The hooks that earn the next two seconds do one of six jobs: they ask a question the viewer already has, state a contradiction, name a number, move before the eye settles, put a face in front of the scroll, or use sound to break a muted feed. Below are 31 concrete examples, grouped by mechanism, with a note on which offer each mechanism suits. Judge them by hook rate, the share of impressions that watch past the first two seconds, not by how clever the line feels in a script doc.
How do you judge a video ad hook?
Use hook rate as the metric. On Meta feed we typically expect 15 to 25 percent; on TikTok in-feed 20 to 30 percent. Composite medians from our 2026 fatigue benchmark put Meta Reels at 28 percent hook rate, TikTok at 33 percent and YouTube in-stream at 22 percent. If a hook holds below those ranges, the next two seconds are the problem, not the CTA.
Most creative is effectively dead within three weeks, according to that benchmark report. That means the hook is not a one-time asset. It is a refresh cadence problem. Brands shipping 15 to 50 creative variants a month see 3 to 5 times longer campaign lifespan than quarterly refreshers. So measure early, cut fast and replace.
Log impressions and three-second views in a hook rate calculator before you debate the creative. A hook that holds 30 percent on TikTok and one that holds 18 percent are not both worth saving. Kill the second even if the first line reads better in a deck.
31 video ad hooks grouped by mechanism
The examples below are written as the first spoken line or on-screen text. They are not industry specific. The mechanism matters more than the vertical.
Question hooks
-
"Are you still running the ad you launched three weeks ago?"
Why it works: it calls out the marketer's own fatigue, which is the first thing a performance buyer feels. -
"What would you do with four extra hours today?"
Why it works: it converts a time saving into a concrete, countable outcome. -
"How many of your best customers found you by accident?"
Why it works: it flips acquisition into a problem the viewer already suspects is true. -
"Why does this moisturiser burn on first use?"
Why it works: it names a negative experience buyers search for, so it reads as honesty rather than objection handling. -
"What if your best ad is already sitting in your camera roll?"
Why it works: it primes the viewer to see the rest of the video as a way to unlock assets they already own.
Contradiction hooks
-
"This ad is not for you. Please keep scrolling."
Why it works: reverse psychology stops the scroll because it violates the expected command. -
"The best time to fix your onboarding was before you launched."
Why it works: it lands as a painful truth for SaaS buyers who ignored onboarding. -
"We raised our prices and churn went down."
Why it works: the claim contradicts the standard pricing playbook and forces a second look. -
"You don't need more traffic. You need fewer clicks that convert."
Why it works: it rejects the metric the viewer is optimising, which creates tension with their current plan. -
"This product looks broken. It is supposed to."
Why it works: it uses visual contradiction, the product's appearance conflicts with its promise, so people stick around to resolve it.
Number hooks
-
"31 hooks. One scroll pattern."
Why it works: a specific count promises a list the viewer can scan and save. -
"We tested 84 versions to find these 3."
Why it works: the large number proves effort and narrows the value to three actionable items. -
"15 creatives a month, or the campaign dies."
Why it works: it turns a cadence benchmark into a threat, which lands for performance marketers. -
"0 to 1,000 signups in 9 days with no audience."
Why it works: the zero-to-result frame promises a repeatable sequence, not just inspiration. -
"3 seconds, 1 decision."
Why it works: it mirrors the viewer's exact situation and reduces the ask to a single binary. -
"We spent £2,300 to learn one thing about our headline."
Why it works: naming a real spend implies the lesson is already paid for, so the viewer gets it free.
Motion hooks
-
"Watch the top right corner. That is the whole ad."
Why it works: it directs the eye to a single moving detail and creates a game the viewer can complete in two seconds. -
"This shift in camera angle is the difference between a click and a scroll."
Why it works: it promises a visual lesson, so the viewer stops to see the change. -
"If this clip makes you blink, the ad lost."
Why it works: it challenges the viewer's own attention and makes them watch more closely. -
"Look what happens when I move the product left."
Why it works: it commands a physical action on screen, which fights the vertical scroll. -
"This transition has a 28 percent hook rate on Meta Reels."
