By 
Hook rate is the one creative metric you cannot choose in advance. You buy it. The number only comes into existence once the impressions are served, which means every hook rate you have ever seen was paid for by somebody, usually you. That shapes everything else worth saying about it: what it actually measures, what good looks like by placement, why the only benchmark that counts is your own, and what to do instead of finding out the expensive way.
What hook rate actually measures
Hook rate is the share of people who watch past the opening few seconds of your video ad. On Meta the standard calculation is simple:
Hook rate = 3-second video plays ÷ impressions × 100
A 3-second video play counts each time your video is played for at least three seconds, and impressions counts the number of times the ad appeared on screen. So a hook rate of 30 per cent means that for every hundred times your ad was shown, thirty people stayed long enough to register it rather than scrolling straight past.
Two details trip people up. First, use impressions as the denominator, not reach. Reach counts unique people, so dividing by it inflates the rate and makes ads look better than they are. Keep it impressions, every time, so your numbers stay comparable. Second, hook rate goes by another name you will see everywhere: the thumb stop ratio. Same metric, same maths. It just describes the moment more literally, the thumb pausing mid-scroll.
One thing worth saying plainly: hook rate is not a native Meta metric. The platform reports 3-second video plays and impressions, but you calculate the ratio yourself, which is exactly why there is no official benchmark for it. Anyone quoting you a hard number as gospel is guessing with confidence.
What counts as a good hook rate
Because there is no official benchmark, every range you read is someone’s directional read, mine included. With that caveat firmly in place, here are the working bands most performance marketers operate to:
| Placement | Weak | Solid | Strong |
|---|---|---|---|
| Meta feed video | Under 20% | 25 to 35% | 35%+ |
| Meta Reels and Stories | Under 15% | 20 to 30% | 30%+ |
The gap between the rows is the point. Fast-swipe placements run lower because the swipe is quicker and the intent is lighter, so a 25 per cent hook in Reels can be doing more work than a 33 per cent hook in feed. Compare like with like, or the numbers will mislead you.
You will notice TikTok is not in that table, and this is where most benchmark posts quietly go wrong. TikTok’s default video views metric counts a view almost the moment playback starts, not at three seconds. So a hook rate built on it is not a lower version of the Meta number, it is a different measurement, and it will come out far higher rather than lower. If you want something comparable, use TikTok’s 2-second video views as the numerator and keep impressions as the denominator. Even then, treat the two platforms as separate scoreboards rather than one league table.
Here is the part that matters more than any table. The benchmark that actually tells you something is your own account’s rolling average for the same placement. Your feed, your audience, your price point and your offer already sit inside that number, which no industry range can account for. A 22 per cent hook rate might be poor for a low-priced impulse product and genuinely good for a considered, higher-priced one. As a rough adjustment, B2B and high-consideration products tend to sit five to ten points below every band above, because the audience has a bigger qualification hurdle to clear before it stops. Pull your last thirty to ninety days by placement, take the median, and judge every new ad against that. Beat your own average and you are improving. Chase a number off a blog and you are optimising against a stranger’s account.
One caution before you act on any of this. Not everything that moves your hook rate is creative. Load time and connection speed affect whether a three-second play registers at all, and changing your placement mix or audience changes the composition of the impressions underneath the number. If your hook rate fell the same week you added a placement, you may be measuring the placement rather than the opening. Check what else changed before you rewrite a hook that was never the problem.
Why the first few seconds carry so much weight
It is tempting to treat hook rate as a vanity metric, a number that goes up without moving sales. It is not, and the research on where attention converts to value is unusually clear.
Nielsen, analysing 173 brand-effect studies run for Facebook, found that viewers who watched three seconds or less accounted for 47 per cent of the total ad recall effect, along with 32 per cent of brand awareness and 44 per cent of purchase intent, with measurable lift registering in under a second. Two caveats belong with that number, since nobody else quoting it seems to carry them: the study is from 2015, and Nielsen included only campaigns that already showed positive lift, which flatters the result. TikTok’s own analysis points the same way, and harder: it puts 90 per cent of ad recall impact and 80 per cent of awareness impact inside the first six seconds. Read that with its provenance in mind. It comes from TikTok’s 2022 Elements of Attention study, published with CreatorIQ in late 2023, which tested 40 creator ads with 1,650 users, so it is platform research on a small base rather than independent evidence. Whether you take the three-second window or the six, the conclusion holds. The opening is not the warm-up. It is where most of the work gets done.
There is a second reason to care, with a caveat attached. In most accounts, the ads with the strongest hook rates also carry the lowest costs per thousand impressions, and a two or three point improvement often shows up in your CPMs rather than just on a chart. What you will not find is either platform publishing a mechanism for that, so treat it as a pattern you can check in your own data rather than a rule about how the auction works. The safe version of the claim is the useful one anyway: a weak hook loses the viewers in front of it, and the ads that lose those viewers tend to be the expensive ones.
Hook rate, hold rate and CTR are three different questions
Hook rate only grades the opening. To read an ad properly you need to see it next to two other numbers, because each answers a different question. Our guide to ad testing covers how these fit into a full creative test; this section is the close-up.
Hold rate measures whether the people who stopped actually stayed. It is the share of your 3-second viewers who go on to reach 15 seconds or the end of a shorter video, which on Meta is the ThruPlay threshold. If hook rate grades the first few seconds, hold rate grades the rest of the ad.
Click-through rate measures whether watching turned into intent to act: link clicks over impressions.
