Instagram launched "Your Algo" topic controls in December 2025 for a small group of users. As of July 2026, it's rolling out to every English-speaking account globally. Not a test. Not a select region. Everyone.
The mechanic is simple: users tap into their settings and tell Instagram which Reels topics they want more of and which they want gone. Fashion, fitness, food, beauty, politics, sports — there's a full category list, and users can tune it however they like.
Meta said the quiet part out loud in their support documentation: these preferences influence ad delivery, not just organic content.
What just changed
"Your Algo" isn't a new idea. TikTok has had interest filters for a while. What changed is scale and category specificity. Instagram now has a formal, user-declared interest graph that it applies to both feed ranking and, explicitly, paid placements.
Before this, Meta inferred what you wanted based on watch time, saves, and interaction patterns. That inference was invisible and passive. Now users are actively opting out of entire content categories with a single tap. The result is a segment of your addressable audience that has formally told Instagram they don't want your content type — and Meta honors that.
Why the CPM number lies to you
Here's what makes this tricky. Meta doesn't let your CPM spike when a portion of your target audience opts out of your category. The system quietly routes your impressions elsewhere — to users who haven't opted out — and the aggregate cost-per-thousand looks normal.
So you open Ads Manager on Monday, see $14 CPM, and assume everything is fine. The dashboard says reach is stable. It is, technically. But the composition of who's in that reach has shifted. The slice of your audience that was previously most receptive — people who followed similar accounts, saved your content type, engaged regularly — may now be partially filtered out.
The leading indicator isn't CPM movement. It's frequency rising in a fixed audience while click rates drift down. If you're spending the same amount but getting fewer clicks at the same CPM, your addressable pool is shrinking and the system is cycling through the remaining users more often.
Who gets hurt most
Categories with high opt-out rates are going to feel this first. Fashion, beauty, fitness, and food are the four most likely to see users actively tune them out — not because people don't like those products, but because the volume of content in those categories is overwhelming and users prune aggressively.
The inverse is actually useful information. Any user who explicitly adds your category to their "more of this" list is the highest-intent person in your audience. They've told Instagram they want it. That's a stronger signal than any behavioral proxy Meta was inferring before.
If you run Advantage+ campaigns, that explicit preference data feeds directly into delivery optimization. Users who've opted into your content category are going to see higher impression allocation. That's good — it means your creative quality matters more now, because the system is routing it to people who already want what you sell.
What Advantage+ does with this
Advantage+ Shopping Campaigns were already using behavioral signals to prioritize delivery. Topic preferences are now a formal layer on top of that. For broad creative with strong hooks, Advantage+ should handle the adjustment automatically — it will route toward opted-in users and find new lookalikes based on similar topic selections.
The risk is in the learning phase. If your category has high opt-out density, Advantage+ needs more time to find enough opted-in users to exit the learning phase cleanly. If you're seeing learning-limited status stretch past two weeks, the opt-out compression in your target category is one possible cause — not just budget or conversion volume.
Manual interest targeting doesn't save you here. The interest category sizes in Ads Manager don't account for "Your Algo" opt-outs. You might see an audience of 8 million in a women's fashion interest set, but the actual deliverable subset after opt-outs is smaller and you can't see the gap from inside Ads Manager.
Four things to check this week
Reels frequency week-over-week. Pull frequency as a standalone metric for your Reels placements specifically. If it's up more than 15% in four weeks without a budget increase, your addressable pool is contracting.
Click rate at stable spend. CTR declining at flat CPM is the most direct signal. It means the impressions are going to people who aren't engaging — classic over-frequency or deteriorating audience fit.
Engagement quality signals. Saves, profile follows, and DM initiation from ads are opt-in behaviors. If users were opting out of your category, they weren't doing these anyway. A drop in these signals is more informative than a drop in link clicks.
Advantage+ learning windows. If you're running ASC and the learning phase is stretching, don't reset the campaign. Give it three to four weeks before drawing conclusions, especially for categories with heavy content volume (fashion, beauty, fitness).
The brands that will hold performance are the ones whose content earns the topic. If someone who said they want less fitness content still watches your ad, clicks, and saves it — that's a real signal. You can't buy your way past a user's declared preference. You can only make creative good enough that they'd pick it anyway.
If you want to see where your Reels frequency is going and how your engagement signals compare to your category benchmarks, the free audit at Gromerce pulls that data in three minutes. You don't need to connect your payment method.
Related reading:
- Meta's Pixel Data Is Now Officially Part of How Your Organic Feed Ranks
- Meta's Andromeda Engine Rewrote How Creative Gets Ranked
- Meta Extended Purchase Audiences to 730 Days
Sources: Instagram official "Your Algo" support documentation, Engadget reporting (July 2026), Social Media Examiner, Influencers Time coverage of global rollout.
