Blog

Best time to post on YouTube, from 2,554 long-form videos

Kevin Wang
Kevin Wang
Founder, AnyAPI · November 16, 2026
youtubedata-studysocial-mediascheduling
Best time to post on YouTube, from 2,554 long-form videos
Key findings
  • Try a weekend upload. Weekend videos got about 11% more views than weekday videos after accounting for each channel's usual views and video age.
  • Saturday did best; Wednesday and Thursday did worst. Those midweek days together were about 9% behind the rest.
  • You don't need to pick a precise daytime hour. Daytime results were close to average; late-night uploads looked better, but the sample was smaller.
  • Likely scheduled uploads did as well as the rest. Publishing at the start of an hour showed no clear advantage or penalty.

Built on AnyAPI

497 scraping and data APIs behind one key.

Every endpoint in this post is one call away. Pay per request in USD from a prepaid wallet, with automatic failover when a provider breaks. No subscription, no card to start.

Get ~160 free requests$0.10 of credit on signup.

You've finished your YouTube video, but the guides disagree on when to post it. Buffer recommends Sunday morning, while SocialPilot recommends weekday afternoons.

I compared 2,554 videos from 255 channels by the day and hour they were posted, separately from the Shorts study. Each video's views were adjusted for age and compared with its channel's usual views, so a channel averaging ten million views counts the same as one averaging five thousand.

Mon 12a-4a ET: 1.23× (n=25)Mon 4a-8a ET: 1.20× (n=70)Mon 8a-12p ET: 1.04× (n=128)Mon 12p-4p ET: 1.01× (n=79)Mon 4p-8p ET: 0.85× (n=55)Mon 8p-12a ET: 1.20× (n=22)Tue 12a-4a ET: 1.15× (n=32)Tue 4a-8a ET: 0.94× (n=65)Tue 8a-12p ET: 0.99× (n=118)Tue 12p-4p ET: 0.93× (n=76)Tue 4p-8p ET: 1.02× (n=56)Tue 8p-12a ET: 1.26× (n=18)Wed 12a-4a ET: 1.02× (n=27)Wed 4a-8a ET: 0.84× (n=73)Wed 8a-12p ET: 1.02× (n=126)Wed 12p-4p ET: 0.98× (n=87)Wed 4p-8p ET: 0.82× (n=65)Wed 8p-12a ET: 1.04× (n=26)Thu 12a-4a ET: 1.02× (n=34)Thu 4a-8a ET: 0.97× (n=59)Thu 8a-12p ET: 0.84× (n=132)Thu 12p-4p ET: 0.92× (n=99)Thu 4p-8p ET: 1.03× (n=65)Thu 8p-12a ET: 0.96× (n=24)Fri 12a-4a ET: 1.13× (n=28)Fri 4a-8a ET: 0.91× (n=51)Fri 8a-12p ET: 0.99× (n=127)Fri 12p-4p ET: 0.88× (n=93)Fri 4p-8p ET: 0.89× (n=55)Fri 8p-12a ET: 1.33× (n=18)Sat 12a-4a ET: 0.94× (n=26)Sat 4a-8a ET: 1.07× (n=61)Sat 8a-12p ET: 1.16× (n=104)Sat 12p-4p ET: 1.05× (n=66)Sat 4p-8p ET: 1.13× (n=32)Sat 8p-12a ET: 1.22× (n=29)Sun 12a-4a ET: 1.20× (n=33)Sun 4a-8a ET: 1.02× (n=55)Sun 8a-12p ET: 1.11× (n=88)Sun 12p-4p ET: 0.99× (n=64)Sun 4p-8p ET: 1.08× (n=49)Sun 8p-12a ET: 1.05× (n=14)MonTueWedThuFriSatSun12a-4a4a-8a8a-12p12p-4p4p-8p8p-12a1.33×1.00×0.82×
Views compared with the sample average by weekday and four-hour block in US Eastern, from 2,554 videos across 255 channels (blue means above average; grey means below). The cells aren't smoothed and hold 14 to 132 videos each (hover for counts), so read the rows rather than choosing a single bright cell.

At a glance

QuestionWhat the data says
Best time to post on YouTubeNo clear daytime winner; daytime results were within about 4% of average
Best daySaturday, about 12% ahead of the rest of the week
Do weekends beat weekdays?Yes, about 11% more views
Worst daysWednesday and Thursday, about 9% behind the rest
Did late-night uploads do better?About 13% more views than other hours, but from a smaller sample
Did likely scheduled uploads do better?About the same views as the rest

Weekend uploads got more views

Mon ET: 1.05× (n=379)1.05×MonTue ET: 1.00× (n=365)1.00×TueWed ET: 0.95× (n=404)0.95×WedThu ET: 0.93× (n=413)0.93×ThuFri ET: 0.96× (n=372)0.96×FriSat ET: 1.10× (n=318)1.10×SatSun ET: 1.07× (n=303)1.07×Sun
Views compared with the sample average by upload weekday, US Eastern, across 2,554 videos. A score of 1.00 matches the sample average; it isn't a comparison with weekdays alone.

