Most teams meet their failed payments the way they meet a fire alarm: something loud happens, everyone runs, and afterwards nobody writes down what they learned. The queue gets worked hard for a week, the numbers improve a little, attention drifts, and the same failures quietly rebuild. Dunning managed as an emergency always feels busy. Dunning managed as an operations review feels almost boring — and boring, repeated weekly, is what actually shrinks a failure rate.
An operations review starts with the same three numbers, read in the same order, every week. First, how many renewals failed, as a share of renewals attempted — not the raw count, which rises and falls with your subscriber base and tells you nothing about health. Second, how many of those failures were recovered, split by who did the recovering: your billing platform's automatic retries, a customer updating their own card, or your recovery process itself. Third, the average age of the cases still open. A queue can post a decent recovery rate while quietly letting its oldest cases rot, and age is the number that catches it.
With those three numbers on the table, the review turns to reasons, because reasons are where the work is. Sort the week's failures by decline type and look for the change, not the total. Expired cards creeping up usually means your pre-expiry warnings slipped or your customer base aged into a replacement cycle. Insufficient-funds declines clustering on particular days of the month usually means your renewal dates drifted away from your customers' paydays. A rise in vague, unexplained declines usually means an issuer changed something and your retry spacing needs to change with it. Each pattern points at a different fix, and none of them is "send more email." Teams that want the full working method will find it in a practical dunning management guide built around a weekly review rhythm, which lays out the cadence, the reason codes worth tracking, and the order to work them in.
The review also needs an honest argument about attempts. Every operation has a house style: some retry quickly and often, some wait and strike once. Neither is right in the abstract. What the weekly review gives you is evidence about your own customers — which attempt number actually recovers money for you, and which attempts mostly generate noise, support tickets, and bank fatigue. When attempt five has recovered nothing for six straight weeks, the review is where that gets said out loud and the schedule gets shortened. Without the review, schedules are inherited habits that nobody remembers choosing.
Finally, the review owns the ending. Cases close in one of three ways: recovered, honestly lost, or handed to a winback motion with fresh terms and fresh timing. What they should never do is linger in a fourth, unspoken state — still technically open, still occasionally retried, long past the point where the customer considers the matter settled. A weekly review forces the endings to happen on schedule, which keeps the metrics honest and the customer experience defensible.
None of this requires heroics. It requires thirty minutes, the same three numbers, a reason sort, and the willingness to change one thing a week based on what the evidence says. Run that meeting for a quarter and the failure queue stops being weather — something that happens to you — and becomes operations: something you run, measure, and steadily improve.
That is the operating model behind Swipe Again: every failure becomes a case the moment it happens, declines are read before attempts are spent, and the counting separates platform recoveries from the tool's own work so the weekly review reads true.
Top comments (0)