The best time to post is the slot where your own posts have repeatedly outperformed your own median, and you find it by bucketing your last 30 to 60 posts by weekday and time of day. Generic charts cannot tell you this, because they average audiences that are not yours. Expect a modest effect: timing usually decides how fast a good post starts, not whether a weak post succeeds.
Key takeaways
- Generic “best time” charts blend niches, countries and formats. Your audience is one specific slice.
- Work per platform and per format. Never mix Reels with carousels or Shorts with long-form in one analysis.
- Normalize every post against your recent median before comparing time slots, or account growth will masquerade as a timing effect.
- Small samples lie. Use broad dayparts until each bucket holds at least five posts.
- History gives you candidates. Only a controlled 4-week test gives you an answer.
- If the test shows no clear difference, post when it is convenient and spend the effort on the content.
Why generic charts fail
A chart that says “Wednesday at 11 a.m.” is an average across thousands of accounts. Three problems follow.
The audience is not yours. A gaming channel with a teenage audience and a B2B Page read by office workers have opposite daily rhythms. Averaged together, they produce a time that suits neither.
The time zone is ambiguous. Many charts do not say whose 11 a.m. they mean. If your viewers are spread across two continents, a single clock time means little.
Cause and effect are tangled. If most professional accounts schedule posts for weekday mornings, weekday mornings will collect the most engagement in aggregate. That shows when people post, not when posting works best. A crowded slot can even be the worst one for you.
The platforms’ own audience-activity reports are a better starting point. YouTube Studio shows when your viewers are on YouTube, Instagram Insights shows when your followers are most active, and TikTok’s analytics show follower activity by hour. Use them to shortlist slots. But they describe when followers are online, not when your posts perform, and much of your reach on short-form video comes from non-followers. So verify with your results.
Step 1: Collect 30 to 60 posts
Choose one platform and one format. Pull your most recent 30 to 60 posts in that format into a spreadsheet. Going back further than about six months adds posts from a different era of your account, and often of the platform.
For each post, record:
- Publish date and time
- Weekday
- An early-performance metric (see step 2)
- A short note on anything unusual: a collaboration, a trending topic, a paid boost, a holiday
YouTube Studio, Meta Business Suite and TikTok’s analytics let you export or view post-level data with publish times. Check which time zone the export uses before you do anything else.
Remove boosted posts and obvious special cases such as collaborations with much larger accounts. They tell you about the boost, not the hour.
Step 2: Normalize by early performance
Timing can only influence the start of a post’s life. Three weeks later, a video’s total is driven by how good it was. So measure early.
The ideal metric is views in the first 24 or 48 hours. YouTube Studio lets you look at a video’s first days directly. On other platforms, start a habit of noting each post’s views at the 24-hour mark. For past posts where you did not capture it, use the total, but only for posts that are all at least two weeks old, so that every post has had the same chance to finish accumulating.
Then turn each number into an index:
Index = post’s early views ÷ median early views of the previous 10 posts in the same format
An index of 1.0 is a typical post. 1.3 is 30% above your recent normal. Using the ten posts before it rather than the overall median removes growth: if your account doubled in the period, later posts would otherwise look better whatever time they went out.
Say a Reel collected 6,200 views in its first 24 hours, and the median for the ten Reels before it was 5,000. Its index is 6,200 ÷ 5,000 = 1.24. (These and all later figures are hypothetical.)
Step 3: Bucket by weekday and time of day
Now group the posts. The temptation is a 7 × 24 heatmap. With 40 posts spread over 168 cells, almost every cell is empty or holds a single post, and a single post tells you about that post.
Use buckets coarse enough to fill:
- Dayparts: morning (6-11), midday (11-17), evening (17-22), late (22-6). Adjust the boundaries to your audience’s day.
- Day type: weekday versus weekend, analyzed separately from daypart.
For each bucket, calculate the number of posts, the median index, and the lowest and highest index.
Minimum sample sizes
These are rules of thumb, not statistical guarantees:
| Posts in bucket | How to treat it |
|---|---|
| 1-2 | Ignore. This is anecdote. |
| 3-4 | A hint. Worth including in a test, not worth acting on. |
| 5-9 | A candidate, if the median is clearly different and the range does not overlap heavily with other buckets. |
| 10 or more | Reasonably trustworthy for a broad daypart. Still confirm with a test before reorganizing your schedule. |
A worked example
Say you bucket 40 Reels and get this:
| Daypart | Posts | Median index | Range |
|---|---|---|---|
| Morning | 9 | 0.92 | 0.61-1.40 |
| Midday | 6 | 0.88 | 0.55-1.22 |
| Evening | 19 | 1.04 | 0.58-1.95 |
| Late | 6 | 1.31 | 0.72-2.90 |
The quick reading is “post late.” The careful reading is different. Late has only six posts and a huge range; one post at 2.90 is doing much of the work. Evening is your habit, with 19 posts and a median close to 1.0, as you would expect from the slot that defines your normal. Morning and midday are slightly below, with ranges that overlap everything.
The honest conclusion: late is a promising candidate, evening is the safe default, and nothing here is proven. That is exactly what a test is for.
Historical data cannot separate “this time is better” from “I happened to post better content at this time.” It produces hypotheses. A test produces answers.
Step 4: Run a clean 4-week test
Pick two slots: your current default (A) and the best candidate (B). Then control everything else you can.
- Fix the format and length. Same type of post throughout.
- Fix the frequency. If you normally post three times a week, keep it at three. Changing volume and timing together makes the result unreadable. The posting frequency planner helps you set a cadence you can hold.
