How the Threads algorithm works
You cannot game it, and you do not need to. Almost everything the algorithm rewards is something a good post does anyway. The useful part is knowing which signals it counts first.
The short version
- Threads tests every post on a small sample first. What that sample does in the first hour decides whether anyone else sees it.
- Replies are the strongest signal. A reply costs a reader more than a like, and the ranking treats it that way.
- Followers are not distribution. The feed is recommendation-first, which is why a 200-follower account can get 40,000 views.
- Dwell time is doing more than you think. A post people stop and read beats a post people tap and leave.
- Consistency beats cleverness. The account posting five ordinary things a week outperforms the one posting a brilliant thing a month.
How does the Threads algorithm decide what to show?
It shows every post to a small sample of people first (some followers, some strangers with matching interests) and watches what they do. Strong early engagement buys a larger sample, which buys a larger one again. Weak early engagement ends the post's life in about an hour, no matter how good it is.
The mental model that helps most is an audition, not a broadcast. You publish, and a few hundred people are shown it. If enough of them reply, stay on it, or open your profile, the post gets a second, bigger audition. Posts that go far have simply passed the same test six or seven times in a row.
Two consequences follow from that, and they are the whole reason this article is worth reading:
- The first hour is nearly all of it. A post that does nothing for ninety minutes almost never recovers. Whatever you were going to do to help it (replying, sharing it, quoting it) has to happen inside that window.
- Follower count is not distribution. Threads is recommendation-first, so it will happily show a stranger's post to someone who follows nobody. This is why new accounts still get reach, and why old accounts still get ignored.
The practical version: treat publishing as the start of the work, not the end of it. Twenty minutes of replying under your own post and under other people's, immediately after you publish, does more than an hour of rewriting the post beforehand.
What signals does it actually count?
In rough order of weight: replies, dwell time, reposts and quotes, profile clicks, then likes. The ranking cares most about actions that cost the reader something (attention or effort) and least about the ones that cost nothing.
Nobody outside Meta has the weights. But the ordering below is consistent with what the platform has said publicly, with how the feed behaves, and with what anyone who posts daily can observe across a few hundred posts.
| Signal | What it costs the reader | How much it seems to matter |
|---|---|---|
| Replies | Thinking of something to say | The most reliable predictor of reach |
| Dwell time | Seconds of attention | Large, and invisible in your stats |
| Reposts and quotes | Their own audience's trust | Large, and rarer than you would like |
| Profile clicks | Curiosity | Matters for followers more than for reach |
| Likes | One tap | Real, but the weakest of the five |
| Follows from the post | A standing commitment | Strong, and hard to engineer |
This ordering is why the standard advice (end a post with a question) works, and also why it stops working when everyone does it. The question is not magic. It is a cheap way to lower the cost of replying, and readers have learned to recognise a question that only exists to farm a comment.
Dwell time is the one people ignore because it is not in the stats. A post that takes fifteen seconds to read and is worth fifteen seconds beats a one-liner that gets more likes. It is the strongest argument for writing posts with three or four short paragraphs instead of a single sentence.
Why do replies outrank everything else?
Because a reply produces a conversation, and a conversation keeps other people on the app. Threads is optimising for time spent, not for your growth: a post that starts twenty threads of discussion is worth far more to it than a post that collects two hundred silent likes.
This works in both directions, which is what most people miss. Replies on your post lift the post. Replies from you, under other people's posts, put your name in front of an audience someone else spent two years building, and a good reply under a large account routinely outperforms your own posts for weeks.
The mechanics of that are a separate subject and worth its own read: the Threads reply strategy covers which accounts to reply under, the window that matters, and what a reply has to actually contain to earn a profile click.
Reply to your own repliers. Every reply you write on your own post is another signal on that post, and it usually pulls the original commenter back for a second round. Ten replies from you in the first hour is not vanity; it is the cheapest reach on the platform.
What quietly limits an account?
Nothing dramatic. There is no secret ban list. What limits most accounts is a set of ordinary habits: posting links with no context, deleting posts that underperform, engagement-bait phrasing, and long silences between bursts of activity.
