What is hook rate?

The usual formula is 3-second views ÷ impressions × 100. If a clip is shown 10,000 times and 3,200 of those viewers are still watching at 3 seconds, its hook rate is 32%. The other 68% scrolled on before the clip had a chance.

It's a number advertisers work out from their ad reports, so the cut-off varies: most use 3 seconds, and TikTok's ad reports count 2-second and 6-second video views instead. Its partner is hold rate, the share of those viewers who keep watching for longer. A high hook rate with a low hold rate means the opening worked and the middle didn't.

For clips you post rather than pay to promote, no app shows a “hook rate.” Each shows its own first-seconds number:

YouTube's Shorts metric is the closest to a true hook rate, and it matters most there: YouTube says it ranks Shorts using the share who chose to view, average view duration and average percentage viewed. The full story is in swipe-away rate. Whatever the app, the number is decided by the clip's first line and first frame, which are set the moment the clip is cut.

None of the apps publishes a target for organic videos, and ad benchmarks don't carry over: an ad is shown to people who didn't choose it, a clip in a feed to people the app expects to like it. The useful comparison is with yourself. YouTube's own advice for Shorts is to compare videos of the same type and to look at the ones that did badly as well as the ones that did well.

• Well under your usual: the opening is the problem. Start later, on the line that matters, or give it a text hook.
• About usual, but few views: the opening held; look at average watch time and the ending.
• Well above your usual: study how it opens, and find more moments that open the same way.

Test one change at a time: two clips of the same kind, two different openings, posted at similar times. Compare the first-seconds number, not the views, which depend on much more than the hook.

A hook rate is measured after you post. A hook score is a prediction made before: AI clipping tools put one on each clip they cut. Clyper's runs from 0 to 100 and rates how likely a clip's opening is to stop someone scrolling and the clip to keep them to the end, read from what's said. It's strict: 75 or more counts as strong, 90 or more is rare. How the tools' scores compare is in what a virality score is.

Use the two together. Post in score order, then note each clip's score beside its first-seconds number. After a few weeks you'll know how far the score predicts your audience, and which kinds of opening beat it. A strong score with a weak hook rate usually points at what the score can't see: a slow first frame, or captions that start late.