Three merchants, three very different Mondays.
The first runs a fitness app. A member's card failed on Friday because the bank flagged an unusual location. By Monday the member has already opened the app twice, annoyed that their plan shows as paused. Nobody has chased them for money they owe in any legal sense — the service simply has not been paid for yet, and the fix is a new card and a quiet retry. That is dunning territory.
The second runs a B2B reporting tool on annual contracts. An invoice is 74 days past due, the customer success manager has had two calls about it, and finance is now discussing whether to involve a third party. The customer is not confused about how to pay. They are choosing not to, or cannot. That is collections territory, and it obeys different rules, different tone, and in many jurisdictions different law.
The third case is the one that causes the damage: a consumer subscription box whose team treats every failed renewal like case two. Their past-due emails read like demands. They threaten account closure on day two. Their recovery rate is poor, and their support queue is full of people who would happily have updated a card if anyone had asked them like a human being first.
The confusion is understandable because the vocabulary overlaps. Both processes follow a failed payment. Both send emails. Both show up in the same billing dashboard. But the underlying situation is opposite. Dunning assumes the customer wants to stay and the payment failed for a mechanical reason: an expired card, a replaced card, a temporary bank hold, a limit hit on payday week. Industry data consistently shows most involuntary failures are exactly that — the card, not the customer, said no. Collections assumes willingness or ability is the problem, and the conversation becomes about obligation.
That distinction changes everything practical. In dunning, speed and helpfulness win: a clear email within hours, a card-update page that works on a phone without a password reset, retries timed around paydays and bank behavior rather than fired on a fixed hourly loop. The tone is service, not pressure — "your card did not go through, here is the one-tap fix" outperforms any language about amounts owed. There is a useful explainer on where dunning ends and collections begins that walks the boundary case by case, including the awkward middle where a long-running dunning sequence quietly becomes a collections problem.
In collections, by contrast, documentation matters more than speed, promises to pay are tracked formally, and consumer-protection rules about contact frequency, wording, and third-party disclosure can apply. Importing collections habits into dunning is not just ineffective, it can create compliance exposure you never needed, because you were never actually collecting a debt — you were fixing a card.
A practical test for your own sequence: read your day-one, day-three, and day-seven past-due emails aloud. If any sentence would make sense in a letter about an overdue loan, it does not belong in a dunning flow for a $20-a-month product the customer still opens every day. Then check the opposite failure: if a genuinely unwilling payer can ride your friendly sequence for 90 days with no escalation path, you do not have a collections boundary at all.
Most subscription businesses need excellent dunning and almost no collections. Building the first well — and knowing precisely where it stops — is the whole job.
We build Swipe Again to handle the dunning side of that line: smart retries and card recovery for subscription businesses, for a flat monthly fee.
Top comments (0)