Median hook rates at a three-second threshold are 28 percent on Meta Reels, 28 percent on Meta feed, 33 percent on TikTok and 22 percent on YouTube in-stream, per our 2026 fatigue benchmark, a synthesis of published platform and agency figures rather than our own measurement. Before you compare your account to any of them, check the threshold: the same creative that reports 28 percent at three seconds reports about 43 percent at two seconds and about 65 percent at one. Most published hook rate benchmarks do not state which threshold they used, which makes them not wrong so much as unusable.
What hook rate is, precisely
Hook rate is three-second video plays divided by impressions, times one hundred. Not divided by reach, and not divided by video plays.
That last error is the common one and it produces a number in the nineties, because almost everyone who triggers a video play triggers a two-second one. If your reported hook rate has ever been above 60 percent, check the denominator before you celebrate. The hook rate definition and the thumbstop rate definition describe the same measurement, and the second name is more common where the threshold is two seconds rather than three, which is the root of most of the confusion in this subject. Log impressions and plays per variation in a hook rate calculator so the denominator is fixed once and never argued about again.
Hook rate benchmarks by placement, at three thresholds
The published medians are three-second figures. To convert them to other thresholds, assume the audience leaves at a constant rate per second, so survival to two seconds is the three-second figure raised to the power two thirds. That is a model rather than a measurement, and it understates the early thresholds, because real curves are steeper at the front. Take the two-second and one-second columns as lower bounds.
| Placement | Hook rate at a 1s threshold | Hook rate at a 2s threshold | Hook rate at a 3s threshold | Ratio, 2s to 3s |
|---|---|---|---|---|
| Meta Reels | 65% | 43% | 28% | 1.53x |
| Meta feed | 65% | 43% | 28% | 1.53x |
| TikTok | 69% | 48% | 33% | 1.45x |
| YouTube in-stream | 60% | 36% | 22% | 1.66x |
Read one row across. A single unchanged file can be honestly described as having a 65 percent hook rate, a 43 percent hook rate or a 28 percent hook rate. Nobody has to be lying for two benchmark posts to disagree by a factor of 1.5.
Two practical consequences. If you find a benchmark quoting hook rates in the forties or above without naming a threshold, it is almost certainly a two-second figure, so divide by about 1.5 before comparing it to a Meta three-second number. And if your agency reports hook rate and your in-house dashboard reports thumbstop rate, confirm they agree on the threshold before anyone explains the gap with a creative theory.
Why broad prospecting reports a lower hook rate than retargeting
Warm audiences hold the opening better than cold ones on identical creative, which means your reported hook rate is partly a media mix number rather than a creative number.
Work it. Suppose in your account the same file holds 40 percent on retargeting and 24 percent on cold prospecting. Those are example inputs; put your own two figures in. The blended number you see in the campaign row is:
blended hook rate = 24 + 16r, where r is retargeting's share of spend
| Retargeting share of spend | Blended reported hook rate | Gap to the 28 percent Meta median |
|---|---|---|
| 25% | 28.0% | 0.0 points |
| 20% | 27.2% | 0.8 points low |
| 15% | 26.4% | 1.6 points low |
| 10% | 25.6% | 2.4 points low |
| 5% | 24.8% | 3.2 points low |
| 0% | 24.0% | 4.0 points low |
Solve it the other way and the finding is sharper: hitting the 28 percent median with those component rates requires a quarter of your spend to sit on retargeting. Any account scaling into broad prospecting walks its reported hook rate down mechanically, and an account that cuts retargeting from 25 percent of spend to 5 percent loses 3.2 points without a single creative change. Teams kill perfectly good cold-traffic creative over gaps that size every week.
The corollary is that the comparison worth making is not against a published median at all. It is against the same creative in the same audience last month.
Your account's benchmark is a weighted average, not a row
Nobody spends on one placement. If the medians differ by placement, then the only benchmark that applies to your account is the average of them weighted by your spend split.
| Spend split | Weighted benchmark | Working |
|---|---|---|
| 40% Reels, 30% Meta feed, 20% TikTok, 10% YouTube | 28.4% | 11.2 + 8.4 + 6.6 + 2.2 |
| 70% TikTok, 30% YouTube in-stream | 29.7% | 23.1 + 6.6 |
| 50% Meta feed, 50% YouTube in-stream | 25.0% | 14.0 + 11.0 |
| 100% Meta Reels | 28.0% | 28.0 |
| 60% Meta feed, 40% TikTok | 30.0% | 16.8 + 13.2 |
Across ordinary media mixes the correct target ranges from 25.0 to 30.0 percent, a five-point spread produced entirely by where the money went. Combine that with the audience-mix effect above and an account can sit four points below one published number and two points above another while nothing about the creative has changed.
