Most short videos are decided before the first sentence ends. We went back through two hundred of our own edits and marked the exact frame where viewers left, then compared the ones that held 80% of their audience with the ones that lost half.

The pattern

The videos that held their audience put the point of the video in the first second, in words and on screen. The ones that lost viewers opened with a greeting, a logo or a slow push in.

• Second zero: the claim, spoken and captioned
• Second one: a cut, a zoom or a change of angle
• Second three: the first proof that the claim is true

What we change first

When a creator sends us a video that is not holding, we rarely touch the middle. We look for the strongest sentence in the whole recording, wherever it sits, and move it to the front. Then we cut everything between that sentence and the first piece of proof.

If the best line is at second forty, the video starts at second forty.

The caption counts as a frame

On a muted feed the caption is the opening shot. One or two words at a time, large, in the upper third where the interface does not cover it. A full sentence on screen reads as a wall and gets skipped.

A test you can run today

Export the same video with two different first lines and post them a week apart at the same hour. Compare the three-second retention, nothing else. In our edits the better opening wins by fifteen to thirty points, and the rest of the graph follows it.