DEV Community

Swipe Again for Swipe Again

Posted on

When Paddle Is the Merchant of Record, Dunning Is a Handoff Problem

Selling through Paddle changes the shape of failed-payment recovery in a way that surprises teams arriving from Stripe. Paddle is the merchant of record: it is Paddle's name on the customer's statement, Paddle's tax handling, and Paddle's first line of retries when a renewal declines. That arrangement removes a great deal of operational burden, and it also means the seller never touches the raw payment the way a direct Stripe merchant does. Dunning, in this world, is less about controlling attempts and more about managing a handoff cleanly.

The first consequence is visibility. A direct merchant watches a charge fail in their own dashboard and can act within the hour. A Paddle seller sees the subscription move through states — past due, paused, cancelled — and the useful work sits in those transitions: knowing a case exists at all, knowing which attempt is next, and never firing a second, competing recovery effort on top of Paddle's own. Double-working a customer whose card is already being retried by the platform is how people get charged twice or emailed twice, and both outcomes cost more trust than a single failed renewal.

The second consequence is the descriptor. Customers who do not recognise the merchant-of-record line on a statement do not open a support ticket to ask about it; they dispute, or they let the subscription die believing it was never theirs. Sellers cannot change Paddle's descriptor, so the recovery work that remains theirs is expectation-setting: the receipt, the renewal reminder, and the account page should all name the billing entity the customer will actually see, long before a failure makes that name matter.

A full map of which parts Paddle runs natively and which parts stay with the seller is laid out in a complete guide to Paddle dunning inside a merchant-of-record setup — the division of labour is the whole subject, because most Paddle recovery failures are really ownership failures: each side assuming the other sent the message, ran the check, or owned the case.

What the seller does own is considerable. Card-update capture: when a customer replaces a card anywhere it can be captured, the case should close and the next attempt should use the new details, not the old ones on a timer. Timing awareness: retries aimed at paydays and away from the dead hours recover more than retries on a fixed interval, and a seller-side layer can hold or release effort around the platform's schedule rather than duplicating it. And case discipline: before any seller-side attempt, the live state of the subscription gets checked, so a customer who already paid through another route is never chased for money they no longer owe.

There is also a quieter accounting point. Because Paddle's reporting is the system of record, seller-side recovery tools must reconcile against it rather than keep a competing count. Total recovered dollars, read from the platform's own ledger, is the only headline that survives an audit; a tool's internal counter is a working note, not a result.

One more seller-owned surface deserves a mention: the pause. When a Paddle subscription enters its past-due window, the customer's access decision — full access, read-only, or a clean pause with a stated return path — belongs to the seller's product, not the payment rail. Generous, clearly explained pauses convert some would-be cancellations into later recoveries, because a customer who feels fairly treated during a money problem tends to fix the card instead of quietly churning. The handoff, in other words, includes the tone of the wait, and that tone is entirely the seller's to set.

None of this makes Paddle a worse place to sell. For many teams the merchant-of-record trade is worth it many times over in tax and compliance relief alone. It simply relocates the recovery craft: away from raw retry mechanics, toward detection, timing, capture, and never working a case the platform has already closed.

Swipe Again supports Paddle sellers on exactly those terms — checking live invoice state before every attempt and counting only what the platform's ledger confirms — because in a handoff system, restraint is a feature.

Top comments (0)