The short version
- Clips that stand on their own run about a minute. The middle clip ran 56 seconds; 80% fell between 41 and 67 seconds.
- The opening carries the most. The first tenth of a video produced 3.0x as many clips as the last tenth, and 62% of all clips came from the first half.
- Longer videos yield more clips, but fewer per hour. Videos under 20 minutes gave 14 clips per hour of video; videos over 3 hours gave about 6 per hour, still 24 clips per video.
- The best clip wins on hook and value, not polish. Against the other clips from the same video, the top clip scored about 12% higher on hook and 10% higher on value, but only 7–8% higher on clarity and production.
- Speech carries it. In the middle clip, someone was talking for 78% of the time.
Where the data comes from
1,543 long videos, 817.5 hours in all, processed by legit2clip between May 29, 2026 and November 20, 2026. For each one, the whole recording was transcribed, every candidate moment was scored on five criteria (hook, value, clarity, production, shareability), and the strongest moments were cut as clips that start and end on whole sentences. 7,559 clips were picked.
The mix: about two-thirds uploaded files and one-third YouTube links, plus some Twitch and Kick broadcasts. 82% of the videos were in English; the rest were mostly Spanish, Hindi, Portuguese, Arabic, Korean, Russian and French. It includes our own runs on public talk archives, such as 156 security talks, so recorded talks are well represented. Videos where moments were placed evenly instead of chosen were left out. No customer content is shown here, only totals.
Finding 1: a good clip runs about a minute
Clips were allowed to run as long as the moment needed, with a whole sentence at each end. The result clustered tightly: the middle clip ran 56 seconds, one in ten was shorter than 41 seconds and one in ten longer than 67.
That is longer than the very short cuts many tips recommend, and it is what a complete thought takes: a setup, the point, and the line that lands it. Shorter cuts usually lose the setup or the payoff.
Finding 2: the opening of a video carries the most
Split every video into ten equal parts and count where each clip starts. The first tenth produced 1,347 clips and the last tenth 452. 62% of all clips started in the first half.
Speakers front-load: the thesis, the story that sets up the talk, the reason to listen. Late in a recording come Q&A, thank-yous and wrap-ups, which rarely stand on their own. If you can only clip part of a long video, clip the first half.
| Part of the video | Clips starting there | Share |
|---|---|---|
| 0–10% | 1,347 | 17.8% |
| 10–20% | 1,061 | 14.0% |
| 20–30% | 837 | 11.1% |
| 30–40% | 730 | 9.7% |
| 40–50% | 688 | 9.1% |
| 50–60% | 636 | 8.4% |
| 60–70% | 591 | 7.8% |
| 70–80% | 648 | 8.6% |
| 80–90% | 569 | 7.5% |
| 90–100% | 452 | 6.0% |
Clip start time divided by video length, in tenths.
Finding 3: longer videos, more clips, fewer per hour
A long stream or service is not ten times as clip-worthy as a short video, because long recordings have more set-up, chat, breaks and repetition. But a 3-hour recording still averaged 24 clips: about a month of daily posts from one recording.
| Video length | Videos | Clips per hour of video | Clips per video |
|---|---|---|---|
| Under 20 minutes | 839 | 14.0 | 2.8 |
| 20 to 60 minutes | 528 | 9.8 | 6.1 |
| 1 to 3 hours | 154 | 6.2 | 9.5 |
| Over 3 hours | 22 | 6.4 | 24.4 |
Clips that stand on their own, as picked; asking for more clips raises these numbers when a video has more good moments.
Finding 4: the best clip wins on hook and value
Each candidate moment was scored on five criteria. Comparing the top-ranked clip in each video with the other clips from the same video shows what made the difference:
| Criterion | Top clip (average) | Other clips (average) | Difference |
|---|---|---|---|
| Hook | 21.9 | 19.6 | +12% |
| Value | 21.2 | 19.2 | +10% |
| Shareability | 17.2 | 15.4 | +12% |
| Clarity | 12.3 | 11.5 | +7% |
| Production | 12.0 | 11.1 | +8% |
Points per criterion; the criteria have different maximums, so compare the difference column.
Finding 5: what the winning moments have in common
Every pick comes with a one-line reason. The words that came up most: story, funny, moment, real, relatable, reveal, payoff, surprising, genuine, explanation, example, reaction. In practice that means a story that ends inside the clip, a surprising fact or reveal, a joke that lands, or an explanation that makes one idea click.
And they are spoken: in the middle clip someone was talking 78% of the time. Moments built on music, silence or on-screen action alone rarely made the cut.
What to do with this
- Aim for about a minute, ending on the line that lands the point.
- Short on time? Clip the first half of the recording first.
- Judge a clip by its first few seconds and by what the viewer learns, not by how polished it looks.
- Pick moments where someone is speaking: a story, a reveal or an explanation.
- From a long video, expect roughly 6–14 strong clips per hour of video.
Using this data
You are welcome to quote these numbers and use the tables, with a link to this page as the source. For questions or a breakdown we haven't published, write to sales@legit2clip.com.
Questions
How long should a short-form clip be?
In this data, clips that stand on their own ran about 56 seconds; 80% fell between 41 and 67 seconds.
Which part of a long video makes the best clips?
The opening. The first tenth of a video produced 3.0x as many clips as the last tenth.
How many clips can you get from a one-hour video?
About 6 to 10 strong ones in this data; shorter videos yield more per hour.
Is this data independent?
No: it comes from our own product, legit2clip, and includes our own runs on public talk archives. The method and its limits are described above.