Why Topical Consistency Beats Posting Volume.

On 10 December 2025, Instagram gave people in the United States a button that shows them the algorithm’s working.

Tap the icon in the upper right of the Reels tab and you get a list of the topics Instagram believes you have been into, with the option to turn each one up or down, and a box for typing in topics of your own.

That changes the question a brand has to answer. Discovery used to depend on a system inferring what a person wanted. Now the person can state it, in words, and adjust it whenever they like.

Which means the accounts that win recommendation are the ones a system can confidently file under a topic somebody has asked for. That rewards a narrow, repeatable content system. It punishes a calendar trying to be about five things at once.

What Your Algorithm actually does

Adam Mosseri, who runs Instagram, announced the tool on 10 December 2025. It is called Your Algorithm, and it opens from an icon in the upper right corner of the Reels tab.

Inside is a plain summary of what the app thinks you have been watching, in the shape of “lately you have been into” followed by a handful of topics. Instagram’s own examples ranged from sports and pop music to thrifting, chess and skateboarding.

Below the summary sits the fuller topic list, and each entry can be turned up or down. There is a field for typing in an interest the system has not spotted, and an option to share the whole snapshot to your Story.

Instagram framed the launch as a control question rather than a ranking one. The tool does not change how recommendation works. It changes who gets to say what it should be working towards.

What shipped, and what was only stated

The distinction matters, because a lot of the coverage blurred it.

What shipped on 10 December 2025 was the Reels tab, in the United States, in English. Instagram’s announcement said it was starting there and planned to bring the same ability to Explore and more places in the app, with a wider English rollout to follow.

Mosseri said Instagram had begun testing similar controls in the search tab, and that the longer-term aim was the main feed and the home tab, with other languages taking priority ahead of that.

So if you open your Reels tab and find nothing in the corner, nothing is broken. Plan against the direction of travel, not against a roadmap.

Discovery has moved from inferred to declared

For a decade, topical relevance was something a platform worked out about you quietly. You watched three videos about running and the system drew a conclusion.

An interest you can read, edit and type into is a different object. It turns a hidden inference into a stated preference, and a stated preference is something a recommendation system is obliged to serve.

TikTok shipped a version of this idea first. Instagram’s is derived from what you actually did rather than from a menu of generic categories, which makes it a more honest picture of the account you have trained.

Most content calendars are a filing problem

Here is the uncomfortable part. If a person turns up “home renovation” and your account posts renovation, recruitment, an office birthday, a product launch and a founder opinion in the same fortnight, the system has to decide whether you are a renovation account.

Building content pillars is the mechanical part of fixing that, and most teams have already done it on paper. The harder question this tool forces is how few pillars you can actually live with.

Five loosely related themes is not a strategy. It is a filing problem with a nice name.

Volume makes a filing problem worse

The instinct when reach drops is to post more. It is the one lever a team can pull without a difficult conversation.

But if the underlying issue is that nobody can tell what your account is about, more output spreads the same ambiguity across more posts. Thirty posts across five topics is five weak signals, not one strong one.

Ten posts about one thing tells a recommendation system exactly where to file you. That is not a smaller strategy. It is a legible one. It is also in step with social media’s wider move from volume to value.

What topical consistency actually means

Topical consistency means a stranger, and a machine, could describe what your account is about in one line after watching three of your videos.

It does not mean any of the following:

• Posting the same video repeatedly with a new caption
• Never covering anything outside one product category
• Abandoning brand personality in favour of a single format
• Publishing on a rigid schedule, which is cadence, not topic

The variety lives in the angles, the formats and the people on camera. The subject stays put. Showing up regularly still matters, and that side of the job is covered in the role of consistency in social media success.

The three video test

Pull the last three Reels you published. Show them to somebody outside the business, in order, and ask four questions.

1. What is this account about, in one sentence?
2. What would you expect the next video to be?
3. Which topic would you tick to see more of this?
4. Would you have known these were the same account?

If the answers are vague, contradictory or all different, that is the diagnosis. Ambiguity is not a content quality problem, so producing better versions of five unrelated things will not fix it. That gap between what you think you publish and what an outsider sees is the whole reason a social audit exists.

How to narrow a calendar that is trying to be five things

The work is subtraction, and it is uncomfortable, because every theme on the board has an internal sponsor.

1. Count the topics you published against over the last eight weeks. Be honest, and count what you actually shipped, not what the strategy deck says.
2. Rank them by what the audience responded to, not by who inside the business asked for them.
3. Keep one lead topic and at most one supporting topic. Everything else comes off the recommendation surfaces.
4. Rebuild the calendar as formats within that topic. A tutorial, a myth to correct, a behind the scenes, a customer question answered, all inside one subject area.
5. Hold it for a full quarter. A topic signal built over eight weeks cannot be judged after ten days.

In our experience the narrowing produces more ideas, not fewer, because the brief finally has an edge to push against.

What to do with the posts that no longer fit

They still have a home. The recruitment post, the office move, the award, the partner announcement, these are real and somebody in the business needs them published.

Our view is that they belong where they do not compete for topical clarity. Stories, a founder’s own account, an email list, a careers page. LinkedIn is usually a better fit for the corporate news than a Reels tab that has been carefully trained to mean one thing.

The test is simple. If a post would confuse a person who asked for more of your main topic, it does not belong in the feed that serves that topic.

What good looks like

A good account is boring to describe and interesting to watch. One subject, understood deeply, expressed through five or six repeatable formats that the team can produce without a fresh strategy meeting each week.

The calendar becomes easier to fill, because the question is no longer “what should we post” but “which format goes on this idea”. Briefs get shorter. Approvals get faster, because nobody is arguing about whether the topic is right.

And when a person types your subject into a topic control and turns it up, there is a body of work sitting there that clearly belongs to it.

How NBK thinks about topical consistency

NBK has argued for a long time that posting more is usually the wrong lever when growth stalls. Your Algorithm gives that argument a mechanism rather than just a principle.

Social run as an operating system means the topic decision sits upstream of everything else. It sets the formats, the shot list, the approval route, the reporting lines and the recruitment brief for whoever fronts the content.

Teams that cannot narrow usually do not have a content problem. They have an ownership problem, where five stakeholders each hold a theme and nobody is allowed to lose one. That is a workflow conversation, not a creative one, and it is where the real reach is trapped.

None of this depends on a feature being live in your market. A brand that is legible under one topic was always the easier brand to recommend. The tool simply made the filing visible to the people doing the filing.

Next step

If your output feels busy but your reach has flattened, the constraint is rarely how much you publish. Start by counting how many things your last eight weeks were about, and be strict about the answer.

If that number is uncomfortable and nobody internally can agree on which topics to drop, NBK can help find the constraint in the system and rebuild the calendar behind it.