The system: three stages, each feeding the next.
The result isn't one lucky campaign. Each stage builds the signal the next one runs on, and both the creative and the read on the data carry through all three.
Warm-up engagement
Engagement campaigns season the pixel before any money is asked of the audience.
Purchase seed conversion
A conversion campaign builds a real buyer signal from the warmed audience.
1% lookalike retargeting
The lookalike layer compounds on everything above. This is where the 36.0x lives.
The numbers, defined.
- Ad spend: $2,433 across the campaign's life.
- Attributed revenue: $87,686, as reported by Meta against this campaign.
- ROAS: $87,686 ÷ $2,433 = 36.0x. Attributed revenue, not profit.
- Cost per purchase: $101 against a $3,600 product.
- Click-through rate: 6.84%. Frequency: 2.01, so the audience wasn't being worn out.
Source: Meta Marketing API, pulled July 23, 2026. Campaign-level, lifetime.
Week by week.
Spend stayed small the whole way. The revenue bars are the story, so both are drawn on the same scale.
| Week | Spend | Revenue | ROAS |
|---|---|---|---|
| Jun 17 | $224 | $16,038 | 71.7x |
| Jun 24 | $342 | $11,179 | 32.7x |
| Jul 01 | $352 | $17,296 | 49.2x |
| Jul 08 | $440 | $24,315 | 55.3x |
| Jul 15 | $861 | $18,858 | 21.9x |
| Weekly rows total | $2,219 | $87,686 |
The five weekly rows total $2,219 in spend; the lifetime figure above is $2,433.
Weekly ROAS ranged from 21.9x to 71.7x. The lowest week (21.9x) was also the highest-spend week ($861), which is the number to watch if this scales.
What this doesn't prove
- One account. Five weeks on one product is a signal, not a proven repeatable system.
- Attributed, not profit. ROAS here is Meta-attributed revenue over ad spend. Margin isn't in it.
- Small spend. $2,433 total. A ratio this high at this spend doesn't tell you what happens at ten times the budget.
- A retargeting layer. This is the bottom of a full-funnel system, which is exactly why it's high. It can't be read apart from the stages above it.
- Not separable from the product. I can't fully split the result from product-market fit or seasonality.
I show the full funnel's economics on the call, not just the best campaign.