Weekend uploads got about 11% more views than weekday uploads. They still led after removing unusually successful videos.

Saturday did best, about 12% ahead of the other days combined. Sunday looked good too, though its uncertainty range included no advantage.

The result held when checked in different ways:

  • Without the videos whose scores hit the cap, the weekend lead was still about 11%.
  • Before adjusting for video age, the lead was larger, about 16%.
  • Weekends led across all channel sizes, including the largest channels.
  • The biggest lead, about 25%, was among channels whose usual video gets between a thousand and ten thousand views.

Only about a quarter of uploads went out on weekends, although most channels posted on a weekend at least once.

Viewers may have more free time, or channels may save better videos for the weekend. You can test a weekend upload without assuming the day itself caused the extra views.

Wednesday and Thursday trailed the rest

Wednesday and Thursday uploads got about 9% fewer views than uploads on the other days. Thursday alone also trailed, while Monday was above average.

Buffer ranks Wednesday and Thursday weakest too. If your schedule's flexible, a weekend is a better place to start testing than either of those days.

Moving a video to another day won't necessarily improve its views. Channels choose what to release as well as when, and this comparison can't separate those choices.

The daytime hour gave little reason to change your schedule

12a ET: 1.10× (n=36)12a1a ET: 1.14× (n=44)2a ET: 1.06× (n=52)3a ET: 1.06× (n=73)3a4a ET: 1.02× (n=66)5a ET: 1.01× (n=65)6a ET: 0.99× (n=115)6a7a ET: 1.00× (n=188)8a ET: 1.00× (n=188)9a ET: 1.00× (n=210)9a10a ET: 1.00× (n=210)11a ET: 0.99× (n=215)12p ET: 0.98× (n=190)12p1p ET: 0.97× (n=127)2p ET: 0.96× (n=135)3p ET: 0.96× (n=112)3p4p ET: 0.95× (n=100)5p ET: 0.97× (n=99)6p ET: 0.97× (n=106)6p7p ET: 0.98× (n=72)8p ET: 1.01× (n=57)9p ET: 1.11× (n=32)9p10p ET: 1.13× (n=31)11p ET: 1.16× (n=31)
Views compared with the sample average by upload hour, US Eastern, smoothed across five hours, from 2,554 videos. Each bar counts only videos from that hour, excluding the other hours used for smoothing; the 9pm, 10pm and 11pm bars have 32, 31 and 31 videos respectively.

Between 8am and 8pm Eastern, the hour barely mattered. The daytime blocks were all within about 4% of average, and none showed a clear difference.

Late-night uploads, between 8pm and 4am Eastern, did about 13% better than uploads at other times. That lead remained after removing channels that mostly posted at night.

But only about 350 videos were posted overnight. A channel outside the US might post in the morning when it's night in Eastern time, so these results aren't enough reason to start uploading at midnight.

Most uploads went out between early morning and early afternoon. Posting when other channels were busiest didn't produce a clear daytime winner.

Likely scheduled uploads did as well as the rest

About a third of videos went live in the first minute of an hour. Those videos were likely scheduled, though a timestamp can't confirm how they were published.

These uploads got about the same views as the rest. If scheduling helps you publish on a day off, the results give you no clear reason to avoid it.

How this compares to what's already published

  • Buffer used median engagement across 1.8 million videos. For long-form videos, it recommends Sunday at 10am and ranks Wednesday and Thursday weakest.
  • SocialPilot studied about 301,000 videos sent through its scheduler. It recommends weekday afternoons, especially Wednesday, Thursday and Friday, in local time.
  • vidIQ compared about 40 million US long-form videos with each channel's usual views. Weekends led, Wednesday trailed, and the exact upload time had little effect on views after a month.
  • OverseerOS compared about 2,600 uploads with their channels' usual performance and included uncertainty ranges. Saturday came first, with advice to test it rather than assume it'll work for everyone.

The weekend result agrees with Buffer, vidIQ and OverseerOS. SocialPilot's recommended midweek afternoons didn't do especially well here.

These methods answer different questions because engagement and views measure different things, and scheduler customers don't represent all channels. Comparing each video with its channel's usual views also keeps a large channel's audience size from deciding the result.

