30 video ad hook templates by industry, with a claim check on each

By Ranjit Bhinge, founder of Blur Studio · Updated November 4, 2026 · 11 min read

Short answer: Pick the hook by what the viewer already knows: people who do not know they have a problem need a different opening than people who are comparing options. Below are 30 fill-in-the-blank hook templates in three groups, each with a first-frame shot and a tag: safe, needs evidence, or regulated. A hook is the first two to three seconds of a video ad, and hook rate is the share of people who watch past the first three seconds. For how to test and rank the hooks you pick, use my hook test plan and scoring rubric.

Writing the opening line is the part most people get stuck on, and it is also the part you can test cheapest. Start from a template, make it true for your business, and film it.

I'm Ranjit, and I run Blur Studio: 15+ years building software, 120 apps taken live, and clients who have gone on to raise $10M+. I build AI systems and processes that help businesses do a lot more with a lot less time. I also work differently from most of what else is out there, more on that further down this page. Say hello.

A hook is the opening of a video ad, the first two or three seconds that decide whether a stranger keeps watching or scrolls on. This page is a swipe file: a bank of ideas you can copy, fill in and film. It is not a second testing guide. I already wrote the testing parts, and I link them near the end so you do not read the same advice twice.

Every template is written as two things: a line of text for the first frame, and a short description of what the camera shows. That is on purpose. Many people watch with the sound off, so the first frame has to carry the hook without a voice. The industry examples are made up to show the pattern. They are not real businesses and not real results.

How to use this list

Do not try all 30. Work like this:

Decide who you are talking to (the awareness group below).
Pick three templates from that group that fit something true about your business.
Fill the brackets with a real detail. A hook with a real detail beats a clever hook with a vague one.
Check the tag. If it says "needs evidence" or "regulated", do the check in the next section before you film.
Film the opening only, then attach it to a body you already have. See "Film one body, swap the openings" below.

The groups follow the idea of viewer awareness, which the video studio Flowjam uses to sort hooks: the less the viewer knows, the more the opening has to earn attention before it can say anything about you. I use three groups: people who do not know they have a problem, people who know the problem, and people comparing options.

Pick the hook by what the viewer already knows

Viewer	What they know	What the hook must do
Does not know there is a problem	Nothing about your category	Show something surprising or familiar, then raise a question
Knows the problem	They feel it, but have not looked for an answer	Name the problem in their words, promise a way to deal with it
Comparing options	They know the solutions, not you	Make the choice easy: show what you get, side by side, with proof

One caution from Elevarus, an agency that writes about Facebook ad hooks: a broad hook can win on hook rate and still attract the wrong people. A hook that stops everyone is not the goal. A hook that stops the right people is.

Hooks that need evidence before you run them

In the United States, the Federal Trade Commission (FTC, the national consumer protection regulator) says advertisers need a reasonable basis for each claim before the ad runs, and its small business guide notes that an ad agency can also be responsible for the claims it helps make. The FTC page is in the sources. This is information, not legal advice, and your country or platform may add rules. Elevarus also notes that hooks which overpromise can get an ad disapproved.

So each template below carries one of three tags:

Safe: no result, price, comparison or promise inside it. You still need to be telling the truth.
Needs evidence: a number, a price, a "best", "most", "fastest", a customer quote, a before and after, or a scarcity claim. Keep the proof in a folder: an invoice, a dated photo, a signed quote, a screenshot of the review.
Regulated: health, money, housing or a guarantee. These areas have extra rules and the platforms look harder. Read the rules for your sector before you film, and use neutral wording.

Ten hooks for people who do not know the problem exists

These viewers are not looking for you. Lead with something they recognise or something odd.

#	Template (first-frame text)	First-frame shot	Tag
1	Most [people like you] never check [hidden thing].	Close-up of the hidden thing	Needs evidence ("most")
2	This is what [common object] looks like after [time].	The worn or damaged object	Safe
3	If you own a [type of place], watch this.	You standing in a place like theirs	Safe
4	I did not know [surprising fact about my trade] until [year].	You in work clothes, looking at camera	Needs evidence if the fact is a claim
5	[Number] seconds of [your craft] you never see.	Close-up of the process	Safe
6	Why does [everyday annoyance] keep happening?	The annoyance, in action	Safe
7	This is what [amount] gets you at [place].	The product on a table, price as text	Needs evidence (price)
8	[Number] mistakes I keep finding in [type of job].	You pointing at the first mistake	Needs evidence (number)
9	Nobody tells you this before you [life event].	Boxes, keys, a new space	Safe
10	Watch this [process] from start to finish.	Fast time-lapse of the whole job	Safe (if it is the real job)

Ten hooks for people who know the problem

These viewers already feel the pain. Name it in their words, then point to a way out.

