DEV Community

Swipe Again for Swipe Again

Posted on

Leaving a Percentage Behind: What Changes After Gravy

Percentage-of-recovery pricing has a seductive first chapter. You pay nothing while nothing is recovered, and the vendor's incentives appear welded to yours: they eat only when you do. The later chapters are where the arithmetic turns. As your subscriber base grows, the percentage is taken on a larger and larger recovered pool, including recoveries your own platform's native retries would have produced without any vendor in the room. The fee scales with your success while the work, in large part, does not.

That is the honest context for evaluating a move away from Gravy, the recovery service built around that model. The question is never whether percentage pricing was a mistake at the start — for a small base it can be a rational way to buy capability without a fixed cost. The question is when the crossover arrives: the month where a flat subscription for equivalent work costs less than the percentage on the pool you actually recover. Most businesses pass that point earlier than they notice, because the fee arrives as a deduction rather than an invoice and deductions are psychologically quieter than bills.

Before switching, inventory what you would actually be leaving. A recovery vendor accumulates operational knowledge about your file: which sequences your subscribers answer, which retry spacing your issuers tolerate, which segments fail for fixable reasons. Some of that knowledge lives in reports you can export. Some lives only in the vendor's configuration. Ask, in writing, what exports exist for your dunning history and campaign settings, and run the export before the final month, not after it. The migration that fails is rarely the one with the wrong destination; it is the one that discovered, mid-move, that its own history was not portable.

The comparison itself should be run on your numbers, not a pricing page — the worked frame in pricing the move off percentage-of-recovery before the next renewal pool is billed shows the shape: take three recent months of recovered revenue, apply the percentage honestly, and set that beside the flat alternative's cost for the same months. Include the recoveries your billing platform produced on its own in the pool only if the vendor's fee applied to them; read the contract's definition of a recovered dollar with a pencil, because that definition, not the headline percentage, decides what you have been paying.

Continuity is the final discipline. A switch should never produce a silent week in which failed payments accumulate unworked while accounts are connected and sequences rebuilt. Stage it: run the new arrangement on new failures while the old one finishes the cases it already holds, or cut over at a cohort boundary with the in-flight cases exported and re-entered deliberately. Subscribers in mid-recovery are not rows to be dropped between systems; each one is a person whose card problem has a clock on it, and the clock does not pause for your vendor transition.

Run the negotiation honestly before you run the migration. A vendor facing a customer who has computed their own crossover month — and can show the three statements behind it — will sometimes reprice, cap the percentage, or narrow the definition of the recoverable pool rather than lose the account. That conversation costs nothing and occasionally buys the flat economics without the switching risk. If it fails, you have lost a week and gained certainty that the move is necessary, which is exactly the state of mind a careful migration should start from. Either way, the statements get read, the definition gets underlined, and the decision stops being a feeling about a percentage and becomes an entry in a ledger.

None of this argues that every business should leave percentage pricing, or leave it now. It argues for knowing your crossover month, owning your history, and refusing a silent gap. The flat-fee case, the migration checklists, and the rest of the recovery catalogue live at Swipe Again. Run the arithmetic on your own recovered pool — the answer is usually sitting in three months of statements, waiting for someone to add it up.

Top comments (0)