Learn/Reading results
How to read hook rate and hold rate
Two ratios most teams use to read a video ad, worked out from Meta’s own columns. What they count, how to compare them fairly, and what they can and cannot tell you about the next ad to make.
The short answerHook rate is the share of impressions where the video played for at least three seconds: 3-second video plays divided by impressions. Hold rate is the share that kept watching: ThruPlays (played to the end or for at least 15 seconds) divided by 3-second plays. Neither is a Meta column; both are worked out from Meta’s. Compare them only against your own ads, in the same placement and of similar length, and use them to decide which part of the next ad to change.
- Meta counts a 3-second video play when the video plays for at least three seconds, and a ThruPlay when it plays to the end or for at least 15 seconds.1
- Hook rate and hold rate are ratios teams work out from those columns. Write down which formula you use and keep it.
- Compare an ad with your own past ads, same placement, similar length. Numbers from other brands are not comparable.
- The plays at 25%, 50%, 75%, 95% and 100% show where people stopped watching.1
On this page
Most teams read a video ad with two numbers: how many people stopped scrolling, and how many stayed. They go by many names (hook rate, thumb-stop ratio, hold rate) and are worked out in slightly different ways in different teams. That is fine, as long as you know what yours counts and compare like with like. Used that way, they point at the part of the ad to change next. Used any other way, they produce arguments.
- Ads Manager, with the video columns added: impressions, 3-second video plays, ThruPlays, and plays at each percentage.
- Your own past video ads in the same placement, for comparison.
- The ad itself, to watch beside its numbers.
Step 1Know what Meta counts
Both ratios are made from Meta’s own video columns, so start with exactly what those columns count.
| Meta’s column | What it counts |
|---|---|
| 3-second video plays1 | Plays of at least 3 seconds, or nearly the whole video if it is shorter; replays are not counted |
| ThruPlays1 | Plays to the end, or for at least 15 seconds |
| Video plays at 25%, 50%, 75%, 95%, 100%1 | Plays that reached that share of the video, including ones that skipped to it |
| 2-second continuous video plays1 | Plays of 2 continuous seconds or more |
From Meta’s Business Help Center, About video ad metrics on Facebook and Instagram.1
Note the 15-second rule in ThruPlays: on a 30-second video, a ThruPlay can be someone who watched half of it. On a 10-second video, it is someone who watched it all. That is one reason ads of very different lengths cannot be compared on hold rate.
Step 2Work out the two ratios, one way, every time
Pick a formula for each, write it down where the team will see it, and never mix them. These are the common ones:
| Ratio | Formula | What it tells you |
|---|---|---|
| Hook rate | 3-second video plays ÷ impressions | How often the opening stopped the scroll |
| Hold rate | ThruPlays ÷ 3-second video plays | Of those who stopped, how many kept watching |
Common formulas, made from Meta’s columns. Some teams divide ThruPlays by impressions instead; choose one and keep it.
Example numbers. An ad with 50,000 impressions, 12,000 3-second video plays and 3,000 ThruPlays has a hook rate of 12,000 ÷ 50,000 = 24%, and a hold rate of 3,000 ÷ 12,000 = 25%.
Wait for enough impressions before reading either. Meta itself does not show its ad relevance diagnostics for ads with fewer than 500 impressions, because the figures are not reliable below that;2 a ratio worked out from a few hundred impressions moves a lot with a few more.
Step 3Compare against your own ads, like with like
A hook rate is only high or low compared with something. Compare it with your own past ads in the same placement, of a similar length, shown to a similar audience, over a similar number of days. Published figures from other brands are measured on different products, audiences and formulas, and say nothing about yours.
| Ad | Placement | Length | Hook rate | Hold rate |
|---|---|---|---|---|
| Last month, ad A | Reels | 15 s | 22% | 28% |
| Last month, ad B | Reels | 15 s | 19% | 31% |
| This month, new ad | Reels | 15 s | 27% | 20% |
Example. Three ads, compared fairly: same placement and length. The new opening stops more people; fewer of them stay.
Step 4Read where people stopped watching
The two ratios say how many; the percentage columns say when. Line them up in order (plays at 25%, 50%, 75%, 95%, 100%) and look for the steepest drop between two of them. That is the part of the ad where people left. Then watch that part of the ad.
| Point in the video | Plays | Share of 3-second plays |
|---|---|---|
| 3 seconds | 12,000 | 100% |
| 25% | 7,800 | 65% |
| 50% | 4,200 | 35% |
| 75% | 3,300 | 28% |
| 100% | 2,900 | 24% |
Example. The steepest drop is between 25% and 50%: whatever happens in the second quarter of this ad is where most people leave.
Watch the ad with the curve beside it and find what happens at the drop: a slow shot, a change of scene, the moment the product or the price appears. Often the drop is not a fault at all. If people leave once the price is shown, the ad has done its job for those who were not going to buy, and the next change is better spent elsewhere.
Step 5Decide which part of the next ad to change
The point of reading the numbers is to choose what to change in the next ad, not to name a verdict on this one. A low hook rate points at the opening: the first frame, the first words, the first movement. A low hold rate with a good hook rate points at what comes after the opening: the moment people saw what the ad was and left. The curve narrows it to a part of the ad.
None of this says in advance how the next ad will do. It tells you which change is worth trying, and the next ad’s own numbers will tell you whether it helped.
Common mistakes
- Comparing your hook rate with a figure from another brand.Different product, audience and formula.
- Mixing formulas between ads.A hold rate over impressions and one over 3-second plays are different numbers.
- Reading the ratios after a few hundred impressions.They move a lot with each new day.
- Comparing a 10-second ad with a 30-second one on hold rate.ThruPlays count 15 seconds, so length changes what they mean.
Questions
Is hook rate a Meta metric?
No. Meta reports 3-second video plays and impressions; hook rate is the ratio of the two, worked out by the team reading them. The same is true of hold rate, which most teams work out from ThruPlays.1
What is a good hook rate?
Better than your own recent ads in the same placement and of the same length. Figures from other brands are measured on different products, audiences and formulas, so they say nothing about whether yours is good.
Why does a ThruPlay not always mean someone watched the whole ad?
Because Meta counts a ThruPlay when a video is played to the end or for at least 15 seconds. On a video longer than 15 seconds, someone can stop before the end and still count.1
Sources
- About video ad metrics on Facebook and Instagram, Meta Business Help Center. 3-second video plays, ThruPlays and plays at 25% to 100%.Checked 26 Oct 2026
- About ad relevance diagnostics, Meta Business Help Center. Quality, engagement rate and conversion rate rankings, and none below 500 impressions.Checked 26 Oct 2026