Creative experiments

The hook bank: 100 ad hooks by type

A hook is the first 1 to 3 seconds of a video ad, or the first line of a text or image ad. Measure it with hook rate: 3-second video plays divided by impressions. Test hooks cheaply by keeping the body of the ad the same and changing only the opening. This bank has 100 fill-in-the-blank templates in 12 types, with an example for a fictional brand, the funnel stage each fits and whether the claim needs proof. The workbook writes all 100 for your product.

Most people decide whether to watch an ad before they have consciously decided anything. The thumb keeps moving, or it stops. The part of the ad that makes it stop is the hook, and it is the cheapest part of an ad to improve: you can keep everything else and change only the opening.

This is a bank of 100 hooks you can use today, grouped into 12 types. Each one is a fill-in-the-blank template with an example for a fictional brand. Below them you'll find how to measure a hook, how to test hooks without remaking whole ads, and which types fit which audiences.

The workbook has all 100 hooks with filters, a sheet that writes all 100 for your product from 14 short inputs, and a hook test log that works out hook rate and compares each hook with your control.

What a hook is

A hook is the opening of an ad:

  • in a video ad, the first 1 to 3 seconds: the first words, the first image, the first movement and any text on screen;
  • in an image or text ad, the first line of copy and the headline on the image.

Its only job is to earn the next few seconds. It does not have to sell. A good hook makes the right person think "that's me" or "wait, what?" and keep watching. The body of the ad does the selling: the product, the proof, the offer and the call to action.

That split is what makes hooks so useful to test. The body takes time and money to make. A new opening takes minutes.

GaaS has made more than 60,000 ads. The 12 types below are the patterns we see again and again across them. They are a vocabulary for thinking and testing, not a ranking: no type wins for every product, and the only way to know which works for you is to test.

Grid of 12 cards, one per hook type, each with what the hook does, a sample line for a fictional brand, the funnel stage it fits and how many templates are in the bank. Question (9 hooks, cold): asks what they answer in their head. Bold claim (8, cold and warm): a strong promise you can prove. Myth-buster (8, cold): names a belief, then corrects it. Price anchor (8, warm and hot): your price next to a bigger familiar cost. Social proof (8, cold and warm): leads with other customers. Problem-agitation (9, cold): names the pain and makes it feel urgent. Curiosity gap (9, cold): opens a question only the ad can close. POV (8, cold and warm): puts the viewer inside a moment. Founder story (8, cold): the builder says why it exists. Comparison (8, warm): you versus the usual alternative. Urgency (8, hot): a real reason to act now. How-to (9, cold and warm): promises useful steps.

Twelve types, 100 templates. Use the types to plan a test; use the templates to write the lines.

How to measure a hook

Hook rate (also called thumb-stop rate)

Hook rate is the share of people who saw your video ad and watched at least the first 3 seconds:

Hook rate = 3-second video plays ÷ impressions

Meta reports 3-second video plays: the number of times your video played for at least 3 seconds, or for nearly its full length if it is shorter than 3 seconds. Replays don't count (Meta Business Help Center: 3-second video plays). Meta does not show hook rate as a column. You add it yourself, as a custom metric in Ads Manager or in a spreadsheet.

Worked example (made-up numbers): an ad gets 10,000 impressions and 2,500 three-second plays. Hook rate = 2,500 ÷ 10,000 = 25%. One in four people who saw it stopped for at least 3 seconds.

On TikTok, the closest columns are 2-second and 6-second video views, which TikTok defines in its help on video play metrics. Pick one and use it consistently.

For image and text ads there is no 3-second play. Use click-through rate (clicks ÷ impressions) to compare first lines, and keep the image the same while you test the words.

Hook rate is the stop, not the sale

Left: worked example with 10,000 impressions and 2,500 three-second video plays, so hook rate equals 2,500 divided by 10,000, which is 25%. Right: Ad A has a 32% hook rate, 0.6% click-through rate and $48 cost per purchase. Ad B has a 22% hook rate, 1.1% click-through rate and $31 cost per purchase. Ad A stops more people; Ad B sells more.

Illustrative numbers. Use hook rate to compare openings, then pick winners on cost per result.