Some of these are ranking effects and some are reader effects, and it does not much matter which is which: they end at the same place.
| Habit | Why it hurts |
|---|---|
| A bare link with no context | Sends readers off the app, and gives the ranking nothing to reward |
| Deleting posts that flopped | You lose the slow tail; posts do occasionally wake up on day three |
| "Comment YES if you agree" | Farms a signal readers have learned to distrust, and looks like every bot |
| Posting nine times in an hour, then nothing for a week | The account has no baseline, so every post is judged cold |
| Hashtag stuffing | Threads allows one topic tag; stuffing reads as spam and adds nothing |
| Reposting your own post to force reach | Rarely works, and burns the followers who already saw it |
The link one is worth expanding, because it is the trade every faceless account has to make. A post with a link in it will usually reach fewer people than the same post without one. That is not a punishment; it is the ranking noticing that the post sends people away. The workaround is not a trick. It is to write the post so it stands alone, and put the link in the first reply, where the people who want it will find it.
How often should you post?
Two to four times a day, every weekday, is the range where most accounts see compounding. One a day works but grows slowly. Ten a day dilutes the sample each post gets tested on, and burns you out inside a fortnight.
The reason a daily baseline matters is that the algorithm has more to work with on an account it sees often. There is a second reason that has nothing to do with ranking: posting daily is the only way to find out which of your ideas people actually want, and you cannot learn that at four posts a month.
| Posts per day | What usually happens |
|---|---|
| 0–1 | Slow, fragile growth. Every post has to be excellent |
| 2–4 | The sweet spot. Enough shots to find the ones that work |
| 5–8 | Fine if you have the material. Watch for repetition |
| 9+ | Diminishing returns, and a visible drop in quality |
Sustaining two to four a day is not an inspiration problem, it is a supply problem, and the fix is to write them in batches rather than in the moment. How to build a Threads post bank is the specific method: twelve post types, where the raw material comes from, and a week written in one sitting.
What should you do when a good post flops?
Post it again in a fortnight, worded differently. A flop usually means the first sample was wrong, not that the idea was, and the sample is close to random. Repeating your best ideas with new framing is the single most underused tactic on Threads.
Nobody remembers your post from two weeks ago. Roughly 90% of your followers never saw it, and the strangers the algorithm tests it on next time are a different group entirely. Treating each idea as one-and-done throws away most of the work.
- Wait ten to fourteen days. Long enough that the overlap in who sees it is small.
- Change the opening line, not the idea. The hook is what was tested and rejected, not the substance.
- Change the format. A list that flopped can work as a story; a story that flopped can work as a single blunt claim.
- Post it at a different time of day. You are sampling a different set of people, which is the entire point.
- Keep a note of what worked. Three months in, your own archive is a better guide to your audience than any general advice about the algorithm.
The uncomfortable truth about virality: the difference between your post at 400 views and your post at 90,000 views is often the first fifteen people it was shown to. You cannot control that. You can control how many times you take the shot.
Frequently asked questions
Does the Threads algorithm punish links?
Not as a penalty, but posts containing links do tend to reach fewer people, because the ranking rewards behaviour that keeps readers on the app.
Do hashtags help on Threads?
Barely. Threads uses a single topic tag rather than a stack of hashtags, and stuffing several into a post reads as spam to both the ranking and the reader.
Is there a best time to post on Threads?
Less than people think. Posting when your audience is awake helps, but the recommendation feed spreads a post over hours, so timing matters far less than it does on a chronological feed.
Why did my follower count stall while views kept growing?
Because reach and followers are different products. Views come from the post; follows come from the profile the post sent people to.
Does editing a post reset its reach?
It does not appear to reset anything, but the practical advice is to fix a typo quickly or leave it. Anything that changes the post substantially after it has been tested is judged against an audience that already moved on.
The system behind this, written down
Everything above is the map. The Income Loop is the work inside it: modules 0–6 from the problem you solve to the offer that pays for it, plus ten traffic paths: the deeper post banks, the content sales systems and the full software build sequence, in one place.
Get The Income LoopNot ready to pay for anything? The Basic Income Loop is free and includes a complete seven-day starter for Threads, Instagram or software, enough to find out which one suits you before spending anything.