Hook Rate & Hold Rate Benchmarks: What Good Looks Like
Observed hook-rate and hold-rate bands by platform, vertical and funnel stage, plus the diagnosis matrix for which half of the ad to recut.
Hook rate tells you whether the first three seconds earned the watch; hold rate tells you whether the rest of the ad kept it. This sheet is the diagnosis table we use on cold UGC: what good looks like by platform, vertical and funnel stage, and what to recut when one number is healthy and the other is not.
Read hook and hold in order, never as blended watch time
Most teams glance at average watch time and argue about whether seven seconds is good. Average watch time mixes two different failures into one number: people who never stopped scrolling, and people who stopped then left at second eight. Hook rate isolates the first failure (3-second views divided by impressions). Hold rate isolates the second (15-second views, or thru-plays, divided by impressions). Until you split them, you will recut the wrong half of the ad. A creative with a dead opener and a brilliant demo looks identical, on average watch time, to a creative with a brilliant opener and a dead demo. The tables below are built so those two cases never share a diagnosis.
The four combinations and what each one actually means
You only need a two-by-two. Weak hook with any hold: the first three seconds lost the auction; recut the opener and leave the body alone. Strong hook with weak hold: they stayed, then the proof, pacing or demo leaked them; recut seconds 3-15 and keep the opener. Strong hook with healthy hold: stop editing the video and read CTR and CPA next — the watch problem is solved. Weak hook with a surprisingly high hold: usually a sample-size issue or warm traffic mixed into a supposedly cold ad set, not a creative miracle. The locked matrix turns those four cells into numbered kill, recut and scale rules, including the impression count at which you are allowed to believe the cell.
Platform, vertical and funnel move the bar, not the order
A talking-head UGC on TikTok is not judged on the same hook band as a 4:5 feed cut of the same product, and a skincare demo will hold longer than an apparel haul of the same length. Funnel stage moves the bar again: warm and hot audiences forgive a slower open because they already know the product. What does not change is the order of operations — always hook, then hold, then click, then purchase. If you skip to CPA you will kill ads whose only problem was a first frame. When the diagnosis says opener, spin hook variants off one product URL in Klip Kanvas instead of reshooting a body that was never the leak.
When the percentage is allowed to count
A dead-looking hook on a few hundred impressions is noise. The same hook on several thousand impressions is a kill. Most premature creative decisions come from reading a real-looking percentage on a sample that cannot support it yet — especially hold rate, which needs more impressions than hook because a 15-second view is a rarer event. The locked sheet includes the impression floors we use before we trust each metric, the early-kill checkpoints that match our testing-budget rules, and the one case where you may kill early: a truly dead opener once a few thousand impressions have landed. Until those floors, the only honest status is too early, which is a different decision from iterate or scale.
6 tables inside: hook-rate bands by platform, hold-rate bands by platform, hook and hold by vertical, both metrics by funnel stage, the hook×hold diagnosis matrix with recut/kill/scale rules, and impression floors plus early-kill checkpoints.
Unlock the full cheatsheet
Enter your email and we'll instantly unlock the complete cheatsheet — plus send you fresh ad playbooks every month.
No spam. Unsubscribe anytime.
Ready to put this into practice?
Create your first AI UGC video ad in minutes — no filming, no actors, no editing.
Try Klip Kanvas freeMore in this section
Win/Loss Creative Readout Sheet
A scoring sheet for creative tests: sample-size floors, win/loss/inconclusive bands by metric, what to scale vs kill vs iterate, and the one-page readout the next sprint actually uses.
TikTok Spark Ads Cheatsheet
Spark Ads as a lookup sheet: authorization codes, post-first rules, inherited social proof, CTA overlay, expiry bands and QA versus cold in-feed uploads.
Ready to make ads like these?
Paste a product link and Klip Kanvas writes the script, casts the creator and renders the ad — no filming, no actors, no editing.