What is a good hook rate on Meta ads?

What hook rate measures

Hook rate is the percentage of people who watched at least 3 seconds of your video ad out of everyone the impression reached. The formula is 3-second video views divided by impressions.

It measures exactly one thing: whether the first moments of the creative stop the scroll. Nothing about the offer, the product, or the landing page has entered the picture yet. That isolation is what makes it useful. When an ad dies, hook rate tells you whether it died at the first frame or somewhere later, and those are different problems with different fixes.

Meta does not surface hook rate as a default column. Build it as a custom metric in Ads Manager: 3-second video plays divided by impressions. Save it once and it is available on every report afterward.

Keep it distinct from hold rate, its natural companion. Hook rate asks whether people stopped; hold rate asks whether they stayed once they stopped. An ad with a strong hook and a weak hold has an opening worth keeping and a body worth reshooting. An ad weak on both is not worth iterating at all.

The thresholds

These are the working lines I use across accounts:

• –Below 25%: kill it. The creative is not stopping the scroll, and everything downstream of the first three seconds is starved of traffic. Iterate on the hook, not the body.
• –25 to 30%: the middle zone. Not bad enough to kill on the hook alone, not good enough to prove the hook works. Read it together with hold rate, CTR, and CPA before deciding.
• –30 to 40%: good. The hook is doing its job. If the ad still underperforms, the problem is later in the funnel: the body of the video, the offer, or the landing page.
• –Above 40%: scale it. Hooks this strong are rare, and they are the raw material for a whole family of variations. Cut new bodies onto the same opening before the account fatigues it.

How to calculate it, with a worked example

Take an ad that has served 100,000 impressions and logged 34,000 3-second video plays. Hook rate is 34,000 divided by 100,000, which is 34%. That is inside the good band: leave the hook alone and judge the ad on what happens after the first three seconds.

Now a second ad on the same product: 100,000 impressions, 21,000 3-second plays, a 21% hook rate. That is below the kill line. It does not matter how good the offer inside the video is, because almost nobody is reaching it. The fix is a new opening, not a new landing page.

The comparison is the point. Two ads can have identical CPAs today and completely different futures: the 34% ad has room to find its audience, while the 21% ad is paying an entry toll on every impression.

When the thresholds do not hold

Caveats, because a benchmark applied blindly is worse than no benchmark:

• –Static images and carousels have no hook rate. The metric only exists for video. Judge statics on CTR and downstream metrics instead.
• –Small impression counts make the ratio noisy. A hook rate computed on a few hundred impressions can swing wildly; let the ad accumulate real delivery before you act on it.
• –Retargeting audiences inflate hook rate. People who already know the brand stop for reasons that have nothing to do with the hook. Hold these thresholds to cold, prospecting traffic.
• –Optimization objective changes the audience. Engagement-optimized campaigns are shown to people who watch everything, so their hook rates read high. These lines assume conversion-optimized delivery.
• –Placement mix shifts the baseline. Sound-off feed scrolling and full-screen Reels autoplay behave differently. Compare hook rates within a placement before comparing across accounts.
• –A high hook rate with a bad CPA usually means the hook is writing a check the product does not cash. That is a message-match problem. The hook did its job; the ad still fails.

What to do with it

Hook rate is a diagnostic, not a verdict. It tells you why an ad is dying, not which ads deserve your attention in the first place. For that, plot the account on a log-spend scatter and let significance bands separate the ads worth reading from the noise, then use hook rate to explain the ones that land in the kill zone.

In practice that means reviewing hook rate at the creative level on a regular cadence, sorting new launches by it once delivery is real, and feeding the winners back into production: every hook above the scale line becomes a brief for the next batch of variations. The metric earns its keep as an input to the creative pipeline, not as a scoreboard.