DEV Community

Swipe Again for Swipe Again

Posted on

The Collections Call That Should Never Have Been a Collections Call

A subscription business I worked with kept a spreadsheet of "collections cases." Every row was a subscriber whose renewal had failed and who had not answered two emails. Once a week, someone worked the list with phone calls that sounded, unavoidably, like debt collection — because the sheet said collections, the calls became collections, with the tone, the urgency, and the quiet threat underneath both.

Almost none of those rows belonged there. One customer's card had been reissued and the new number was sitting in the vault, unpresented. One had failed on the Friday before payday and cleared on the Monday retry nobody had scheduled. One had never seen the emails at all; they were in a promotions folder, unread, next to a discount offer from the same company. These were not people refusing to pay. They were people whose payment machinery had hiccuped while their intention stayed intact.

That distinction is the whole subject of where dunning ends and collections properly begins: dunning is the work you do on a payment that failed, for a customer who still wants the thing. Collections is the work you do on a debt, for a customer who has decided not to pay, or vanished. Confusing the two is expensive in both directions. Treat a dunning case like a debt and you manufacture the cancellation you were trying to prevent. Treat a real debt like a dunning case and you send polite reminders into a void for ninety days.

The practical test is evidence of intent. A subscriber who opens the app, clicks the update link, replies to the email, or pays late-but-eventually is showing intent; their case is a mechanics problem and stays in dunning, where the remedies are timing, credential refresh, and one clear message. A subscriber who has cancelled in writing, disputed the charge, or gone silent across every channel for a full cycle has crossed the line; that case leaves the recovery queue and follows whatever collections or write-off policy the business actually has.

Most queues I audit have the line drawn in the wrong place for a simpler reason: time. Anything older than thirty days gets stamped collections, as if age alone changed the customer's mind. Age changes the odds, not the category. A sixty-day-old failure with a fresh login last week is still a dunning case. A six-day-old failure from a customer who emailed "cancel this and stop charging me" is not, no matter how recent.

There is a third category that muddies the spreadsheet further, and it deserves its own row: the administratively stranded. Their payment method was a company card reassigned by an employer, a joint account closed in a separation, a prepaid card that cannot be reloaded. Intent is present, ability is tangled, and neither a retry schedule nor a collections script fits. These cases need a human sentence — "we can see this is complicated; here is a link to change the payment method whenever you are ready" — and then patience measured in weeks, not attempts. Businesses that create this third row stop forcing their strangest cases into scripts written for somebody else, and their recovery numbers improve partly because the first two categories are finally being worked correctly.

The spreadsheet itself should record the judgment, not just the balance. Which category, decided when, on what evidence, reviewed on what date. Without that, every staff change resets the line to wherever the new person's instincts sit, and customers experience the wobble as inconsistency: a gentle reminder one month, a stern notice the next, for the identical situation. Consistency is a form of fairness customers can feel even when they never learn the taxonomy behind it.

Draw the line by intent, review it weekly, and two things happen. The recovery work gets gentler and more precise, because it is aimed at people who want to stay. And the genuinely gone stop consuming the attention that the merely stuck were owed all along.


Published by Swipe Again — failed payment recovery at a flat $179/mo.

Top comments (0)