Instagram's "Your Algorithm" topic controls went live globally for English-speaking users across Feed, Reels and Explore on June 10, 2026, giving anyone the ability to see and edit the interest categories the app's AI has assigned them. The catch, confirmed by Instagram's own engineering documentation and a University of Maryland study published the same month: remove a topic you've actually been watching, and your scrolling behavior will quietly out-vote your stated preference. The feature works as advertised only when you're correcting a misclassification, not when you're trying to override a habit.
What Is Instagram's "Your Algorithm" Feature?
Announced by Instagram head Adam Mosseri in a June 10, 2026 Threads post, Your Algorithm shows users the topics — "trail running," "vintage fashion," "plant-based cooking" — that Instagram's recommendation system has inferred from their watch time, taps, saves and DM shares. Users can remove topics, add ones they want more of, or simply use the panel to understand why certain content keeps surfacing. The tool launched first on Reels in October 2025, expanded to Explore, and is now unified across all three surfaces: removing a topic in one place removes it everywhere. Mosseri framed the update as a response to a structural problem: "Who you follow used to be a meaningful tool people had for shaping their own experience, and as recommendations took over the main feed that tool quietly stopped working."
Why Doesn't Removing a Topic Actually Remove the Content?
Because the control sits on top of, not instead of, Instagram's engagement-ranking engine. Per Meta's own engineering documentation, Instagram runs a two-stage system: a Two-Tower neural network retrieves candidate posts by matching a user embedding against content embeddings, then a multi-task ranking model scores those candidates on predicted engagement — click, save, rewatch, and especially DM shares, which Instagram's confirmed 2026 ranking signals weight at roughly fifteen times a like. Removing a topic adds a negative signal to your user embedding. But your real-time scrolling behavior updates that same embedding continuously, and a 2026 University of Maryland study on stated versus revealed preference found the behavioral signal typically wins. Pause on a reel for even a second while "scrolling past," and the system logs it as engagement, not rejection.
What Does This Mean for Brand and Discovery Reach?
Topical Clarity Is Now a Structural Ranking Advantage
Instagram's recommendation engine classifies an account largely from its most recent nine to twelve posts. An account that mixes founder content, product shots, event recaps and lifestyle posts in no particular order creates classification ambiguity — it's harder for the AI to slot into a topic bucket, and harder for a user's topic list to match it. A boutique hotel, wellness studio or lifestyle brand posting a consistent content category earns a cleaner match to users who've actively added that topic. This is a bigger lever for non-follower reach than caption keywords or hashtags.
Users Can Now Hard-Block a Category — Regardless of Past Engagement
The inverse risk is real too: a user who removes a topic your brand's content falls into will stop seeing it, even if that user previously engaged with your posts at a high rate. Engagement history no longer guarantees future reach if a viewer has explicitly opted out of the category. That's a new failure mode brands haven't had to plan for before.
Why the Timing Makes This More Than a Feature Update
Five days before Your Algorithm's global rollout was reported in detail, on July 10, 2026, the European Commission published preliminary findings that Instagram and Facebook's use of autoplay, infinite scroll and "highly personalized recommendations" breaches the EU's Digital Services Act, calling the design "addictive" and citing a potential fine of up to 6% of Meta's global annual revenue — roughly $12 billion. Meta disputes the findings. Instagram is, in effect, simultaneously offering users a dashboard to edit its recommendation engine and defending that same engine in a regulatory proceeding. eMarketer projects average time spent on the app will hold flat around 17 minutes a day through 2027 regardless — a signal that the feature's actual pull on user behavior, and therefore on how much brands should reweight strategy around it, remains unproven.
What Hospitality, Wellness and Lifestyle Brands Should Do Now
Don't treat Your Algorithm as a reason to chase "add me to your topics" prompts — Instagram gave users no mechanism to browse and follow brand-specific topic categories, and the tool doesn't replace the Following tab. The actionable move is upstream: audit your last nine to twelve posts for topical consistency, because that window is what the classification model is reading right now. If a wellness studio's grid alternates between class reels, founder vlogs and product try-ons with no organizing thread, tighten it before worrying about anything else in this update. And keep weighting content toward saves and DM shares over likes — that hierarchy hasn't moved, and it's still the fastest way to earn distribution to people who've never followed you.
One Media Society tracks how platform ranking changes affect where hospitality, wellness and lifestyle brands actually get discovered — not just what a feature announcement says it does. Talk to us about auditing your content mix for the algorithm you're actually being ranked by.