X algorithm update 2026: mutuals, replies, and reach

Short answer

X made mutuals matter again.

On July 13, 2026, TechCrunch reported that X had adjusted its algorithm to increase visibility for "mutuals," meaning accounts you follow that also follow you back. Nikita Bier, X's head of product, said the missing mutual-follow signal had made friends show up less in replies and made reply sections feel more hostile.

For creators, the lesson is simple: relationships are distribution.

Not fake engagement pods. Not random follow-backs. Not "reply to 100 people a day" advice from people who do not read what they reply to.

The useful version is narrower. If you want more reach on X in 2026, build a recognizable place in a real niche. Follow people you would actually talk to. Reply where you have something useful to add. Publish posts that your peers, buyers, collaborators, and future fans would want to answer.

Then measure whether that attention is worth having.

Build your X growth system in PostFury if you want one place to collect ideas, draft stronger posts, schedule the best ones, and review which topics brought real replies.

Why this article uses X and Twitter

The platform is X. A lot of people still search for Twitter.

That matters for SEO, but it also matters for how creators talk. Someone may ask "how does the Twitter algorithm work?" while meaning the X For You feed. Someone else may search "X algorithm update 2026" after seeing a post from Nikita Bier.

This article uses both terms where they help the reader. The main target is the current platform: X.

What changed in July 2026

The July update was not presented as a total rebuild of X. It was described as a smaller ranking tweak.

The important part is what it points to.

TechCrunch reported that X was boosting visibility for mutuals: people you follow who also follow you back. Bier said X had noticed this data was missing from the algorithm, which made friends appear less often in replies. He also said the change should help clusters form around interests.

That last phrase matters.

If X wants more interest clusters, creators should stop thinking only in terms of broadcasting to strangers. The feed still recommends posts from outside a user's network, but a creator's network now looks more important to the shape of distribution.

Here is the practical reading:

There is no guarantee that a mutual will see every post. There is no public formula saying "mutual equals reach." But the direction is hard to miss: X is trying to make parts of the platform feel more like connected circles instead of random drive-by attention.

How X says recommendations work

X's own recommendation documentation says the platform uses many signals and that no single signal has a fixed, permanent weight. Those signals include accounts and Topics a user follows, posts they like, repost, share, reply to, media they watch, and content popular in their network.

That is not a hackable checklist. It is a reminder that the feed learns from behavior.

X also says For You can show posts from accounts a user does not follow. The For You timeline includes posts from accounts and topics a user follows, plus recommended content selected through multiple signals. The Following timeline is different: it only shows accounts the user follows.

So the creator question is not "How do I trick the For You feed?"

The better question is:

What behavior would make X understand who should see my posts?

That question changes the work. You stop optimizing every post like a lottery ticket and start creating clearer signals.

The open-source algorithm still teaches the right mental model

X's public algorithm repository is not a full live map of every proprietary ranking change. Treat it as architecture, not a cheat sheet.

Still, it is useful.

The repository describes X's Recommendation Algorithm as services and jobs that serve feeds across surfaces like For You, Search, Explore, and Notifications. It lists components such as user action streams, user signals, graph features, reputation models, topic identification, candidate sources, rankers, and visibility filters.

In plain English: X is not ranking posts in one simple pass.

It is gathering possible posts, scoring them, filtering them, and mixing them into surfaces where a user is likely to engage. That means a creator needs more than a hook. A creator needs a pattern of behavior that helps the system place the account.

Three signals matter from a workflow point of view:

You do not need to understand every model name to use this. You need to publish and interact in a way that makes your account easy to classify and worth recommending.

Mutuals are not an engagement pod

This is the part people will probably ruin.

Any time a platform says a social signal matters, marketers turn it into a tactic. Mutuals will become the new version of comment pods, follow trains, and "drop your handle below" threads.

That is a bad reading.

A mutual is useful because there is already some shared context. You follow them because their posts are useful or relevant. They follow you because your work is relevant to them. When they see your post, there is a higher chance that the reply is informed, specific, and visible to the right circle.

That is very different from random mutual farming.

Bad mutual strategy:

• Follow hundreds of people hoping they follow back.
• Reply with empty praise.
• Join groups where everyone boosts everyone.
• Write broad takes that please a crowd but teach nobody anything.
• Treat your timeline like a distribution spreadsheet.

Good mutual strategy:

• Follow the 50 to 200 people whose work actually shapes your niche.
• Reply when you can clarify, disagree, add an example, or ask a useful question.
• Publish posts that those people would recognize as specific to the niche.
• Turn good replies into follow-up posts.
• Review whether those interactions bring the right followers.

One path creates noise. The other creates a network.

What this means for small creators

Small creators should like this update.

If X gives more attention to mutuals and interest clusters, the advantage is not only follower count. A smaller account with a sharp niche can become visible inside a useful circle faster than a generic account trying to entertain everyone.

The mistake is trying to manufacture closeness.

