Pro Playbook

Combinatorial Ad Engine Playbook

A repeatable system for producing 150–750 unique ad variations per week using a Hook × Meat × CTA combinatorial framework. Volume creates data. Data creates winners.

Output150–750 ads/week
CycleWeekly (repeating)
Components50 Hooks × 3–5 Meats × 1–3 CTAs
ProductionSingle filming session

How This Works

Most advertisers create ads one at a time. One hook, one body, one CTA — a single creative that either works or doesn’t. When it doesn’t, they start over. This is slow, expensive, and gives the algorithm almost nothing to work with.

The Combinatorial Ad Engine flips that model. Instead of building complete ads, you build components — hooks, meats, and CTAs — then combine them mathematically. Fifty hooks multiplied by five meats multiplied by three CTAs gives you 750 unique ad variations from a single production session. That volume feeds the algorithm the variety it needs to find winners fast.

50 Hooks Proven openings across awareness 3–5 Meats Educational body copy & proof 1–3 CTAs Clear action buttons × 150–750 Unique Ad Variations per production cycle
50 Hooks × 3–5 Meats × 1–3 CTAs
= 150 to 750 unique ad variations per week

The Three Components

Every ad is made of exactly three parts. Understanding the job of each part is the foundation of the entire system.

ComponentJobQuantity
Hook Stop the scroll. Earn the first 3 seconds. Match the viewer’s awareness level so they think “this is for me.” 50 per cycle
Meat Educate the viewer. Explain the offer, demonstrate the product, reveal the solution, or reframe the problem. 3–5 per cycle
CTA Tell the viewer exactly what to do next. Remove ambiguity. Make the next step feel easy and obvious. 1–3 per cycle
Why this matters for the algorithm: Platforms like Meta and YouTube reward ad accounts that give them variety. More variations means more data points, faster learning, and lower cost per result. A single “perfect” ad can’t compete with 500 variations finding 5 winners you’d never have predicted.
1

Write 50 Hooks

The hook earns the first 3 seconds. You need 50 of them — sourced from five proven places.

The hook is the single highest-leverage component of any ad. A great meat with a weak hook never gets seen. A great hook with a mediocre meat still gets watched. Your 50 hooks don’t need to be original — they need to be proven. That’s why every hook comes from one of five research sources, not from brainstorming in a vacuum.

The Five Hook Sources

1

Winning Hooks from Previous Ads

Pull your top-performing ads from the last 90 days. Extract the opening line or visual. These have already been validated by your audience — they’re the safest starting point.

2

Hooks from Your Free Content

Review your organic posts, videos, reels, and newsletters. Which openings got the highest engagement? If it stopped the scroll organically, it’ll work in paid.

3

Winning Hooks from Competitor Ads

Use the Meta Ad Library, YouTube ad transparency, and TikTok Creative Center to study what’s working for competitors. Don’t copy — extract the pattern and adapt it to your offer.

4

Winning Hooks from Competitor Content

Browse competitor organic content (LinkedIn posts, YouTube intros, podcast clips). High-performing free content reveals what language resonates with the shared audience.

5

Platform-Specific Ad Libraries

Meta Ad Library, TikTok Creative Center, Google Ads Transparency Center, and third-party tools like Foreplay, AdSpy, or Minea. Search by keyword, competitor, or industry to find proven hooks at scale.

Action

Aim for 10 hooks from each source to reach your 50. If one source is thin (e.g. you don’t have previous ad data), over-index on the others. The goal is 50 hooks total, not 10 from each perfectly.

Spread Hooks Across Awareness Levels

Not all prospects are at the same stage. Eugene Schwartz identified five levels of buyer awareness, and your 50 hooks should cover all of them — or deliberately target just one, depending on your campaign objective.