How I measured it

The final sample contains 2,554 videos from 255 channels. Every video was longer than 3 minutes and between 2 and 23 days old when collected.

  1. Find channels. Run 155 niche searches through youtube.search with uploadDate set to this_month, returning 3,099 videos from 2,393 channels.
  2. Get recent uploads. List 800 channels through youtube.channel_videos and keep the 691 whose first page reached past the start of the study window.
  3. Get exact times. Use youtube.video for each video's upload time to the second and view count, because the channel listing gives relative times such as "4 days ago".
  4. Keep the same age window. Keep videos at least 2 but less than 23 days old, a 21-day window covering each weekday 3 times, and exclude the video used to find each channel.
  5. Remove short videos and sparse channels. Drop 146 videos lasting 3 minutes or less and keep channels with at least 5 eligible videos, leaving 2,554 videos from 255 channels.
  6. Account for age. Older videos have had more time to collect views, so adjust each video's views for its age before comparing it with its channel's median.
  7. Limit extreme scores. Cap scores at 6x above or below the channel baseline so unusually successful or unsuccessful videos don't dominate; 142 videos reached a cap.
  8. Compare upload times. Convert timestamps to US Eastern, take geometric means of scores by weekday, hour and four-hour block, and estimate uncertainty with 1,000 resamples of whole channels.

Collecting the data in early November took 5,226 successful calls through AnyAPI's YouTube endpoints: 155 searches, 800 channel listings and 4,271 video lookups. The run cost $4.13.

This request gets a video's exact upload time and views. The recorded response includes only the relevant fields:

curl -sS -X POST https://api.getanyapi.com/v1/run/youtube.video \
  -H "Authorization: Bearer $ANYAPI_API_KEY" \
  -H 'Content-Type: application/json' \
  -d '{"url":"https://www.youtube.com/watch?v=g6Die8gVFn0"}'
{"output":{"found":true,"data":{"id":"g6Die8gVFn0","title":"OLED iPad Mini is The Most Apple Product They've Ever Appled","channel":"Marques Brownlee","createdUtc":1793714491,"views":10450652,"durationMs":780000}},"costUsd":0.0008}

The limits of this data

  • Videos had different amounts of time to gather views. The age adjustment helps compare them, but it doesn't replace measuring every video at the same age or tell you how it did on its first day.
  • All times use US Eastern here. Converting the times accounts for the clock change in early November, but it doesn't tell us where each channel's viewers live.
  • Views are the only outcome. Likes, watch time and subscribers gained could give different answers.
  • Video quality could explain some of the weekend lead. Channels might save better videos for days they already believe work best.
  • The window is short. It covers uploads from mid-October into early November, including Halloween; another season could look different.
  • Search shaped the sample. English niche searches favor the channels they find, and requiring several recent uploads favors active channels.
  • The busiest channels were left out. A channel's first page had to reach beyond the whole window, so channels posting too frequently to fit weren't included.
  • The sample has fewer large channels. Only 42 channels had typical views above 100,000, so their result is less certain.

With a different set of channels, the estimated weekend lead could range from about 3% to 20%.

What to do with it

  1. Try a weekend release. Compare it with similar weekday videos before changing your whole schedule.
  2. Give Saturday a fair test. Saturday did best; Wednesday and Thursday were weaker days to try first.
  3. Pick a convenient daytime hour. The daytime differences don't justify rearranging your day to post at an exact time.
  4. Check your viewers before trying overnight uploads. YouTube Analytics' "When your viewers are on YouTube" card helps you choose a time for your actual audience.
  5. Schedule if it helps. Likely scheduled uploads did about as well as other uploads.
  6. Repeat the comparison for your channel with AnyAPI's youtube.channel_videos to list uploads, then youtube.video for exact times and views at $0.00072 per lookup. Use the same age window and adjustment before comparing weekend and weekday scores against your channel's median.

Frequently asked questions

What's the best time to post on YouTube?
Try a weekend upload: weekend videos got about 11% more views than weekday videos in this sample. No daytime hour showed a clear advantage.
What's the best day to post on YouTube?
Saturday, with about 12% more views than the rest of the week in this sample. Test it with your own videos before changing your schedule.
Do these times apply outside the US?
Every upload time was converted to US Eastern, even for channels elsewhere. Check when your own viewers are online before using an Eastern time to plan your uploads.
Does this show when to post for more first-day views?
No, views were counted at different ages within roughly the first three weeks and adjusted for age. They weren't measured at the end of each video's first day.

Compare the separate results for YouTube Shorts, X posts and TikTok videos if you manage those too. AnyAPI's YouTube endpoints let you make this comparison for your own channel or a client's.

Use data responsibly and follow AnyAPI's Acceptable Use Policy.