Key Takeaways

  • Ad platform dashboards mislead founders by double-counting sales. Meta reported customer acquisition cost (CAC) is typically 20% lower than true acquisition cost.
  • Spreading marketing spend across multiple platforms too early destroys capital efficiency. Founders should focus on one channel until its performance is fully proven.
  • Growth teams should operate with weekly budget unlocks tied to a strict CAC ceiling, scaling spend only while unit economics hold.
  • Geographic A/B tests run every six months reveal actual channel contribution by cutting ad spend in half of two identical growth markets.
  • Eight Sleep operates this playbook through the Eight Sleep's Single-Channel Scaling and Incrementality Testing Framework.

The Eight Sleep's Single-Channel Scaling and Incrementality Testing Framework

Step 1: Single-Channel Focus

Start with one primary platform (such as Meta) rather than allocating budget simultaneously across multiple channels. Do not move to a new channel until the first is proven.

Step 2: Capped CAC Scaling

Unlock a small weekly testing budget. Set a strict CAC ceiling. Allow the growth team to scale ad spend up to maximum volume as long as the blended CAC remains below that ceiling.

Step 3: Bi-Annual Incrementality A/B Testing

Every six months, select two geographically comparable markets with similar baseline year-over-year revenue growth. Turn off ad spend entirely in one market while leaving it active in the control market.

Step 4: True CAC Calibration

Calculate the delta in revenue between the test and control markets to determine the true incremental revenue generated by the channel, correcting for platform over-attribution.

When This Works (and When It Doesn't)

This framework works once a consumer startup has clear product-market fit and enough sales volume to run statistically valid geo-split experiments. If you generate $20,000 a month in revenue, cutting ad spend in Ohio will not give you clean data. You need stable baseline sales across regions before turning off ad dollars will register as signal instead of random noise.

It breaks down when your sales cycles stretch beyond 90 days or when your brand has heavy offline distribution. When retail stores or wholesale partners drive the bulk of checkout volume, turning off Meta ads in one state can cause delayed retail sales drops that take quarters to track down. For pure direct-to-consumer businesses with high purchase velocity, however, direct geographic blackouts are the only reliable check against platform attribution models.

Franceschetti points out that ad platforms take credit for customers who would have bought anyway. “If you look at the CAC on Meta, it is probably 20% lower than what the true CAC is,” Franceschetti notes. Founders who rely purely on ad manager dashboards end up overspending, which artificially inflates top-line growth while destroying margins. When they return to the venture market to raise capital, their blended CAC explodes because they spent months buying phantom demand.

What to Do With This

Pick two states with near-identical revenue and growth rates over the last six months, such as Colorado and Washington. Starting next Monday, zero out paid social ad spend in Colorado for four weeks while maintaining your normal budget in Washington. Compare the revenue gap between both states at the end of the test period against your paid social spend to find your real CAC, then update your financial model before allocating next quarter's budget.