Ad Hooks: 18 Scroll-Stopping Examples & Fill-in-the-Blank Templates

Ad hooks are the first one to three seconds of a video ad: the opening line, the first frame, the thing that runs before any pitch. They exist to do one job, which is to buy the rest of your ad a viewer. Everything downstream — your offer, your script, the CPM you already paid — only matters if the hook clears the bar. Below are eighteen ad hooks you can steal, each with a fill-in-the-blank template, the psychology behind it, and the tested scroll-stop rate where one exists.

The data backs the obsession. TikTok's own analysis found that 63% of all successful TikTok ads convey their main message inside the first three seconds, and 45% of people who watch those three seconds stay for at least 30 more. The hook is not the warm-up. It decides whether the warm-up gets watched at all.

Why three seconds decides the whole ad

The brain triages before it reads. It starts processing an ad within about 100 milliseconds, allocates most visual attention by 400ms, and reaches a skip-or-stay decision by one second. At that stage you are not persuading anyone. You are passing a reflex test for novelty and relevance before the conscious mind even gets a vote.

The stakes show up in the retention curve. On TikTok, videos that hold 70 to 85% retention in the first three seconds get 2.2x more total views than baseline, and 85%+ retention earns a 2.8x multiplier. The hook gates your message and the algorithm's decision to keep distributing the ad.

On Meta the same idea has a name and a number. Hook rate is three-second video plays divided by impressions. A healthy account benchmarks at 20 to 25%, with top performers above 30%, and the standard advice is to fix hook rate before hold rate, because viewers who scroll never reach the body. A hook rate benchmark under 20% means you are paying premium money for impressions that never convert.

What changing only the opener actually does to CPM

One teardown of 50 hooks found 300 to 500% variance between variations of the same ad with only the first three seconds swapped. That variance is not abstract; it flows straight into delivery cost. Picture a winning body that draws a 15% hook rate. Meta's auction rewards engagement, so a weak opener forces the system to buy attention you should be earning, and your effective CPM drifts up. Lift that same ad to a 32% hook rate by re-opening it, and the algorithm starts handing you cheaper impressions because viewers are staying. Same product, same offer, same body. The cheapest high-leverage asset you can produce is not a new ad. It is a new first three seconds bolted onto one that already converts.

How to use these ad hook templates

Five rules before the list, because a template used wrong is worse than no template.

One driver per opener. The opening should carry a single, clear attention driver with no competing elements in the first frames. A hook with two ideas has none.
Specific beats clever. Fill the brackets with the most concrete version you can defend. "Most founders" is weaker than "founders running their first paid campaign under $1k a month."
Freeze everything after second three. When you test, change only the opener and keep the body identical, so the hook rate reads cleanly as the effect of the hook and nothing else.
Write for muted playback. A large share of paid social plays silent. On LinkedIn, 80% of viewers watch with sound off, so every spoken hook needs an on-screen text version. TikTok's overlay guidance runs 5 to 10 words per second.
Never open a gap you cannot pay off. A hook that promises a payoff and never delivers produces a sense of betrayal and damages brand sentiment. Every curiosity hook below has to be resolved in the body.

Curiosity hook examples (hooks 1 to 4)

The mechanism is precise. Information Gap Theory, formalized by George Loewenstein at Carnegie Mellon in 1994, holds that curiosity fires when someone becomes aware of a gap between what they know and what they want to know. You do not make people curious by being interesting. You do it by showing them a hole in their own knowledge.

Hook 1: The unnamed thing

Positions the answer as a withheld secret. Template: "Most people don't know this about [topic]." The word "this" does the work, naming a payoff without revealing it. Pay it off fast or you have built a false gap.

Hook 2: The counterintuitive outcome

Creates a logical contradiction the brain wants resolved. Template: "The reason [common thing] keeps failing you isn't what you think." Variant: "We tested [X] vs [Y] for 90 days. The results surprised us." The implied contradiction, you assumed A but reality is B, is the gap.

Hook 3: The numbered breakdown

A list promises a contained, finite payoff, which lowers the perceived cost of watching. Template: "[Number] things I wish I knew before [doing the thing]." The curiosity-gap listicle stopped the scroll 31% of the time in one controlled test, lower than flashier hooks but a strong converter because it pre-frames the whole video.

Hook 4: The cascading gap

For longer formats, resolve one gap while opening another. Cascading gap structures combat the mid-video retention drop-off that kills ads around the 15-second mark, where TikTok's algorithm wants to see 60%+ retention. Template: "The fix isn't [obvious answer], and the part nobody mentions is [next gap]."

Pattern interrupt hook examples (hooks 5 to 8)

Pattern interrupts exploit automatic novelty detection. The feed trains a viewer to expect a rhythm, and you break it. Polish is often the pattern worth breaking: one DTC analysis found 73% of ecommerce video ads fail in the first three seconds because they look overly polished and read instantly as ads.

