Hold rate is the share of people who keep watching a video ad after its opening stopped them: 15-second video views divided by 3-second video views. Hook rate measures the first three seconds; hold rate measures the body of the ad, the part that has to deliver on what the hook promised. Read together they split creative diagnosis in two: the hook earns attention, the hold proves the story deserves it. A strong hook with a weak hold means the opening is writing cheques the story cannot cash.
A worked example. An ad serves 40,000 impressions. 10,000 people watch at least three seconds, a 25% hook rate. 1,800 of them reach 15 seconds, so the hold rate is 1,800 ÷ 10,000 = 18%: a hook doing its job and a body holding inside the usual target. Multiply the two (25% × 18%) and 4.5% of everyone who saw the ad watched 15 seconds.
Finding it in Meta Ads Manager. There's no hold rate column, so build one: add a custom metric in the column settings that divides ThruPlays by 3-second video plays, and save it into the view you check every week. Ads of 15 seconds or less only log a ThruPlay on a full watch, so on short cuts hold rate is really a completion rate. Compare short and long cuts separately rather than in one list.
Using it. This is what makes modular creative work. When hold rate is healthy but the hook has fatigued, keep the proven body and cut new openings onto it: three fresh variants the same afternoon, off footage you already own. When the hook works and the hold doesn't, leave the opening alone and fix the body. The full creative benchmarks are in Section 14 and the module system in Section 17.
Benchmark. On Meta, about 15-20%+ is a good target, and below about 10% the narrative is losing people after the hook did its job. It's a guide, not a rule: format, category and ad length all move it. Measure it per body and per angle, not just per ad, so you learn which modules are pulling the weight.
Common questions
What is a good hold rate?
On Meta, a hold rate of about 15-20% or higher is a good target when you measure it as 15-second views divided by 3-second views. Below about 10%, the body of the ad is losing people the hook had already stopped. Treat it as a guide rather than a pass mark: format, category and ad length all move it, so your own best ads are the fairest comparison.
How do you calculate hold rate?
Divide 15-second video views by 3-second video views. If 10,000 people watch the first three seconds and 1,800 of them reach 15 seconds, the hold rate is 18%. On Meta, use ThruPlays as the 15-second figure.
What is the difference between hook rate and hold rate?
Hook rate is 3-second video views divided by impressions: how many people the opening stopped mid-scroll. Hold rate is 15-second views divided by those 3-second views: how many of the people the hook stopped the body kept. A weak hook needs a new opening. A strong hook with a weak hold needs a better body, because the opening is promising something the rest of the ad doesn't deliver.
How do I find hold rate in Meta Ads Manager?
Meta doesn't report hold rate directly, so create a custom metric in the column settings: ThruPlays divided by 3-second video plays. For ads longer than 15 seconds a ThruPlay is a 15-second view. For ads of 15 seconds or less Meta only counts a ThruPlay when the video is watched to the end, so compare short and long cuts separately.
Why do hold rate benchmarks vary so much between sources?
Because there's no official definition. Some tools divide by impressions rather than 3-second views, which gives a much smaller number. Others use ThruPlays on ads shorter than 15 seconds, where a ThruPlay means a full watch, so short ads read higher. Check the denominator and the ad length before you compare yourself with any benchmark, including this one.
How do I improve a low hold rate?
Check the hook rate first. If the hook is stopping people and the hold is weak, the body isn't paying off what the opening promised: get to the product or the demonstration sooner, make the body answer the hook directly, or test a different body under the same hook. Tag every ad by hook, body and angle, so you can see which bodies hold across every ad they appear in rather than judging one ad at a time.
Does hold rate matter more than CPA or ROAS?
No. Hold rate is a diagnostic, not the outcome. Cost per new customer and contribution margin decide whether an ad is worth running; hook rate, hold rate and click-through tell you where a weak ad is breaking, so you know what to change.