A collections call works because the person answering owes you money and knows it. There is an invoice, a purchase order, a relationship with a credit limit attached. The call is awkward, but the obligation is not in dispute. Almost everything collections does well — the firm tone, the escalation ladder, the sense that this is now a formal matter — depends on that shared understanding.
A subscriber whose card failed has a completely different relationship with you. They did not decide to stop paying. Most of them do not know anything happened. Their bank declined a charge at 3 a.m. for reasons that may have nothing to do with their balance, and the first time they hear about it is your email. Treating that moment like collections — formal, escalating, faintly accusatory — does not recover the payment. It teaches a happy customer that your company is unpleasant to owe money to, and it converts a billing accident into a cancellation decision they had not been planning to make.
This is the line the industry keeps blurring, and it is worth drawing precisely. I keep a working version of it in where dunning ends and collections begins, but the short form fits in a paragraph. Dunning is everything that happens while the relationship is intact and the failure is presumed accidental: retries timed to the decline type, clear emails that make updating a card take thirty seconds, a grace period that matches the product's real usage cycle. Collections begins when the failure is presumed deliberate or the relationship is over — and for a subscription business, that presumption is wrong often enough that building your whole recovery motion on it is expensive.
Three practical differences fall out of the line. First, tone. A dunning email can be warm without being weak: "your card was declined, here is the one-click fix, your account stays active until Friday" outperforms "FINAL NOTICE" language with subscribers because the subscriber is not a debtor dodging you; they are a customer with a logistics problem. Second, timing. Collections escalates on a schedule set by the creditor. Dunning should escalate on a schedule set by the decline: insufficient funds waits for payday patterns, a generic decline wants a fresh read of the invoice before any retry, an expired card wants a message, not a retry at all. Third, the exit. Collections ends in write-offs and agencies. Dunning should end, at worst, in a respectful pause — because a subscriber who leaves over a billing failure and is treated well comes back at a rate no winback campaign aimed at angry ex-customers ever matches.
None of this means being soft about money. The firmest thing in a good dunning sequence is its clarity: what failed, what happens next, exactly when access changes, and what the customer can ignore if they have already fixed it. Clarity is firmness that does not require threat. The businesses that recover the most failed payments are rarely the ones with the most aggressive sequences; they are the ones whose sequences a reasonable person would describe as helpful — sent at the moments the bank's answer suggested, and stopped the moment the payment landed.
If your current process was borrowed from B2B collections — or from a template written by someone imagining a debtor — it is worth re-reading it as your best customer would: mid-workday, card declined for a reason they cannot see, deciding in four seconds whether your email is help or hassle. Everything recoverable in that account rides on which of those two words they pick.
For context on how we approach this side of the line, see Swipe Again's failed payment recovery — flat-fee recovery built for subscribers, not debtors.
Top comments (0)