Unaware
Problem Aware
Solution Aware
Product Aware
Most Aware
LevelWhat They KnowHook Angle
Unaware Don’t know they have a problem. Scrolling mindlessly. Pattern interrupt. Curiosity. “I can’t believe I used to do X.”
Problem Aware Know the pain, don’t know the solution exists. Name the pain directly. “Tired of X?” “Here’s why X keeps happening.”
Solution Aware Know solutions exist, haven’t chosen one. Differentiate. “Most people try X. Here’s what actually works.”
Product Aware Know your product, haven’t bought yet. Proof and urgency. Testimonials. Objection handling. Time-limited offers.
Most Aware Ready to buy, need a final push. Direct offer. Discount. Bonus. “Last chance.” “Here’s the link.”
Why awareness levels matter for hooks

A hook aimed at “Most Aware” prospects (“Get 20% off today”) will fail completely on someone who’s “Unaware” — they don’t even know they have a problem, let alone want your product. Conversely, a curiosity-based hook aimed at the unaware will frustrate someone who already knows your product and just wants the link.

Spreading your 50 hooks across all five levels means the combinatorial system serves the right message to the right person, regardless of where they are in their journey. The algorithm does the matching — you just supply the variety.

2

Write 3–5 Meat Scripts

The meat educates the viewer on the offer, product, solution, or problem. Five proven formats cover every angle.

The meat is the substance of the ad — the 15–60 seconds between the hook and the CTA where the actual persuasion happens. Unlike hooks (where you need volume), meats need depth. You’re writing 3–5 scripts, each taking a different angle on the same core message.

The job of the meat is to do one or more of the following: educate on the offer, demonstrate the product, explain the solution, or reframe the problem. The best meats do all four.

The Five Meat Formats

Format 1

Demonstration

Show the product or service in action. Screen recordings, before/after reveals, live walkthroughs. Proof over promises.

Format 2

Testimonial

A customer tells their story. The transformation, the specific results, the emotional shift. Third-party credibility that you can’t manufacture.

Format 3

Educational

Teach something useful. Share a framework, explain a concept, reveal a counterintuitive insight. Value-first selling — the viewer learns something and wants more.

Format 4

Story

A narrative arc: the problem, the turning point, the result. Stories bypass scepticism because the viewer processes them as experience, not pitch.

Format 5

Faceless

No presenter on camera. Text overlays, stock footage, screen recordings, voiceover. Lower production barrier, often performs surprisingly well for technical or introvert-led brands.

Action

Write one meat script per format. Start with Demonstration and Testimonial (highest conversion historically), then add Educational, Story, and Faceless. Each script should be 30–90 seconds when spoken aloud.

Common mistake: Writing meats that only describe the product. The viewer doesn’t care about features — they care about what changes for them. Every meat should answer: “What does this make possible for me that wasn’t possible before?”
3

Write 1–3 CTAs

Tell and show the viewer exactly what to do next. A good CTA communicates five things in under 10 seconds.

The CTA is the shortest component but the one most ads get wrong. A weak CTA (“check it out”, “link in bio”) squanders all the work the hook and meat did. A strong CTA removes every ounce of ambiguity — the viewer knows exactly what to do, what they’ll get, and why they should do it now.

The Five Elements of a Strong CTA

#ElementWhat It Does
1The actionWhat exactly to do. “Click the link below.” “Tap the button.” “DM me the word FREE.”
2The destinationWhere it takes them. “You’ll land on a short form.” “It opens a free training.”
3The rewardWhat they get. “You’ll get the full playbook.” “I’ll send you the template.”
4The proofWhy it’s worth it. “Over 5,000 people have already downloaded it.” “It’s the same system I used to grow from $0 to $10M.”
5The urgencyWhy now. “This comes down Friday.” “Only 50 spots.” Or implicit urgency: “Every day you wait is another day stuck.”
Action

Write 1–3 CTAs that each include all five elements. Keep each under 15 seconds when spoken. Vary the tone: one direct and urgent, one soft and value-led, one social-proof-heavy.

Why multiple CTAs matter

Different meats pair better with different CTAs. A testimonial meat feels natural leading into a proof-heavy CTA (“Join the 5,000 people who already…”). An educational meat pairs better with a value-led CTA (“Get the full framework free…”). Writing 2–3 CTAs gives the combinatorial system more surface area to find winning pairings.

4

Film & Edit in One Session

Batch-produce all components in a single session. Film every hook, every meat, and every CTA separately — then assemble in post.

This is where the system becomes genuinely efficient. Instead of filming complete ads one at a time, you film components — all 50 hooks in one sitting, all 3–5 meats back-to-back, all 1–3 CTAs in a row. The edit team (or you, with a basic editor) then mixes and matches them into every possible combination.

