Prepaid perfume, no return leakage, a 20×+ return.

Prepaid changes the entire equation. In cash-on-delivery markets a chunk of “sales” evaporate at the door as refused deliveries, so reported ROAS is always inflated. Prepaid strips that out: if the pixel fires, the money is in the bank. That let us optimize aggressively for purchase value without fearing return leakage, and judge every campaign on revenue we knew was real. It also reshapes the creative brief: instead of chasing impulse clicks, you sell the reasons a buyer will commit money before the box arrives, trust, guarantee, and a clear, specific reason the fragrance is worth paying for upfront. The filter is brutal, a shopper who won’t pay upfront never converts, but the buyers who clear it are worth far more.

• Kept the account lean and offer-led, a tight set of campaigns rather than sprawl, so budget concentrated on what actually worked.
• Tested the offer structure hard: bundles versus singles, a longer-guarantee “one-year” angle, discount framing, the offer, not just the creative, was the lever.
• Mixed ABO for controlled testing with CBO for scaling the winners, plus a “secure-decision” retargeting angle to close hesitant prepaid buyers.
• Fed the algorithm clean prepaid purchase signal, so it optimized toward people who actually pay upfront.

Blended return cleared 20×+, with individual winning campaigns climbing far higher, roughly 570 clean, prepaid purchases from a base of well over a million people reached. Because it’s prepaid, that return is unusually honest: not gross-before-returns, but collected revenue. (On a number this high we pressure-test the tracking before we celebrate, but even discounted, it’s a category outlier.)

Prepaid is a feature, not a handicap. When every conversion is real money, you can optimize without flinching, and the ceiling on return is far higher than any COD account will ever show you.