Your Audience Can Now Switch Your Topic Off.

Instagram now lets people edit the topics behind their own recommendations.

The feature is called Your Algorithm. It shows a person the subjects the system thinks they care about, then lets them ask for more of one, less of another, add a topic that is missing, or take one off the list entirely.

Since 10 June 2026 it has covered Feed, Reels and Explore as one system, so a change made in Reels applies in Explore and in Feed too.

For brands, that turns a planning label into a distribution mechanic. The topic your content gets filed under now helps decide whether a viewer keeps being shown you at all.

What Instagram actually shipped

Your Algorithm launched on Reels on 10 December 2025, first in the United States and then to English language users everywhere. It reached Explore in April 2026 and the main Feed on 10 June 2026. It is available globally in English, and Instagram has not announced a wider language rollout.

Inside the control, a person can:

• read a plain language summary of the topics the system associates with them
• ask for more or less of any topic on that list
• type in a topic that is not listed and add it
• remove a topic they no longer want
• adjust interests straight from the topic pills at the top of Explore

Instagram also labels the outcome. When a post appears because of one of those declared interests, it carries a tag naming the interest, in the style of “From Running”.

One thing to hold loosely. Instagram is testing further ways in, including pulling down in Feed to open the menu and per Reel buttons for more or less like this. Adam Mosseri has said plainly that some of it is testing, some is coming soon, and some may not work out. Plan against what is live, not against the tests.

The mechanic, stated precisely

Most reactions to this feature get the mechanic wrong, so be exact about it.

Removing a topic is not unfollowing you. It is not blocking, hiding or reporting your account. Someone who follows you still receives your posts through the connected path, ranked the way they always were.

What changes is the recommended lane. Declared topic preferences feed the systems behind Explore, Reels and suggested posts, sitting alongside the behavioural signals Instagram already used: what a person watches, what they engage with, who they follow.

So a removed topic is a signal to show less of that subject, not a ban on your account. Instagram has not published how heavily that signal weighs against everything else, and nobody outside the company can tell you.

Why the recommended lane is the half you cannot afford

Instagram talks about reach in two halves. Connected reach comes from people who already follow you. Unconnected reach comes from people who do not.

Growth lives almost entirely in the second half, and recommendations now carry a large share of what an average person sees in Feed. Following an account used to be how people shaped their own experience. As recommendations took over the surface, that lever quietly stopped doing much.

Which means the risk here is not that your followers stop seeing you. It is that the people who have not followed you yet stop being offered you. We have already made the performance case for staying in one topic rather than chasing volume. The new part is that the audience now holds a switch.

Your topic is now visible to the viewer

The “From Running” style label is the quiet change in all of this. The platform tells a viewer which interest surfaced you before they have watched a second of the post.

That sets an expectation, and expectations get judged. A post served under a topic it does not deliver reads to the viewer as a mistake by the system, and the fix the app offers for a mistake is a button that says less of this.

Suppose a fitness brand posts a recruitment video. It gets classified into something, shown to someone under a label like strength training, and it is plainly not strength training. Nothing dramatic follows. One person taps less of that topic, and a subject they genuinely wanted from you goes quieter.

What drift costs now

Posting off your lane used to cost you one weak post. The reach was poor, the numbers were poor, the team shrugged and moved on.

The cost now has three parts, and only the first shows up in your reporting:

1. The post itself underperforms.
2. It teaches the classifier that your account also belongs to that topic, which changes who your next post is offered to.
3. It hands part of your audience an obvious reason to switch a topic off.

Instagram does not report topic removals back to the account, so nothing in your insights will ever name the cause. You get a quieter quarter and a set of numbers that look like general decline.

Content pillars stop being a planning convenience

Pillars are usually sold to teams as an organising idea, a way to stop the calendar becoming a list of whatever anyone thought of on Monday. That framing undersells them now.

A pillar is the topic pool you are asking to compete in. Keep them clear, consistent and legible to a machine, and you accumulate strength in a small number of pools and keep appearing for people who declared an interest in them.

Keep them vague and every post starts from scratch somewhere new. This is exactly the question a social audit and strategy engagement exists to settle, because most teams already have pillars on a slide and cannot tell you which of the last thirty posts belonged to which one.

Name your pillars the way a viewer would

A pillar written for an internal deck, something like “brand world” or “community moments”, is useless as a topic. It has to survive contact with two readers now: a classifier, and a person choosing what they want more of.

Test each one:

• Would a real person type this phrase into a “see more of” box? If not, it is a theme, not a topic.
• Could someone name the subject from the first two seconds with the sound off?
• Do the caption, the on screen text and the spoken words all point at the same thing?
• Would three consecutive posts in this pillar look like they came from one account?
• If the post carried a label naming this topic, would the viewer agree with it?

That last question used to be a thought experiment. It is now literally what happens on the screen.

Put the topic question inside the approval

Most drift is not a strategy decision. It arrives as a favour: a request from another department, a founder’s idea on a Friday afternoon, a format somebody wants to try because a competitor did.

The fix is procedural rather than creative. Add one compulsory field to whatever brief or approval step you already run:

• Which pillar is this post in?
• If the honest answer is none, where does it go instead?
• Who signs off an exception, and how many are allowed in a month?

A team that cannot answer the first question in five seconds has just found the reason its reach is uneven.

Where off-lane content should actually live

You will still need to publish things that do not fit. A hire, an office move, a charity day, a partner announcement. None of it has to go into the surfaces this control governs.

Instagram named Feed, Reels and Explore. It did not name Stories, and Stories are mostly a followers surface anyway, which makes them the right home for company news. The people who want that update have already opted in by following you.

The rule is not never post off-lane. It is that off-lane content does not go where discovery happens.

What good looks like

• Three to five named pillars, written in the words an outsider would use.
• Every scheduled post tagged to one of them before it is approved.
• A standing home for off-lane content that is not the recommendation surfaces.
• One named person who answers when someone asks for an exception.
• A quarterly check on whether the pillars still match what the audience watches, rather than what the team enjoys making.

None of that needs new tools. It needs the decision made once, written down, and enforced by the process instead of re-argued every week.

How NBK thinks about topic discipline

A large share of the reach problems we get asked to fix are not content quality problems. They are classification problems. The work is good, and it is spread across so many subjects that nothing compounds.

The remedy is unglamorous and it holds. Decide the lane, make the lane legible, and build the workflow so an off-lane request gets routed rather than published. If your pillars are still a slide rather than a rule, building them properly is where to start.

Instagram has made the consequence of skipping that work explicit. The audience can now see the topic. The audience can now turn the topic off.

Next step

If your output feels busy but your reach keeps sliding with no obvious cause, the constraint usually sits upstream of the content. Start by auditing what your last quarter of posts is actually teaching the platform about you.

If that is a job you would rather not run internally, NBK can help find the constraint in the system.