Hook rate tells you how many people kept watching your video ad past the first few seconds instead of scrolling away. Learning how to measure hook rate on video ads correctly is the single biggest lever for improving it, because most teams either grab the wrong numbers or compare metrics that aren't built the same way across platforms. This guide walks through the exact formula, where to pull the raw numbers on each ad platform, and how to avoid the comparison mistakes that quietly ruin creative testing.
What Hook Rate Actually Measures
Hook rate isolates the very first moments of your ad — the part responsible for stopping a scroll. It is not the same as overall watch time, click-through rate, or conversion rate. A video can have a strong hook rate and still convert poorly if the offer or the middle section falls flat. That's why hook rate should be measured and judged on its own, as a diagnostic for your opening three to five seconds, before you draw conclusions about the rest of the ad.
The core formula is simple:
- Hook Rate = (Viewers who watched past the threshold ÷ Total impressions or plays) × 100
- Threshold is usually 3 seconds, though some teams test 5 seconds depending on ad length
- The denominator can be impressions or video plays — these are not the same number, and mixing them up is the most common measurement error
How to Measure Hook Rate on Video Ads: Step-by-Step
Step 1: Decide your denominator and lock it in
Some platforms report hook rate against impressions (everyone the ad was served to, including people who never pressed play if autoplay was off). Others report it against video plays (people whose feed actually triggered the video). Decide which one you're using before you pull any numbers, write it down in your testing doc, and never switch mid-campaign. A hook rate of 35% against impressions is not comparable to a hook rate of 35% against plays — the second number is almost always higher because it excludes people who never saw the video start.
Step 2: Pick a consistent time threshold
Three seconds is the most common starting point because it reflects the window where a viewer decides whether to keep watching or scroll on. If your ads run longer than 30 seconds, you can also test a 5-second threshold to see how many people survive past the hook into the problem or benefit section. Whichever threshold you choose, apply it to every ad in the test so the numbers mean the same thing across your account.
Step 3: Pull the raw numbers from native reporting
Go into the ads manager for the platform you're running on and pull two numbers for the exact date range and placement you care about: total impressions (or plays) and plays past your threshold. Avoid blended or cross-platform dashboards for this step — they often normalize metrics in ways that hide which denominator was used.
Step 4: Calculate and segment
Divide plays past the threshold by your denominator, multiply by 100, and you have your hook rate. Do this separately for each placement (feed, Stories, in-stream) and each format (9:16 vertical, square, landscape) before comparing ads. A hook can perform very differently in a vertical Reels placement than in a landscape in-stream slot, so blending placements together will give you a number that doesn't represent any single viewing experience.
Step 5: Log it in a swipe file
Keep a running sheet with the hook's opening line or visual, the platform, placement, format, threshold used, impressions, plays past threshold, and the resulting hook rate. Over several tests this becomes your internal benchmark — far more useful than any industry-wide number, because it reflects your actual audience and product category.
Where to Find the Raw Numbers on Each Platform
| Platform | Where to look | Typical denominator | Notes |
|---|---|---|---|
| TikTok Ads Manager | Columns menu, add video metrics for 3s/6s views and impressions | Impressions or 6-second views | Check whether you're viewing by ad, ad group, or campaign level — totals can differ |
| Meta Ads Manager | Breakdowns and custom columns for ThruPlays or 3-second video plays vs impressions | Impressions | Meta's reporting can lag; pull the number after the attribution window closes for stability |
| YouTube Ads (Shorts/in-stream) | Video campaign reporting, views and impressions columns | Impressions or views, depending on skippable vs non-skippable format | Skippable and non-skippable ads behave differently in the first few seconds, so don't mix them in one test |
Hook Rate vs Thumb-Stop Rate vs Retention Rate
These three terms get used interchangeably, but they answer slightly different questions. Hook rate measures survival past an early time threshold (plays past 3 seconds divided by impressions or plays). Thumb-stop rate measures whether the ad interrupted a scroll at all, often tied to the very first frame or first second. Hook retention rate looks at the shape of the drop-off curve across the whole video, not just one checkpoint. If you want the full breakdown of each, see how to measure thumb-stop rate on video ads and how to measure hook retention rate. Once you have a number, compare it against a reasonable range in what is a good hook rate for video ads before deciding whether a hook needs to be replaced.
How to Use Hook Rate Once You Have It
A hook rate number is only useful if it changes what you do next. Build a simple decision rule before you start testing, so you're not making judgment calls ad by ad:
- If hook rate is well below your account average, treat the opening line or visual as the problem first, not the offer or CTA
- If hook rate is strong but conversion or click-through is weak, the hook is working — look at the middle and the offer instead of touching the hook
- Run hook variants against the same audience, budget and placement so the only thing that differs is the opening few seconds
- Give each variant enough impressions before comparing — a handful of plays can swing a percentage wildly in either direction
- Retire a hook only after it has underperformed across more than one placement or audience, not from a single data point
Common Mistakes When Measuring Hook Rate
A few errors show up again and again in creative testing workflows, and each one can make a perfectly good hook look bad or a weak hook look strong:
- Comparing a hook rate measured against impressions to one measured against plays, which almost always makes the plays-based number look artificially higher
- Judging a new hook after very little spend or very few impressions, before the number has stabilized
- Blending placements or formats together, so a strong vertical-feed result gets diluted by a weak landscape in-stream result
- Changing more than the hook between variants — if the middle, offer or CTA also changed, you can no longer attribute the result to the opening seconds alone
- Measuring hook rate once and never again, instead of treating it as an ongoing check every time a new creative batch goes live
How FrameNotion Fits Into a Hook-Testing Workflow
Once you know how to measure hook rate, the bottleneck usually shifts to producing enough hook variants to actually test. FrameNotion is built for that part of the workflow: paste a product link and FrameNotion AI writes and renders a 30-second vertical video ad from scratch, including the hook, in about 10 to 20 minutes. On Pro and Agency plans, you can generate A/B hook variants of the same ad without rebuilding it from zero, which keeps every other element — problem, benefit, proof, offer, CTA — identical so a hook rate comparison is actually clean. Every ad also ships with word-by-word captions and is made to work on mute, which matters because a hook that only works with sound can quietly underperform in placements where autoplay starts silent. You can see finished examples at examples made with FrameNotion or check how FrameNotion works before testing your first batch of hooks.
Frequently asked questions
Is hook rate the same across TikTok, Meta and YouTube?+
No. Each platform can define the metric against a different denominator — impressions versus video plays — and use a different default time threshold. Always check which one you're pulling before comparing numbers across platforms.
What time threshold should I use to measure hook rate?+
Three seconds is a reasonable starting point for most short-form ads. If your ad runs longer than 30 seconds, test a 5-second threshold as well to see how many viewers survive into the problem or benefit section, not just the opening frame.
How much impression volume do I need before trusting a hook rate number?+
There's no fixed number, but treat early results as directional until the ad has run long enough and reached enough people that the percentage stops swinging noticeably each time you refresh the report. A single day of low spend is rarely enough to retire or declare a winning hook.
Should I measure hook rate separately for each placement?+
Yes. A hook can perform very differently in a vertical feed placement versus a landscape in-stream slot, so blending placements into one hook rate number hides which viewing environment is actually working.
Can a good hook rate still lead to poor conversions?+
Yes. Hook rate only measures whether people kept watching past the opening seconds. If the middle of the ad, the offer, or the call to action is weak, a strong hook rate can still end in a poor conversion rate — those are separate problems to diagnose separately.