The marketplace had a 4.2× blended ROAS and a growing sense that something was wrong. Revenue was up, contribution margin was flat, and no one could explain why scaling spend was not scaling profit. The answer was in the customer mix, they were optimising for volume and profiting on a subset of what they were acquiring.
The CAC payback problem no one was measuring
ROAS and CPA are output metrics, they measure what happened, not whether it was profitable. CAC payback period is the metric that actually governs cash flow: how many months of revenue does it take to recover the cost of acquiring a customer? At 18 months payback, every new customer is a cash drain for a year and a half. On a marketplace with high churn and thin per-transaction margins, the business can be growing fast while the acquisition engine is destroying value. The diagnostic question: at our current contribution margin per customer per month, how many months until we recover the acquisition cost?
Contribution margin as the North Star
Contribution margin per customer = revenue minus the variable costs directly attributable to that customer: fulfilment, payment processing fees, marketplace platform fees, and the marginal customer support load. On this marketplace, contribution margin varied by a factor of 4× between customer segments. B2B buyers who ordered in bulk had high average order value, low return rate, and low support cost. Consumer buyers who ordered small and returned frequently had low contribution margin or were margin-negative. The campaigns that drove the highest volume were disproportionately acquiring the second type. Blended ROAS was good. Unit economics were not.
Restructuring the acquisition channels
Step 1: Tag every customer at acquisition with their buyer type, based on first order characteristics (order value bracket, product category, B2B versus consumer account type). Step 2: Feed these tags back into Meta and Google as offline conversion events with value weighted by contribution margin, not revenue. Step 3: Restructure campaigns by segment, separate campaign structure for B2B buyers on LinkedIn and Google with value-based bidding against contribution margin, separate for high-value consumer buyers using Meta lookalike audiences built from the top decile by LTV. Step 4: Kill the campaigns that drove high volume at low margin. Reported CPA went up. Actual payback period dropped from 14 months to under 3.
The measurement model that replaced ROAS
Stop reporting blended ROAS as the primary acquisition metric. Report CAC payback period instead: (customer acquisition cost) divided by (monthly gross margin per customer). A customer with ₹2,000 CAC and ₹800/month contribution margin has a 2.5-month payback. A customer with ₹1,200 CAC and ₹200/month has a 6-month payback. The ₹2,000 CAC customer is dramatically better despite the higher reported cost. When the team and the leadership both see this metric, the conversation changes from "our CPA is too high" to "how do we find more customers who look like the ₹800/month margin cohort." That is a more useful question, and the answer, better audience targeting, better channel mix, better offer for the right buyer, is available in the data you already have.