- Assign slots before you make the content. If you decide the slot after seeing the finished video, you will unconsciously give the stronger pieces to the slot you believe in.
- Rotate across weekdays. Week 1: Monday A, Wednesday B, Friday A. Week 2: Monday B, Wednesday A, Friday B. Repeat for weeks 3 and 4. Each slot gets every weekday twice and six posts in total.
- Measure at a fixed age. Record views at 24 hours and again at 7 days for every post.
- Avoid distorted weeks. Do not run the test across major holidays or during a launch or collaboration.
Reading the result
Compare the median index of slot A and slot B. Then compare the matched pairs: the Monday A post against the Monday B post, and so on.
Call it a win for B only if both are true: B’s median is higher by more than the typical gap between two ordinary posts of yours, and B wins most of the pairs. If B’s median is higher only because of one runaway post, it is a draw.
Look at the 7-day numbers as well. If B leads at 24 hours but the two are level after a week, timing changed the speed of delivery and nothing else. That still matters for time-sensitive content, and hardly at all for evergreen content.
A draw is a useful result. It tells you timing is not your constraint, and you can stop thinking about it.
Time zones
Schedule for your audience’s clock, not yours. Check the audience location report on each platform first.
- One dominant region: convert everything in your sheet to that zone and analyze in it.
- Two major regions: look for the overlap, for example an evening in one region that is midday in the other, and test it against a slot tuned to the larger region alone.
- Daylight saving changes: regions switch on different dates, and some do not switch at all. Around those dates a fixed slot in your scheduler shifts by an hour for part of your audience.
- Tool settings: exports, in-app reports and scheduling tools may each use a different time zone setting. Confirm all three before trusting a comparison.
Platform nuances
YouTube long-form
Timing matters most for videos that rely on subscribers and the home feed in the first hours: news, commentary, series episodes. For videos found through search and suggested videos over months, the upload hour is close to irrelevant compared with click-through rate and retention. Many creators publish a little ahead of their audience’s peak so the video is fully processed and available when viewers arrive. For premieres and live streams, timing is decisive, because the audience must be present at that moment. Which numbers to watch is covered in YouTube analytics metrics that matter.
YouTube Shorts
Shorts are distributed mainly through the Shorts feed and can pick up views days or weeks after upload. Timing effects tend to be weak. Analyze Shorts separately from long-form; their viewers and their clocks are not the same.
Stories disappear after 24 hours, so they are the most time-sensitive format. Feed posts and carousels lean on followers and benefit from being posted when followers are active. Reels are recommended to non-followers over a longer period, which dilutes the effect of the hour. Run a separate analysis for each. The meaning of each number is in Instagram Insights explained.
TikTok
Most views typically come from the For You feed rather than from followers, and a video can take off well after posting. Follower activity is a reasonable first guess for a slot, but check your traffic sources: the larger the For You share, the less the hour is likely to matter. See TikTok analytics explained for how to read that report.
Facebook Pages
Page posts depend heavily on followers and on early reactions, comments and shares. Timely posts, such as event reminders, news and offers, gain the most from a well-chosen slot. Analyze video and non-video posts separately.
When timing does not matter
Be realistic about the size of the prize. Timing is a second-order lever in these situations:
- Your results vary wildly between posts. If your best posts get ten times the views of your worst, the hook and the topic are doing that, not the clock. Fix openings first; the hook generator can help with alternatives.
- Your content is evergreen and found by search. It will collect views for months whatever hour it went live.
- Your audience is small. With very low numbers, random variation between posts is larger than any timing effect you could detect.
- Your schedule is irregular. A slot you can hit every week beats a theoretically better slot you hit half the time.
If timing sits low on your list of problems, a broader review such as the 8-pillar audit framework will show what sits higher.
Put it into practice
- Pick one platform and one format.
- Export your last 30 to 60 posts with publish times, and convert them to your audience’s main time zone.
- Remove boosted posts and one-off special cases.
- Calculate each post’s index against the median of the ten posts before it.
- Bucket by daypart, and separately by weekday versus weekend. Note post count, median index and range.
- Discard buckets with fewer than five posts as conclusions; keep them as candidates.
- Choose your default slot and one challenger. Run the alternating 4-week test.
- Compare medians and matched pairs at 24 hours and 7 days.
- Adopt the winner, or, after a draw, the more convenient slot. Repeat in three to six months or after a clear change in your audience.
CreatorsMetrics runs a best-time-to-post analysis on your own post history from connected YouTube, Instagram, Facebook and TikTok accounts. Sign-in is read-only, and a content calendar is included for planning your schedule.
Frequently asked questions
What is the best time to post on social media?
There is no universal best time. It depends on when your own audience is active and on the format, so the reliable answer comes from bucketing your past posts by time slot and testing the top candidates.
How many posts do I need to find my best time to post?
Use your last 30 to 60 posts in one format, and as a rule of thumb do not trust any time slot with fewer than five posts in it. With a small sample, use broad dayparts instead of single hours.
Does posting time matter on YouTube?
It matters most for content that depends on subscribers in the first hours, such as news, live streams and premieres. For searchable evergreen videos, the title, thumbnail and retention matter far more than the upload hour.
Which time zone should I use when scheduling posts?
Use the time zone where most of your audience lives, not your own. Convert every publish time in your spreadsheet to that single zone before you analyze anything.
How long should I test a new posting time?
Run it for about four weeks with at least six posts in each slot, alternating slots across weekdays. Shorter tests mostly measure the difference between individual posts, not between times.