Most video ads don't fail at the end. They fail in the first second.
Someone's thumb is moving. Your ad has a moment to make it stop. Miss that moment and nothing else in the ad matters.
Hook rate is how you measure that moment.
The short version
Hook rate is the percentage of people who saw your video ad and watched at least three seconds of it.
Hook rate = 3-second video plays ÷ impressions × 100
Say an ad gets 10,000 impressions and 2,800 three-second plays. Its hook rate is 28%.
You'll also hear it called thumb-stop ratio. Same thing.
Why it's the first number I check
Cost per purchase is the number that pays the bills. But it's slow. An ad needs real spend before its cost per purchase means anything.
Hook rate is fast. Within a day or two you know whether people are stopping.
I call this the early read: a leading number you can trust long before the lagging one is ready.
An ad with a weak hook rate almost never recovers. The best offer in the world can't sell to someone who already scrolled past.
Its partner: hold rate
Hook rate tells you if people stopped. Hold rate tells you if they stayed.
Hold rate = ThruPlays ÷ 3-second video plays × 100
A ThruPlay is Meta's count of plays that reached 15 seconds, or the whole video if it's shorter than that.
Some people divide ThruPlays by impressions instead. Either works. Pick one and stay consistent, or your comparisons will mean nothing.
Reading the two together
Put the two numbers side by side and they tell you what to fix:
- Strong hook, weak hold: the opening promised something the rest of the video doesn't deliver. Keep the hook. Rework the body.
- Weak hook, strong hold: people who get past the start like it. Keep the body. Test new openings.
- Both weak: it's the concept, not the cut. Move on to a new angle.
- Both strong, still no sales: the ad is doing its job. Look at the offer and the landing page.
The numbers don't just grade the ad. They tell you which part to fix.
How to add it in Ads Manager
Hook rate isn't built in, but it takes a minute to add:
- In Ads Manager, open Columns, then Customize columns.
- Click Create custom metric and name it "Hook rate".
- Set the formula to 3-second video plays ÷ Impressions and the format to percentage.
- Repeat for hold rate with ThruPlays ÷ 3-second video plays.
- Save the column set so it's there every time.
What's a good hook rate?
Here's the honest answer: the best benchmark is your own account.
Rules of thumb get passed around in the ecommerce world. Roughly: under 20% is weak, 25 to 30% is solid, and above 30% is strong. Treat those as folklore, not research. Hook rate swings with placement, audience, and how much the platform autoplays.
So compare like with like. Reels against Reels, cold traffic against cold traffic, this month's ads against last month's.
The question isn't "is 28% good?" It's "is this better than what we usually get?"
The trap: hooking the wrong people
Hook rate is easy to game.
A loud noise, a shocking image, or a bait headline will stop plenty of thumbs. Most of those people were never going to buy.
A hook that stops everyone is worse than a hook that stops buyers. You pay for the attention either way.
The right people for three seconds > everyone for three seconds
That's why I always read hook rate next to click-through rate and cost per result, never alone.
How to raise it
1. Make the first frame do the work. Movement, a face, or on-screen text that names the viewer's problem. A logo is not a hook.
2. Start in the middle. Cut the intro. Open on the most interesting second of the video.
3. Design for sound off. Many people watch muted. Put the hook in text on screen, not just in the voiceover.
4. Break the pattern. Feeds are a wall of polish. Something that looks different gets a second look. I collected the patterns that keep winning in what hundreds of ad creatives taught me about hooks.
5. Test hooks, not whole ads. Keep the body, swap the first three seconds, and run three to five versions. It's the cheapest test in advertising, and the brief for it takes minutes with this creative brief template.
Pool rule 010Never run an ad you'd scroll past.
Quick answers
Is hook rate the same as thumb-stop ratio? Yes. Same formula, different name.
Does hook rate work for static ads? No. Statics don't have plays. Use click-through rate as the equivalent first read.
Why is it three seconds? Because Meta counts a "3-second video play". It's the closest standard measure of "they stopped scrolling".
Does a high hook rate guarantee sales? No. It means people watched. Whether they buy depends on the rest of the ad, the offer, and the page.
Is hook rate different on Reels? The formula's the same. The numbers often aren't, because full-screen autoplay behaves differently from the feed. Compare Reels with Reels.
The one-line version
Nobody buys from an ad they didn't watch. Hook rate tells you whether they watched.