A shocking or funny opening can stop lots of people who will never buy. So read hook rate together with what happens next:

MetricWhat it tells youFormula
Hook rateDid the opening stop people?3-second video plays ÷ impressions
Hold rate (optional)Did the body keep them?ThruPlays ÷ 3-second video plays
Click-through rate (CTR)Did they want to know more?Link clicks ÷ impressions
Cost per resultDid it pay?Spend ÷ purchases (or leads)

Meta defines a ThruPlay as a play to the end of a video, or for at least 15 seconds (Meta Business Help Center: ThruPlays). Hold rate, like hook rate, is a formula you build, not a Meta column.

We deliberately don't quote a "good" hook rate here. It varies with format, placement, audience and category, and published benchmarks rarely say how they were measured. Your best benchmark is your own control ad.

How to test hooks cheaply

The trick is to change only the opening. Keep the body, offer, audience, budget and dates the same. Then any difference in hook rate comes from the hook.

Four versions of one 30-second ad. Each has a different first 3 seconds (question, social proof, POV and myth-buster) followed by the same body: product, proof, offer and call to action. Keep identical the body, offer, audience, budget per version and dates. Example budget: 4 versions at $50 each, $200 for a first read.

Record the body once. Record several openings. Stitch each opening onto the same body.

  1. Pick 3 to 5 hooks from different types. Different types teach you more than five versions of the same question.
  2. Record the openings. Same person, same setting, same lighting, so the only change is the hook.
  3. Stitch each opening onto the same body. Any video editor works. If you're comfortable with a terminal, FFmpeg can join two clips that share the same size and frame rate: see the command below.
  4. Launch them side by side in the same campaign and ad set, with the same budget per version where your setup allows it.
  5. Wait for enough impressions. The workbook's test log has a minimum (2,000 impressions by default) before it gives a verdict. That's a rule of thumb, not a statistical test.
  6. Keep the winner as your new control, then test the next batch of hooks against it.
# Join a 3-second opening (hook1.mp4) to the shared body (body.mp4).
# Both clips need the same resolution and frame rate.
ffmpeg -i hook1.mp4 -i body.mp4 \
  -filter_complex "[0:v][0:a][1:v][1:a]concat=n=2:v=1:a=1[v][a]" \
  -map "[v]" -map "[a]" version1.mp4

For more on designing a clean hook test, read test ad hooks without changing the whole story.

Which hook for which audience

Cold audiences have never heard of you; they need a reason to stop. Warm audiences know you but haven't bought; they need a reason to believe. Hot audiences are ready; they need a reason to act now.

Matrix of the 12 hook types against cold, warm and hot audiences. Question, myth-buster, problem-agitation, curiosity gap and founder story fit cold audiences. Bold claim, social proof, POV and how-to fit cold and warm. Price anchor fits warm and hot. Comparison fits warm. Urgency fits hot.

An editorial rule of thumb, not a law. Your own tests decide.

Hook typeUse it whenWatch out
QuestionThe problem is common and easy to recognise.Yes/no questions with an obvious 'no' end the scroll.
Bold claimYou have proof: a guarantee, a test, a number you can show.Needs substantiation. Unproven claims break ad policies and trust.
Myth-busterYour category has a popular belief your product disagrees with.The myth must be real and the correction must be true.
Price anchorYou are cheaper than the usual alternative, or worth more than it looks.Compare like with like; don't invent the 'usual' price.
Social proofYou have real reviews, customer counts or customer footage.Use real reviews and real numbers only, with permission.
Problem-agitationThe pain is clear and frequent.Agitate, don't shame. Keep it about the situation, not the person.
Curiosity gapYou have a genuinely surprising answer.The payoff must be real, or the click becomes a bounce.
POVShort-form video on Reels or TikTok; a vivid everyday moment.Keep it specific; a generic POV feels like an ad.
Founder storyThere is a real, personal reason the business exists.Must be true and in the founder's own words.
ComparisonThere is a clear, visible difference.Compare against a category ('the gym'), not a named competitor, unless legal says yes.
UrgencyA real deadline, real stock limit or a seasonal moment.Fake deadlines break trust and can break consumer law.
How-toYou can teach something useful in the ad itself.Deliver the steps; don't hide them behind a click.

