Hook Rate Explained: The Metric That Sets Your CPM

Most buyers stare at ROAS and CPA all day and never look at the one metric that explains them. Hook rate, explained simply, is the percentage of impressions that turn into a 3-second view. It's the earliest signal of creative quality you get — and the platform uses it to decide how cheaply to keep showing your ad.

The formula

Hook rate = 3-second video views ÷ impressions. If 1,000 people were served your ad and 280 watched past three seconds, your hook rate is 28%. That's it. (It's effectively the same idea as the thumb-stop ratio — different name, same math, depending on which dashboard you live in.)

Why it controls your CPM

Auction-based platforms reward creative that holds attention because attention keeps users in-app. A strong hook rate signals "people like this", so the algorithm distributes the ad more cheaply. A weak one signals "people scroll past this", and you pay a premium for every impression. In practice, fixing a bad hook can drop your CPM more than any amount of audience tinkering.

What's a good hook rate?

Benchmarks move by platform, placement and vertical, so treat these as rough guideposts rather than gospel:

• Under 20%: the opener isn't working — fix it before touching anything else.
• 25–30%: solid, scalable territory for most cold-traffic offers.
• 35%+: a genuine winner; pour budget in and build variants around it.

Your own historical average matters more than any public number. Beat your account's baseline and you're winning.

Hook rate vs hold rate vs CTR

Don't confuse the three. Hook rate measures whether they stop. Hold rate (a.k.a. retention) measures whether they stay. CTR measures whether they act. A high hook rate with terrible hold means a great opener bolted to a weak body — you stopped the scroll and then bored them. Reading all three together tells you exactly where the ad breaks. We go deeper on the full set in the ad creative metrics that actually matter.

How to lift a low hook rate

The lever is almost always the first frame, not the audience. Things that move the number:

• Open on motion or a result, never a logo or a slow pan.
• Make frame one legible muted (most feeds autoplay silent).
• Swap the opener across 4–5 variants and let data choose — see the best video hook types for a starting menu.
• Make it look native, not produced. Polished kills hook rate on cold audiences.

Make hook testing cheap

The reason most buyers never optimize hook rate is that producing five new openers is expensive and slow. It doesn't have to be. Pulling fresh, ready-to-run hooks from a library turns hook testing into a weekly habit instead of a quarterly project — and that habit is what keeps CPMs low as you scale. If hook testing is your bottleneck, a subscription pays for itself in saved production time alone.