On TikTok, most videos lose the viewer in the first two seconds. If the hook fails, nothing else in the ad matters: not the offer, not the product, not the landing page. That is why hook rate is the first metric we read on any new creative.
What hook rate is
Hook rate is the percentage of impressions that become 3-second video views.
Hook rate = 3-second video views / impressions
It tells you how many people stopped scrolling long enough to see the start of your video. It is available within a few thousand impressions, which makes it the fastest creative signal you have.
Why not just use CPA?
CPA is the right final judge, but it is slow and noisy. To read CPA with confidence you may need 30 or more conversions per creative. At a $40 CPA that is $1,200 per variant. Hook rate gives you a useful read at $50 to $100 per variant.
| Metric | Needs | Tells you |
|---|---|---|
| Hook rate | 5,000 to 10,000 impressions | Did the opening earn attention? |
| Hold rate | Same | Did the body keep it? |
| CTR | Same to 20,000 impressions | Did the offer earn the click? |
| CPA | 30+ conversions | Does it pay back? |
Benchmarks to use as rules of thumb
Benchmarks vary by category, so treat these as starting points and calibrate on your own account.
- Below 15%: the hook is weak. Rewrite the first two seconds.
- 15 to 25%: average. Test a stronger opening before spending more.
- 25 to 35%: good. Move to body and offer tests.
- Above 35%: strong. Promote to scaling and protect it from fatigue.
Five hook types worth testing
1. Problem callout
Name the frustration directly. Example structure: a person holds the problem object to camera with a line that says what is wrong with it.
2. Demo first
Show the product doing the surprising thing in the first second. No intro, no logo.
3. Social proof
Open with a real comment or review on screen and a creator reacting.
4. Pattern interrupt
Start in the middle of an action, with unusual framing or a fast cut. The aim is to break the scroll rhythm.
5. Curiosity gap
Pose a question the viewer needs answered. Avoid clickbait the video does not pay off.
A simple hook testing loop
- Pick one angle and one body. Keep them identical across variants.
- Film or edit six different hooks, each two to three seconds.
- Launch all six in one test ad group with equal budget, about $50 to $100 each.
- Read hook rate at 5,000 to 10,000 impressions per variant, usually within 48 to 72 hours.
- Take the top one or two and read hold rate and CTR.
- Promote winners to the scaling campaign. Write down why the losers lost.
When hook rate is good but nothing sells
Read the next gate. A high hook rate with low hold rate means the body does not deliver on the promise of the opening. A high hold rate with low CTR means the offer or call to action is weak. A good CTR with poor conversion points to the landing page, which is often a CRO issue rather than a creative one.
A worked example
Say you test six hooks on one body, each with $75 of spend. After 72 hours the results look like this.
| Hook | Impressions | 3-sec views | Hook rate | Decision |
|---|---|---|---|---|
| A. Problem callout | 7,400 | 2,368 | 32% | Keep |
| B. Demo first | 7,100 | 2,201 | 31% | Keep |
| C. Social proof | 7,800 | 1,326 | 17% | Iterate |
| D. Pattern interrupt | 7,000 | 1,050 | 15% | Kill |
| E. Curiosity gap | 7,300 | 1,022 | 14% | Kill |
| F. Question | 7,200 | 864 | 12% | Kill |
You carry A and B into a second test that measures hold rate, CTR and cost per purchase. The four losers go in the log with a note on why. Next week, you write six new hooks that build on what A and B have in common.
How to brief creators for hooks
- Ask for three alternative openings for every video, not one.
- Give a clear first-frame rule: no logo, no greeting, the subject must be in motion by second one.
- Ask for raw footage so you can re-cut hooks without a new shoot.
- Agree usage rights and Spark authorisation codes before filming.
Common mistakes
- Chasing hook rate alone. A hook that attracts the wrong people produces cheap views and expensive purchases. Check conversion later in the funnel.
- Changing several variables. Test hooks against one fixed body so you know what moved the number.
- Editing live ads. It resets learning and muddies data. Publish a new variant instead.
- Ignoring fatigue. Strong hooks decay in one to two weeks at scale. Keep new ones in the pipeline.
We build this loop into every TikTok Ads engagement alongside a weekly creative sprint. Want a hook-rate read on your current ads? Request a free audit.