The TikTok hooks that work are the ones that arrive mid-thought, with the camera already pointing at something and the sentence already in progress. Eighteen of them are written out below, grouped by the only thing that actually constrains a TikTok hook: the second at which the product is allowed to appear. A hook that promises a verdict on an object can show that object at 0:02. A hook that promises a story cannot, and putting the product in at 0:02 anyway is what makes an ad read as an ad.
The native-pacing rule, and what polish costs on TikTok
TikTok gives you more attention than Meta and does less with it. Our 2026 fatigue benchmark, a synthesis of published platform and agency figures rather than our own measurement, puts TikTok's median hook rate at 33 percent against 28 percent on Meta feed, and TikTok's median video CTR at 0.84 percent against 1.62 percent. Divide those: TikTok hands you 1.18 times the attention and 0.52 times the clicks.
That ratio is the whole brief. Stopping someone on TikTok is comparatively easy, and converting the person you stopped is comparatively hard, so the expensive part of a TikTok ad is not frame one. It is the join between the hook and the pitch. Every hook below is written to make that join survivable, which mostly means not promising more than the next four seconds can pay off.
Four operating rules we brief against, none of them mysterious:
- No establishing shot, no fade, no title card. Frame one is already the middle of something.
- The first spoken word is not the first word of a sentence. Start at word four.
- Something changes in the frame every 1.2 to 2 seconds, including in the hook.
- Vertical, hand-held or hand-held-looking, and one subject. Two subjects in frame one is a scene, and a scene needs setting up.
The polish penalty is not about production value in the abstract. Colour grading and a clean mix cost nothing. What costs is the grammar of a commercial: the wide shot, the reveal, the voiceover that starts at the beginning.
18 TikTok hook ideas, grouped by when the product can appear
Each is written as the on-screen text plus the first spoken line, because on TikTok those two land together and reviewing one without the other is how bad hooks get approved.
Product can appear at 0:02
These promise a verdict on a specific object, so the object has to be there almost immediately. Best for repeat purchases and low prices, where the decision is small.
- On screen: "day 4" / Spoken: "so I have been using this four days and I need to talk about the smell."
- On screen: "wrong size" / Spoken: "ordered the wrong size, keeping it anyway, and here is why that was lucky."
- On screen: "fourteen euros" / Spoken: "fourteen euros. That is the entire reason this video exists."
- On screen: "before you buy" / Spoken: "if you are about to buy one of these, give me six seconds first."
- On screen: "it actually does it" / Spoken: "did not believe this part worked, so I filmed it happening."
- On screen: "two good, one bad" / Spoken: "there are three things in this box and one of them is not good."
Product can appear at 0:04
These promise a problem you recognise, and the product arrives as the answer to it. Best for considered purchases and first-time buyers, where the problem needs naming before the object means anything.
- On screen: "this is not normal" / Spoken: "your hairbrush should not look like this after one week."
- On screen: "three hours" / Spoken: "three hours of my Sunday. Every Sunday. For two years."
- On screen: "watch the corner" / Spoken: "keep your eye on the bottom left, it happens fast."
- On screen: "I stopped in March" / Spoken: "quit doing this the proper way six months ago and nothing bad happened."
- On screen: "inspection is Friday" / Spoken: "four days to fix this before the landlord walks in."
- On screen: "second ingredient" / Spoken: "the second one on this list is the one nobody warns you about."
Product can appear at 0:07 or later
These promise a person changing their mind, which takes longer to set up and pays off in trust. Best for higher prices, subscriptions and anything where the objection is scepticism rather than money.
- On screen: "I was the problem" / Spoken: "blamed the water for two years. It was not the water."
- On screen: "he told me not to" / Spoken: "my physio said stop doing this and I ignored him for a month."
- On screen: "twenty-three of them" / Spoken: "twenty-three of these, bought with my own money, ranked worst to best."
- On screen: "the shop version" / Spoken: "this is the one in the shop. This is what is actually inside it."
- On screen: "nobody asked me to" / Spoken: "nobody asked me to talk about this, which is the only reason I am."
- On screen: "week 1 vs week 6" / Spoken: "same face, same light, same time of morning, five weeks apart."
Write the on-screen text at four words or fewer. It is being read in the same second as the first spoken clause, and anything longer is still being read when the product should be arriving.
When to reveal the product, and what each timing costs
| Hook family | What the opening promises | Product appears at | Runtime it needs | Suits |
|---|---|---|---|---|
| Product-first, hooks 1 to 6 | A verdict on this specific object | 0:02 | 12 to 18s | Repeat purchase, low price, impulse |
| Problem-first, hooks 7 to 12 | A problem the viewer recognises | 0:04 | 15 to 25s | Considered purchase, first-time buyer |
| Story-first, hooks 13 to 18 | A person changing their mind | 0:07 | 25 to 40s | Higher price, subscription, sceptical buyer |
The reveal timing is a reach decision, and it is worth being honest about what we can and cannot put a number on. The benchmark gives a 33 percent median at the three-second mark and nothing beyond it, so the share of impressions still present at 0:07 is unknown. It is strictly lower than the share present at 0:02. A story-first hook therefore only pays when the smaller audience is worth more per person, which is why the third row maps to higher prices and the first row does not.
