Open the CSV in Excel, or import it into Google Sheets and choose comma-separated values. Save your working copy in a controlled location. The CSV contains the rows below and space for your notes; it has no macros, formulas or live account connection. The separate Excel workbook contains editable example inputs and formulas that update the forecast.
How to use the template
Draft with AI from a source
Run the website prompt on your product, reviews and FAQ pages, or the winning-ad prompt on an ad that has run for 30 days or more in the Meta Ad Library. Both prompts tell the model to write MISSING instead of inventing facts and to model the reference ad's structure, never its words or visuals.
Fix the facts
Check every claim against its source page. Make sure the offer is live, the proof can be shown and any customer quote is real and used with permission. Fill or cut every MISSING field.
Score before production
Score the best of the three hooks, Brand fit and Survival from 1 to 5. The weighted score is (Hook x 0.4 + Brand fit x 0.3 + Survival x 0.3) x 20. Produce at 75 or more, revise at 60 to 74 and rework below 60, or whenever any score is 1 or Brand fit is 2 or lower.
Record what happened
Add each brief to the library with its status. After the ad reaches the spend in 'success looks like', write the result and what the next brief should change.
Preview the worksheet
Every row shown here is included in the download. Scroll the table horizontally on a small screen.
| Field | What to write | Example (fictional Corvel Mobile Detailing) |
|---|---|---|
| Audience | One specific person: who they are, where, what they've already tried | A busy suburban parent with a family SUV who has tried drive-through washes |
| Problem (their words) | The problem in the customer's own words, ideally quoted | "I don't have three hours on a Saturday to clean crushed snacks out of the back seat." |
| Angle | The one idea this ad argues: why this, why now | We come to your driveway; you get your Saturday back |
| Hook 1 (question) | Opening for the first 1-3 seconds | "When did you last see the floor of your back seat?" |
| Hook 2 (bold, provable claim) | Opening for the first 1-3 seconds | "We clean your car in your driveway while you're at work." |
| Hook 3 (visual demonstration) | Opening for the first 1-3 seconds | A hand lifts a car seat to reveal crumbs; hard cut to the same spot spotless |
| Proof | What the viewer sees that makes the claim believable | Real before-and-after of one customer's car (with permission), timestamped |
| Offer | What they get and why act now; must be live | First interior detail $129 (normally $159), booked this month |
| Format | Placement, aspect ratio, length, who's on camera | 9:16 video, 20 seconds, phone-shot, owner on camera for the last 5 seconds |
| Script or shot list | Beat by beat: visual and audio | 1) Crumbs reveal 2) Van arrives 3) Three cleaning shots 4) Spotless seat 5) Owner: in your area Tue/Thu 6) Offer card 7) Book a slot |
| CTA | Exact button and action | Book now, to the booking page with the offer applied |
| Success looks like | One metric, a target, and spend or time before judging | Cost per booking under $40 after $300 spend; at least 8 bookings |
| Modelled on (link) | Reference ad link, run length, what you borrow and change | Illustrative: a local detailer's before-and-after reveal reel, active 70+ days. Borrow the reveal; change the setting and lead with time saved |
| Hook score (best of 3) | 1-5: 5 stops a stranger in under 2 seconds; 1 opens on the brand or logo | 5 |
| Brand fit score | 1-5: 5 only you could run it and every claim is provable; 1 off-brand or unprovable claim | 4 |
| Survival score | 1-5: 5 reference ran 60+ days and variants are easy; 1 no reference or one-off stunt | 4 |
| Weighted score and verdict | (Hook x 0.4 + Brand fit x 0.3 + Survival x 0.3) x 20. Produce 75+, revise 60-74, rework under 60 or on a hard stop | 88: Produce |
Worked review example
Illustrative brief for a fictional mobile car detailer: the best hook is a visual reveal of crumbs under a car seat (Hook 5), the claim of cleaning in the customer's driveway is provable (Brand fit 4), and the reference reveal-format ad had run for more than 70 days (Survival 4).
The weighted score is (5 x 0.4 + 4 x 0.3 + 4 x 0.3) x 20 = 88, so the brief goes to production. A fictional software brief whose hooks all open on the product name scores Hook 2 and 58 overall, so it is reworked from the customer's problem line.
What this template does and does not establish
Scores are a structured judgement, not a prediction of results. A reference ad's run length is a signal that a format may be working for that advertiser, not proof of profitability. The Meta Ad Library API returns commercial ads only for ads that reached the EU; outside that scope use the Ad Library website. Model a reference ad's structure only; never copy its words, music, people or visuals.
Primary references
- Meta Ad Library
- Meta: Ad Library API (ads_archive) reference
- Meta: Ads MCP server tools, including Ad Library search
- marketingskills: ad-creative and research skills for AI agents
- Anthropic Agent Skills, including skill-creator
