Hook rate is the share of ad impressions where someone watched at least the first three seconds of your video: 3-second video plays divided by impressions, times 100. Most published ranges put a solid Meta hook rate between about 20% and 40%, and one ad we've logged ran at 45.9%.
A high hook rate tells you the opening stopped the scroll.
It doesn't tell you the ad sold anything.
This page covers the hook rate formula, its partner metric hold rate, what good looks like in our archive compared with published benchmarks, whether hook rate predicts CPA, and what has actually raised it.
The hook rate formula: 3-second views ÷ impressions
The formula most media buyers use, and the one Motion's glossary publishes, is:
Hook rate = (3-second video plays ÷ impressions) × 100
Hook rate is also called thumb-stop ratio or thumbstop rate.
They're the same number: how often your opening stopped someone long enough to count as a view.
Worked example: a video ad gets 40,000 impressions and 11,200 3-second video plays.
11,200 ÷ 40,000 = 0.28, so the hook rate is 28%.
Out of every 100 times the ad showed up, 28 people watched past the first three seconds.
Some things to know about the inputs:
- What Meta counts as a 3-second play. Meta defines 3-second video plays as the number of times your video played for at least three seconds, or for nearly its whole length if it's shorter than three seconds. Each impression is counted separately and replays are excluded.
- It isn't a default column. Ads Manager won't show "hook rate." Build a custom metric (3-second video plays ÷ impressions) or export both columns and divide in a sheet.
- TikTok uses different cutoffs. TikTok Ads Manager reports 2-second and 6-second video views, not 3-second plays. A TikTok "hook rate" built on 2-second views will run higher than a Meta one built on 3-second plays, so don't compare them directly.
- Use the same window for both numbers. Same date range, same ad, same placement mix. Mixing a Reels-only number with a feed-only one gives you a meaningless average.
If you track hook rate across dozens of ads, doing the division by hand gets old fast.
Our roundup of creative analytics tools covers tools that calculate hook rate and hold rate per ad and line them up against spend and conversions.
Hold rate: what happens after the hook
Hook rate measures the first three seconds. Hold rate measures whether people kept watching after that. The common version is:
Hold rate = (ThruPlays ÷ 3-second video plays) × 100
Meta counts a ThruPlay when a video is played to completion, or for at least 15 seconds if it's longer than that.
So hold rate answers: of the people your hook caught, what share stayed to about the 15-second mark?
Continuing the example: if those 11,200 3-second plays produced 2,240 ThruPlays, hold rate is 2,240 ÷ 11,200 = 20%.
Not everyone defines hold rate the same way.
Some teams divide 15-second views by 3-second plays, as in this Jetfuel benchmark roundup.
Others divide ThruPlays by impressions. Pick one definition, write it down, and keep it the same everywhere you report it.
Hook rate vs hold rate, in short:
- Low hook, any hold: the opening is the problem. People scroll past before the ad starts.
- High hook, low hold: the opening works but the body doesn't keep the promise. People stop, then leave.
- High hook, high hold, weak sales: the video is watchable but isn't selling. Look at the offer, the call to action and the landing page.
Atria makes the same point in its guide to creative analysis: when the hook is strong but conversions are weak, "the fix is often mid-creative, not the first three seconds."
What a good hook rate looks like: real entries vs published ranges
Published benchmarks for a good hook rate on Facebook ads mostly fall in the same band:
- Motion says hook rates "vary wildly by industry and audience, but 20-40% is generally solid performance."
- Jetfuel's 2026 roundup, citing AdSights and Taylor Sicard, puts the median hook rate around 23% for Meta feed prospecting and around 30% for Reels, with 30%+ counted as strong. For hold rate it gives a feed median around 18% (a 12-25% range) and about 23% on Reels.
Treat those as rough guides, not targets. They're agency and tool benchmarks, not platform numbers, and Jetfuel itself notes they vary by vertical, placement, audience temperature and objective.
