Watch time

Watch time collapses two things into one figure: how many people a video reached, and how long each of them stayed. A clip seen by ten thousand people for five seconds and one seen by a thousand for fifty seconds produce identical watch time, which is precisely the point — both held the same amount of human attention.

Why platforms optimise for it

Attention is what a platform sells. Watch time is the closest direct measure of how much of it a piece of content generated, so it is the number recommendation systems are tuned to increase. Views alone can be inflated by a strong thumbnail or an aggressive hook that does not deliver; watch time cannot be inflated the same way.

Watch time and retention are not the same

• →Retention is a percentage — how far through, on average.
• →Watch time is an absolute — how many hours in total.
• →A short clip can have superb retention and modest watch time.
• →A long video can have poor retention and enormous watch time.
• →Growing either one grows the other, which is why they get confused.

What it means for clipping

Clips are a watch-time strategy applied at scale. A single long episode has a ceiling on how many people will start it. Twenty clips from that episode each get their own chance in a feed, and the aggregate attention they collect usually exceeds what the source could have gathered alone — which is the entire economic argument for clipping.