"How many a day" is really three questions
The first is capacity: how many can you finish in a day? That depends on shooting, editing and copywriting time, not on the platform. If one video takes 40 minutes including the shoot, posting three a day means borrowing from tomorrow.
The second is rhythm: do you publish everything in one window, or spread it across the day? That shapes how often viewers see you, and how manageable the work feels.
The third is platform limits: is there a daily cap, a cooldown, behaviour that gets flagged as abnormal? This one changes, so check the current official announcement rather than a claim someone made in a group six months ago.
Ask all three at once and you will never get a usable answer. First work out which one is blocking you, then you know whether to fix output, scheduling, or something else.
Work out capacity before frequency
The practical approach is to work backwards: time one video end to end, then decide how many you post per week. 20 minutes shooting, 25 editing, 10 for titles and tags, 5 to upload and check — 60 minutes per video. 8 hours a week buys you 8 videos.
The number is useful because it cannot lie to you. Set a target of three a day and you have committed to 21 videos and 21 hours a week against 8 hours of real time. The only possible outcome is one burst day followed by three empty ones — which, from the viewer's side, is no rhythm at all.
Backlog is the other variable. Shooting 20 videos in one session and publishing them over four weeks feels nothing like shooting and posting daily. With a backlog you can hold a steady rhythm; without one you can only post whenever there is time.
- Time one video end to end: shoot, edit, copy, upload
- Divide weekly available hours by that time to get a sustainable weekly output
- Keep your target at or below 1.2x that number, so there is slack
- Fix one day as a production day and publish only on the others
Batch everything, or spread it out?
Batching means one session at the desk instead of a whole day tethered to it; the cost is that too much lands in one window, viewers compete with each other for attention, and a repair means rescheduling the batch again if a few videos fail.
Spreading gives you more even exposure and a smaller blast radius when something goes wrong; the cost is that you actually have to be free at that hour, or you need scheduling to cover it.
In practice, semi-spreading works best: split the day into two windows, say midday and evening, and publish 1 to 3 videos in each. You are not on call all day, and nothing piles up in a single hour. If your content has an obvious time-of-day fit — breakfast, commute, bedtime — follow that instead of forcing it into a fixed slot.
Picking time slots: avoid the dead zones first
The problem with "best time to post" articles is that the answer belongs to someone else's account, in someone else's category, with someone else's audience. For a new category or a new account, finding the obvious dead zones is far more useful than finding the best slot.
The method is simple: for two weeks, put the same kind of content into morning, midday and evening slots and record how each one performs. After two weeks one slot will be consistently low — that is your dead zone, so avoid it. Fine-tune the rest later.
One caveat: do not draw conclusions from a tiny sample. A single video's performance is heavily driven by its topic, so you need two or three videos on the same topic in each slot before a comparison means anything.
Scheduling a large backlog
The most common mistake after a 20-video shoot day is publishing all of it that night. Four weeks of content lands in one day, and the next three weeks are empty.
Treat the footage as inventory instead: lay out a list first, assign a date and a window to every video, then work from the list. The same list doubles as your not-yet-published queue — mark each item once it runs, and you will not double-post or lose track.
Backlog has a second benefit: you can run an entire series inside a short window so viewers see the whole theme at once instead of across three months. Series coherence matters a lot in short video, and it only exists if you schedule for it.
Let the published record decide, not your memory
Rhythm collapses for lack of memory, not lack of discipline. 30 videos, four series, three accounts — two weeks in, you are no longer sure whether video 14 ever went out.
So every publish needs a recorded result: which video, when, success or failure. That record does three jobs — it tells you what to post tomorrow, it prevents double posting, and two weeks later it shows which slot actually performed better.
If your record is only a daily total like "posted 5 today", it cannot help you. You need per-video granularity to answer a question as simple as "did this one already go out?"
Three common rhythm mistakes
The first is the burst. Two weeks of accumulated content goes out in a day, then nothing for three weeks. Viewers feel no rhythm and you cannot sustain it either.
The second is posting to hit a number. Chasing a daily quota pushes out work that is clearly not ready. Short video recommendation leans heavily on the first few seconds, so one weak clip drags down the average for the whole account.
The third is keeping no record. What went out, what failed, what was re-posted — all from memory. This is the hardest to notice, and it makes the first two unfixable, because you have no data to compare against.
Three practical scheduling rules
First, build in slack. Do not pack the schedule tight enough that a single failure forces a full rebuild. 10 to 15 minutes between items leaves room to react.
Second, retry only what failed. If 3 of 30 videos did not go through, handle those 3. Rerunning the batch turns the 27 successes into duplicates and makes it harder to see what actually broke.
Third, separate scheduling from production. Production days are for production; publish days are for confirmation only. Do both at once and you will be writing copy while thinking about uploads, and neither gets done well.
- Leave 10-15 minutes between each item
- Retry failures only, never the whole batch
- Fix one day for list clean-up and scheduling; the rest are execution days
- Spend 10 minutes a week on the record and decide whether to shift any slots