Stop optimizing for length. Optimize for completion rate.

It's easy to fixate on hitting a specific duration — '30 seconds for TikTok' — as if the number itself is what the algorithm rewards. It isn't. What every major short-form platform actually measures is what percentage of viewers watch a clip all the way through, and how many rewatch it.

A 45-second clip that 70% of viewers finish will consistently outperform a 15-second clip that only holds 40% of its audience, even though the shorter clip technically demanded less commitment. Completion rate is the signal; length is just one of several levers that influences it.

This reframes editing decisions in a useful way: the question isn't 'how do I cut this down to fit a target length,' it's 'where does the story actually finish, and does anything after that point exist only to pad runtime.' A clip should end when the idea ends, not when a stopwatch says it can.

Scoring a clip on predicted completion — not just checking it against a length target — is the more useful check to run before publishing, since it catches the actual failure mode (a slow middle, a weak payoff) instead of a proxy for it.