How often to cut in a short vertical ad, why the fast-cut default backfires on demonstrations, the shot length that matches how long a viewer needs, and a cut plan by runtime.
HighQualityUGC Team
Editorial
We run UGC ad tests daily and publish what holds up: real credit costs, real hook rates, no vendor fluff.
Somewhere along the way "fast cuts" became a synonym for "good short form editing", and a lot of ads now cut every second for the entire runtime. They feel energetic and they underperform, because energy is not the thing being measured.
What is being measured is retention, and retention comes from the viewer continuously getting something. A cut is only useful when it delivers new information. A cut that delivers nothing is just motion.
One shot every two to three seconds is a sensible starting rhythm. In a 20 second ad that is roughly 6 to 10 shots.
The reason it works is not pace. It is that roughly two to three seconds is how long a viewer needs to read a simple visual: see the thing, understand it, absorb one label. Cut before that and they got nothing. Hold much longer with no new information and attention drifts.
So the real rule is: cut when the shot has finished delivering, not on a timer.
| Shot type | Hold for | Why |
|---|---|---|
| Hook, face or interrupt | 2 to 3 seconds | Long enough to register a person and a claim |
| Demonstration, the actual proof | 4 to 8 seconds, uncut | A cut mid-demo reads as hiding something |
| Detail crop | 1.5 to 2 seconds | One thing, quickly understood |
| Text card | 1 second per 3 words | Reading speed, not taste |
| Reaction | 1 to 2 seconds | Only useful if it is genuine |
| Closing ask | 2 to 3 seconds | Time to register the product and the action |
If your ad exists to prove something works, the proof shot must be continuous. This is the single most important editing decision in the format.
A viewer watching a stain lift, a lid seal, a tool cut through something, is doing one thing: checking whether it is real. Cut away halfway and the answer they take is no. The edit itself is the claim that something was hidden, whether or not anything was.
This is why fast-cut editing hurts UGC specifically more than it hurts brand video. Brand video is not trying to prove anything, so its cuts cost nothing. A demonstration ad is entirely a proof, and every cut inside the proof subtracts from it.
Practically: identify the one shot in your ad that is the evidence, and protect it. Everything around it can move quickly.
For context on runtime, marketer usage clusters in the middle: about 18% run 11 to 20 seconds, 31% run 21 to 30 seconds and 33% run 31 to 60 seconds. Mid-length, roughly 15 to 30 seconds, is the reliable default for performance.
A related finding worth holding onto: a 12 second video watched in full often outperforms a longer one that loses viewers halfway. Completion is doing more work than duration.
Inside that window, two things:
Hard cuts for everything, with two exceptions.
No crossfades, no zoom transitions, no motion graphics wipes. If a transition draws attention to itself, it has cost more than it added.
Too fast:
Too slow:
The retention graph in your ads manager will point at the exact second. That is a more useful editing note than any general rule here.
Around one cut every two to three seconds as a default, so roughly 6 to 10 shots in a 20 second ad. But cut when a shot has finished delivering rather than on a timer, because a cut that adds no new information is just motion.
No. If the ad exists to prove something works, the proof shot must run continuously, usually 4 to 8 seconds. A cut in the middle of a demonstration reads as hiding something, and the viewer's conclusion is that the product does not do what it claimed.
Hook and closing ask 2 to 3 seconds, demonstration 4 to 8 seconds uncut, detail crops 1.5 to 2 seconds, text cards about one second per three words, reactions 1 to 2 seconds. Around two to three seconds is how long a viewer needs to read a simple visual.
Only when each cut delivers something new. Fast cutting hurts UGC more than brand video because a UGC demonstration is entirely a proof, and every cut inside the proof subtracts from it. A 12 second video watched in full often outperforms a longer one that loses people midway.
Hard cuts for almost everything, plus match cuts where the composition stays and the state changes, which works well for before and after inside a demonstration. Avoid crossfades, whips and graphic wipes, since a transition that draws attention to itself signals that someone edited this.
