Meta Brings 'Your Algorithm' to Facebook Feed and Reels

Meta launched "Your Algorithm" on Facebook this Tuesday, letting users tell the feed ranking system — in plain words — what they want to see more or less of across Feed and Reels. It's the same AI-powered tool Instagram got first, and it marks the most direct hand users have ever had on Facebook's recommendation engine.

The mechanics are simple on the surface. Instead of reverse-engineencing the algorithm, users type what they want: more cooking videos, fewer rage-bait posts, whatever it is. Meta says the platform may ask follow-up questions to sharpen its understanding, and changes to recommendations take effect in real time. That last claim is worth watching — "real time" promises a lot, and Meta hasn't published data on how quickly or how completely preferences actually reshape a feed.

Facebook's three-dot menu got a matching update. From any individual post, users can now open "Your Algorithm" directly or hit "Interested" / "Not interested" — a shortcut that existed in fragments before but now sits one tap away from the content itself.

Is this new, or a rebrand of old controls?

Partly both. Facebook has offered interest signals for years; what's new is the natural-language layer and the promise that the system converses back. A Meta description of the feature says it's "designed to make content recommendations more relevant." Relevance, of course, is Meta's metric — not necessarily the user's.

The rollout follows a clear sequence across Meta's apps:

• Reels (Instagram): "Your Algorithm" first launched December 2025
• Instagram Feed and Explore: extended in June
• Facebook Feed and Reels: launched this week
• Threads: runs a related feature called "Dear Algo"

That cadence suggests Meta is treating algorithm feedback as a core product line, not an experiment. It also suggests the company sees personalization prompts as a way to keep users from drifting toward streaming-style services, where viewers explicitly shape what gets recommended.

What does this change for creators and marketers?

In the near term, not much — and that's the honest answer. Meta hasn't said how heavily stated preferences weigh against engagement signals in ranking, so there's no basis yet for claiming a new optimization playbook exists. Anyone promising "prompt-optimized content" this week is selling speculation.

The plausible medium-term risk runs the other way. If enough users actively suppress a topic or format, reach for that content could erode in ways that don't show up in standard performance dashboards. Watch for Meta releasing adoption or impact numbers; the absence of any usage data in this week's announcement is itself telling.

The deeper shift is philosophical. For years, recommendation systems decided what users saw and users adapted. This move — like "Dear Algo" on Threads — nudges the model closer to Netflix-style explicit personalization, where the viewer's stated taste is an input the system must honor.

Whether users actually bother is the open question. Preference menus have historically seen thin uptake, and Meta will need to prove that conversational feedback changes that pattern. Expect the next signal in how aggressively Meta surfaces the feature in-feed — and whether it ever shares numbers on how many people use it.

By Grace Kim