#	Template (first-frame text)	First-frame shot	Tag
11	Tired of [specific problem]? Here is what I do.	The problem on screen	Needs evidence (that it works)
12	If your [thing] is doing [symptom], check this first.	The symptom up close	Safe (not for health topics)
13	The real reason your [thing] keeps [failing].	The failed part held up to camera	Needs evidence (cause)
14	Before you hire anyone for [job], ask these [number] questions.	A handwritten list on a table	Safe
15	[Problem] can cost about [amount] a [period]. Check yours.	A bill or a meter	Needs evidence (amount)
16	I get this call every [season]: [customer's problem].	A phone ringing on a real job site	Safe (if true)
17	Do not do this when [situation].	The wrong way, shown plainly	Safe
18	A customer asked me [real question]. Honest answer:	The question on a phone screen	Safe (question must be real)
19	Still doing [old way]? There is a quicker one.	The slow way, then a cut	Needs evidence ("quicker")
20	From this to this in [time].	Real before shot, hard cut to real after	Needs evidence; regulated for cosmetic or health work

Template 18 is only safe if a customer really asked. Do not invent the customer. If the question is a composite, say so in the caption or choose another template.

Ten hooks for people comparing options

These viewers are shopping. Make the choice easier and show proof.

#	Template (first-frame text)	First-frame shot	Tag
21	[Option A] or [Option B]? How I choose.	Both options on a table	Safe
22	What [amount] gets you at three different [places].	Three receipts or three menus	Needs evidence (fair, dated comparison)
23	Why customers switch from [type of provider] to us.	A real job with your van or shopfront	Needs evidence ("customers switch")
24	Real review: "[exact quote]" from [first name], [town].	Review text over a real photo of the job	Needs evidence (permission, genuine review)
25	[Years] doing [job] in [town]: what I would check before you buy.	You on site	Needs evidence (years, if stated)
26	Here is exactly what is in the [package].	Contents laid out on a table	Safe
27	[Product A] next to [Product B], side by side.	Split screen, same lighting	Safe if the comparison is fair
28	Our [guarantee]: [terms in one line].	The written terms on screen	Regulated (guarantees)
29	[Number] [slots/tables/appointments] left this week.	Your real calendar	Needs evidence (scarcity must be real)
30	Ask me anything about [offer]. First question:	A real message on screen	Safe

Examples by industry

Here is how the templates look when filled in. These are invented to show the shape. Swap in your own facts.

Industry	Template	Filled in (illustrative)	Tag and what to keep on file
Trades (plumber)	8	Three mistakes I keep finding in old bathrooms.	Needs evidence: dated photos of each mistake from real jobs
Trades (electrician)	17	Do not do this when a fuse keeps tripping.	Safe, but keep safety advice general and point to a qualified person
Restaurant	7	This is what 12 dollars gets you at lunch.	Needs evidence: the current menu price
Restaurant	26	Here is exactly what is in the family box.	Safe: film the real box
Retail (shop)	27	Our two best-selling lamps, side by side.	Safe, if "best-selling" is true; otherwise drop the phrase
Retail (shop)	29	Six gift boxes left this week.	Needs evidence: a real stock count
Property	9	Nobody tells you this before you sign a rental lease.	Regulated sector (housing): describe the property and process, not the kind of person who should live there. See my property manager video page
Property	21	Renting or buying in [area]? How I compare the monthly cost.	Regulated: use real numbers with a date
Clinic	12	If your [body part] is aching, what to ask your doctor first.	Regulated (health): general advice, no cure or result claims, check your regulator and the platform's health ad rules
Clinic	14	Before you book any [type of treatment], ask these four questions.	Safe if the questions are neutral
B2B (accounting firm)	15	Late filing can cost small firms about 500 a year. Check yours.	Needs evidence: cite the official penalty page you used
B2B (software)	19	Still copying orders by hand? There is a quicker way.	Needs evidence: time two real orders, both ways

Hooks to leave out

Flowjam lists hooks that tend to backfire, and I agree with each one:

Fake notifications or fake app screens. Viewers feel tricked, and it can read as deceptive.
A "point of view" opener with no point-of-view shot. If the line says "POV: you just found the best coffee in town", the picture has to match.
An AI avatar pretending to be a customer. It is a fake testimonial. It also brings in disclosure rules. TikTok's help page says a label is required on realistic AI content, and a New York law on synthetic performers in ads took effect on 9 June 2026, with a $1,000 first violation and $5,000 for later ones, according to the bill text. Check whether these apply to you.
Fake urgency. A countdown that resets every day is a claim that is not true.

The simplest rule: if the hook only works because the viewer is misled, it is not a hook, it is a risk.

Make the first frame carry it (sound off)

Write the hook as text on the first frame, big enough to read on a phone, with the picture doing the explaining. A good check: mute the video, look at the first frame for one second, and ask what a stranger thinks the video is about. If the answer is "nothing", the hook is in the voice only, and many viewers will never hear it.

Keep text out of the screen areas that platform buttons cover. Each platform sets its own safe zones, so check the current ad specs before you finalise the layout.