The fourth reason: you are comparing a median to a specific week
A benchmark median is a cross-section. It averages creative in its first week with creative in its fourth, which means it already includes fatigued ads.
Our benchmark report puts CTR decline at 15 to 20 percent across a creative's first two weeks, with week three at 45 to 70 percent below the launch baseline, and finds most creative is effectively dead within three weeks. Those are CTR figures rather than hook rate figures, and we are not going to convert them for you. What follows is only the direction: a launch-week number compared to a mixed-age median flatters you, and a week-four number compared to the same median does the opposite. Meta internal research cited in the report puts the CTR drop at 45 percent after a fourth exposure to the same creative, and weekly frequency of 2.5 on Meta prospecting is where decline begins.
So state the creative's age with the number, or the number means little. From experience, TikTok in-feed runs 20 to 30 percent and Meta feed 15 to 25 percent across the full life of a creative, which is wider and lower than the medians above precisely because it includes the tired weeks.
How to write down a hook rate so somebody can use it
Four labels, and a hook rate without all four is a rumour: the threshold in seconds, the placement, the audience temperature, and the creative's age in days. "31 percent" is not a benchmark. "31 percent at a three-second threshold, Reels, cold prospecting, day four" is one, and it happens to be a modest winner against a 28 percent median.
Keep your own four-label history and you stop needing external benchmarks for anything except sanity checks. Your account's own distribution from last quarter tells you more about whether 31 percent is good than any composite median can, because it already has your audience mix, your placement mix and your threshold baked in.
Where Genyad fits, and what it will not do
Genyad is our product, so read this as disclosure. It does not measure hook rate and it will not tell you a benchmark. It produces the variations you measure: upload footage once, it transcribes and tags every clip, then each variation is a fresh script, shot selection, voiceover, caption set and export drawn from that library rather than a re-cut of one timeline. One variation costs 1 credit, the free plan gives 5 with no card, Starter is EUR 29 for 15 credits and Growth is EUR 99 for 65 credits, and credits never expire.
What it does not do, stated plainly: no AI avatars or synthetic presenters, no static banner formats, no product-URL import, no product-feed or CSV-driven template rendering, no direct publishing to Meta or TikTok, and no predicted performance scores. That last one is deliberate and relevant here. A predicted score is a model's guess at a number you can measure for about EUR 6 of media per variation, and a guess with no threshold, no audience and no placement attached is exactly the kind of benchmark this post is arguing against. If a predicted score is what you want, AdCreative.ai was listed at $39 per month in August 2026. Prices move.
Frequently asked questions
What is a good hook rate?
At a three-second threshold, compare against 28 percent on Meta Reels and Meta feed, 33 percent on TikTok and 22 percent on YouTube in-stream. Weight those by your own spend split before treating any of them as your target, because ordinary media mixes shift the correct figure between about 25 and 30 percent. Across a creative's whole life, 20 to 30 percent on TikTok in-feed and 15 to 25 percent on Meta feed are the ranges we plan around.
Why is my hook rate lower than published benchmarks?
Four reasons, in the order worth checking. The benchmark may use a two-second threshold, which inflates it by about 1.5 times. Your audience may be more heavily cold prospecting than the benchmark's. Your placement mix may lean toward the lower-median placements. And your creative may be older than the mixed-age average that any median represents.
Does a two-second hook rate mean the same thing as a three-second one?
No, and the gap is large enough to invalidate most cross-source comparisons. Under a constant-decay model, a creative holding 28 percent at three seconds holds about 43 percent at two seconds and about 65 percent at one. Always state the threshold next to the number.
Should I compare hook rate across prospecting and retargeting?
Only if you want to measure the audience rather than the creative. Warm pools hold the opening better on identical files, so a blended campaign number moves whenever the spend split moves. Compare a hook against other hooks inside the same audience and placement, and compare audiences separately.
Is hook rate the same as thumbstop rate?
They are the same measurement, three-second or two-second plays over impressions, under two names. The reason it matters is that the thumbstop label is more often attached to a two-second threshold, so a dashboard using one term and an agency report using the other can differ by half again on the same creative.