Videos · Glossary
What is thumb-stop ratio?
Thumb-stop ratio is the percentage of a video ad's impressions in which someone stopped scrolling long enough to watch, usually measured as three second video plays divided by impressions. It is the same calculation many teams call hook rate, named after the thumb that pauses mid swipe on a phone feed.
Advertising teams have coined several names for one simple division, and reports that mix them confuse everyone who reads them. Knowing the aliases lets you confirm that two numbers are really the same number before comparing them.
Nuwan Madhusanka · Co-founder
6 min read · Published
| Name | Typical formula | Where you usually see it |
|---|---|---|
| Thumb-stop ratio | 3-second video plays divided by impressions | Paid social reports for feed placements |
| Hook rate | 3-second video plays divided by impressions | Creative testing and UGC ad reviews |
| Scroll stop rate | 3-second video plays divided by impressions | Agency summaries written for clients |
| 3-second view rate | 3-second video plays divided by impressions | Reports built directly from Ads Manager column names |
| 2-second view rate | 2-second video views divided by impressions | TikTok reports, with a shorter threshold |
| 6-second view rate | 6-second video views divided by impressions | TikTok reports, with a stricter threshold |
Where the name comes from
The phrase describes a physical moment: a thumb dragging a feed upward, then pausing because something on screen was worth a look. Mobile feeds made that gesture the first decision any video ad has to win, and the ad industry needed a word for it. No platform reports thumb stops as such, so the moment is approximated with the closest thing the reporting does offer, a short watch threshold. On Meta placements that threshold is three seconds. The name stuck because it is vivid in a creative meeting in a way that 3-second video plays divided by impressions is not, but the vivid name and the dull formula describe exactly the same thing.
The calculation, briefly
Take the 3-second video plays for an ad and divide by its impressions for the same date range. Meta counts a 3-second play when a video plays for three seconds, or for 97 percent of its length when it is shorter, and counts each impression separately without replays. TikTok Ads Manager has no three second count, so a TikTok version is built on 2-second video views, which need two seconds in one impression session, or on 6-second video views, which also count a complete play of a shorter video or an engagement in the first six seconds. An ad shown 25,000 times with 6,500 plays that passed three seconds has a thumb-stop ratio of 26 percent. The hook rate page goes further into how each platform counts the early view, what a low or high figure suggests and why published benchmarks rarely compare like with like, so none of that is repeated here.
When the aliases cause trouble
The trouble starts when two reports arrive with different names and different numbers. An agency reports a thumb-stop ratio of 28 percent; the in house team reports a hook rate of 31 percent for the same ad. Before anyone argues about which creative is better, check four things. The threshold: three seconds on Meta, or two or six seconds on TikTok. The denominator: impressions or reach. The level: a single ad, an ad set or a whole campaign. The date range: the same days for both. Most disagreements between two early attention figures dissolve once those four are lined up. Write the chosen definition into the report template itself, so the next person to open it does not have to rediscover it.
When the ratio is worth tracking
Thumb-stop ratio is worth tracking wherever a video starts on its own in a scrolling feed: Reels, Stories, TikTok and the main feeds on phones. It is at its most useful during creative tests, when several openings share the same body and offer, because it arrives first and needs fewer impressions to settle than a sale does. It is much less informative where somebody chose to press play, such as an embedded product video, and for a countdown or deadline ad whose opening frame already carries the whole offer, it effectively measures the offer itself.
Common mistakes
Treating thumb-stop ratio and hook rate as two separate targets in one dashboard doubles the reporting without adding information. Building the ratio on video plays, which counts every start, rather than on 3-second video plays inflates the figure with starts that autoplay produced on its own, which says little about the opening. Comparing a Meta thumb-stop ratio with TikTok's 2-second view rate mixes thresholds. Confusing it with click through rate, which divides clicks by impressions, is common in reports written for people outside the team. And judging an ad on a thumb-stop ratio from a few hundred impressions mistakes chance for a finding.
The opening seconds of a video from the builder
The video builder reports no views or impressions, so thumb-stop ratio is measured in Meta Ads Manager or TikTok Ads Manager once the exported MP4 runs as an ad. The builder shapes what the thumb meets. Vertical videos export at 1080 by 1920 for full screen feeds. A promotional scene runs 2.5 to 4 seconds with 5 to 8 words of narration, so the opener is usually one scene carrying one idea. Captions are on by default for every scene, which matters when the first seconds play silently. Where those captions and the offer sit against platform overlays is covered in the safe zones guide, and the UGC generator keeps videos under 30 seconds with at most three talking scenes.
Questions people ask
Is thumb-stop ratio a Meta metric?
Not by that name. Meta Ads Manager reports 3-second video plays and impressions, and advertisers divide one by the other to get the ratio. Some third party reporting tools add a thumb-stop column that does the division automatically, so check the tool's definition before assuming it matches a figure calculated by hand from Ads Manager exports.
Can thumb-stop ratio be used for image ads?
No. It depends on a watch time threshold, and a still image has no playback to measure. For image ads the nearest signals are engagement and click through rate. If a campaign mixes images and videos, report the thumb-stop ratio for the video ads only rather than averaging it across creatives that cannot produce a play.
Does watching without sound count?
Yes. The three second count records playback time, not audio, so a viewer who watches silently for three seconds counts exactly the same as one with sound on. That is one reason on screen text and captions carry so much weight in the opening seconds of a feed ad: they are what a silent viewer actually reads.
What counts as an impression?
An impression is recorded each time the ad is shown on a screen, so one person scrolling past the same ad several times creates several impressions. That is why the ratio uses impressions rather than reach: both the plays and the impressions are counted per showing, which keeps the two sides of the division consistent with each other.
Should thumb-stop ratio be the main goal of an ad?
No. It is an early diagnostic, useful for comparing openings, but an ad is paid for clicks, sign ups or sales. A video can stop plenty of thumbs with a surprising first frame and then lose everyone. Use the ratio to choose between openings, then judge the winner on hold rate, completion and the outcome the campaign exists for.
Is thumb-stop ratio useful for YouTube videos?
Less so. YouTube analytics describe viewing through average view duration and average percentage viewed rather than a three second count, and many YouTube views begin with a deliberate choice to watch. For YouTube, the retention curve and the share of viewers still watching after the opening give a clearer picture of whether the start is working.
Make one with videos
The button opens the generator with this use case already described. Change the wording to match your own.
Create a video with OneCraftRelated questions
- What is hook rate?Hook rate is the share of video ad impressions that play for three seconds. The formula, where each platform reports its inputs, and how to read it.
- What is hold rate?Hold rate is the share of hooked viewers who keep watching a video ad, usually ThruPlays divided by 3-second plays. How to calculate and read it.
- Average view duration vs watch timeAverage view duration vs watch time: one is minutes per view, the other is minutes summed across every view. A worked example and what each one shows.
Step by step in the builder: Export one video for YouTube, Reels and TikTok.
Written and checked by the OneCraft team. Last checked .