The 100 hooks

Square brackets are the blanks. Fill them with your own words:

BlankWhat to put in itExample
[product]Product or brand name. What you sell, as you'd say it out loud.Northpaw
[audience]Who it's for (plural). A group noun: 'dog owners', 'new freelancers'.dog owners
[problem]Main problem (a noun phrase). What they want gone: 'itchy skin', 'late invoices'.itchy skin
[outcome]Result they want (a noun phrase). What they want instead: 'a calmer, shinier coat'.a calmer, shinier coat
[old way]The usual alternative (an -ing phrase). What they do now: 'trying a new food every month'.trying a new food every month
[time]A realistic timeframe. '30 days', 'two weeks', 'one weekend'.30 days
[price]Your price, as you'd say it. '$2 a day', '$29 a month', '$89'.$2 a day
[old price]What the usual alternative costs. '$60 a month', '$80 a session'. Must be true.$60 a month
[number]A true number (customers, reviews). Only a number you can prove: '4,000'.4,000
[myth]A common belief that's wrong (a clause). 'grain-free is always better'.grain-free is always better
[founder]Founder name and role. 'Maya, a vet nurse'.Maya, a vet nurse
[offer]Your offer (a noun phrase). '20% off your first box'.20% off your first box
[deadline]When the offer ends. 'Sunday', 'midnight', '1 January'. Must be real.Sunday
[review]A short real customer quote. Paste a real review, with permission.He stopped scratching within a month

The examples use four fictional brands: Northpaw (dog food sold online), FixRight Plumbing (a local plumber), Tallyo (an invoicing app for freelancers) and Liftwell (online strength coaching for people over 40). Their numbers are made up. Hooks marked "proof" make a claim: only use them if you can show the evidence.

Question hooks (9)

Use when: the problem is common and easy to recognise.

#TemplateExample (fictional brand)StageProof
1Still [old way] to deal with [problem]?Still trying a new food every month to deal with your dog's itchy skin?Cold
2Why does [problem] keep coming back?Why does your kitchen drain keep clogging every few weeks?Cold
3Do you know what [problem] is really costing you?Do you know what late invoices are really costing you?Cold
4Is [problem] normal, or a sign something needs to change?Are stiff mornings after 40 normal, or a sign something needs to change?Cold
5What if you could get [outcome] without [old way]?What if you could get stronger without paying for a gym you never use?Cold
6How much did [problem] cost you last year?How much did emergency call-outs cost you last year?Cold
7Would you pay [price] to stop worrying about [problem]?Would you pay $12 a month to stop worrying about unpaid invoices?Warm
8Still thinking about [product]? Here's what most people ask first.Still thinking about Northpaw? Here's what most people ask first.Warm
9Did you know [audience] can get [outcome] in [time]?Did you know beginners over 40 can learn a safe deadlift in two weeks?ColdYes

Bold claim hooks (8)

Use when: you have proof: a guarantee, a test, a number you can show.

#TemplateExample (fictional brand)StageProof
10This is the easiest way to get [outcome].This is the easiest way to get a shinier coat: change one bowl.ColdYes
11[outcome] in [time]. No more [old way].Paid on time in 30 days. No more chasing clients by email.ColdYes
12We fix [problem] the first time, or you don't pay.We fix your leak the first time, or you don't pay.HotYes
13[product] does in minutes what [old way] takes hours to do.Tallyo does in minutes what a spreadsheet takes an afternoon to do.ColdYes
14[product] is made only for [audience].Liftwell is made only for people starting strength training after 40.ColdYes
15You can get [outcome] without [old way].You can get invoices paid on time without hiring a bookkeeper.ColdYes
16Stop putting up with [problem]. Start enjoying [outcome].Stop putting up with half-eaten bowls. Start seeing empty ones.ColdYes
17Most cases of [problem] can be fixed. Here's how we do it.Most slow drains can be fixed in one visit. Here's how we do it.WarmYes

Myth-buster hooks (8)

Use when: your category has a popular belief your product disagrees with.

#TemplateExample (fictional brand)StageProof
18Think [myth]? Not true. Here's why.Think grain-free is always better? Not true. Here's why.ColdYes
19Stop believing that [myth].Stop believing that you're too old to lift heavy.ColdYes
20The biggest myth about [problem]: that [myth].The biggest myth about late payments: that chasing harder works.ColdYes
21Everyone says [myth]. Here's what they're missing.Everyone says chemical drain cleaner is fine. Here's what they're missing.ColdYes
22Believing [myth] could be why you still have [problem].Believing cardio is enough could be why you still feel weak on the stairs.ColdYes
23You've been told [myth]. Here's what actually helps with [problem].You've been told invoicing software is for big companies. Here's what actually helps freelancers get paid.ColdYes
24For years I believed [myth]. Then I tried [product].For years I believed all kibble was basically the same. Then my dog tried Northpaw.WarmYes
25No, [problem] isn't something you just have to live with.No, a gurgling drain isn't something you just have to live with.ColdYes

Price anchor hooks (8)

Use when: you are cheaper than the usual alternative, or worth more than it looks.

