Hooks That Stop the Scroll: 30 Openers From Top-Performing TikTok Ads
We pulled 30 hook formulas from TikTok ads that ran 21+ days with rising spend. Here's every opener, why it stops the thumb, and how to adapt it without copying.
By
Kinnari Ashar
Read time: 6min
Most TikTok ads don’t lose on the offer. They lose in the first second, before anyone hears it. We went through thousands of long-running TikTok ads in the Winning Hunter Ad Library and pulled the 30 openers that kept showing up on ads with rising spend.
Why the first second decides your CPA
TikTok’s algorithm reads early retention as a quality signal. If viewers swipe before the two-second mark, delivery gets more expensive, fast. On the ads we tracked, creatives with a hook rate above 30% (3-second views divided by impressions) ran at a CPA 38% lower on average than the account median.
That makes the hook the highest-leverage part of the ad. A few things we saw over and over:
The product appears late, or not at all, in the first second. The hook sells the next ten seconds, not the item.
Text on screen does half the work. Around 70% of top ads had a caption in frame one, because a lot of people scroll with the sound off.
Faces beat packshots. A person talking to camera or reacting outperformed product-only openers in every niche we looked at.
Winners get re-shot, not retired. The same hook often ran in five or six variations from the same store.
How we picked these 30 hooks
We didn’t sort by views. Views are easy to buy and easy to fake. We looked for hooks that advertisers kept paying for, which is a much stronger signal.
Filtered the Ad Library to TikTok, last 90 days, in the US, UK, Canada and Australia.
Kept ads live for 21 days or more with an upward estimated spend curve.
Grouped similar openers with Magic AI‘s similar-ad search, so one hook copied by 40 stores counts once.
Kept the 30 patterns that appeared across at least three unrelated niches.
The Ad Library filters we used
Two filters did most of the work. If you want to rebuild the list for your own niche, start with these.
Longevity over spend
A big spend number on a three-day-old ad often means a test that’s about to be killed. Longevity tells you the ad survived several rounds of budget decisions. Sort by days active first, then check the spend trend.
The 21-day rule
Under three weeks, you’re mostly looking at tests. Past 21 days with spend still climbing, the advertiser has seen enough data to keep scaling. That’s the pool we pulled from.
Cross-platform proof
We cross-checked every candidate against Meta. A hook that also runs on Reels or Facebook is usually a real mechanism, not a quirk of one platform’s algorithm.
Same hook, different wrapper
Stores rarely post the identical file twice. Look for the same opening line with a new creator, a new background or a new product angle. Reverse search makes these clusters easy to spot.
Hooks 1–10: pattern interrupts
These break the rhythm of the feed, visually or verbally. They work best in impulse niches: gadgets, beauty, home and pets.
“Stop scrolling if you have [problem].” Blunt, but it still works as long as the problem is specific.
“I wasn’t going to post this, but…” Suggests the viewer is seeing something they shouldn’t.
“This is the weirdest thing I’ve bought this year.” Curiosity first, product second.
A product dropped, squeezed or spilled in frame one, with no voiceover for the first second.
“POV: you finally found a [product] that actually works.” Native to TikTok and easy to shoot.
“Don’t buy [category] until you watch this.” Positions the ad as advice, not a pitch.
A before/after split screen that resolves in under two seconds.
“My [family member] thought I was crazy for buying this.” Adds a second character and a bit of tension.
An extreme close-up that only makes sense once the camera pulls back.
“Things TikTok made me buy, part 14.” The part number suggests a series worth following.
Hooks 11–20: problem and proof
These lead with the pain or with evidence. They take a beat longer to land but convert better on considered purchases above $40.
“If your [skin / room / dog] does this, it’s not your fault.” Removes blame, then offers a fix.
“I tried every [category] out there so you don’t have to.” A comparison without naming anyone.
“Here’s why your [routine] isn’t working.” Diagnosis before the product.
“We’ve sold 40,000 of these and here’s the one complaint we get.” Pre-empts the objection with a number.
A screenshot of a customer review, read aloud by the founder.
“Day 1 vs day 30.” Works for skincare, fitness and anything with visible progress.
“My trainer told me to stop doing this.” Borrowed authority, used carefully and honestly.
“This fixed my [problem] in 10 minutes.” A concrete time frame beats a vague promise.
“3 signs you need a [product].” List formats hold attention because viewers wait for number three.
“I returned the expensive one and kept this.” Price anchoring in one sentence.
We stopped writing hooks from scratch. Every Monday we pull ten long-running openers from the Ad Library, rewrite them for our product, and let spend decide by Friday.
That’s how Marco Ferrand, media buyer at a mid-size home goods store, describes his weekly routine. It’s less glamorous than brainstorming, and it works better.
Hooks 21–30: curiosity and identity
These speak to who the viewer is, or open a loop the brain wants to close. They’re strong for apparel, accessories and anything tied to self-image.
“Nobody talks about this, but…” Opens a loop and implies insider knowledge.
“If you’re a [identity], you need to see this.” Nurses, new dads, runners, dog moms: the tighter the better.
“I found the dupe.” Works without naming the original, which also keeps you on the right side of ad policy.
“Rating every [product] I own.” Promises a verdict and keeps people to the end.
“The $19 thing that replaced my $200 [tool].” The price contrast does the selling.
“Things that just make sense in your 30s.” Age-based identity hooks travel well across niches.
“Wait for it.” Overused, but still strong when the payoff lands before second five.
“Replying to @[user]: does it actually work?” The comment-reply format feels organic and answers doubts up front.
“Packing orders for my small business.” Behind-the-scenes content that sells without looking like an ad.
“You’re using your [product] wrong.” Corrects the viewer, and people hate being wrong enough to keep watching.
How to adapt a hook without copying it
Copying a competitor’s ad frame by frame is lazy and risky. What you want is the mechanism: the reason the hook stops the thumb. Once you have that, you can rebuild it around your own product.
Save five to ten hooks from your niche into a board in the Ad Library.
Label each one with its mechanism: interrupt, problem, proof, curiosity or identity.
Ask the AI agent to “Write a script based on this hook” for your product, then rewrite it in your own voice.
Shoot three variations of each and launch them in one batch.
The hook remix workflow
The routine above takes about two hours a week. Two details decide whether it produces winners or noise.
Keep the mechanism, change the words
If the original uses a price contrast, keep the price contrast and change the numbers, the product and the setting. The viewer should feel the same tension without seeing the same ad.
Swap the specifics
Change the creator, the location, the first line and the product angle. Keep the rhythm: how fast the payoff arrives and where the text sits on screen.
Red flag
If you can’t explain why a hook works in one sentence, don’t adapt it yet. You’ll copy the surface and miss the reason.
Test hooks in batches of five
Put the same body and offer behind five different hooks. That isolates the variable and gives you a clean read in three to five days at modest spend.
Same body, same CTA, same landing page.
One ad group, equal budget per creative.
Kill anything under a 20% hook rate after 2,000 impressions.
Read the right metric
Judge the hook on hook rate and hold rate, not ROAS. A great hook on a weak body still loses money, but it tells you what to keep for the next round.
Metric to watch
Hook rate is 3-second views divided by impressions. Above 30% is strong on TikTok, 25–30% is workable, and under 20% means the opener needs another pass. Once you have a winner, use Brand tracking to see whether competitors pick it up, and go back to our creative testing framework for the next batch.
