Facebook's 'Your Algorithm': Feed Control in Plain Words

Meta has announced Your algorithm, a feature that lets Facebook users add and manage the topics they want to see in Feed and Reels by typing them in plain language rather than picking from a fixed list of categories.

What the feature does

Direct interest input: you can add and manage the topics you want to see in Feed and Reels from a dedicated space, reached through the three-dot menu on any post.
Conversational feedback: after tapping Interested or Not interested, Facebook may ask follow-up questions so you can explain in your own words what you want more or less of.
Real-time adaptation: Meta says recommendations update within the same session, rather than over days.
Updated menu: the three-dot menu on every post has been changed to make these tools easier to reach.

Why this is a real change, not a settings reshuffle

Facebook’s ranking systems have historically inferred your interests from behaviour: what you like, comment on and watch, among thousands of other signals. That approach is good at detecting patterns and poor at accepting instruction. If you watch a video out of curiosity, the system learns you like that topic, and there has been no practical way to correct it.

Meta’s own explanation of the change is that large language models make it possible to incorporate “the interests, context, and intent behind natural language,” producing a ranking system that responds to stated preferences rather than inferred ones.

The combination is the interesting part. The system keeps behavioural signals and adds freeform text on top, so a stated interest can override a misleading inferred one.

What it means for people who use Facebook

You can now correct a bad inference directly. This is the single most useful change for anyone whose feed has drifted toward content they never intended to signal interest in.
Feedback is now a conversation. Explaining why you are not interested should carry more weight than a single tap, because the system has more to work with.
Stated interests are visible to the platform. Typing your interests in natural language gives Meta a more explicit record of your preferences than behavioural inference alone. That is worth knowing before you fill it in.

What it means for brands and creators

If recommendations respond to explicit statements, the value of content that clearly signals its topic rises. A video that is unambiguously about a subject is easier for a ranking system to match to a stated interest than one that is thematically vague.

It also creates a new question about how audiences are counted. A user who has explicitly asked to see a topic is a different kind of viewer from one who arrived through behavioural inference, and the two may respond differently to the same content. For now, that is a hypothesis worth testing rather than a settled fact.

The context

Meta has been rolling out similar user-control features across its apps, and this brings an approach already seen on Instagram to Facebook. The direction of travel across the industry is toward giving users more explicit control over ranking, partly in response to regulatory pressure in multiple markets.

For related coverage, see our report on Meta’s Reels trending ads expansion and Instagram Edits’ new audio links and kits.