My first pilot was a free test for a friend who runs an anime fight-analysis channel. I cut ten shorts from one of his long videos, sent them over, and got a clear answer: too short. I had thought I was being sensible. I was wrong, and the length of those clips has kept changing ever since, because he keeps telling me what works for his channel.
This post is the story of that number. I’ve already written about why I default to 30–45 seconds. This is the follow-up: what happened when one real creator kept moving the target.
Version one: 11 to 23 seconds
My reasoning for the first batch was simple. Short-form is fast, attention is scarce, so shorter must be better. I cut the clips at roughly 11 to 23 seconds.
The creator’s reaction was that they were too short. They had cut off the setup. His channel tells fights like boxing commentary: he builds a small argument, then lands the verdict or the joke. Remove the build-up and what’s left is the verdict with nothing under it. A stranger doesn’t know why it matters.
The lesson there wasn’t really about seconds. I had treated length as a style choice when, for this kind of content, it was a measure of how much context survives. If you cut for speed first, the context is the first thing to go.
Version two: 30 to 45 seconds
For the second batch I kept the setup in. Every clip ran 30 to 45 seconds. His feedback was that this length was right, that anything shorter felt too short, and that in the past he had tried going longer and felt it was too much.
That is the batch where I stopped asking “how short can this be?” and started asking “what does a stranger need to understand this?” The answer was usually a sentence or two of setup, the turn, and a payoff with a beat after it.
Version three: 35 to 55 seconds
Once he approved the first full set from the paid pack, with no revisions, he nudged the range up to 35 to 55 seconds for the rest of it. It was a small change, and it was his call. I note it because it shows something I’d underestimated: the right length isn’t one number you find and keep. It drifts as the creator sees more of their own clips in the wild.
Version four: 61 to 85 seconds, for a reason that has nothing to do with craft
The next change had a different cause. He told me he’d been looking at TikTok’s Creator Rewards Program. According to TikTok’s help pages, the program rewards eligible creators for qualified views on videos that meet its requirements, and the length requirement is that videos need to be longer than one minute. One of his shorts had also done well on TikTok, so he wanted to see whether longer clips were worth it.
So we ran a trial. I recut one video’s ten clips at 65 to 85 seconds. They didn’t have to be the same moments: I could extend a clip at the front and back, join two neighbouring moments into one arc, or cut a new one. He wanted to keep posting the shorter versions to YouTube during the trial, and the plan was that if the long ones worked, he would use one file everywhere rather than juggle per-platform edits.
His worry going in was that longer clips would hurt the quality. After he watched the trial set, his reply was that making them longer hadn’t messed up the quality at all, and that from then on the range would be 61 to 85 seconds per clip.
I want to be careful about what that does and doesn’t show. It shows that this creator, with this kind of material, was happy with longer clips. It doesn’t show that longer clips earn more. I don’t have performance data that proves it, and I won’t pretend I do. The decision was his, based on his own goals for TikTok.
What it took to make longer clips actually work
Longer is not the same as stretched. When I extended clips from the 35–55 range, the problems were predictable:
- Padding. Some of my earlier clips couldn’t reach 65 seconds without including filler. Rather than pad them, I dropped them and found other moments. A couple of the earlier clips didn’t survive the trial.
- Dead stretches. Over a minute, a flat patch is far more noticeable. I started checking each clip for long gaps where nothing was happening.
- A second hook. A short that runs 70 seconds needs a reason to keep watching halfway through. Where the old ending used to be, I looked for a line that re-hooks: a new claim, a turn, a question.
- A verdict ending. The clip still has to end on the point, not trail off after it.
This is also why I don’t think of “longer” as an upgrade. With the same source video, a 75-second clip is usually a different edit, not a stretched version of the 40-second one.
The newest batch: set by the creator, not by me
For his latest batch of videos the range is 61 to 75 seconds. Before cutting anything, I send a plan listing more candidate moments than he asked for, each with a timestamp, a rough length and a few sentences on why it works, so that he picks from them. That plan is the main thing I’d recommend to anyone working with an editor, and it fits the idea behind briefing a shorts editor in five minutes: decisions like length and moment choice should be written down before the cutting starts.
What I take from the whole sequence
- Ask for the length, then keep asking. My first guess was wrong by a wide margin. Each later change came from the creator seeing real clips, not from me predicting.
- Length follows what the clip needs. Context, a turn and a payoff set the minimum. A platform’s rules, like TikTok’s one-minute threshold, can set a different one. YouTube now allows Shorts up to three minutes, so the ceiling isn’t the issue; the question is what you put in the room.
- One creator is not a rule. What works for a commentary channel that builds arguments may not suit a channel that posts quick reactions. I treat every number here as this creator’s preference.
- Don’t pad to hit a target. If a clip can’t reach the range with real content, it’s the wrong clip.
- Plan first. A short list of candidate moments, with lengths, catches most mismatches before they cost any editing time.
If you’re deciding what length to ask for, start from your own best-performing clips and your own goals for each platform, then adjust after the first batch. And if you’d like to see what your own video looks like cut down, I can make a free short from one of your long-form videos. You keep it either way.