# How to read hook rate and hold rate

> Read hook rate and hold rate together to see where a video ad loses people. Formulas from Meta's own metrics, a worked example and what to fix first.
- **Author**: Marxx
- **Published**: 2026-10-23
- **Category**: Creative
- **URL**: https://marxx.ai/library-guides/how-to-read-hook-rate-and-hold-rate

---

Hook rate shows how many people stayed past the first three seconds of your video, and hold rate shows how many of those kept watching to 15 seconds or the end. Read hook rate and hold rate together and you know which part of the video to fix.

## Get the numbers

Ads Manager reports the counts you need, and you work out the two rates yourself. Open Ads Manager, click **Columns**, and add **Impressions**, **3-second video plays** and **ThruPlays**.

- A 3-second video play counts when the video played for at least 3 seconds. Replays are not counted ([Meta](https://www.facebook.com/business/help/743427195703387), checked 2026-10-20).
- A ThruPlay counts when the video played to the end, or for at least 15 seconds. For videos shorter than 15 seconds, it counts at 97% played ([Meta](https://www.facebook.com/business/help/471190536725647), checked 2026-10-20).

## The formulas

- **Hook rate = 3-second plays / impressions x 100**
- **Hold rate = ThruPlays / 3-second plays x 100**

Some teams divide hold by impressions instead. Pick one version and keep it, or your ads stop being comparable.

## A worked example

Say two Reels ads each get 1,00,000 impressions. The numbers are made up.

| | Ad A | Ad B |
|---|---|---|
| 3-second plays | 22,000 | 38,000 |
| ThruPlays | 9,900 | 11,400 |
| Hook rate | 22% | 38% |
| Hold rate | 45% | 30% |

Ad A has a weak opening, but the people it does stop stay. Ad B stops many more people, then loses most of them before 15 seconds.

The fix is not "make A more like B". Put B's opening on A's video and test that. You change the first three seconds and nothing else.

## What each pattern means

| Hook rate | Hold rate | Fix first |
|---|---|---|
| Low | Any | The opening. Show the result in the first frame and put words on screen |
| High | Low | The middle. Cut slow parts and show the product by second five |
| High | High | Look at clicks and sales next. The video is doing its job |

For opening ideas, see [how to write ad hooks](/guides/how-to-write-ad-hooks).

## Do not over-read it

- **Compare like with like.** A 10-second ad and a 60-second ad hold differently. Compare ads of similar length and placement.
- **Wait for enough views.** Judge an ad after a few thousand impressions, not a few hundred.
- **Rates are not sales.** A high hook rate with no clicks means the ad stops people for the wrong reason.
- **Watch the trend.** If hook rate falls week by week on the same ad, check for [ad fatigue](/guides/how-to-fix-ad-fatigue).

Marxx reads your connected Meta ad account in full and tags each creative by hook, angle, format and offer, so you can group your ads by hook type and compare them. See [creative intelligence](/features/creative-intelligence).

## FAQ

### What is a good hook rate?

No single number fits every account. Compare each new ad with your own median from the last 90 days. The [hook rate glossary page](/glossary/hook-rate) explains more.

### Is hook rate the same as thumbstop ratio?

In this guide, yes. Hook rate is 3-second plays divided by impressions, and thumbstop ratio is the name some teams use for that same figure. See [thumbstop ratio](/glossary/thumbstop-ratio).

### Do these rates work for static ads?

No. Both need video plays. For a static ad, use click-through rate to see if it stopped people.

---
- [More Creative articles](https://marxx.ai/library-guides/category/creative)
- [All articles](https://marxx.ai/library-guides)