ShapelessAI← Back
Best time to post

Nobody agrees when to post. Here is the disagreement.

What the major 2026 timing studies actually say, per platform, aggregated into one heatmap of where they agree - and where they contradict each other, which is more often than any of them admits.

24 studies read

instagram4 studies plotted, 5 read2 documented conflicts

Best time to post on Instagram

Four studies covering more than 40 million posts name weekday late mornings through evenings, and they disagree by up to sixteen hours on the single best slot.

tiktok4 studies plotted, 5 read2 documented conflicts

Best time to post on TikTok

Weekday afternoons draw the most agreement, but the four studies split hard on the weekend: one calls Saturday the best day of the week, another calls it the quietest.

facebook_page3 studies plotted, 4 read2 documented conflicts

Best time to post on Facebook

Three studies, three different answers on the same weekday: 5am, 9am and midday. Facebook has the least settled advice of any platform here.

linkedin4 studies plotted, 5 read2 documented conflicts

Best time to post on LinkedIn

Every study agrees weekends are worst - and that is the only thing they agree on. Their best hours run from 4am to 5pm, and the one study that measured the size of the effect found it small.

youtube2 studies plotted, 2 read2 documented conflicts

Best time to post on YouTube

YouTube is the one platform where the advice is not to publish at the peak. Both studies say upload one to two hours before it, so the algorithm has time to start testing the video.

x3 studies plotted, 3 read2 documented conflicts

Best time to post on X

Mid-morning wins on X, and the largest study finds it wins on every single day of the week - including both weekend days. This is the platform where the day matters least.

Why this matters less than the rest of the page suggests

Every study on this page is published by a company that sells scheduling software, and most of them tell you plainly that timing is a small lever. We think that is the single most useful thing in the whole category, so we put it here rather than at the bottom.

Only one study we found measured the size of the effect instead of just ranking hours. AuthoredUp, across more than 3 million LinkedIn posts, found the gap between the best and worst day is under 10% for personal profiles, and that weekday against weekend engagement lands within roughly 7%. Company Pages see a larger swing, about 1.7 times between Wednesday and Sunday.

Set against that, the studies here disagree with each other by up to sixteen hours on the same platform. When the disagreement between sources is far larger than the effect any of them measures, precision to the hour is false precision.

The honest version: post in the broad window your audience is awake and reachable, keep the cadence steady, and spend the effort you saved on the post itself. Then replace all of this with your own analytics as soon as you have enough of your own posts to read.

  • “Timing can't save a weak post. Think of it as the cherry on top of your strategy, not the cake itself.”
    Buffer
  • “Posting at 4 p.m. on a Wednesday won't rescue a post that isn't relevant or valuable to your audience.”
    Buffer
  • “The difference between the best and worst format dwarfs the difference between the best and worst posting hour.”
    AuthoredUp
  • “Being consistent and creating engaging content is important too - if your posts don't resonate, your posting time won't make much of a difference.”
    Later

Questions

What is the best time to post on social media?
There is no single answer, and the studies that claim one disagree with each other by up to sixteen hours on the same platform. Weekday late mornings through early evenings draw the most agreement across platforms, LinkedIn is the most consistently weekday-bound, and X shows almost no day-of-week effect at all. Each platform page here shows the overlap between studies rather than picking a winner.
Why do best-time-to-post studies disagree so much?
Because they measure different populations with undisclosed methods. Sprout measures profiles run by its own customers, Buffer and Later measure posts scheduled through their tools, and Hootsuite normalises across 118 countries. None of them publishes how that normalisation works, so the gaps cannot be reconciled from outside. On Facebook, one publisher's two pages disagree by four hours on the same dataset.
Are these times in my timezone?
Yes. Most of these studies now publish windows already normalised to the reader's local time, so we plot them as published rather than converting them - converting an already-local window is the usual error in this category. Later's TikTok windows are published in PST, so those we convert to your clock with daylight saving resolved for today.
How much does posting time actually change engagement?
Less than this category implies. The only study we found that measured the effect rather than ranking hours is AuthoredUp, across 3 million LinkedIn posts: the gap between the best and worst day is under 10% for personal profiles, and weekday against weekend lands within roughly 7%. The disagreement between studies is larger than the effect any of them measures.
What should I do instead of following a generic schedule?
Use these windows as a starting point while you have no data of your own, keep a steady cadence, and switch to your own analytics as soon as you have enough posts to read. Your own audience beats every industry average on this page.

How we built this

We read each study's own page, transcribed its day-by-day windows, and recorded the clock it publishes on. The heatmap shades each hour by how many independent studies name it. It is a count of agreement, not an engagement score: we never blend numbers from studies that measured different things in different ways.

A study only reaches the map if we can put its hours on a clock. Metricool measures the largest samples in this category and states no timezone on any page we read, so its hours are quoted in the text and left off the grid. HubSpot surveys marketers about what they believe rather than measuring posts, so it is left off too. Both are listed with the rest, with the reason attached.

Most publishers now normalise their windows to the reader's local time, so we plot those as published rather than converting them - converting an already-local window is how these pages usually go wrong. Later's TikTok windows are the exception, published in PST, so those we convert to your clock with daylight saving resolved for today.

Who we read, on 2026-08-25

8 publishers, 24 pages. Each platform page lists the studies behind its own map, with the sample size each one states.

  1. Sprout SocialMedian engagement per hour across the profiles of 30,000+ Sprout customers.Reports in local time. Sprout's 2026 refresh dropped the old US Central Time hub: "schedule your posts for the hours listed below in their respective local time zones".5 pages read
  2. BufferThe time slots with the highest median engagement.Normalised to the reader's local time, so no conversion is needed - Buffer says its data team makes the recommended timezones "universally applicable", without publishing the method.6 pages read
  3. HootsuitePosts analysed with the data-science agency Critical Truth, localised across 118 countries.Normalised across 118 countries and stated as time-zone agnostic. The normalisation method is not published.5 pages read
  4. LaterMedian engagement per hour across Later's scheduled posts.Normalised to the audience's local time zone.2 pages read
  5. Metricool · not plottedPosts published through Metricool, measured for views and interactions.States no timezone on any page we read, so its hours cannot be placed on a clock.3 pages read
  6. AuthoredUpEngagement by day and hour, and the only study here that publishes the size of the timing effect itself.Local time.1 page read
  7. SocialPilotViews and engagement on videos scheduled through SocialPilot.Local time.1 page read
  8. HubSpot · not plottedA survey of marketers, not measured engagement.Unspecified except for TikTok, where it references EST.Left off every heatmap: It asks marketers what they believe rather than measuring posts, states no sample size, and is the oldest page we read.1 page read