The failure mode in each direction is specific. Reveal too early on a problem-first hook and the ad answers a question the viewer has not been asked yet, so it reads as a pitch. Reveal too late on a product-first hook and the promise of a verdict goes unpaid, so completion collapses while hook rate looks fine.
Why TikTok needs a weekly refresh, worked from the frequency ceiling
TikTok's frequency ceiling is about 3.0 a week, and the benchmark report cites Meta internal research putting the CTR drop at 45 percent after a fourth exposure to the same creative. Divide the ceiling by seven to get exposures per day, then divide four by that.
| Placement | Weekly frequency ceiling | Exposures per day | Days to the fourth exposure | Refresh cadence that fits |
|---|---|---|---|---|
| TikTok | about 3.0 | 0.43 | 9.3 | Weekly |
| Meta Reels | 2.5 | 0.36 | 11.2 | Fortnightly, running about three days long |
| Meta feed | 2.5 | 0.36 | 11.2 | Fortnightly, running about three days long |
| YouTube in-stream | 3 to 7 per week | 0.43 to 1.0 | 4.0 to 9.3 | Weekly at the top of that range |
Nine days. That is where the weekly TikTok cadence comes from, and it is not a preference. It also lines up with the fastest vertical in the report, food and beverage, which reaches a 40 percent CTR decline in 9 days. Fashion takes 12 to 14 days, beauty and DTC about 18, electronics about 21, and B2B SaaS about 28, so a slow-cycle B2B account on TikTok can stretch the cadence and a food brand cannot.
Now turn the 9.3 day life into a production requirement. A typical active campaign holds 8 to 20 live variations, and each one dies after about 1.33 weeks.
| Live variations in rotation | Replacements per week | Replacements per month |
|---|---|---|
| 8 | 6 | 26 |
| 12 | 9 | 39 |
| 16 | 12 | 52 |
| 20 | 15 | 65 |
The benchmark's 15 to 50 variants a month is 3.5 to 11.5 a week, which covers everything up to about 16 live variations. Past that the arithmetic asks for more than the top of the observed range, which is a fair explanation of why 20 live variations on TikTok is a large-account behaviour rather than a default. Eighteen hooks, as listed above, is roughly two weeks of a 9-variation rotation.
Where Genyad fits, and what it will not do
Genyad is our product, so read this as disclosure. The list above is free and the rotation arithmetic is the expensive part: 9 finished 9:16 files a week is not something a single editor produces alongside anything else. You upload footage once, it transcribes and tags every clip, then each variation comes out as a fresh script, shot selection, voiceover, caption set and export drawn from that library rather than a re-cut of one timeline. One variation costs 1 credit. At Growth rates that is about EUR 1.52 per variation, so a 9-per-week rotation runs at roughly EUR 14 a week, and a 65-credit pack covers about 13 weeks of it. The free plan gives 5 variations with no card. Scripts are written natively in English, German, French, Spanish, Italian or Hindi rather than translated, which matters for hooks specifically, since a translated opening line is as easy to spot on TikTok as a fade from black. Openings are what the hook variation generator and the TikTok ad maker routes are for.
What it does not do. No AI avatars or synthetic presenters, so hooks 13 to 18 need a person you have actually filmed; if you have no footage of a human being, Arcads was listed at EUR 100 per month for 10 AI actor videos in August 2026 and is the better fit. No static banner formats. No product-URL import and no product-feed or CSV-driven template rendering. No predicted performance scores, so nothing will rank these eighteen hooks before they run. And no direct publishing to TikTok, so you export the files and upload them in TikTok Ads Manager yourself.
Frequently asked questions
What makes a TikTok hook work rather than a Meta one?
TikTok gives about 1.18 times Meta feed's median hook rate and about half its video CTR, so attention is cheap there and the transition to the pitch is expensive. Practically that means the opening should promise less: a verdict on one object or one recognisable problem, paid off inside four seconds, rather than a broad curiosity gap you have to earn your way out of.
When should the product appear in a TikTok ad?
At 0:02 if the hook promised a verdict on that object, at 0:04 if it promised a problem, and at 0:07 or later if it promised a change of mind. Revealing earlier than the promise justifies makes the video read as a pitch, and revealing later than it justifies leaves the promise unpaid and collapses completion rate.
How often should TikTok ad creative be refreshed?
Weekly. At a frequency ceiling of about 3.0 a week, a creative reaches its fourth exposure in roughly 9 days, and that is where the CTR drop of 45 percent lands per Meta internal research cited in our benchmark report. Fortnightly is defensible on Meta feed, where the same arithmetic gives about 11 days, and it is three to five days too slow on TikTok.
How many hooks do I need in rotation on TikTok?
If you hold 12 live variations and each lasts about 9 days, you are replacing about 9 a week, or 39 a month. Eight live variations needs about 6 a week. Those figures sit inside the 15 to 50 variants a month that our benchmark associates with 3 to 5 times longer campaign lifespan than a quarterly refresh.
Can these hooks be reused across verticals?
The structures transfer and the specifics do not. Hook 7 works because a hairbrush after a week is a recognisable and slightly unpleasant object, and swapping in an abstract noun kills it. Rewrite each one around something in your own footage that a viewer can name in half a second.