Now the real numbers. The clearest hook rate in our database is a 45.9% hook-rate ad logged by @alexgoughcooper: a Meta video ad for a beauty brand built around a makeup teacher hook, which ran at a 45.9% hook rate and a $47.13 CPA.
That sits above the top of Motion's 20-40% "solid" band and well above the Reels median above.
It's also one of only a few results in our archive that report a hook rate next to a cost result.
Most people who post wins share CPA, ROAS or spend, not attention metrics.
So we can't yet give you a median hook rate from our own data, and we won't invent one.
What we can say: when a hook rate does get posted with a result, it's usually because it was unusually high.
Useful rules of thumb when you judge your own number:
- Compare against your own account first. Your median hook rate across the last 20-30 video ads is a better baseline than any published range.
- Split by placement. Reels and Stories usually run higher than feed. A blended number hides that.
- Wait for enough impressions. A hook rate on 800 impressions can swing a lot. Judge it once the ad has had a few thousand.
Hook rate vs CPA: does it predict results?
Partly. Hook rate is an early signal, not a result.
It tells you an ad has a chance, but plenty of ads with great hooks never convert.
The reason is simple math. Your CPA depends on what you pay for impressions, how many people click and how many of those buy.
Hook rate only touches the first step: getting someone to pay attention.
A shocking opening that has nothing to do with the product can win the scroll and still lose the sale.
(For the full breakdown of what goes into the number, see our guide to CPA.)
The results in our archive point the same way.
The ads that cut acquisition costs did it with hooks that were tied to what the product solves, not just eye-catching openings:
- The makeup teacher video from @alexgoughcooper paired its 45.9% hook rate with a $47.13 CPA. High attention and a workable cost, from the same ad.
- A test shared in the Janky Patel newsletter moved an 8-figure DTC brand toward problem-led hooks with broad targeting and logged −10.5% CAC and +20.5% ROAS over 75 days. The lever was what the hook said, not how loud it was.
- MHI Media (Kamal Razzak) reported −62% CPA from a founder car-selfie video: a plain, native-looking opening rather than a polished one.
- inBeat's case study for Hurom logged −65% CPA and +300% ROAS from UGC built around a health-benefit angle.
Only the first of those reports a hook rate, so they don't prove hook rate predicts CPA.
They do show that the openings which lowered costs were about the buyer's problem.
Use hook rate to sort out openings that fail, then let CPA decide the winner.
A practical way to use both: if a new ad's hook rate is far below your account median after a few thousand impressions, it rarely recovers, so cut or re-edit it.
If hook rate is high but CPA is bad, keep the opening and rework the body, offer or landing page.
How to raise your hook rate
Hook rate lives almost entirely in the first one to three seconds: the first frame, the first line, the first movement.
These are the changes that show up most often in our hook tests with results:
- Lead with the problem. Open on the pain the product fixes, said in the buyer's words. The problem-hook test above is the clearest example in our archive.
- Look native, not produced. Founder-shot, selfie-style and UGC openings blend into the feed. The founder car-selfie video is one example.
- Put the payoff in the first frame. Show the result, the product in use or a before-and-after right away, not a logo or a slow build-up.
- Add on-screen text to the first second. Many people watch with the sound off. A short text line that states the hook means the opening works silently.
- Test hooks, not whole new ads. Keep the body the same and swap only the first three seconds. You'll learn which opening works without changing everything else. Our creative testing guide covers how to set that up.
- Reuse footage you already have. @alexgoughcooper's old-footage B-roll mashup became a top spender at a $58.61 CPA on $107k spend. New hooks don't always need new shoots.
- Refresh when it decays. Hook rate usually drops as the same audience sees an ad again. A falling hook rate on a steady ad is often an early sign of ad fatigue.
For more on which changes moved the number, see what raised hook rate across the results we've collected, and our list of the best ad hooks.
If you'd rather have someone who writes and tests hooks every week do this for you, we can match you with a media buyer or creative team.