Hook rate is the share of your video ad's impressions in which someone watched at least three seconds: 3-second video plays ÷ impressions × 100. On Meta, the largest recent dataset puts the average at about 25%. Below about 20% the opening is not stopping people; 30% or more is a strong opening. Hook rate judges the first three seconds only, so read it next to hold rate, CTR and cost per purchase before you call an ad a winner.
We read hook rate on every ad we make, one row per hook, because it is the earliest signal a video ad gives you. This guide covers the formula, how to add it to Meta Ads Manager (it is still not a standard column), what counts as good in 2026, and how to test hooks so the number means something.
What is the hook rate formula?
Hook rate = 3-second video plays ÷ impressions × 100. If an ad served 40,000 impressions and 10,400 of them turned into 3-second plays, the hook rate is 26%.
It has three close relatives. Keep your definitions fixed, because two people using the same word for different formulas is how bad decisions get made:
| Metric | Formula (Meta) | What it tells you |
|---|---|---|
| Hook rate | 3-second video plays ÷ impressions | Did the opening stop the scroll? |
| Hold rate | ThruPlays ÷ 3-second video plays | Of the people who stopped, how many stayed? (A ThruPlay is 15 seconds, or the whole video if it is shorter.) |
| Link CTR | Link clicks ÷ impressions | Did the video make people act? |
| Cost per purchase or lead | Spend ÷ results | Did it make money? The final word. |
Many teams use "thumbstop rate" as another name for hook rate. Some measure it on a shorter view. Either is fine as long as everyone reading the report uses the same one.
How do you add hook rate in Meta Ads Manager?
Meta still does not ship hook rate as a ready-made column, so you build it once as a custom metric:
- In Ads Manager, open Columns → Customize columns.
- Choose Create custom metric and name it "Hook rate".
- Build the formula: 3-second video plays ÷ Impressions, and set the format to percentage.
- Repeat for "Hold rate": ThruPlays ÷ 3-second video plays.
- Add both columns next to spend, link CTR, CPM and cost per result, and save the set as a preset so every report opens the same way.
Read it at the ad level, not the campaign level. A campaign average hides the one hook that is carrying the rest.
What is a good hook rate on Meta in 2026?
The most useful public number comes from Billo's analysis of 88,329 sales-objective Meta video ads in the first half of 2026: the average hook rate was 25.44%. In that data, anything below 21% was lower than every industry's average, and 30% or more was above every industry's average.
Here is how we read it in practice:
| Hook rate | What it usually means | What to do |
|---|---|---|
| Under 20% | The opening is not stopping people | Replace the first three seconds; keep the body if it was never seen |
| 20–25% | Below or around average | Test new openings against it |
| 25–30% | Around average | Fine if CTR and cost per purchase hold up |
| 30%+ | Strong opening | Now check that the body and offer convert |
Two cautions. First, benchmarks move with placement, audience and objective: retargeting audiences who already know you stop more easily than cold ones, and Reels and Feed behave differently. Second, the best benchmark is your own account's median for the same placement and audience. An ad at 24% in an account where everything else sits at 18% is a winner.
Why can a high hook rate still lose money?
Because hook rate only measures the opening. Billo found that hook rate did not predict CTR on its own: a strong opening earns you the chance to sell, it does not make the sale. A shock opening that has nothing to do with the product will post a great hook rate and a terrible cost per purchase.
Read the numbers in order, top to bottom, and fix the first one that breaks:
| What you see | Likely cause | What to change |
|---|---|---|
| Low hook rate, normal CPM | The first three seconds | New opening: a different first visual and first line |
| Good hook rate, low hold rate | The promise is not paid off quickly, or the middle drags | Show the product sooner, cut slow lines, tighten the edit |
| Good hook and hold, low CTR | The offer or the call to action is unclear | Clearer offer, proof before the ask, one specific CTA |
| Good CTR, high cost per purchase | Landing page, price, or audience | Usually not a creative problem; check the page and the offer |
How do you test hooks properly?
Change only the opening. If two ads differ in hook, body, offer and music, a better hook rate tells you nothing about the hook. The clean test is one body, several openings: same product demo, same proof, same CTA, with only the first three seconds different.
That is why every master video we produce ships with three hooks. Each month a plan delivers six master videos with three hooks each, so you get eighteen ad variations where the body is held constant inside each set of three. Then:
- We read hook rate and CTR for every hook variation by day 7.
- Winning hooks are identified within 7–14 days.
- Cost per purchase and ROAS confirm the decision once enough conversions have come through.
- The monthly report has one row per live hook, with a decision for each: scale, keep, kill or refresh.
Give each variation enough delivery before you judge it. With a few hundred impressions, a hook rate can swing ten points on noise; wait for a few thousand per variation. The full loop is on how we work, and the four-week version is in our UGC test plan for D2C brands.
What makes a strong hook in the first three seconds?
The opening has to work as a still frame, with the sound off, in under three seconds. The patterns that hold up across categories:
- The problem, in the customer's words. "My hair fall got worse every monsoon" beats "Introducing our new serum".
- The result first. Open on the after: the clean counter, the finished dish, the dashboard with the number.
- Call out who it is for. "If you run a clinic in a tier-2 city…" makes the right people stop and lets the wrong ones scroll.
- A claim the body proves. Contrarian is fine, as long as the next ten seconds back it up.
- Motion in the first frame. A hand, a pour, a product being used, not a logo on a plain background.
- Text on screen. Many feed plays are muted, so the first line should also be readable as on-screen text, inside the safe zone.
More worked patterns are in 12 UGC hook patterns for Meta Advantage+.
When should you refresh a hook?
When an ad's hook rate falls week on week while its frequency rises, the audience has seen the opening too many times. That is fatigue, and the cheapest fix is usually a new opening on a body that still converts, not a whole new video. For retainer clients we turn around a single hook-refresh variation for a fatigued ad in 5–7 days.
Frequently asked questions
Is hook rate the same as thumbstop rate?
Usually, yes. Most teams use both names for 3-second video plays ÷ impressions. Some define thumbstop on a shorter view. Pick one definition and use it in every report.
Does hook rate apply to image ads?
No. It needs video plays, so it only exists for video ads. For images, link CTR is the closest early signal.
My hook rate is high but sales are low. What is wrong?
The opening works and something after it does not. Check hold rate, then CTR, then the landing page, in that order. A hook that promises something the video never shows will also produce this pattern.
How many hooks should each video have?
We ship three per master video. That is enough to find a clearly better opening without spreading the budget so thin that no variation gets a fair read.
How soon can I judge a new hook?
Hook rate and CTR give a usable read within about a week if each variation has had a few thousand impressions. Cost per purchase needs longer, because it waits for conversions.
Where to go next
Every plan is six master videos with three hooks each, delivered in 9:16, with the first set in 15–20 days and every set after in 15 days. See pricing, look at results from live campaigns, or book a strategy call and bring your current hook rates.