Hook 5: The stop command

A direct interrupt that contradicts intent. Template: "Don't buy [product category] until you watch this." Telling someone not to do the thing they were about to do is the opposite of a normal ad, which is exactly why it scans as not-an-ad. Pairs well with the disqualifier (hook 12).

Hook 6: The confession

Template: "I wasted $[amount] on [topic]. Here's what I'd do differently." A specific dollar figure plus an admission of failure breaks the polished-brand pattern. This is the skeptic conversion arc, which scroll-stopped 39% in testing.

Hook 7: The accusation

Template: "You've been [doing common thing] wrong your whole life." A controlled provocation: mild enough to play, sharp enough that the viewer needs to know whether it is true about them. Specificity keeps it from reading as an insult.

Hook 8: The visual experiment

Not every interrupt is verbal. A show-and-fix opener, an unexpected on-screen action with no narration, works on muted feeds and exploits motion-based attention. Template (a shot, not a line): open mid-action on [surprising visual], reveal the product as the cause at second two.

Direct-address hook examples (hooks 9 to 12)

Direct address answers the only question a scroller asks: is this about me? The more precisely you name the persona, the higher your hook rate among the people who matter. On TikTok, opening with a face beats product-only openings, with face cam performing 31% better, partly because a face is itself a relevance signal.

Hook 9: The persona call-out

Template: "If you're a [specific role] at a [specific type of company], this is for you." The narrower the bracket, the louder it lands for the right person. "If you're a founder" is noise; "if you're a solo founder running paid ads with no creative team" is a hand on the shoulder. This is the spine of any founder-story ad that needs to feel personal.

Hook 10: The painful-situation match

Template: "If [specific painful situation] sounds familiar, keep watching." You describe a scene from their week so accurately they assume you have lived it. The test: could the viewer have said this exact sentence to a friend last week?

Hook 11: The direct question

Template: "Still [doing the inefficient thing] by hand?" A question forces a silent internal answer, which is engagement before the click. Keep it to one question; a direct-question hook dies the moment it stacks.

Hook 12: The disqualifier

Reverse psychology that filters and intrigues at once. Template: "This isn't for you if [trait]." Excluding people makes the included feel chosen. The reverse-psychology disqualifier scroll-stopped 34% in controlled testing.

Social-proof hook examples (hooks 13 to 16)

Social-proof hooks borrow credibility and trigger FOMO before you make a claim of your own. They work even when the proof is modest, as long as it is specific. You do not need a famous customer. You need a believable result.

Hook 13: The transformation stat

Template: "[Specific customer type] grew from [X] to [Y] in [timeframe] using this." The numbers do the persuading, so your job is to make them defensible. Never invent the figure, because a single fabricated number poisons the rest of the ad and eventually the brand. If you have no result yet, build a testimonial-style ad without real customers rather than faking one.

Hook 14: The earned-authority line

Template: "After [number] years doing [thing], here's what actually works." Time in the trenches is its own credential. The phrasing sets up a contrarian payoff, so the body should genuinely contradict conventional wisdom.

Hook 15: The platform-native reply

Framing the ad as a response to a real comment or DM. This platform-native reply hook scroll-stopped 38% because it mimics organic content rather than advertising. Template: "Someone asked how I [achieved result] without [resource], so here's the actual answer." This is the backbone of convincing AI UGC ads that read like real user content.

Hook 16: The industry secret

The highest scroll-stop rate in one 50-hook test belonged to the industry secret, at 47%. Template: "The thing [industry] doesn't want you to know about [topic]." Use it sparingly, and only when there is a real insight behind it. Overuse teaches your audience to discount you.

Consequence and urgency hook examples (hooks 17 to 18)

The last category establishes personal stakes: the cost of doing nothing. This is not fear-mongering. It is plain math on what inaction costs.

Hook 17: The ongoing cost

Template: "If you're [doing thing], you're losing [specific thing] every [time period]." Concrete and recurring beats vague and one-time. "You're losing customers" is air; "you're losing about three hours every week you'll never bill for" lands. The consequence-and-urgency category in the 40-hook library leans entirely on this kind of personal stake.

Hook 18: The number-one reason

Template: "[Common mistake] is the #1 reason [target customer] fails to achieve [goal]." It promises a diagnosis and, by implication, the fix. The body has to deliver that fix. Open this gap and leave it open and you have burned trust.

The 18 ad hooks ranked by tested scroll-stop rate

Where a controlled test produced a number, here it is in one place. Scroll-stop percentages for hooks 1, 6, 12, 15, and 16 come from the 50-hook teardown; the rest are marked "test it" because their performance depends too heavily on the bracket you fill. Sort by the right-hand column and start at the top.

Hook 16, industry secret: 47% scroll-stop. The current ceiling, but it burns out if you cry wolf.
Hook 6, confession: 39% scroll-stop. Specific dollar loss plus admitted failure.
Hook 15, platform-native reply: 38% scroll-stop. Reads as organic, not ad.
Hook 12, disqualifier: 34% scroll-stop. Exclusion as inclusion.
Hook 3, numbered breakdown: 31% scroll-stop. Lower stop rate, strong downstream conversion.
Hook 9, persona call-out: test it. Stop rate tracks how narrow your bracket is.
Hooks 1, 2, 4, 5, 7, 8, 10, 11, 13, 14, 17, 18: test it. Strong patterns whose number is decided by the specifics you load in.

