Hook rate formula and how to calculate it
Last updated October 10, 2026. Column names checked against Meta Ads Manager on that date.
Hook rate = 3-second video plays ÷ impressions × 100. It is the share of people served a video ad who were still watching after three seconds. An ad with 40,000 impressions and 9,200 three-second plays has a hook rate of 23%. Both numbers come from Ads Manager, for the same ad and the same date range.
In short
- The formula has two inputs and Meta ships both as columns; it does not ship the ratio.
- Read both inputs at the ad level with one fixed date range, or the result drifts.
- Hold rate divides ThruPlays by the 3-second plays the hook earned; read the two together.
- For what counts as a good hook rate, see the benchmark guide linked below. This page is the formula.

The formula, with a worked example
Take the ad's impressions for the period you care about. Take the number of times a play lasted at least three seconds in the same period. Divide the second by the first and multiply by 100.
| Input | Ads Manager column | Example |
|---|---|---|
| Denominator | Impressions | 40,000 |
| Numerator | 3-second video plays | 9,200 |
| Result | 9,200 ÷ 40,000 × 100 | 23% |
Three details change the answer, so fix them before you compare two ads. Use the ad level, not the ad set or campaign, because a campaign aggregates ads with different hooks. Use one date range for both columns. And use impressions, not reach: an impression is each time the ad was served, and the three-second play is counted per serving, so the two match.
To run these formulas on your own ad, open the hook rate calculator: paste 3-second plays and impressions from Ads Manager to get hook rate, hold rate and cost per three-second play.
Where the two numbers sit in Ads Manager
Meta does not ship a column called hook rate. You build it from two columns it does ship.
- Open Ads Manager at the ad level and set the date range you want to measure.
- In the Columns menu, choose the Video engagement preset. It exposes the video play columns.
- Read 3-second video plays and Impressions for the same ad, then divide and multiply by 100.
Meta documents what counts as a video play in the Meta Business Help Center, and the counting rules differ by placement and by whether sound is on. Check there if a number looks wrong. Column names checked on October 10, 2026.

Hook rate, hold rate and thumb-stop ratio
These three names are used interchangeably in benchmark posts, which makes their numbers hard to compare. None of them is an official Meta metric. Each is built from Meta's video play columns, and different teams build them differently. Check which numerator a source used before comparing your own number with it.
| Metric | Common formula | On the example ad | What it tells you |
|---|---|---|---|
| Hook rate | 3-second plays ÷ impressions | 9,200 ÷ 40,000 = 23% | Whether the opening frame stopped the scroll |
| Hold rate | ThruPlays ÷ 3-second plays | 3,600 ÷ 9,200 = 39% | Whether the rest of the video kept the people the hook won |
| Thumb-stop ratio | Usually identical to hook rate | 23% under another name | Same signal; check the source's denominator |
The ThruPlay figure above is a worked example, not a benchmark. A ThruPlay is a play to completion or to 15 seconds, whichever comes first. Some sources divide ThruPlays by impressions instead, which gives 9% on this ad and folds the hook into the hold number; check the denominator before comparing.

Reading hook and hold together is more useful than either alone. A high hook rate with a low hold rate means the opening promised something the rest of the ad did not deliver. A low hook rate with a high hold rate means the few people who stayed were the right people, so the creative is fine and the opening frame is not. The ranges that count as low and high, with each source named, are in what is a good hook rate on Meta ads. That guide also covers the four-quadrant reading of hook against hold. This page stays with the formula.
Why the first three seconds carry so much
The three-second mark is where a viewer decides to keep watching or to scroll. Everything after it, the offer, the proof, the call to action, is only seen by the share of people the hook kept. That is why a weak hook wastes spend regardless of how good the rest of the ad is: the budget buys impressions, and most of those impressions never reach the part of the ad that sells.
It is also the most testable part of an ad. The hook can be changed without touching the offer or the landing page, so two versions of the same ad that differ only in their first three seconds give a clean read. A Meta creative testing workflow turns that into a repeatable process, and founder-led selfie video is the cheapest format to test new hooks in.
Three mistakes that produce a wrong hook rate
| Mistake | What goes wrong | Fix |
|---|---|---|
| Reach as the denominator | Reach counts people, plays count servings; the ratio inflates as frequency rises | Use impressions |
| Campaign-level columns | Ads with different hooks are averaged into one number | Read at the ad level |
| Mixing placements | Reels, Stories and Feed start the play clock under different conditions | Break down by placement before comparing |
How to raise a low hook rate
Change the first three seconds, keep everything else fixed, and run the variants side by side. The 2026 Meta ads A/B testing guide covers test structure and when a difference is real. AdRiseLab's Hook Diversity Matrix generates 8 hooks across 5 angles per creative, which gives 40 opening variations to test instead of one guess at a time, and the weekly performance report shows which opening kept people watching. Plans from $39/mo.
Frequently Asked Questions
What is the hook rate formula?
How do I calculate hook rate in Meta Ads Manager?
What is the difference between hook rate and hold rate?
What is the hook rate and hold rate formula for the same ad?
Is thumb-stop ratio the same as hook rate?
Does hook rate apply to image ads?
Why does my hook rate differ between placements?
Related concepts
Founder & CEO, AdRiseLab
Performance marketer turned product builder focused on Meta advertising, creative workflows, and measurement. Founded AdRiseLab to reduce the research-to-publish bottleneck in Meta advertising.