The campaign that broke even on day 19 — and why that was fine

The campaign that broke even on day 19 — and why that was fine

The setup. Subscription VPN offer, $40 payout, monthly rebill share. Search-arbitrage traffic. Illustrative figures.

The move. Front-end CPA didn't cover acquisition cost. Day 1 I was paying $52 to acquire a $40 first payout. On a CPA mindset, dead on arrival.

The numbers. Cohort of 1,000 subs cost $52,000. First payout $40,000 — underwater $12,000. But rebill at 62% month-2, 48% month-3, decaying. Revenue share kicked $19/rebill. By the day-19 mark of month 2 the cohort crossed break-even; by month-4 it had returned $71,000 on $52,000.

The lesson. Rebill offers invert the math. You're not buying a conversion, you're buying a decaying annuity. Judging them on first-payout ROI guarantees you kill the profitable ones, because the profit lives in months 2-4 that your day-1 dashboard can't see.

What I'd do differently. Model the retention curve from the network's historical data BEFORE buying, and set a max payback-period tolerance (mine's 35 days). If a network won't share cohort rebill data, I assume the curve is bad and price accordingly.