Hook rate is one of the most confidently quoted numbers in paid social, and no benchmark for it exists anywhere with a disclosed methodology. Every target you have been given, including the eighteen to twenty eight percent range that circulates constantly, was made up by whoever needed a figure for their article.
I do not track it. That is not a stylistic preference, and the reasoning below is worth more than the metric.
This is for you if somebody has told you your hook rate is too low and you would like to know against what.
What hook rate is supposed to measure
It is usually three second video views divided by impressions, as a percentage.
The idea is reasonable. It attempts to isolate whether the opening of your video stopped somebody, separately from whether the rest of it held them.
Two problems arrive immediately. The definition is not standardised, so some calculations use impressions, some use reach, and some use a different view threshold entirely. Two people comparing hook rates are frequently comparing different quantities.
And a three second view is a very low bar. It captures a great deal of accidental attention, autoplay, and scrolling that happened to pause. Whether that constitutes being hooked is a genuine question that the metric quietly assumes away.
Why there is no hook rate benchmark
Meta has never published one. That is the beginning and the end of the sourcing problem.
Search for a target and you will find plenty of pages offering one. Follow the citation chain on any of them and it terminates within two or three steps in another article with no source, usually on a domain that publishes benchmark content across dozens of unrelated topics.
The research behind this cluster catalogued that pattern across sixteen sub-topic searches in this space. On hook rate specifically, the entire visible first page was vendor content citing other vendor content, with precise-looking numbers and no traceable origin.
So the honest position is that no published, methodologically disclosed hook rate benchmark exists. Saying that plainly is worth more than repeating a number, and it is what nobody ranking for this term will tell you.
Why I do not track hook rate
Because hold rate already contains the information.
I track retention instead, defined as more than half the video viewed and more than three seconds watched. The logic is simple. If somebody stayed that long, the hook worked. It has to have worked, or they would not still be there.
So hook rate becomes a partial measurement of something I already measure completely. Adding it gives me another number and no additional decision.
There is a second reason. A strong hook with weak retention describes an opening that grabbed attention and then lost it. That usually means the opening was misleading rather than good. Optimising towards hook rate rewards exactly that, and I would rather not point an incentive at bait.
What a hook rate figure can legitimately tell you
Relative comparison inside your own account, on matched conditions. That is the whole legitimate use and it is not nothing.
If two creatives run to the same audience, in the same placement, at the same time, and one produces a materially higher hook rate, that is a real observation about your own adverts.
What you cannot do is compare that figure against a published standard, because there is no standard. And you cannot compare it against another account, another placement or another period, because too much moves underneath it.
So use it descriptively and never as a target. A metric with no reference point can rank your own options and it cannot tell you whether you are doing well.
The pattern this metric is an example of
Hook rate is the clearest case of something that recurs across this whole topic.
A platform declines to publish a threshold. The query has commercial demand. Content fills the silence with a specific-sounding number, other content cites it, and within a couple of years the figure has the texture of an established fact without ever having been measured.
The same shape appears in fatigue frequency thresholds, where Meta analysed 2,439 Brand Lift campaigns and deliberately published no threshold number, while the internet publishes several.
Once you recognise the pattern you will see it repeatedly, and the recognition is more valuable than any individual correction. When a platform will not name a number and everybody else will, the everybody else is inventing.
What to watch instead
Three things, in the order I look at them.
Hold rate, on the definition above, compared across your own creatives on matched conditions.
The three Meta rankings, quality, engagement rate and conversion rate, which separate a creative problem from an audience problem from an offer problem.
And cost per result, which is the only one that pays anybody. Every upstream metric is a diagnostic for explaining that number, not a substitute for it.
A creative with a modest hold rate and a good cost per result is working. Nothing about a video metric overrides that, and I have watched accounts pause profitable adverts for failing a benchmark nobody could source.
What to do when somebody quotes you a target
Ask where it came from, plainly.
That question is not confrontational and it is remarkably effective. In most cases the person quoting it has never checked, and the checking takes minutes.
Follow the chain yourself once, on any hook rate benchmark you find. Watching it terminate in an unsourced article is a more durable lesson than being told, and it will change how you read every benchmark afterwards.
Then apply the same test to the other numbers in your reporting. Several will not survive it, and knowing which is a real advantage over almost everybody competing with you.
Why the hook rate gap has not been filled
It is worth asking why no credible source has published one, because the answer is instructive rather than lazy.
A benchmark would have to hold across wildly different businesses, placements and video lengths. Feed and stories produce completely different retention for identical creative. A single figure would be wrong for most advertisers on the day it was published.
There is also no obvious party to produce it. Meta has the data and no reason to invite advertisers to optimise towards a number that would immediately get gamed. Independent researchers have no access to the raw figures.
So the silence is structural rather than an oversight. What fills it is content answering a question with commercial demand and no available answer, which is the exact condition that produces invented statistics.
Recognising that condition is the transferable skill. When a platform holds data it will not publish, and the query has money behind it, expect fabricated numbers and check every one you are given.
What to change this week
Three steps.
Check whether any target in your reporting is a hook rate figure, and if so, find out where the target came from before your next review.
Then set up hold rate on the definition above, so you have the measurement that actually contains the information.
Then stop comparing any video metric against an external standard, and compare your creatives against each other on matched conditions instead.
Fund the ad set and read what the algorithm chose is the position underneath, creative volume covers what the distribution of winners actually looks like, statistical significance in ad testing covers how large a difference has to be before it is real, and most cost per lead benchmarks are recycled is the same pattern in a different topic. On the search side the equivalent invented metrics show up in Google Ads management conversations constantly. To have your reporting checked for numbers nobody can source, book a teardown.