YouTube posting schedule: best times

Each row is one published analysis, read from its own page on 2026-09-23. They measure different samples over different periods, so they are listed side by side rather than merged into one number.

With a single large study there is nothing to agree or disagree with, which is a limit rather than a finding: we did not find a second large 2026 study to check Buffer’s numbers against. What Buffer’s own data does show is that the two formats behave differently, so schedule long-form and Shorts on their own patterns rather than one.

Both large studies express times as the audience’s local time: Sprout Social records every time in local time (a 9 a.m. window is 9 a.m. in Los Angeles, London and Sydney alike), and Buffer says to “consider all the recommendations local time”. If your audience is spread across time zones, schedule for the zone where most of your followers are.

Every figure here is an aggregate across many accounts and industries. It tells you where to start testing, not when your audience is online.

Buffer’s roundup recommends about one long-form video a week and one to three Shorts, but it attributes the long-form figure to a single cited experiment and gives no source for the Shorts figure, so read both as loose guidance. HeyOrca’s manager survey has no YouTube figure.

YouTube’s help centre describes scheduled publishing as a way to schedule a private video to go public at a specific time: you upload first, then set the video to scheduled. That is a release timer on an upload you have already made, not a way to queue uploads.

Shorts are the same upload flow: YouTube categorises a video as a Short when it is square or vertical and up to three minutes long, so there is no separate “Short” scheduler.

YouTube Studio has built-in scheduling (an upload set to scheduled or private, released at a chosen time). It works one video at a time and needs the upload done in advance.

YouTube posts through SkedCast follow the same rules as the composer: up to 5,000 characters, with at least one image or video required.

SkedCast holds a scheduled post and publishes it at the time you set, from one calendar shared with every other connected account, with the same per-platform limits checked when you schedule rather than at publish time. Scheduling never changes when your audience is online; it only makes sure the post goes out in the window you chose.

A study average is a starting hypothesis. Your account’s own results are the evidence. A simple test:

SkedCast’s best-times feature does the “find your own window” step for you where the data exists: a daily job reads each connected account’s own publish history and the engagement analytics SkedCast has ingested for it, and recommends the day-and-hour slots where that account’s posts did best, in the account’s own time zone. It never invents a slot — an account with no analytics data yet simply has none — and how much it can derive depends on what each platform exposes to third-party apps. The same slots are available through the API (GET /v1/best-times, scope besttimes.read) and the MCP get_best_times tool. Best times is included from the Solo plan.