meta-ads-launch-plannerlisted
Install: claude install-skill ua5er/user-acquisition-skills
# Meta Ads launch and scaling planner
Output: a launch plan the team can follow without a meeting, and (if access is given) the campaign built in Ads Manager, switched **off** so the owner launches it.
## 1. Inputs
Ask once, in one message:
- Target or historical CPA, and currency.
- Optimization event (Purchase, StartTrial, Lead) and daily events the pixel sees from all traffic.
- Total budget the client accepts for the test period.
- Proven creatives from history (or run `meta-ads-learning-audit` first) and new creatives to test.
- Constraints: special ad category, countries, exclusions (recent purchasers), who approves the launch.
## 2. Budget math (show the numbers)
```
learning_budget_per_day = target_CPA × 50 / 7
day1_budget = learning_budget_per_day / 2 (when the client wants a cautious start)
checkpoint_events = 5 purchases after day 1 (or after ~2× target CPA × 5 spent)
max_exposure = spend until the first real stop decision
```
Example: CPA $38 → $271/day to learn → round to ~10 purchases/day = $380. Cautious start: $190 on day 1, $380 from day 2 if 5+ purchases. Worst case before stopping: ~$380 over 2 days.
Use `scripts/budget.py` (below) to print the table for several CPA scenarios. Always show the CPA sensitivity: purchases/day at CPA ±30%.
## 3. Structure (default, explain trade-offs only if asked)
- **One campaign, CBO, one broad ad set** (country, 18 to 65+, all genders, Advantage+ audience), static and video together. All