Facebook adds written interests to Feed and Reels recommendations
Facebook adds topics and written feedback to recommendations, without establishing new advertiser targeting or proven brand reach gains.
Meta is bringing more direct control over recommendations to Facebook Feed and Reels. Its October 13 announcement introduces Your algorithm, a place for people to add and manage topics they want to see, alongside written feedback on posts.
For brand publishing teams, the change makes explicit audience interests part of the recommendation conversation. It does not provide a new marketer-controlled targeting field, and Meta has not supplied evidence that the feature will increase a brand’s reach. The practical question is how clearly a piece of content fits an interest a person actually wants to pursue.
Key Takeaways
- Facebook is rolling out topic management and written feedback for Feed and Reels recommendations.
- Meta says large language models help ranking systems interpret those requests and respond within the same session.
- The announcement does not establish new paid-ad targeting controls, geographic availability or measurable brand reach gains.
Table of contents
Jump to each section:
- Users can give the feed a clearer instruction
- Explicit interests join behavioral signals
- Brand relevance still needs to be earned
Users can give the feed a clearer instruction
Your algorithm gives users a dedicated space to add and manage topics in Feed and Reels. Meta says it can be accessed through the post menu, which it has updated to make the new customization tools easier to find.
Users can also indicate interest or disinterest directly from a post. Meta says it may follow up with a request for more detail, allowing people to describe what they want to see more or less often in their own words. That puts written input alongside the familiar act of reacting to a particular piece of content.
The announced controls operate at different points in the experience:
| Control | What Meta says it does |
|---|---|
| Your algorithm | Lets users directly add and manage topics for Feed and Reels |
| Post-level feedback | Lets users indicate interest or disinterest in what they see |
| Written follow-up | May ask users to explain what they liked or disliked in their own words |
Meta says recommendations respond in real time, including within the same session. That is the company’s description of the product, rather than an independently measured responsiveness result. The announcement does not specify a country-by-country rollout or a complete implementation timetable.
Current reporting on the release places Facebook’s feature alongside Your Algorithm on Instagram and the related Dear Algo product on Threads. They reflect a similar direction of user input, but their names and interaction models should not be treated as interchangeable.
Explicit interests join behavioral signals
Meta’s existing Feed ranking explanation describes a mix of connected content and recommended content. Connected posts can come from friends, followed Pages and joined Groups; recommendations can introduce content from people or sources a user does not already know.
The new announcement adds a more expressive way to state an interest. Meta says large language models can interpret the context and intent in written feedback alongside what its systems already understand from a person’s activity. The company is not saying that previous behavioral signals disappear.
That distinction matters for content strategy. A person watching a post can mean several things, while a written request can be more specific about the topic they want to explore. The useful change is another input into personalization, rather than a public rule that tells publishers exactly how every post will be ranked.
It also limits what brands can conclude. A team cannot assume that a popular keyword in a caption will override individual preferences or guarantee recommendation. Meta has not disclosed the relative weight assigned to these new inputs, nor a marketer-facing method for inspecting them.
The announcement concerns people shaping what Facebook shows them. It does not establish that those descriptions become a new advertising audience segment available to buyers. Paid-ad planning should continue to rely on the actual controls and documentation available in the advertising product.
Brand relevance still needs to be earned
For a publishing team, the most useful response is to examine whether its content has a clear topic and delivers what it promises. A post framed as practical travel advice should offer useful travel information, rather than use the label as a wrapper for a loosely related promotion. That is an editorial evaluation, not a disclosed ranking formula.
The same idea applies to a series. Consistent subject matter can help a team understand what audience need each post serves, while a shift in topic should be deliberate. The new controls make relevance a reasonable testing question, but they do not prove that any particular publishing pattern will produce better reach.
Teams should also separate a visible audience response from an explanation of the algorithm. A change in views or engagement can have many causes, including the subject, creative, distribution and broader viewing conditions. Without additional evidence, it should not be attributed automatically to Your algorithm.
A useful reporting practice is to record the intended topic and compare results across genuinely comparable posts. That can support decisions about what the audience finds useful without claiming access to a user’s private written preferences. The announcement does not give brand teams that access.
Facebook’s change creates a more direct channel between users and its recommendation systems. For marketers, it reinforces the value of clearly useful content while leaving the central commercial questions open: who receives the feature, how it affects distribution and whether measurable brand outcomes change.