Film one body, swap the openings

The fastest way to use a swipe file is to film one body (the part after the hook) and then film several openings for it. Heath Media, an agency that sells Meta ad services, describes filming a set of visual hooks and a set of spoken hooks on a fixed body and mixing them. You do not need that scale. Three openings on one body is already a real test, and it costs one extra half hour of filming.

AI can help here, but only in one place. If the hook needs a scene you cannot film cheaply, you can generate just the opening scene with an AI video tool such as Higgsfield (an AI platform for images and video) or Runway (an AI video generator) and keep the rest real. Do not use AI people as fake customers, and keep the generated scene honest about the product. My page on testing 20 ad variations and my batch variants guide cover how to handle the volume without losing track of what was changed.

Hook rate in one paragraph

Hook rate is a ratio you work out yourself. One tool vendor, Vaizle, defines it as 3-second video plays divided by impressions (the number of times the ad was shown), times 100. Hold rate is ThruPlays divided by 3-second plays, where a ThruPlay is the view count Meta reports for people who watch a video to a set length. Neither is a default column in Meta Ads Manager (Meta's tool for running Facebook and Instagram ads), so you add them as custom metrics. Check how each tool defines its denominator, because some use impressions, some use reach, some use video starts.

Made-up example: 2,800 three-second plays out of 10,000 impressions is a 28% hook rate. Now compare two hooks with made-up numbers. Hook A: 40% hook rate and 15% hold rate, which multiplies to 6%. Hook B: 28% hook rate and 30% hold rate, which multiplies to 8.4%. Hook A stopped more thumbs, but Hook B kept more of the people it stopped, so B can be the better ad.

On benchmarks, vendor pages disagree, with figures from 20 to 30%. The only figure with a stated sample is from Billo, a company that sells creator-made video ads: its report (published 6 August 2026) says 88,329 Meta sales video ads that ran from January to June 2026, each with more than 1,000 impressions, averaged a 25.44% hook rate, defined as 3-second plays divided by impressions. That is an average across many categories, not a target. Compare against your own account first. Do not compare numbers across platforms either, because each platform may count a "view" differently.

Test the best three

This page stops at ideas. For the test itself, use these:

Pick your three best templates.
Run them on one ad body with equal budgets. The steps, the 7-day plan and the claim log are in Ad Hook Tests.
Score the results with the rubric in How to Rank Video Hooks.
If you are comparing whole videos, not just openings, see Five Video Versions, One Budget.

How I'd approach this

I would not write 30 hooks. I would write six: two from each awareness group, each with a true detail from the last month of real work. I would tag each one, collect the proof for the ones that need it, and film all six in one afternoon on a phone. Then I would attach them to the one ad body that already works and let the numbers pick the winner. The tags are there so that the first thing you learn is not a disapproved ad.

If you run this every month for several clients, the slow part is not the ideas, it is tracking which hook went on which body, what proof was filed and who approved it. Blur Studio builds the repeatable workflow, approval portal and shared asset library around that. If that is your problem, have a look at what I build.

Questions people ask

What is hook rate and how do I calculate it?

Hook rate is the share of people who watch past the first three seconds. One vendor, Vaizle, defines it as 3-second video plays divided by impressions, times 100. For example, 2,800 plays out of 10,000 impressions is 28%. Check how your tool counts the denominator, since some use reach or video starts, and note that platforms may not count a view the same way.

What is a good hook rate?

There is no single number. Vendor pages quote figures from about 20% to 30%. The only one with a stated sample is Billo's first-half 2026 average of 25.44% across 88,329 Meta sales video ads. That is a cross-category average, not a target. Compare against your own past ads, and read hook rate together with hold rate and cost per result.

How many hooks should I make?

Start with three on one ad body. That is enough to learn something and small enough to review properly. If you film a fixed body once, extra openings are cheap, so six is fine. Beyond that you spread the budget thin and each version gets too little data to compare. The hook test plan explains how to set equal budgets.

Can AI write my hooks?

Yes, an AI writing tool can draft options from these templates. You still need to check every claim, price and number against your own records, and cut anything you cannot prove. AI drafts also tend to sound alike, so replace generic words with a detail only you would know, like a street name or a real job.

Do I need to disclose AI in an ad?

It depends on where you advertise and what the ad shows. TikTok's help page says realistic AI content must be labelled. New York has an ad disclosure law for synthetic performers (people who are not real) that took effect on 9 June 2026. Meta says it applies its own AI info label in many cases. Check the current rules for each platform and place. This is information, not legal advice.

Why did my best hook rate ad not sell?

A hook can stop people without stopping the right people. A broad or curious hook can win on hook rate and still bring viewers who never wanted the product. Look at hold rate, click rate and cost per sale, not hook rate alone, and check whether the hook promises something the rest of the ad does not deliver.

More in this topic: Higgsfield: step-by-step guides (12 guides) →