#TemplateExample (fictional brand)StageProof
26[old way] costs [old price]. [product] costs [price].A personal trainer costs $80 a session. Liftwell costs $29 a month.WarmYes
27[price]. Less than your morning coffee.$2 a day. Less than your morning coffee.Warm
28Why pay [old price] for [old way] when [product] is [price]?Why pay a bookkeeper $150 a month to send invoices when Tallyo is $12?WarmYes
29What [price] gets you: [outcome].What $89 gets you: a camera inspection and a straight answer.Hot
30Fixing [problem] now: [price]. Ignoring it: [old price] later.Fixing a slow leak now: $150. Ignoring it: a ceiling repair later.WarmYes
31[price] for [outcome]? Here's what's included.$29 a month for a coach in your pocket? Here's what's included.Warm
32Try [product] for [price]. Keep it if it works for you.Try Tallyo for $1 for your first month. Keep it if it works for you.Hot
33Under [price] for [outcome]? Here's the catch.Under $2 a day for a dog who finishes his bowl? Here's the catch.Warm

Social proof hooks (8)

Use when: you have real reviews, customer counts or customer footage.

#TemplateExample (fictional brand)StageProof
34[number] [audience] switched to [product]. Here's why.4,000 dog owners switched to Northpaw. Here's why.ColdYes
35"[review]." A real review from one of our customers."Showed up in 40 minutes and fixed it first time." A real review from one of our customers.WarmYes
36Why [audience] keep recommending [product] to friends.Why people over 40 keep recommending Liftwell to friends.ColdYes
37We asked [number] customers what changed after [product].We asked 500 freelancers what changed after Tallyo.WarmYes
38[number] five-star reviews from [audience]. Here's what they say.1,200 five-star reviews from local homeowners. Here's what they say.WarmYes
39Watch [audience] try [product] for the first time.Watch three picky dogs try Northpaw for the first time.ColdYes
40[time] ago, this customer had [problem]. Now: [outcome].Six months ago, Dave couldn't carry the shopping upstairs. Now he deadlifts twice a week.WarmYes
41Don't take our word for it. Here's what [audience] say about [product].Don't take our word for it. Here's what freelancers say about Tallyo.WarmYes

Problem-agitation hooks (9)

Use when: the pain is clear and frequent.

#TemplateExample (fictional brand)StageProof
42If you deal with [problem], this is for you.If your drain gurgles every time you shower, this is for you.Cold
43[problem] isn't just annoying. Here's what it's really costing you.Late invoices aren't just annoying. Here's what they're really costing you.Cold
44[old way] hasn't fixed [problem]? You're not alone.Three new foods haven't fixed the scratching? You're not alone.Cold
45Tired of [problem]?Tired of feeling stiff every morning?Cold
46Every week you put off fixing [problem], it gets harder.Every week you put off fixing a slow leak, it gets harder (and pricier).WarmYes
47That moment when [problem] ruins your day again.That moment when a client says the invoice "must have gone to spam" again.Cold
48Nobody warns [audience] about [problem].Nobody warns new freelancers that chasing payments becomes a second job.Cold
493 signs [problem] is about to get worse.3 signs your water heater is about to get expensive.ColdYes
50You deserve better than [problem].Your dog deserves better than a bowl he walks away from.Cold

Curiosity gap hooks (9)

Use when: you have a genuinely surprising answer.

#TemplateExample (fictional brand)StageProof
51The one thing [audience] get wrong about [problem].The one thing people over 40 get wrong about warming up.Cold
52I didn't expect this to fix [problem].I didn't expect a food change to fix the 3am scratching.ColdYes
53Why [outcome] has nothing to do with [old way].Why getting paid on time has nothing to do with chasing harder.Cold
54The simple fix for [problem] most people skip.The $5 fix for a running toilet most people skip.ColdYes
55Watch what happens at the end.Watch what Bella does when she hears the bag.Cold
56Don't keep [old way] until you've seen this.Don't renew that gym membership until you've seen this.Cold
57What the pros check first when they see [problem].What plumbers check first when they see a damp ceiling.Cold
58There's a reason [problem] keeps happening, and it's not your fault.There's a reason invoices get paid late, and it's not your fault.Cold
59Nobody's talking about this way to get [outcome].Nobody's talking about this way to get stronger at home.Cold

POV hooks (8)

Use when: short-form video on Reels or TikTok; a vivid everyday moment.

