as such, and more towards entertainment – which is why TikTok has repeatedly noted that it’s actually an entertainment network, not a social app.

And by highlighting the top performing content from topical categories, Facebook and Instagram are now moving in that direction as well.

So what does that mean for social media strategy, and how you can maximize your reach and resonance in each app?

Well, really, it should provide some level of advantage for brands, because while people are viewing more content on Facebook, they’re also posting fewer updates themselves, with personal sharing increasingly shifting into DMs instead. That leaves more room for the best content to get greater reach in the app – but the challenge then is that you’re no longer looking to maximize social engagement or interactive value as such, you’re instead focusing more directly on entertainment value.

Create good, entertaining content, especially in Reels (Meta says that Reels has seen a 24% increase in total time-spent on Instagram), and you stand a good chance to get more reach in either app, with its algorithm now looking to showcase more content to people outside your direct audience.

Topical, thematic content, aligned with your audience’s interests could actually see more reach as a result, with Meta also noting that users are now resharing Reels more than 2 billion times every day, a 100% increase over the last six months.

Of course, creating entertaining content isn’t necessarily easy. It’s one thing to say ‘create content that your audience will love’ and another thing to do it. But the focus of Meta’s AI recommendations is to show users more content that they’re likely to be interested in, which is based on the content that’s generating the most views and time spent within each topic.

That could present new opportunities. Worth considering in your approach.