Best time to post on Instagram, and why the studies are twelve hours apart

If you want the short answer that the evidence actually supports: post on a weekday afternoon, between about 12pm and 6pm, and put your best work on Tuesday or Wednesday. Three of the four studies below agree on that much. The fourth says to post at five in the morning, and the reason it says that is more useful than the number itself.

What the four studies actually found

Every figure below was copied from the study's own page on 8 September 2026. The overlap column is our reading of where they agree, not anybody's published finding.

Read across the Wednesday row. Sprout says the window runs from midday to nine at night. Hootsuite says five in the afternoon. RecurPost says eight to nine in the evening. Later says three in the morning. Those first three overlap enough to act on. The fourth is in a different half of the day.

Twelve hours apart on the same platform

Later's headline is that the best time to post on Instagram is 5am, and it is not a typo or a rogue outlier. Its per-day figures run 5am, 5am, 3am, 4 to 5am, 3 to 6am, 5am, and midnight. Every single day of the week lands before breakfast. Sprout Social, looking at nearly two billion engagements, does not put a single peak window before midday on any day.

Two studies of the same platform cannot both be right about that, so it is worth understanding what makes them differ, because the answer changes how you should read all four.

• Later normalises to the audience's local time zone. Its own framing is that 5am means 5am where your followers are, not where you are. Once you convert a global dataset into each viewer's local clock, a post published overnight in one region is being counted as an early morning post rather than a late night one, and the peak moves accordingly.
• Later's window is January to October 2024. As of today that is roughly two years old, presented on a page that reads as current guidance. Sprout's window is 27 November 2025 to 27 February 2026. Instagram's ranking has been rebuilt more than once in between.
• Later counts posts, Sprout counts engagements. A study weighted by engagement is dominated by the accounts that get engagement, which are larger and more professional. A study weighted by posts gives equal say to every account that published, most of which are small.
• Sprout's population is its own customers, which skews to brands and agencies with a marketing function. Hootsuite localises across 118 countries with a research partner. These are different populations of Instagram, not different measurements of one.

None of that makes Later wrong. It makes Later an answer to a different question: given a global dataset normalised to local time and weighted by volume, when did posts perform best two years ago. If you are a single business posting to followers in one city, the question you are asking is closer to the one Sprout and Hootsuite answer. The same split explains most of the disagreement in the four Facebook studies and in the TikTok ones, where the gap on a single day ran to seven hours.

The study that stopped publishing

Buffer is not in the table because it no longer answers the question. Its Instagram page opens by observing that if you search this term you will find countless articles quoting studies of a few million posts, and then declines to be one of them. Its position is that the best time to post is when your own followers are online, and what it offers is a feature that works that out for your account.

The company whose 14 million post dataset anchors half the articles in this category has decided that the general answer is not worth publishing.

That is a commercial move and it is also the most defensible position on this page. Every number in our table is a median of an enormous, wildly heterogeneous pool: a nail bar in Bratislava and a software company in Seattle contribute to the same figure. The median of those two audiences describes neither. We made the same point at more length in the Facebook piece, and Instagram makes it sharper, because Instagram's audience skews younger and its usage peaks are correspondingly less like the working day.

What Reels change

Only one of the four separates formats, and the separation is large enough that applying a feed number to a Reel is a real mistake. Later reports analysing 975,000 Reels specifically, and puts their best slot at midnight on Mondays, against 5am on Tuesdays for carousels. Its Stories recommendation is different again, early morning between 6 and 8am and late evening between 7 and 9pm.

Read that with the time zone caveat above and the useful part is not the hour, it is the direction: Reels peak later than static posts, and Stories split into two windows because they are consumed in short bursts at the ends of the day. If most of what you publish is Reels, and for most accounts it now is, none of the feed post figures in the table were measured on your format. That matters more than choosing between 2pm and 5pm, and if Reels are your main output then what actually goes into them is a better use of an afternoon than the schedule.

Why four serious studies still cannot agree

It would be easy to read the table as three good studies and one strange one. It is more accurate to read it as four honest answers to four slightly different questions, and the differences are structural rather than sloppy.

• Different populations. Sprout measures its customers, Hootsuite measures across 118 countries with a research partner, Later measures a global pool weighted by post count. Instagram is not one audience and none of these samples it the same way.
• Different units. Engagements, posts, and scheduled posts are three different denominators. A ranking built on one does not convert to another.
• Different clocks. Local time zone normalisation moves peaks by hours. Two studies that both say 5pm may not mean the same 5pm.
• Different years. Later's data ends in October 2024 and Sprout's ends in February 2026. On a platform that has changed its ranking in between, that is not a detail.
• Everything is a median. Every figure pools industries that behave nothing like each other, which is the objection Buffer has now built a product around.

The honest conclusion is that the overlap column in the table is worth more than any individual cell in it. Where three independent studies with different methods land in the same few hours, that is a signal. Where one of them stands twelve hours away, that is a methodology difference wearing the clothes of a finding.

What to actually do on Monday morning

Start with a weekday afternoon, somewhere between midday and six, and favour Tuesday and Wednesday. That is the only recommendation on this page supported by more than one dataset, and it costs nothing to adopt.

Then stop reading articles like this one and read your own account, which is the part Buffer is right about. Instagram Insights tells you when your followers are active, and for an account with a real audience in one place that number beats every median above. Fifty of your own followers is a better sample for your business than two billion engagements belonging to somebody else.

And keep the timing question in proportion. Across every study in this cluster, the gap between the best hour and an average one is small next to the gap between posting consistently and not. An account that publishes three good posts a week at the wrong time will beat one that publishes once a month at the perfect time, which is why how often to post is the more consequential question and what to post about is more consequential still. If the honest problem is that nothing gets published at all, the schedule is not the bottleneck, and Instagram post ideas for small business or story ideas will move the numbers further than any of the hours above.

One last practical note. Whatever time you choose, the post has to be the right shape when it arrives, and Instagram has changed what that means twice in two years. The current figures are in Instagram post size and Instagram story size, and getting the ratio wrong costs more attention than getting the hour wrong. Timing is one input among many, and what Meta documents its models actually predicting puts the hour in perspective.