How fast do viral Shorts cut? A 54-video study
Viral YouTube Shorts change shot about every 4.7 seconds, and the first cut usually lands before the 2-second mark. We measured 54 of them frame by frame to find out how fast short videos really move.
Published by the Viibeo team
Key findings
- 4.7 s
- Median shot length
- 2.8 s
- Shot length in the fastest quarter of videos
- 1.9 s
- Typical time to the first cut
- 47 of 54
- Videos with at least one cut (87%)
Viral pacing next to a typical AI video
Most AI video tools hold one image or clip for a whole sentence of narration. In our own videos before this study, scenes ran 12 seconds at the median (11 to 14 seconds for the middle half), more than twice as long as a viral shot.
- Fastest quarter of viral Shorts2.8 s
- Median viral Short4.7 s
- Typical AI video scene, no cuts12 s
What this means for your edits
- Change the picture within the first 2 seconds.
- Cut or reframe every 3 to 5 seconds. A tighter crop of the same shot counts as a cut for the viewer.
- Put cuts on pauses in the voice, so the picture and the sentence move together.
- Keep text inside the Shorts safe zone so new shots never hide it behind buttons.
We built this into Viibeo: the Craft edit style reframes each scene about every 3 seconds, lands cuts on voice pauses and makes the first cut at 1.4 seconds. For limits and the right length per format, see our guide to YouTube Shorts length. Try it free.
How we measured
- 54 YouTube Shorts with high view counts from our inspiration library, across several niches.
- For each video we compared consecutive storyboard frames and counted a shot change when the picture changed sharply. Frames were 0.3 to 0.6 seconds apart.
- Shot length is the time between two changes. We report the median per video, then the median across videos.
Limits
- The frame spacing means cuts closer than about half a second apart can be merged, so the fastest edits are slightly understated.
- Slow zooms and camera moves are not counted as cuts.
- 54 videos is a small sample chosen for performance, not at random. Treat the numbers as a guide, not a rule.