The Best TikTok Length Inverts With Follower Count (4,264 Posts)

"Keep it short" is the most repeated rule in short-form video. Across 4,264 real TikTok posts it is correct for most creators and wrong for the ones with the largest audiences — and the crossover is sharp enough to plan around.

Below 10,000 followers, videos of 15 seconds or less out-reach one-to-two-minute videos by a factor of 5 to 7. Above 100,000 followers that ordering reverses and the longer videos win by 2.2x. The advice is not wrong so much as it is stated without the condition that makes it true.

Duration against account size

Median plays, every cell drawn from at least 59 posts:

Read it as a gradient rather than a switch. The advantage of going short is 7.2x under 1,000 followers, 4.9x from 1,000 to 10,000, 2.0x from 10,000 to 100,000, and then it crosses over: above 100,000 the long bucket wins by 2.2x. Four bands, one direction of travel, no zig-zag.

Saves invert too, and harder

If this were an artefact of how the feed distributes reach, it should not show up in a deliberate act like saving. It does, more sharply:

A sub-1,000-follower account gets 5 times more saves from a 15-second video than a two-minute one. A 100,000+ account gets 3.6 times more from the two-minute video. Two metrics, same inversion, same crossover point.

What we think is happening, and what we cannot prove

The honest position is that this data shows a pattern, not a mechanism. The reading we find most plausible is that length is a request for trust. Two minutes of someone's attention is a real ask, and it is an ask a stranger declines. Large accounts are not strangers to the people being served their videos, so the ask lands.

A competing explanation we cannot rule out: large accounts may simply be better at making long videos, having survived long enough to learn. That would produce the same table without length being causal at all. We have no way to separate those two with a snapshot.

What we would not conclude is that a small account can buy reach by posting longer. Nothing here supports that, and the sub-10,000 rows argue hard against it.

By vertical, with a caveat that matters

Comparing posts of 30 seconds or less against posts over 60 seconds, within each vertical that has enough of both:

The caveat: this table is not controlled for follower count, and the two are entangled. A vertical whose accounts skew large will look like a vertical where long videos work. We are showing it because the SaaS reversal is interesting and because hiding a weaker analysis would be worse than labelling it — but treat the follower-band tables above as the finding and this one as a hint.

31 to 60 seconds is the length to avoid

One band loses everywhere. Posts of 31 to 60 seconds underperformed at every account size on plays, on shares and saves, across twelve band-level comparisons with not one in its favour.

Our reading is that it is long enough to need a real structure and short enough that most people do not give it one — a 20-second idea with 30 seconds of padding. If a cut lands there, the useful question is whether it is a tight video wearing extra weight, or a long video that has been truncated.

What to do

1. Under 10,000 followers: go short. 15 seconds or less beat everything else in both bands, by 4.9x to 7.2x on reach.
2. Over 100,000 followers: stop cutting your best material down. The 61-120 second band was the strongest in both plays and saves.
3. Between 10,000 and 100,000: you are at the crossover. Short still wins on reach by 2.0x but 16-30s already leads on saves. This is the range where testing actually pays.
4. Never land at 31-60 seconds by accident. Go under 30 or commit past 60.

If you are cutting long videos down, our AI clip generator picks boundaries with these bands in mind, and the free video length calculator will work the arithmetic out for a specific upload. The reach-side companion to this piece is what actually goes viral on TikTok.

Limits

TikTok only. We hold no Instagram Reels or YouTube Shorts corpus, and the platforms weight watch time differently enough that copying these numbers across would be a guess in a table's clothing.

It is hashtag-sampled, so it leans toward posts that already received distribution, and it is a snapshot rather than a tracked cohort. SaaS is 38% of the corpus by design. And follower count is measured at capture, not at posting time, which blurs the bands for accounts that grew quickly.

Frequently asked questions

How long should a TikTok be?

It depends on your follower count and the answer inverts. Below 10,000 followers, 0-15 second videos out-reach 61-120 second videos by 4.9x to 7.2x. Above 100,000 followers, the 61-120 second videos win by 2.2x.

Do short TikToks get more views?

For small accounts, by a wide margin. For accounts over 100,000 followers, no: 0-15 second posts median 169,900 plays there against 379,300 for 61-120 second posts.

Why would longer videos work better for big accounts?

We cannot prove the mechanism. The pattern fits the idea that length asks for trust an established audience already extends, but we cannot rule out that larger accounts are simply better at making long videos.

What is the worst length for a TikTok?

31 to 60 seconds. It underperformed at every account size across plays, shares and saves — twelve comparisons, none in its favour.