#TemplateExample (fictional brand)StageProof
60POV: you finally have [outcome].POV: every invoice this month got paid on time.Cold
61POV: it's Sunday night and [problem] is back.POV: it's Sunday night and you're chasing invoices again.Cold
62POV: you stopped [old way] and switched to [product].POV: you stopped buying supermarket kibble and switched to Northpaw.Warm
63POV: you're one of the [audience] who's done with [problem].POV: you're 47 and done with feeling weak when you lift the kayak.Cold
64POV: your first [product] order just arrived.POV: your first Northpaw box just arrived and the dog heard it.Hot
65POV: you tried [product] for [time].POV: you tried Liftwell for 30 days and the stairs got easier.WarmYes
66POV: someone asks what you changed, and the answer is [product].POV: the vet asks what you changed, and the answer is Northpaw.WarmYes
67POV: [problem] used to be your normal.POV: stiff, sore mornings used to be your normal.Cold

Founder story hooks (8)

Use when: there is a real, personal reason the business exists.

#TemplateExample (fictional brand)StageProof
68I started [product] because I had [problem] too.I started Northpaw because my own dog wouldn't stop scratching.Cold
69I'm [founder]. I dealt with [problem] for years before building [product].I'm Priya. I chased invoices for six years before building Tallyo.Cold
70Why I gave up on [old way] and started [product].Why I gave up on big-chain call-out jobs and started FixRight.Cold
71We almost didn't launch [product]. Here's what changed our minds.We almost didn't launch Liftwell. Here's what changed our minds.Cold
72The day [problem] got bad enough that I built [product].The vet bill that made me build Northpaw.Cold
73I'm [founder]. Here's the promise I make to every customer.I'm Tom, FixRight's owner. Here's the promise I make to every customer.Warm
74I built [product] for [audience] like me.I built Liftwell for people like me who started lifting at 45.Cold
75What I learned from [number] [audience].What I learned from talking to 300 freelancers about getting paid.ColdYes

Comparison hooks (8)

Use when: there is a clear, visible difference.

#TemplateExample (fictional brand)StageProof
76[product] vs [old way]: which one gets you [outcome]?Liftwell vs the gym: which one gets you stronger after 40?WarmYes
77Left: [old way]. Right: [product].Left: supermarket kibble. Right: Northpaw. Same dog, two bowls.ColdYes
78Same [problem], two ways to fix it.Same blocked drain, two ways to fix it: chemicals or a camera.ColdYes
79We put [product] up against [old way] for [time].We put Tallyo up against a spreadsheet for 30 days.WarmYes
80Why [audience] are switching from [old way] to [product].Why dog owners are switching from supermarket kibble to Northpaw.WarmYes
81[old way] or [product]? The honest answer.Personal trainer or an online coach? The honest answer.WarmYes
82Before [product]: [problem]. After: [outcome].Before Tallyo: 14 unpaid invoices. After: 2.WarmYes
83[old price] vs [price]: what you actually get.$80 a session vs $29 a month: what you actually get.WarmYes

Urgency hooks (8)

Use when: a real deadline, real stock limit or a seasonal moment.

#TemplateExample (fictional brand)StageProof
84[offer] ends [deadline].20% off your first box ends Sunday.HotYes
85Last chance: [offer].Last chance: a free inspection with any repair this month.HotYes
86Last few spots before [deadline].Last few spots in the January group. Doors close Friday.HotYes
87Fix [problem] before [deadline].Get your pipes checked before the first frost.Warm
88Been waiting for a sign? [offer] until [deadline].Been waiting for a sign? Your first month is $1 until Friday.HotYes
89Prices go up after [deadline]. Lock in [price] now.Prices go up after 1 January. Lock in $29 a month now.HotYes
90Still in your cart? [offer] until [deadline].Still in your cart? 20% off your first box until midnight.HotYes
91[problem] won't wait. Neither will [offer]: it ends [deadline].Winter won't wait. Neither will our free boiler check: it ends Friday.HotYes

How-to hooks (9)

Use when: you can teach something useful in the ad itself.

