What watch time actually measures
Watch time tells you how many seconds of a video people gave you. Ad platforms report it two ways: added up across everyone who saw the ad, and divided per view as an average. Seconds are harder to flatter than views, because a view can be credited after two or three seconds while watch time keeps counting for as long as somebody stays. For a brand buying creator video, it is the clearest read on whether the content held attention or merely interrupted it.
A view and a second are not the same unit. Each platform sets its own bar for what counts as a view, and those bars have moved over the years, so read the current help documentation before you trust a comparison. Hook rate uses that short threshold on purpose, to judge the opening alone. Watch time asks a different question: once someone stayed, how long did they stay.
Keep the two versions straight. Total watch time is a volume number that grows with spend, so it belongs in a reach summary and nowhere near a creative comparison. Average watch time per view is the one media buyers argue about, because it holds steady when budget changes and moves when the edit changes. Both are usually available as columns you add yourself.
Seconds against percentage watched
Divide average watch time by the length of the cut and you get the share of the video watched, which is where short vertical video plays tricks on you. A nine-second clip only has nine seconds to lose, so its percentage looks strong almost by default. A long demo can hold somebody for four times as many seconds and still report a smaller fraction. The one you quote should match what you are trying to learn.
Here is an illustrative pair, not a benchmark. Version A runs ten seconds and averages six, so sixty percent of the video got watched. Version B runs forty seconds and averages twelve, so thirty percent. B looks weaker on the percentage and yet each viewer spent twice as long with the product. If you are selling something that needs explaining, B is probably the better asset. If you want cheap attention at scale, A is.
What the number changes in your brief
Watch time is most useful as a production instruction. If the average sits at three or four seconds, the opening is the problem and nothing later in the video is being seen, so rewrite the first line rather than the offer. If the average lands halfway through, people are leaving at a specific moment: usually a claim that needs proof, a pause, or a second feature nobody asked for. Find that moment in the drop-off chart and cut it.
It also settles arguments about length. If nobody reaches second twenty, a thirty-second cut wastes ten seconds of paid delivery; end the video where the attention ends and put the ask just before it. Book creators for material rather than for one render: alternate opening lines, a longer take of the demo, and B-roll you can re-cut into shorter versions. A tight script template beats another round of polish.
Where the number misleads you
Be careful with a creator's organic numbers. Watch time on their own posts comes from an audience that chose to follow them, often with sound on, in a feed where the algorithm already decided those viewers would like it. Your ad meets strangers who chose nothing. Read their retention the way you read engagement rate on a media kit: useful context about whether they can hold a room, not a prediction of how the same face performs against cold traffic.
Treat the figure as comparable only against itself. Reporting definitions differ between platforms and get revised, so seconds from one buying tool and seconds from another do not belong in the same ranking; the last month or two inside your own ad account is the baseline worth beating, and the bar shifts with vertical, placement and video length. Then pair it with a money number such as cost per acquisition, because a video can hold people for ages and sell nothing.
How it's used
Average watch time on Priya's demo is nine seconds against six on the studio cut, same placement, same audience, so I'm moving most of tomorrow's budget behind her version.
In the brief: keep the product on screen from about second four to second ten, because that is where most people are still watching before they drop off.