You cannot fake being part of a niche for long. People can tell when you only show up to harvest attention. They can also tell when you have been doing the work, reading the same discussions, making the same mistakes, and thinking through the same problems.

If you are under 1,000 followers, focus on this:

That is enough for most people. The hard part is doing it without turning into a reply machine.

What this means for established creators

Established creators have a different problem.

They usually have reach, but the audience can drift. A broad account may get impressions from people who like drama, jokes, or easy agreement while the actual target audience becomes a smaller share of the crowd.

The mutuals update is a reason to clean up the network.

Ask:

• Who do I want seeing my replies?
• Who do I want replying to me?
• Which topics create useful peer conversation?
• Which topics create reach but pull the account away from its business?
• Which recurring formats bring the right people back?

If the account has become too broad, do not panic. You do not need to delete your personality. You need stronger recurring signals.

For a founder, that might mean more posts about product decisions, customer conversations, pricing, distribution, hiring, or lessons from shipping. For a consultant, it might mean sharper teardown posts and fewer generic productivity takes. For a creator selling a product, it might mean more proof, more process, and more replies to people who already talk about the problem.

Use PostFury to review your X content patterns if your account gets attention but you are not sure which posts bring the right people.

The reply section matters more than people admit

Most creators treat replies as distribution fuel.

That is too shallow.

Replies are also positioning. They tell new readers who your post attracted, how people interpret your point, what objections come up, and whether the conversation feels worth joining.

If your replies are full of random arguments, the post may have reached people, but it did not necessarily build the right room. If your replies include smart objections, extra examples, buyer questions, or people tagging peers, the post did more than get views.

After the mutuals update, reply quality deserves more attention.

Use this quick review:

Do not just count replies. Read the room you created.

What to post after the mutuals update

The safest answer is also the least useful: "post quality content."

Fine. But what does that mean on X?

After this update, useful posts often do one of five jobs:

If you need prompts, start with what to post on Twitter when you have no ideas. Use the formats, but fill them with something real from your work.

A 7-day posting plan for the new X algorithm

Do not rebuild your entire strategy overnight. Run a one-week test.

Here is a simple plan:

This is not a magic schedule. It is a diagnostic week.

The goal is to see which topics create the right conversation, which accounts keep showing up, and which posts make people curious enough to visit your profile.

If you want a deeper system after the test, read how to grow on X without posting like a bot. That guide is about keeping consistency without sanding off your voice.

How to track whether the update helps you

The worst way to measure this update is to stare at impressions every hour.

Impressions matter, but they are not the whole story. A post can reach a large crowd and still move the wrong people. A smaller post can bring better replies, new mutuals, stronger profile visits, and future collaborations.

For one month, track these:

Pair those numbers with notes. Write down the post topic, format, hook style, and who replied. After four weeks, patterns will start to show.

For a cleaner review process, use the framework in the X analytics metrics that matter for creator growth.

What not to do

The mutuals update will tempt creators into bad shortcuts.

Avoid these:

• Do not follow people only because you want a follow-back.
• Do not reply with empty compliments.
• Do not join engagement groups that make your account look active but teach you nothing.
• Do not pivot your whole account after one high-impression post.
• Do not copy viral formats without understanding why they worked.
• Do not turn every post into a debate prompt.

The update does not remove the need for good writing. It raises the value of being known by the right people.

That is slower than a hack, but it compounds better.

Source-backed notes for creators

The safest way to think about the X algorithm in 2026 is to separate confirmed platform guidance from creator folklore.

What X says:

• X recommendations use many signals, not one fixed public weighting.
• For You includes accounts you follow, topics you follow, and recommended content from outside your network.
• X says recommendations consider user behavior such as likes, replies, reposts, shares, watched media, follows, topics, lists, and network popularity.
• X says replies may be ranked out of chronological order, including factors like whether the original author replied, whether a reply is from someone you follow, and whether the person has X Premium.

What the public repository suggests:

• X recommendations involve candidate sources, rankers, user signals, graph features, reputation models, topic signals, and visibility filters.
• The public code is useful for mental models, but it should not be treated as a complete live ranking recipe.

What the July 2026 reporting added:

• X adjusted the algorithm to boost mutual visibility.
• The stated goal was to make replies feel less dominated by unfamiliar accounts and help interest clusters form.

What creators should infer:

• Build a real network in your niche.
• Publish posts that start specific conversations.
• Track audience quality, not only reach.
• Use systems to stay consistent, but keep judgment in the loop.

The bigger lesson

The X algorithm changes all the time. Creator behavior should not.

You still need a clear topic. You still need posts with a point. You still need replies that sound like a person read the post. You still need analytics that tell you whether the right audience is getting closer.

The mutuals update is useful because it makes an old truth more visible: the best reach often starts inside a real circle before it escapes into the wider feed.

Build the circle first.

Then give it something worth passing around.

Start building that system in PostFury when you want an X workflow that connects ideas, drafts, scheduling, replies, and weekly review instead of treating every post like a one-off gamble.