← All posts
Jun 17, 2026 · ShortRemix Team · 3 min read

Fast cuts stopped working when everyone started using them

The 'cut every 1.5 seconds' advice was correct in 2023 and is now actively hurting your retention. A real example from a video I cut down.

I cut eight jump cuts out of a 30-second short last week. Retention at second 5 went from 51% to 67%.

The thing I was sure about going in was that this video had a pacing problem. The pacing problem turned out to be that there was too much pacing.

Where this "cut every 1.5 seconds" thing came from

Somewhere around 2022, somebody — I think it was an editor working on Iman-Gadzhi-adjacent stuff — noticed that shorts with cuts every 1.5 to 2 seconds outperformed shorts with longer takes. They wrote a thread. Everyone copied it. It became gospel.

In 2022-2023 the advice was correct. Most short-form video at the time had been ported over from horizontal YouTube editing rhythms, which are much slower. Fast cuts were a real competitive advantage because the bar was low.

That bar is now a ceiling. Every creator cuts every 1.5 seconds whether or not there's a reason to. The result is the visual equivalent of someone tapping you on the shoulder every two seconds while telling you a story. It's exhausting and the viewer can feel it before they can name it.

What broken pacing actually looks like

Open one of your shorts. Watch it with the sound off. Count the cuts. Then ask, for each one: did the shot before it run out of information? Was new information arriving?

In most of the shorts I look at right now, the answer to both questions is no. The cuts are decorative. They're there because the editor knew there should be cuts, not because there should be a cut here.

This is the editing version of the second-beat retention problem. The editor has nothing to say but is saying it anyway, just faster.

The eight cuts I deleted

The video was about glute training. The hook was strong — the kind that names a thing the viewer's been doing wrong. The body was four points, edited as four 5-second segments, with about two jump cuts per segment.

When I rewatched it without sound I noticed that every cut landed on a word boundary instead of an idea boundary. They were following the speech, not the content. Cutting that way feels like punctuation, which is fine in long-form where you want rhythm. In a 30-second short it just chops the message into syllables.

So I deleted every cut that wasn't necessary — meaning every cut that didn't either introduce a new visual, skip dead time, or punch a real punchline. I went from 14 cuts to 6.

The video is now slower in a frame-by-frame sense and faster in an information-per-second sense. Retention reflects it.

When fast cuts still work

Listicles. Comparisons. Anything where the video is structurally a sequence of different things shown one after another. Each cut introduces a new entity, which is information.

Also: the very beginning. The first 1.5 seconds can use a fast cut to signal "this is a high-energy video, stay" — even if the rest of the video is slow. That's almost the only thing the opening cut is doing.

Everywhere else: ask if the cut is for you (the editor, because you're proud of finding the right b-roll) or for the viewer (who needed new information right now).

If it's for you, delete it.