Every subscription business already has dunning software. It lives inside the billing platform, retries on a schedule somebody set during onboarding, sends whatever template shipped as the default, and quietly cancels whoever is left at the end. The question a growing business eventually faces is not whether to have dunning. It is whether dunning is a setting or a system — a default nobody owns, or a managed process with a person, a review rhythm, and a written counting rule.
The difference shows up first in classification. A setting treats every failure as the same event: retry, email, retry, cancel. A system sorts before it acts. Expired and replaced cards need an update page that works without a login fight. Timing failures need spacing aimed at paydays, not persistence aimed at the bank. Closed accounts and revoked cards need a fast, clean stop. Fraud flags need the customer brought in to confirm with their issuer, once, politely. None of this is exotic. All of it is invisible to a default schedule, because a schedule cannot read.
It shows up second in sequence. In a managed system, the message and the retry are designed together: the customer hears about a failure before the second attempt, in words that name the amount, the date, and the one place to fix it. The card-update page is tested like a checkout page, because economically it is one — it is where a meaningful share of renewal revenue either completes or dies. And cancellation, when it comes, comes at the end of a sequence a human being reviewed, not at the end of a timer nobody remembers setting.
Third, and most neglected, is counting. A system has a written recovery rule: what counts as recovered, what a pause does to the scoreboard, whether a payment the platform rescued on its own native retry belongs in the tool's column or the house's. Teams skip this because it feels like bookkeeping. It is actually the management layer — the thing that lets a monthly review say anything truer than a percentage the vendor chose. The guide making the case for treating dunning management as a system with an owner puts the counting rule alongside classification and sequence for exactly this reason: the three are one discipline, and buying software only buys help with the first two.
None of this requires a large team. It requires an owner — one named person who reads the failure queue weekly, re-scores the sequences quarterly, and can say, from the business's own case log, what was recovered and why. For a small subscription business that person might spend an hour a week on it, much of it reading. What they cannot do is be nobody. Dunning run by nobody degrades exactly as fast as card files age, message templates stale, and pricing changes ripple through renewal amounts — which is to say, constantly, and silently.
The review rhythm is what keeps the system a system. A monthly half-hour with three artefacts on the table — the failure queue sorted by family, the recovery count under the written rule, and one sequence that underperformed — catches almost everything that degrades quietly: a template that stopped matching a price change, a platform update that reset a retry setting, a card file ageing into an expiry bulge two months out. None of those announce themselves. All of them are cheap to fix in the month they appear and expensive in the quarter after. Management, here, is mostly this: looking on purpose, on a schedule, before the numbers force the look.
So the practical order of operations is unfashionable: assign the owner, write the counting rule, sort the queue by family, then evaluate software against the system you have described. Tools chosen that way tend to fit, because the requirements were real. Tools chosen first tend to become the next default — slightly more expensive, equally unowned. The software matters. It is just the third decision, not the first.
Published by Swipe Again, the flat-fee failed payment recovery service for subscription businesses.
Top comments (0)