Best Time to Post on X (Twitter)

Every number in the grid above comes from studies of millions of posts across tens of thousands of accounts. That is what makes it a reasonable place to start and exactly what makes it wrong for you. An average built from everybody describes nobody: it blends a recruiter posting to people at desks with a football account posting to people on a sofa, and the hour it returns is a compromise neither of them should take.

The audience weighting in the tool is an attempt to unblend it, and it is honest about being judgement rather than measurement. Tilting the grid toward a working day or toward the evening gets you closer than the raw average does, but it is still a guess about your readers made by a page that has never seen them.

The specific ways it will be wrong for you are worth knowing. If your followers are concentrated somewhere other than where you live, the default timezone is wrong and the whole grid is off by hours. If you post in a niche with its own rhythm, live sport, market open, school pickup, the general curve has nothing to say about it. And if your account is small, a single post that travelled will dominate your own data so heavily that it looks like a pattern when it was one good Tuesday.

Treat the top three slots as a hypothesis. Post into them for a fortnight, then check them against the next section, and keep whichever survives.

This takes about ten minutes and it beats every benchmark on the internet, including this one. You do not need Premium for it: post-level analytics are available on a free account.

Open X on desktop and go to your profile, then click through to any of your own posts and open View post analytics. The panel that opens has an Account overview link at the top, which takes you to the full analytics view. Set the range to the last 28 days, because seven days is not enough posts to see anything and ninety days reaches back past whatever you changed.

Now sort your posts by impressions and take the top ten. For each one, note the day and the hour it went out. Do the same for the bottom ten. You are looking for a cluster, not a winner: if eight of the top ten went out between 8am and 11am on a weekday and six of the bottom ten went out after 7pm, you have your answer and it took ten minutes.

Two traps. First, impressions are driven far more by what the post said than by when it went out, so a single viral post will swamp the signal. Throw out the top one before you look. Second, if every post you have ever made went out at 9am, your data says 9am is best and it is not evidence, it is your own schedule reflected back. Deliberately post a third of them somewhere else for two weeks before you trust the pattern. While you are in there, the numbers on that screen are also what the engagement rate calculator runs on.

It matters, and the reason has changed. On a chronological timeline the logic was obvious: a post went out, it sat at the top of the feed for a few minutes, and everyone asleep missed it forever. Timing was the whole game. That feed is gone. Your post can surface in somebody's For You tab a day later, so the cost of a bad hour is no longer total.

What replaced it is a test with a short fuse. X shows a new post to a small slice of your followers and watches what happens: replies count heaviest, then reposts, then likes, and a profile visit or a long dwell counts too. If that slice reacts, the post gets pushed to a wider one, and the whole decision is made on early signal. Post when your readers are awake and the test runs on a real sample. Post at 3am and it runs on whoever happens to be up, which is a thin sample that rarely earns the push.

Which puts timing in proportion. It is worth getting roughly right and not worth agonising over: the difference between 9am and 10am is noise, the difference between 9am and 3am is not. Consistency beats precision, because an account that posts on most days trains the system on what its audience responds to, and an account that posts three times in one afternoon and then nothing for nine days never gets a stable read.

The other half of it is the post. A thread posted into a good slot outperforms a single post in the same slot, mostly because every part gets its own chance to be shown, which is what the thread maker is for.