Short video is not published to your followers and left there. Every clip is shown to a small group first, and what that group does decides whether it is shown to a larger one. A clip that stops at a couple of hundred views was not hidden, penalised or filtered: it was tested, and the test came back flat. That is a much easier problem than a punishment, because the test measures things you can change.
All three big short-video feeds work the same way in outline, and all three describe it in their own documentation. A new clip goes to a sample of people. Their behaviour is measured. If it looks good against everything else competing for that slot, the clip goes to a bigger sample, and so on. Nothing about that first sample depends on how many followers you have, which is why a new account can be shown to strangers on day one and why a large account can still post something that goes nowhere.
What gets measured is behaviour, not approval. In rough order of how much the platforms themselves emphasise it:
Described as of September 2026 from what the platforms publish themselves. Read them at the source: YouTube on how Shorts are recommended, TikTok on the For You feed and Instagram on ranking.
The platforms do restrict reach, and they say so. What they describe is specific: content that breaks a written rule, content they mark as not recommendable to strangers, accounts with a standing strike. In each of those cases there is normally something to see - a notice, a status page in the app, a removed post.
What people call a shadowban is usually none of that. It is the ordinary case of a clip being sampled and not passed on, which looks identical from the outside: no message, no explanation, just a view count that stops. The tell is simple. If your next clip does noticeably better with no appeal, no waiting period and nothing else changed, nothing was ever switched off.
There is one exception worth knowing, because it is real and it is fixable: reused material can be excluded from recommendation without any notice on your account. That is a different mechanism from copyright, and we wrote it up separately - what actually happens to a clip of somebody else's recording.
In the order they cost you views.
Greetings, a logo, "so as I was saying". The sample group decides during exactly those seconds. Start where the interesting thing starts, not where the file did.
A letterboxed clip with black bars reads as something taken from elsewhere before a word of it is heard. Faces end up too small to hold anybody.
Most short video is watched with the sound off. Without captions the clip is not slow, it is silent, and there is nothing to hold somebody scrolling in public.
The moment lands at eighteen seconds and the clip runs to fifty. Everything after the point is measured as people leaving.
A single clip is one sample of a noisy process. Two clips from the same recording can differ tenfold for reasons nobody can see. One result is not a signal.
Take the clip that stalled and watch the first three seconds with the sound off. That is the version most of the sample group got. If those three seconds do not show a face, a caption or something moving that explains why to stay, you have found the problem, and no amount of hashtags was going to fix it.
Then check the obvious mechanics in this order, because they are cheap and they are decided before anybody hears a word: is it filling the screen, are the captions readable at arm's length, does it end at the end of the point.
Each has its own page here: where a clip should start, what a wide recording loses going vertical, how long a clip should be and captions that get watched.
SnapDub does that pass for you: it picks the moments out of a long recording, frames each one on whoever is speaking, burns in captions and hands you the clips as files. It cannot make anybody watch. It can stop you losing the sample group to a black bar and a silent first second, which is where most stalled clips are actually lost.
Start free with 500 credits a month - no card required. Same material, opened where it starts and framed for a phone.
Start clipping freeBecause it was shown to a sample of that size and not passed on to a bigger one. Two hundred views is what an untested clip gets, not a penalty - the feeds show every new clip to a small group and decide from what that group does. The number to look at is not the views, it is how many of them kept watching past the first seconds.
Almost certainly not. Restrictions that platforms do apply usually come with something visible: a removed post, a notice, an account status page in the app. A view count that stops with no message is the ordinary outcome of a clip that was sampled and not passed on. If your next clip does better without you appealing anything, nothing was switched off.
Barely. They give the feed a hint about what the clip is about, which can affect who is in the first sample. They cannot make that sample watch. A clip with perfect hashtags and a slow opening loses to a clip with none and a good one.
No, and it costs you whatever the original had gathered. If the clip genuinely deserved better, the fix is a different cut of the same material - a different moment, a tighter opening - which is a new clip, not the same file posted twice.
Because each clip is sampled separately and the samples are noisy. A single result, good or bad, is one measurement. What is worth reading is a run of them: if the ones that hold attention share an opening style or a subject, that is a signal you can act on.
There is no honest public number for this, and anybody quoting one is guessing. What is true is that follower count does not decide the first sample, so a new account is not starting from zero reach - it is starting from zero history, which is different.