Read together, hook and hold form a quick diagnostic that tells you where an ad is actually breaking:
| High hold | Low hold | |
|---|---|---|
| High hook | Strong creative. Stop fiddling and scale it. | The opening overpromised. People stopped, then bailed when the body did not deliver. Fix the middle, or the hook is writing a cheque the ad cannot cash. |
| Low hook | A good ad almost nobody stops for. The problem is the opening frame, not the story. Rework the first two seconds. | A concept problem, not an execution one. No single tweak saves this. Rework the idea or kill it. |
One caution here. A high hook rate on its own can be a warning sign, not a win. Clickbait openings, a fake notification sound, a “wait for it” that never pays off, can spike your hook rate while your hold rate and return on ad spend quietly collapse. Never celebrate a hook number without checking what happened to the people it caught.
How to improve your hook rate
Once you know your own baseline, improving the opening is fairly concrete. The tactics below are ordered roughly by impact.
Open on the payoff, not the preamble. Cut the logo sting, the slow establishing shot and the three-second ramp. Given how much value lands in the opening seconds, your first frame should already be mid-action, showing the product doing the interesting thing or the moment that makes someone pause. Front-load the most arresting second of your ad; do not save it for the reveal.
Put a person in the first couple of seconds. In the TikTok and CreatorIQ study of creator ads, showing a person early increased “hooking power” by 50 per cent and lifted ad recognition by 32 per cent. A face reads as human before the brain processes anything else, and it stops the scroll.
Talk to the viewer, early and directly. The same research found that having the creator greet the audience directly drove a 112 per cent uplift in brand recall, and saying “you” within the first five seconds lifted purchase intent by 128 per cent. “You need to see this” outperforms “Our new range is here”, and it is a one-line change to the script.
Design for the sound off. Most feed video autoplays muted, so a hook that depends on a spoken line lands on no one. Put the promise on screen as text in the opening frame and treat audio as a bonus, not the delivery mechanism.
Then test several hooks against one concept. The opening is where the biggest performance swings live, so the highest-value thing you can test is three or four different openings for the same underlying idea, which most ad testing tools are built to handle. Do not test the concept and the hook in the same breath, or you will not know which one moved the number.
The catch: you pay to learn your hook rate
Here is the honest limitation of hook rate as a metric. It is an in-flight number. You do not get it until the ad is live, the impressions are bought and the money is spent. By the time you can see that a hook rate is 12 per cent, you have already paid to find out. Run four openings this way and three of them are, by definition, tuition.
That is a fine way to confirm a winner once something is already working. It is an expensive way to choose which opening to launch in the first place. The better sequence is to screen the hooks before they touch your media budget, then use the live hook rate to confirm what scales rather than to discover what fails.
This is the job TestFeed is built for. You put your hooks, or the ad concepts behind them, in front of an audience modelled on your buyers, and you get back a purchase-intent read, a plain verdict, the shoppers’ reasons in their own words, and a clear next move, in days rather than weeks. The point is to spot the weak openings before you pay to serve them, so only the strongest ideas reach the auction. Be clear-eyed about what it is: a pre-spend, directional signal, not a market forecast or a guaranteed result. It tells you which hooks are worth putting money behind, and it does not replace watching real behaviour once they are live. Used that way, pre-testing does not compete with your in-flight numbers. It decides what deserves to earn them.
If you want the wider frame around this, from concept to claim to creative, our guide to choosing a concept testing platform covers how to screen ideas before launch, and the piece on building a test audience explains how to get a read from the right buyers rather than a random crowd.
Track your hook rate, absolutely. Judge it against your own average, read it next to hold rate, and never let a shiny opening distract you from what happened after the scroll stopped. Just remember that the cheapest hook rate to fix is the one you caught before you paid for it.
Frequently asked questions
What is a good hook rate?
There is no official benchmark, so treat any number as directional. On Meta feed video, roughly 25 to 35 per cent is a solid working range and above 35 per cent is strong, while under about 20 per cent usually signals a weak opening. Fast-swipe placements such as Reels and Stories tend to run lower because the swipe is quicker, so a 25 per cent hook there can be as good as a 33 per cent hook in feed. TikTok needs separate treatment, because its default video views metric counts a view almost immediately rather than at three seconds, so a TikTok hook rate is not directly comparable to a Meta one. The benchmark that matters most is your own account’s rolling average for the same placement, because it already controls for your audience, offer and price point.
How do you calculate hook rate?
Hook rate is 3-second video plays divided by impressions, multiplied by 100. On Meta, 3-second video plays counts each time your video is played for at least three seconds, and impressions counts the number of times it was on screen. Use impressions rather than reach as the denominator, because reach counts unique people and will overstate the rate. Keep the same definition every time so your numbers stay comparable.
Is hook rate the same as thumb stop ratio?
Yes. Thumb stop ratio is another name for hook rate: the share of impressions where someone stopped scrolling long enough to watch past the opening few seconds. The two terms are used interchangeably. Whichever you use, the underlying calculation is the same, 3-second video plays over impressions.
What is the difference between hook rate and hold rate?
Hook rate measures whether people stop to watch your opening: 3-second video plays over impressions. Hold rate measures whether the ones who stopped stayed: the share of 3-second viewers who go on to reach 15 seconds or the end of a shorter video, which on Meta is the ThruPlay threshold. Hook rate grades the first few seconds; hold rate grades the rest. A high hook with a low hold usually means the opening overpromised and the body did not deliver.
How do I improve my hook rate?
Open on your strongest visual or the payoff, not a logo or a slow build, because most of an ad’s value lands in the first few seconds. Put a person or face in the opening moment, address the viewer directly, and design for a muted autoplay with clear text on screen. Then test several different hooks against the same concept, because the opening is where the biggest swings in performance live. To avoid burning budget finding out, pre-test the hooks with target buyers before you launch.