#TemplateExample (fictional brand)StageProof
92How to get [outcome] in [time].How to do your first full push-up in 30 days.ColdYes
933 ways to get [outcome] without [old way].3 ways to get paid faster without chasing clients.Cold
94How to tell if [problem] is getting worse.How to tell if a small leak is getting worse.Cold
95Read this before you spend another cent on [old way].Read this before you spend another cent on a food your dog won't eat.Cold
96Give us [time]: the plan for [outcome].Give us 30 days: the plan for stronger legs and easier stairs.WarmYes
97How [audience] get [outcome] with [product].How freelancers send an invoice in 60 seconds with Tallyo.WarmYes
98Fix [problem] in 3 steps.Unblock a sink in 3 steps, no chemicals.Cold
99The right way to get [outcome] (most people skip step 2).The right way to switch your dog's food (most people skip step 2).Cold
100Watch how [product] works in under 60 seconds.Watch how Tallyo works in under 60 seconds.Warm

Turn a template into your hook

  1. Fill in the blanks in plain words. Write the problem the way a customer would say it, not the way your website does. "My dog won't stop scratching", not "dermatological discomfort".
  2. Say it out loud. If it sounds like an ad, rewrite it until it sounds like a person.
  3. Check the claim. Any number, review, guarantee, deadline or comparison has to be true and provable. Ad platforms reject unsupported claims, and customers notice.
  4. Match the visual to the words. The first frame should show the problem, the person or the surprise the words promise.
  5. Keep it short. Most of these hooks take 2 to 3 seconds to say. If yours takes six, cut it.

If you'd rather start from your own website, the free ad hook generator writes cold-open hooks for TikTok and Reels from your URL.

Tools and repos that help

ToolWhat it's forWho should use it
GaaS ad hook generatorFree: hooks for TikTok and Reels from your website URLAnyone who wants a first draft in a minute
FFmpeg/FFmpegThe open-source video tool behind most editors. Joins each opening to the same body with one commandAnyone testing several hooks on one video
openai/whisperOpen-source speech-to-text. Transcribe your existing ads to log exactly what each one says in its first secondsTeams with a back catalogue of ads to label by hook type
SYSTRAN/faster-whisperA faster re-implementation of WhisperThe same job, on a laptop, in less time
pipeboard-co/meta-ads-mcpAn MCP server that lets an AI assistant read Meta ad insights. Open source under the Business Source License, with a hosted versionMarketers who want to ask an assistant for hook rate by ad
meta/facebook-python-business-sdkMeta's official Python SDK. In the Marketing API, 3-second plays come back in the actions field as the video_view action typeDevelopers building their own hook-rate report
anthropics/skillsExample Agent Skills for ClaudeTurning "write 10 hooks from this template bank" into a reusable skill

Transcribing your own ads with Whisper takes one line once it's installed:

whisper my-ad.mp4 --model small --output_format txt

Mistakes that waste hook tests

  • Changing the body too. If the opening and the body both change, you can't tell which one moved the numbers.
  • Testing five questions. Test different types first. Fine-tune wording later.
  • Judging after 300 impressions. Small numbers swing. Wait for your minimum.
  • Crowning the highest hook rate. Check CTR and cost per result before you call a winner.
  • Clickbait. A curiosity gap must pay off in the ad. If the body doesn't deliver, people leave, and the platform notices.
  • Unprovable claims. Bold claims, social proof, comparisons and urgency need to be true. Fake deadlines and invented reviews break trust and can break consumer law.

Or let the agent test hooks for you

GaaS writes, makes and tests ads for you from your website, a spend limit and a one-line goal, and asks for your approval before anything spends. If you'd rather do it by hand, everything in this post and the workbook works without it.

Sources

Test ad hooks without changing the whole story

Design an ad hook test that isolates the opening idea, preserves the offer and narrative, and measures downstream response alongside early attention.

Write an ad creative hypothesis that can be tested

Turn a creative idea into a testable advertising hypothesis with a customer concern, message mechanism, expected behavior, control, and falsifying evidence.

The Ad Library heist: reverse-engineer your competitors' winning ads

Use the free Meta Ad Library, Google Ads Transparency Center and TikTok Creative Center to find the ads your competitors keep running and keep re-making, build a pattern table, and turn their best angles into original ads of your own. Includes AI prompts and a tracker workbook.

The brief that writes itself: an AI creative brief template

A one-page ad creative brief that AI can fill from your website or from a proven ad in the Meta Ad Library, scored on Hook, Brand fit and Survival before you pay for production. Copy-paste prompts, three worked fictional briefs and a scoring workbook.

Have a correction or a question about the workflow? Contact GaaS. Read our editorial standards for sourcing and example conventions.