Meta has provided an updated guide of how its different feed algorithms function, and just how it's leveraging sophisticated AI systems in order to try and match the right content for interested users in-stream, which could help better make sense of why you're seeing what you're seeing on Facebook and Instagram.
And for marketers, may give you a better handle on it the same, in order to help you better connect with your target audiences.
In a new explainer, Meta President of Global Affairs Nick Clegg says that while it is true that people use Meta's services, there are myriad other contributors who affect their feed as well, particularly in terms of activity.
"The AI systems predict how valuable a piece of content might be to you, so we can show it to you sooner," Clegg said. For example, our systems will use a pretty good predictor that you may share a post if you tend to share posts, in part as an indicator that you found the post interesting. As you might imagine, no single prediction is a perfect gauge of whether a post is valuable to you. So we combine a large range of predictions to get as close as possible to the right content, but some are based on behavior, and some are based on user feedback received through surveys."
Meta has provided similar descriptions of its algorithms before, intended to explain why people see what they see in their feed.
The critical considerations-or factors-that the system looks at, according to those notes are:
Where the post comes from – How often a user engages with a profile or person
When it was posted – The time it was posted, and initial post response.
How likely that it'll drive engagement – The system will optimize to best lean into each users' specific behaviors, including likelihood to comment or share.
With the integration of AI, Meta can now double down on those core elements while fine-tuning the user experience in real time for everyone.
To explain exactly how all these different factors play into that, Meta has published a new series of 22 'system cards' outlining how its systems rank content.
Every card gives an overview of how Meta's feed algorithms work, which can assist in improving the understanding of both what is seen inside the apps themselves and what determines reach.
It's a great resource to draw on to build up your understanding of the system, and could potentially be a very effective way to maximize content performance – though much of the explainers is fairly vague and intentionally, so as not to give too much guidance for people to game the system.
Meta has also stated how it is applying AI particularly within its ranking process, given a new overview explaining its improved systematic content understanding, which can now interpret the 'semantic meanings of content holistically across different modalities (such as image, text, audio, or videos)'.
"These production models provide capabilities like visual recognition, object detection, text extraction, and audio recognition. They also enable us to do more application-specific tasks, such as topic/genre classification, hashtag prediction, similarity matching, and clustering."
In short, Meta's systems are better at understanding what is in each element of your posts – including objects in images and videos – to help show the right content to users based on their interests.
Well, TikTok is actually also making use of the same kind of logos, and in that case you will have even more content being pulled out and simply visual from outside of the hashtags or key words in the description. So that makes the TikTok feed even more interesting and Meta's now trying to bring that out to Reels, which has been driving the growth of Facebook and Instagram engagement over the past year.
But such secrets aren't revealed here. Meta's not publishing a talisman that will explain how to boost your reach across its apps, but it is trying to provide a better overview of its ranking system in order to help users understand the many considerations that factor into what they're seeing, and how they can influence such, both through their activity and manual controls.
On the latter, Meta is also looking to provide more insight, with an update to its 'Why Am I Seeing This?' element in Reels, both on Facebook and IG, that will give more information about how your previous activities have informed the Reels you are seeing.
Meta is also bringing new content control options to Facebook and Instagram where users will have more control over the content they see in each app.
" You can visit your Feed Preferences on Facebook and the Suggested Content Control Center on Instagram from the three-dot menu on relevant posts, as well as through Settings.
It's also rolling out new 'interested' indicators on Reels, so you can tell the system you want to see more like this type-done in much more explicit form than Likes.
Again, no magic formula at play here: Meta's not opening up its black box and letting you into all of its algorithmic secrets. But the new transparency tools certainly give you a better look at all of its ranking models and general factors at play when weighing how it goes about shaping each user's experience.
From a marketing point of view, it's the real value in knowing which elements Meta is weighting more at any given time, but, one, it's always changing, and two, giving people a map of how to game the system is probably not the best exercise.
But if you want to know how Meta's systems work and how they're improving, perhaps take some time over the long weekend to go through these explainers and notes.