How to test ad hooks without re-shooting

Templates are inputs. The system around them produces winners. Here is the loop.

Pick one body that already converts. If nothing converts yet, run this on your single best guess so you are isolating the opener.
Write five hooks from five different categories, not five curiosity gaps. Different categories fail differently, so you learn faster.
Keep everything after second three identical: same body, same CTA, same captions. The recommended method is precisely this, to launch the same ad with five different opening hooks and measure hook rate.
Read hook rate first, hold rate second. Target 30 to 40% for a good hook rate, 40%+ for excellent, and 60%+ hold rate. A high hook rate that cliffs right after means the opener over-promised relative to the body.
Change one variable per cycle and read CTR, view rate, and conversion together. Two changes at once teach you nothing clean.
Re-hook the winner weekly. High performers drop about 37% after seven days, and a 20%+ decline warrants archiving. The fix is usually a fresh opener on the same body. This is the practical core of any creative testing system for video ads.

Video ad hook examples: dead opener to scroll-stopper

Same product, a project tool for agency owners. Watch the opener move from brand-first to scroll-stopping.

Before (brand-first, near-zero stop): "Acme is the all-in-one platform for modern agencies."
After, hook 10 (painful-situation match): "If you've ever found out a deadline slipped from the client instead of your own team, keep watching."
After, hook 6 (confession): "I lost a $40k retainer because a status update lived in five places and none of them agreed."
After, hook 17 (ongoing cost): "Every week your status lives in spreadsheets, you're eating hours you can't bill and risk you can't see."

None of the rewrites mention the product. Each names a situation the viewer recognizes and opens a gap the body closes. The opener's companion is a tight body, covered in the four-block ad script framework.

Calibrating ad hooks per platform

The patterns are universal; the calibration is not. Each platform's mechanics reward a different opener, and these are the TikTok hook ideas, Meta plays, and B2B angles worth matching to the surface.

TikTok and Reels

Native and fast wins. TikTok-first content drives 3.3x more actions than repurposed content, and the platform recommends getting your proposition in within three seconds and your hook inside six. Lead with the confession, the visual experiment, or a trend-aware open, since 77% of users like it when brands use trends and memes. Specifics live in the TikTok ad specs guide and the Reels format rules.

Facebook and YouTube

Facebook tolerates a longer, more educational hook and rewards early branding, which boosts ad recall by 23%. The numbered breakdown and persona call-out travel well; full setup in the Facebook structure breakdown. On YouTube the skip button is your deadline: ads average a 31.9% view-through rate across nearly 4,900 tested ads, and the best ones put a brand visual and the product on screen inside five seconds. Front-load curiosity or consequence; the YouTube Shorts guide covers the vertical case.

Visual-first and muted by default. A visible logo in the first two seconds lifts CTR by 17%, captions carry the message, and LinkedIn's own advice is to lead with the key message rather than build to it. The persona call-out and earned-authority line fit best; detail in the B2B LinkedIn playbook.

FAQ

What is a good hook rate benchmark on Meta?

Hook rate is three-second video plays divided by impressions, times 100. A healthy account benchmarks at 20 to 25%, with top creatives above 30%. Fix hook rate before hold rate, because if the opener does not earn the three seconds, nobody reaches your offer. Anything under 20% means you are paying for impressions that never convert.

How many ad hooks should I test at once?

Five per round is a workable default, ideally one from each category, with everything after second three held constant. The recommended method is to launch the same ad with several openers and compare hook rate. TikTok separately suggests running 3 to 5 creatives per ad group, which lines up with the five-hook test.

Why do my ad hooks stop working after a week?

Creative fatigue. High-performing hooks lose around 37% of performance after seven days, and a sustained 20%+ decline is the signal to archive. The cheap fix is a fresh opener on the same body rather than a brand-new ad. The full pattern is in the guide to spotting and fixing ad fatigue fast.

Do AI-written ad hooks work as well as ones I write myself?

They work when they are specific and you test them like any other variant. The trap is generic output: "Most people don't know this!" with no real payoff is a false gap that damages brand sentiment. The leverage is not AI writing one hook; it is generating eighteen, filling the brackets with defensible specifics, and letting hook rate pick the winner. Full comparison in AI vs human ad creators.

Sources

The reason most teams test two hooks instead of twenty is that rendering each opener into a finished, captioned video by hand takes an editor and days. Aitachyon exists to make re-hooking a proven winner a five-minute job: keep the body that already converts, generate five fresh openers, and render them so hook rate can pick the next champion before this week's creative fatigues. It fits founders and indie hackers running several products at once and small teams shipping a new opener per ad set. Start free and run your first five hooks against each other.