How hook rate is calculated
Take the number of times a video was played for at least three seconds and divide it by the number of impressions the ad served. Meta does not ship this as a default column, so you build it yourself: create a custom metric in Ads Manager using the three-second video plays column and the impressions column, or export both and calculate it in a spreadsheet.
Because impressions sit in the denominator, hook rate measures the whole audience the ad reached, not just the people who chose to engage. That is the point of it. It answers one narrow question — of everyone whose feed this ad appeared in, how many did not scroll straight past.
Why hook rate matters
Feed advertising is decided in the opening moment. If the first frames do not register, nothing later in the video gets a chance, and the budget spent on that impression is gone. Hook rate isolates that moment and separates it from everything else the ad is trying to do, which makes creative testing far quicker.
It is also the metric that stops teams blaming the wrong thing. When leads dry up, the instinct is to rewrite the offer or change the audience. If hook rate is weak, neither will help — almost nobody reached the offer. Fixing the first seconds is cheaper than either, and it can be tested with a new opening cut rather than a new production.
Common mistakes with hook rate
The first is dividing by reach or by video plays instead of impressions, which produces a different, flattering number that cannot be compared with anyone else’s. Agree the formula once and keep it fixed.
The second is comparing hook rate across placements. Reels and Stories are full-screen and sound-on by default, feed is not, and each placement produces a different band of results for the same creative. Read the figure inside a placement breakdown, not as a single blended number.
The third is optimising the hook in isolation. A shock opening, a fake notification sound or a misleading first frame will lift hook rate and damage everything after it, because the people it stops were never the right people. Always read it beside hold rate and cost per result.
How to act on it
Treat the opening as its own testable asset. Keep the body of the video fixed and cut several different first seconds against it: the product in use, the customer stating the problem, an on-screen question, a bold caption. Run them together and let hook rate tell you which stopped more people, then check that the winner did not lose ground further down the funnel.
Practical fixes that usually move it: show something recognisable in the first frame rather than a logo animation, burn in captions so a muted video still makes sense, and shoot vertically so the ad fills a phone screen. In markets like Nepal where nearly all impressions are mobile and often on a slow connection, a heavy opening sequence costs you viewers before the message starts. If you are rebuilding creative from scratch, the production side is covered under Meta ad creative.
