Your best time to post is already in your numbers

Search for the best time to post on Instagram and you will get a heatmap within four seconds. It will look authoritative, it will be built on millions of posts, and the next one you open will tell you something different. Not slightly different. One of the big studies puts the best hour at five in the morning and another puts it at nine at night.

Both of those numbers are real, both were measured carefully, and neither of them is about your account. That is the whole subject in one sentence, and the rest of this is what to do about it.

Four studies, four different Wednesdays

Start with the disagreement, because it is the fastest way to stop trusting any single chart. Take one day, Wednesday, and ask four of the largest published analyses what hour to post at.

None of those publishers is being careless. Every one of them ran a larger sample than you will ever have. Look at what each was actually measuring and the disagreement starts to make sense.

Three things drive that spread. The first is who is in the sample. Sprout Social measured social profiles belonging to its own customers, most of them businesses. Later measured posts scheduled through Later. Hootsuite measured business accounts across 118 countries. Every one of those is a different population from a lifestyle creator in Mumbai with eighteen thousand followers.

The second is the window. Later's data runs to October 2024. Sprout Social's runs to February 2026. Instagram changed what it shows people between those two dates, more than once. A study is a photograph of a period, and the periods do not overlap.

The third is the definition. Some of these rank by engagement rate, some by reach, some by raw engagement volume. An hour that wins on total engagement can lose on engagement rate, because the hour with the most people in it also has the most posts competing.

What your own posting history can actually prove

Your account has something no study has: the same person, the same followers, the same subject, measured across months. That controls for almost everything the published charts cannot control for.

Take the last twenty or thirty posts, write down the hour each went out and the reach each got, and group them. You are looking for one thing only, which is whether the hours cluster. If your 7pm posts sit in a band and your 10am posts sit in a clearly lower one, and there are at least five or six posts in each, that is a signal worth acting on.

Two rules stop this from producing nonsense. Compare reels with reels and carousels with carousels, because format moves reach far more than hour does. And throw out your best post and your worst before you average anything: one runaway post can drag an hour's average up on its own, and the reason it ran away almost certainly was not the clock.

proof: in Instagram Insights, open the last thirty posts, sort by reach, and write the posting hour beside each one. If no hour appears three times you do not have a sample yet, you have a list

You have no evidence at all for the hours you never posted at

This is the part nobody writing about this says out loud, and it undoes half of the standard advice. Your data can only ever rank the hours you have actually used. If you have posted at 7pm for a year because a chart told you to in your first month, your numbers will confirm that 7pm works, and they will be telling you nothing about 6am, because you have never been there.

So the loop closes on itself. You pick an hour from a general chart, you post at it, your own data agrees with it, and you conclude your data proved the chart right. What your data proved is that you were consistent.

This is why the answer is a test rather than an audit. An audit of your history can only find the best of what you have already done. Finding out whether something better exists means posting somewhere you would not normally post, on purpose, and being willing to lose a little reach while you find out.

A six week test you can actually run

This is the smallest version that produces something you can believe. It assumes you post around three times a week. If you post less, extend it rather than shrinking the sample.

1. 01Weeks one and two: record what you already do. Change nothing. Write down the hour and the format of every post, and the reach and saves after seven full days. You are establishing what normal looks like, which you cannot do retrospectively because you were not writing it down.
2. 02Pick one challenger hour, for a reason you can say out loud. One, not three. Take it from your follower activity graph, or from the published study whose sample looks most like your audience, or from an hour you have always suspected. Write down the reason before you start so you cannot rewrite it afterwards.
3. 03Weeks three to six: alternate, and keep the format fixed. Post one at your usual hour, the next at the challenger hour, back and forth. Every one of these should be the same format. Twelve alternating reels is a test. Six reels and six carousels is two overlapping tests that will not separate.
4. 04Give every post the same measuring window. Seven days, then read it. Comparing a post from yesterday to one from last month is comparing a post that is still growing to one that has finished, and it will make whatever you published most recently look worse than it is.
5. 05Compare the middles, not the totals. Line up the six challenger posts and the six usual ones and take the middle value of each group. The median ignores the one that ran away, and the one that ran away is the thing most likely to fool you.
6. 06Read the result as a lean. Six posts against six is enough to notice a large difference and nowhere near enough to prove a small one. If the gap is under about 15%, treat the hours as equal and go with whichever you can hit reliably. If it is large and consistent, move your default hour and start the next test somewhere else.

The last line matters more than it looks. Reliability beats optimality here. An hour you can hit every week without rearranging your evening will beat a marginally better hour you miss twice a month, because a missed post scores zero and no timing advantage recovers from that.

When your followers are online is a different question

Instagram will show you a graph of when your followers are most active. It is the most commonly misread number in the app, because it looks exactly like an answer to this question and is answering a different one.

That graph counts people who have the app open. It does not know how much competition there is for their attention at that moment. The hour when the most of your followers are scrolling is also the hour when the most of everybody else's posts were published, which is why some studies keep finding value in hours when almost nobody is awake.

proof: Insights, Total followers, Most active times: note the hour the graph peaks. That is a candidate for the test above, and on its own it is not a result

It is still useful, in two specific ways. It tells you which time zone your audience actually lives in, which is worth knowing if you assumed and were wrong. And it gives you a sensible challenger hour for the test above, because posting shortly before a peak gives a post time to gather its first signals before the crowd arrives.

What it will not do is rank hours by how well posts perform in them, because it does not contain a single number about your posts.

Timing is a small lever, and it is worth saying so

Having argued for six weeks of testing, here is the honest scale of the prize. Posting hour is one of the smaller things you control, and it is a long way behind what the post is and how often you post at all.

Buffer's separate 2025 study, over 2 million posts from more than 100,000 accounts, found that accounts posting three to five times a week got around 12% more reach per post than accounts posting once or twice, and more than double the follower growth rate. That is a measured effect from a change in frequency. The timing studies cannot even agree on which hour to compare against which.

None of which makes the hour worthless. It is one of the few things in this job you can settle once with a real test and then stop thinking about, and a question you have answered properly is a question that stops costing you attention every Sunday.

Questions people ask about this