Why it works: it names the metric the viewer is already judged on, so it reads as benchmark data rather than claim.
Face hooks
-
"I am recording this at 11pm because the campaign just broke."
Why it works: it suggests a live, unfiltered moment, which earns more than polished narration. -
"Founder here. We screwed up the launch date."
Why it works: a direct confession from a named person lowers the ad defence. -
"This is the face of someone who spent £4,000 on creative that flopped."
Why it works: it shows the cost of failure on a human face and hooks people who have felt that loss. -
"I am not an actor. I am the person who packs your order."
Why it works: real operational faces beat synthetic presenters for trust, especially in DTC. -
"My co-founder said this hook would never work."
Why it works: it creates an internal conflict the viewer watches to see who wins.
Sound hooks
-
"Turn the sound on. The next three seconds matter."
Why it works: it converts silent autoplay into active sound-on, which raises retention and conditions the viewer. -
"That click you just heard is the product locking."
Why it works: a product sound gives physical proof without needing a demo. -
"This audio had a 33 percent hook rate on TikTok."
Why it works: it cites the platform median as social proof for sound, not just visuals. -
"Hear that? That is the sound of a feed stopping."
Why it works: it defines the moment the ad interrupts the scroll and makes the viewer aware of it. -
"The first second is silent. The second one is not."
Why it works: it teaches the viewer to listen for a pattern, which keeps them past the two-second mark.
Which hook mechanism fits which offer
| Mechanism | Best offer type | Why it works |
|---|---|---|
| Question | SaaS, services, low awareness offers | Opens a known problem you can solve in the next ten seconds. |
| Contradiction | DTC with strong differentiation, mature feeds | Breaks the expected pattern when the viewer has seen too many ads. |
| Number | B2B SaaS, lead gen, performance offers | Quantifies proof and makes a promise feel measurable. |
| Motion | Beauty, food, physical products | Shows texture, transformation or mechanics without explanation. |
| Face | Founder-led DTC, agencies, high-trust purchases | Humanises the offer and cuts the distance between brand and buyer. |
| Sound | TikTok and Reels, app install, impulse offers | Turns a muted scroll into an active listening moment and lifts hook rate. |
How do you turn these hooks into a testing cadence?
Write ten of these hooks, then cut the bottom seven before you shoot. Most accounts need 8 to 20 live variations per active campaign, so the bottleneck is not writing one hook. It is producing enough finished versions to test.
That is where Genyad's hook variation generator fits. You upload your footage once. Genyad transcribes and tags every clip, then each variation is a fresh script, shot selection, voiceover, caption set and export built from that library rather than a re-cut of the same timeline. One standard variation costs 1 credit. The free plan gives you 5 video ad variations plus one AI-generated video with no card, so you can test hooks on real footage before paying.
Genyad does not generate AI avatars or synthetic presenters. If you need a synthetic face reading a script, Genyad is the wrong tool. Genyad also does not score predicted performance. Hook rate is the score you collect yourself.
Frequently asked questions
What is a good hook rate for a video ad?
From experience, TikTok in-feed typically runs 20 to 30 percent and Meta feed 15 to 25 percent. Composite medians from our 2026 fatigue benchmark put Meta Reels at 28 percent hook rate, TikTok at 33 percent and YouTube in-stream at 22 percent. Treat these as bands, not guarantees. If a variant sits below the band, change the hook before you change the landing page.
How many hooks should I test at once?
Most active campaigns need 8 to 20 live variations in rotation. You are not testing one hook against another in isolation. You are testing hooks against fatigue. A hook that clears 30 percent today can be below 20 percent in two weeks. Brands shipping 15 to 50 creative variants a month see 3 to 5 times longer campaign lifespan than quarterly refreshers.
Do I need to write different hooks for each platform?
Yes. A TikTok hook often uses sound and a faster cut. A Meta feed hook can carry more text. The mechanism table above is a starting point. The same footage can produce a Reels version and a feed version without reshooting.
Can Genyad write hooks for me?
Genyad writes a fresh script for each variation natively in English, German, French, Spanish, Italian or Hindi, not translated. It does not generate AI avatars or predicted performance scores. You still judge the hook rate using a hook rate calculator.