Search for a good hook rate on Meta and you will find a confident number. It will be different on every page you open, and none of them will show you where it came from. That is worth knowing before you build an alert, a creative brief or a client report around one.
This piece does three things. It pins down what the metric measures, says plainly what is and is not known about a "good" value, and shows how to build a benchmark from your own account that you can defend in a meeting.
What hook rate measures
Hook rate is the share of impressions on a video ad that turned into a three second view on Meta. Some people call it thumbstop rate. Motion's own documentation describes the goal of the metric as hooking people in to start watching, and puts the threshold at 3 seconds for Meta and 2 seconds for TikTok.
Both numbers come from the ad platform and both are conventions rather than laws. A three second view on a muted autoplay feed is a much weaker event than a three second view that someone chose to open in full screen, and the metric does not tell them apart. Keep that in mind every time someone quotes you a percentage.
Is there an official "good" number?
No. We could not find a public, audited benchmark for hook rate, from Meta or anyone else. What circulates is practitioner rules of thumb, a single percentage repeated from article to article. Treat these as folklore until someone shows the sample, the placement mix, the industries and the period.
Even in a perfect dataset a single number would mislead, because hook rate moves with things that have nothing to do with how good your opening is:
- Placement. Full screen vertical placements and in feed placements behave differently, and an ad set that shifts its delivery mix will shift its hook rate without a single frame changing.
- Audience temperature. People who already know the brand will stop for it more readily than people who have never seen it.
- Format and length. A fifteen second ad and a ninety second ad are not competing on the same terms in the first three seconds, and a static image has no hook rate at all.
- Sound and captions. Whether the opening makes sense muted is a creative choice that shows up directly in this number.
The practical conclusion is that a hook rate is only meaningful against something comparable: other ads from your own account, in the same placement, for the same kind of audience, over the same dates.
Build your own benchmark
Motion's advice on the question is the right instinct: pull your videos into one report and use the account average as the yardstick for which ads are above or below it. Make it sturdier with four rules.
1. Compare like with like
Split by placement group and by prospecting versus retargeting before you take an average. A blended average across both will flatter your retargeting ads and punish your cold ones for reasons that are not creative.
2. Use the median, not the mean
A handful of tiny ads with a freakishly high rate on a few hundred impressions will drag the mean around. The median of ads with enough volume is steadier. See the next section for what "enough" means.
3. Set a floor before you read anything
Hook rate is a proportion, so how precisely you know it depends on impressions. At a true rate of 30%, the standard error is about 1.5 percentage points on 1,000 impressions and about 0.5 points on 10,000. That is simple binomial arithmetic, and it ignores the fact that ads are shown repeatedly to overlapping people, so the real uncertainty is a little wider. The consequence is that a 2 point gap between two ads on a thousand impressions each is inside the noise.
4. Read it as a rank, not a grade
| Where the ad sits against your median | What to do |
|---|---|
| Well above, on enough volume | Study the opening. Which of its first two seconds could be reused in other ads? |
| Around the median | Nothing to learn yet. Look at hold rate and the outcome metric instead. |
| Well below, on enough volume | Re-cut the opening only. The rest of the ad may be fine, which you can only tell from hold rate. |
| Anywhere, on thin volume | Wait. Say it is too early to call. |
A high hook rate is a measure of attention, not of value. An ad can stop thumbs with a shock cut that has nothing to do with the product and then lose almost everyone by second five. That is why the pair matters: hook rate says whether the doorway worked, hold rate says whether the room was worth walking into.
What the research adds
The most useful finding for anyone chasing a higher hook rate is not about the number at all. VCCP Media and Dr Karen Nelson-Field tested 72 digital video ads across more than 20,000 views and found that 1.5 seconds of genuine attention was enough to encode memory, provided the brand was present. Well branded ads were 2.5 times more effective in low attention settings. We cover that study in full in the 3-second hook, deconstructed.
Put that next to the metric. A hook rate counts the people who stayed three seconds. The research suggests the useful question is whether those three seconds carried the brand. Two ads with identical hook rates can be worth very different amounts, and no ratio of plays to impressions will tell you which is which.
Knack grades hook rate against your own account rather than a number carried in from elsewhere, reads it together with hold rate, and holds back any verdict on ads that have not yet earned one. It also tags what the opening actually is, such as a question, bold claim, relatable moment or direct address, from the creative itself, so you can ask which kind of opening works on your account and not only which ad scored highest.
Book a demo to see this run on your own account.
What to do this week
- Export the last 90 days of video ads with impressions, 3 second plays and ThruPlays.
- Drop everything under a few thousand impressions.
- Split by placement group and by prospecting versus retargeting.
- Take the median hook rate for each group. That is your benchmark.
- Review the top and bottom quarter of ads side by side, looking at the first two seconds only.
Sources
- Motion Help Center: Tell stories and build hypotheses with metrics. Definitions of hook and hold metrics and the advice to benchmark against your own account average. The page gives no numerical benchmark.
- VCCP Media and Dr Karen Nelson-Field / Amplified (May 2025). Hacking the Attention Economy. 20,000+ views of 72 digital video ads, each tested against a version with brand elements removed.
- The standard error figures are the binomial formula sqrt(p(1 minus p) / n), shown for p of 0.30.