Completion rate

The measure is simple: completed views divided by total views. It became central to short form because feeds autoplay the next video the moment one ends, so a completed view is both proof the content held and the natural point at which a viewer rewatches, shares, or reads the caption. Long form analytics report the same idea as the final value of the retention graph.

It matters because recommendation systems treat a finished clip very differently from an abandoned one. Two clips with the same average retention can diverge if one loses everyone at the last moment and the other carries most viewers over the line, and the one that finishes tends to be shown more widely. For clippers that makes the ending a decision, not an afterthought.

The nuance is in the counting. Looping clips register rewatches that can push related figures past full, very short clips complete almost by default, and platforms label the metric differently: TikTok reports the share who watched the full video, while YouTube Shorts leans on viewed versus swiped away. A high completion rate on a seven second clip says less than a moderate one on a sixty second clip.

In practice the two levers are length and ending. Cutting a clip to end on the payoff line, rather than running into the next topic, is the single biggest improvement, and clipping tools that snap the out point to the end of a sentence do most of that work. Ending on a line that invites a rewatch or a reply adds the engagement that completion alone does not.