Hook rate vs hold rate: formulas and what each one tells you

Hook rate tells you who stopped. Hold rate tells you who stayed. Together they show where a video ad loses people, so you fix the right part instead of remaking the whole thing.

The short answer

Hook rate is the share of impressions that watched at least the first 3 seconds of your video. It measures the opening: the first frame, the first line, the first movement.

Hold rate is the share of viewers who kept watching after that opening. It measures the body of the ad: pacing, proof, and whether the video delivers what the first seconds promised.

Neither is an official Meta metric. Ads Manager gives you the raw counts, and you build both ratios yourself. That is why every article defines them a little differently, so pick one definition and never change it mid-test.

The hook rate and hold rate formulas

Here is the version most Meta buyers use. It needs only three columns from Ads Manager: Impressions, 3-Second Video Plays and ThruPlays.

A ThruPlay is a play of at least 15 seconds, or a play to the end if the video is shorter than 15 seconds. Keep that in mind, because it affects how you read hold rate.

Example numbers, not benchmarks: an ad gets 10,000 impressions, 2,800 three-second plays and 700 ThruPlays.

• Hook rate = 3-second plays ÷ impressions. Example: 2,800 ÷ 10,000 = 28%.
• Hold rate = ThruPlays ÷ 3-second plays. Example: 700 ÷ 2,800 = 25%. This is hold among people who were hooked.
• Some teams use ThruPlays ÷ impressions instead (700 ÷ 10,000 = 7%). That blends hook and hold into one number and hides which part failed. Avoid it for diagnosis.
• Another option is average watch time ÷ video length. It is useful for long videos where the 15-second mark says little.

How they differ in practice

The two metrics answer different questions and point to different parts of the video. Mixing them up is how teams end up rewriting a script that was fine and keeping an opening that was not.

Hook rate is about the scroll stop. It moves with the first frame, the opening line, on-screen text, sound, the creator's face and the movement in the first second. It tells you nothing about the rest of the video.

Hold rate is about the middle. It moves with how fast you get to the point, whether the proof arrives on time, whether the story keeps tension, and whether the video is simply too long for the promise it makes.

• Hook rate is a ratio of impressions. It is affected by placement, audience and the ad's first frame.
• Hold rate is a ratio of people who already stayed. Its denominator is smaller, so it swings more on low volume.
• A change to the first 3 seconds should move hook rate. A change after second 3 should move hold rate. If an edit moves the wrong one, something else changed too.

Read them together: four patterns

Look at the two numbers side by side, against your own account's usual range. A 28% hook rate means nothing alone. It means something when your last ten videos in the same format sat between 20% and 25%.

These are the four patterns, with the fix that fits each.

• Low hook, decent hold: the video is good but nobody starts it. Keep the body and test three new openings. This is the cheapest fix there is.
• High hook, low hold: the opening earns the stop and the body loses it. Either the hook over-promises, or the middle is slow. Cut the dead seconds or move the proof earlier.
• High hook, high hold, weak sales: people watch, but they do not buy. Look at the offer, the call to action and whether the landing page continues the ad. The video is not the problem.
• Low hook, low hold: the concept itself is not landing. Do not polish it. Go back to the desire and the angle and test something different.

What distorts both numbers

Most bad decisions come from reading these numbers too early or across things that do not compare.

Volume matters first. An ad with a few hundred impressions can swing wildly. Wait until each ad has at least a few thousand impressions before you compare hooks. That is a rule of thumb, not a law, so set your own floor and stick to it.

Video length matters second. A ThruPlay on a 12-second video means finishing it. On a 60-second video it means reaching second 15. Hold rate on short videos looks better, so compare hold rate only inside the same length bucket.

Placement matters third. Feed, Stories and Reels behave differently, and the mix shifts as Meta delivers. If one ad ran mostly in Reels and another mostly in Feed, break the numbers out by placement first.

Last, do not chase hook rate alone. A shocking opening that has nothing to do with the product will lift it and then lose people at second 4 and at checkout.

Where hook and hold fit in a test

Hook rate and hold rate explain why an ad behaved the way it did. They do not decide whether it worked. An ad can have a great hook and still never earn spend, and a plain one can carry a campaign.

That is the split Hitrate is built around. You plan each test as a numbered batch, one concept in one ad set, and Hitrate writes the ad names so every result maps back to the batch, its desire, angle, awareness stage, format and creator. That is what lets you compare hooks inside one angle instead of across a pile of unrelated ads.

After 7 days the batch gets a label from spend and ROAS against its campaign: Scaler, Spender, Efficient or Miss. Hook and hold never set the label. You use them afterward to see why a Miss missed, and the hit rate by angle and format shows which kinds of ideas keep winning.

Related guides

• Ad concept vs angle: what each one is and how to test them
• What creative fatigue means in Meta ads (and how to spot it)
• Thumbstop rate formula: how to calculate and read it
• What is creative testing? A plain definition for D2C brands
• Ad creative hit rate: what it means and how to calculate it

Updated 2026-10-02