DEV Community

Swipe Again for Swipe Again

Posted on

Three Maps of the Same Failure Queue

Every failed-payment queue can be drawn three ways, and each drawing argues for a different purchase. Draw it by decline code and the queue looks like a banking problem: so many insufficient funds, so many expired cards, so many do-not-honors. Draw it by age and it looks like a patience problem: most value sits in the first week, quietly expiring. Draw it by customer and it looks like a communication problem: people who would have paid, had anyone told them in time and in plain language. A vendor demo will happily show you whichever map flatters its product. The work of choosing is deciding which map describes your losses.

Start with the decline-code map, because it sets the ceiling. Sort last quarter's failures into three piles: timing declines, where the money exists but not on the day you asked; credential declines, where the card itself changed, expired, or was replaced; and terminal declines, where the account is closed or the card is reported lost. Only the first two piles are recoverable by any tool, and they recover by completely different means — timing by aiming attempts at paydays, credentials by asking the customer for a new card early and making that ask painless. A solution that treats all three piles with one retry ladder is not being simple. It is being incurious about two-thirds of its job.

The age map is where demos get quiet. Plot recovered dollars against the day they were recovered and a honest pattern appears: a strong first seventy-two hours, a long shallow middle, and a tail that costs more in customer goodwill than it returns in cash. That shape argues for a system that front-loads intelligence rather than volume — live checks before the early attempts, spacing that widens instead of repeating, and a deliberate end. If a vendor's map shows recoveries spread evenly across a month of attempts, ask what their denominator excludes, because real queues do not behave that way.

The customer map is the one nobody draws, and it may matter most. Mark which failures belonged to customers in their first ninety days, which to annual plans, which to accounts that had already survived one failure. Each group needs a different voice: the new customer needs reassurance that the product is worth fixing billing for, the annual payer needs concierge clarity because the amount is large, and the repeat survivor needs honesty about what happens if this one is not resolved. Templates that ignore the map read as debt collection to exactly the people you can least afford to alienate.

Choosing, then, is a mapping exercise before it is a shopping exercise. Bring your three maps to every evaluation and score each candidate against the pile it claims to solve: does it aim retries at real paydays, does it detect a card updated elsewhere before spending an attempt, does it close each case with a reason your future self can audit? The tools separate quickly once the questions are that concrete, because most products are excellent at one map and silent about the other two.

For a structured way to run that comparison, a coverage-first guide to dunning solutions that scores each option against the failure piles it actually solves walks through the decline, age, and customer maps in the order an operator would draw them on a whiteboard.

Draw your maps first. Vendors become much easier to grade once you know which territory your money is actually lost in — and much harder to impress with a demo drawn on theirs.

That is the standard Swipe Again publishes for itself: payday-aimed retries for the timing pile, live card-state checks for the credential pile, honest endings for the terminal pile, and one flat monthly price across all three maps.

Top comments (0)