Production Day Structure

BlockDurationWhat You Film
Block 1: Hooks 60–90 min All 50 hooks. Each is 3–8 seconds. Rapid fire. Don’t overthink — the algorithm will decide which ones win.
Block 2: Meats 60–90 min All 3–5 meat scripts. Each is 30–90 seconds. Multiple takes per script. Change angle, energy, or framing between takes.
Block 3: CTAs 15–30 min All 1–3 CTAs. Each is 5–15 seconds. Film in multiple tones — confident, casual, urgent.
Block 4: B-roll 30 min Product shots, behind-the-scenes, screen recordings, lifestyle footage. Used as visual variety across combinations.
Production tip: Film hooks and CTAs standing, sitting, walking, and in different locations. Visual variety across the same script text gives the algorithm even more combinations to test without writing a single extra word.
Action

Schedule a single 3–4 hour production session. Film all hooks first (get them done while energy is highest), then meats, then CTAs, then B-roll. Label every file by component type and number (e.g. HOOK-037, MEAT-02-DEMO, CTA-01-URGENT).

5

Assemble & Deploy

Combine all components into 150–750 unique ad variations and push them live.

This is the multiplication step. Your editor takes the filmed components and creates every viable combination: Hook 1 + Meat 1 + CTA 1, Hook 1 + Meat 1 + CTA 2, Hook 1 + Meat 2 + CTA 1… and so on across all permutations.

You don’t need to deploy all 750 on day one. Start with a subset — the combinations you’re most confident in — and feed more into the ad account as you gather data.

Assembly Workflow

StepDetail
1. Priority combinationsStart with your top 10 hooks (from previous winner data) paired with each meat and CTA. That’s 10 × 5 × 3 = 150 variations as a launch batch.
2. Render and exportUse batch editing tools (CapCut, Descript, Premiere Pro templating, or custom scripts) to speed up assembly. Name each export clearly: HOOK-012_MEAT-DEMO_CTA-URGENT.mp4
3. Upload and structureOrganise in your ad manager: separate ad sets by awareness level. Cold audiences get Unaware/Problem Aware hooks. Retargeting gets Product Aware/Most Aware hooks.
4. Launch in wavesDeploy 30–50 variations per wave. Give each wave 48–72 hours before assessing. Don’t kill ads too early — the algorithm needs time to optimise.
Naming convention matters: When you have 500+ ad variations, being able to quickly identify which hook, meat, and CTA combination is winning (or losing) is everything. Invest the 20 minutes upfront to name files properly. You’ll thank yourself on review day.
6

Review & Iterate

Every week: identify winners, cut losers, recycle what works into the next production cycle.

The Combinatorial Ad Engine isn’t a one-off campaign — it’s a weekly operating rhythm. Every week you review performance data, extract lessons, and feed them back into the next cycle. Winners get recycled. Losers get cut. The system compounds.

Weekly Review Cadence

DayActivity
Day 1 (Monday)Deploy new wave of combinations from the current batch.
Day 3 (Wednesday)Early read: pause obvious losers (high spend, no results). Don’t touch anything else yet.
Day 5 (Friday)Full review. Identify top 5 hooks, top 2 meats, top CTA. Cut bottom 50% of performers.
Day 6–7 (Weekend)Mine new hooks using the 5 sources. Write new meats if needed. Prep for next week’s production session.

What Gets Recycled

The power of this system is the feedback loop. Here’s what flows back into the next cycle:

Winning Hooks

Top-performing hooks get re-used with new meats and CTAs. A hook that works once will often work again in a new combination.

Winning Meats

A demonstration meat that converts well gets paired with all 50 new hooks next week. Don’t retire what’s working.

Winning Hook Patterns

If “question + surprising stat” hooks consistently win, write 20 more in that pattern next week. Double down on what the data says.

Non-Performers

Any component that doesn’t perform across multiple combinations gets cut permanently. If Hook 037 fails with every meat and CTA, the hook is the problem.

The compounding effect: After 4 weeks, you’ll have a library of proven hooks, meats, and CTAs. Week 5 doesn’t start from zero — it starts from a curated bank of validated components plus fresh ones. Each cycle gets faster and more effective than the last.