Hook rate: how many people stop in your first second

Hook rate is the share of people who stayed through the first three seconds of your video, out of everyone the video was shown to. It separates "the ad was delivered" from "someone stopped to look". In an infinite feed, those three seconds decide the rest.

Formula

Hook rate = 3-second video plays ÷ impressions × 100.

A video with 12,000 impressions and 3,600 three-second plays has a 30% hook rate. Both columns come from Ads Manager; the division is yours, because Meta does not ship the metric ready-made.

Two caveats that change the reading:

• Use impressions, not reach. Reach counts people, impressions count deliveries, and plays happen per delivery.
• Placements behave differently. Reels, with autoplay and sound on, produces a number that does not compare to in-feed video. A report that mixes both hands you an average describing neither.

What an acceptable number looks like

Above 30% signals an opening that works. Between 20% and 30% is common ground. Below 20%, the problem is almost always the first frame, not the offer.

But hook rate alone proves nothing commercial. It measures attention, not buying intent. A video that opens with a shock and hits 45% while selling nothing is a worse outcome than a 22% hook that sells. Always read it next to CTR and the campaign's cost per result.

What makes the number drop

• The first frame says nothing. Logo, black screen, animated intro: each one spends the second you had.
• The ad depends on sound. Most of the feed plays muted. Without captions or on-screen text, your opening does not exist for the person scrolling.
• Wrong format for the placement. Horizontal video in Reels loses half the screen, and with it half the chance of stopping the thumb.
• The audience has seen it. Rising frequency drags hook rate down before it touches any other metric. It is the first symptom of creative fatigue.

How to use it

Hook rate is a creative-testing metric, not a daily tracking one. The right use is comparing openings: three versions of the same video, changing only the first three seconds, running at the same time against the same audience. Whichever holds more people becomes the default, and the next round tests something else.

When the hook is high and sales are not coming, the problem moved. Look at hold rate to see whether people stay long enough to hear the argument, and at the landing page's conversion rate to see whether the destination delivers what the video promised.

To build the full video reading, with what to check before and after the hook, see the paid media reporting page.