Most subscription businesses talk about dunning as if it were a setting. Turn on smart retries, write three emails, and the failed payments are supposed to take care of themselves. Anyone who has watched a renewal queue for a month knows it does not work that way. Dunning is a stack of separate decisions, and a weakness in any one layer quietly caps what the layers above it can recover.
The bottom layer is the card itself. A meaningful share of declines are not about money at all: the card was replaced, the billing address changed, the issuer wants a different format. Account updater services fix some of this, but only some, and only when the retry actually uses the updated credential. A retry fired with the old card details is not a recovery attempt. It is a ritual.
Above that sits timing. Issuers decline for reasons that expire: daily limits reset, paydays land, temporary holds lift. A retry schedule built on a fixed clock — every 24 hours, say — ignores all of that. The same card that declines on a Tuesday evening may approve on a Friday morning for no reason anyone logged. Timing is not a detail in recovery work. In many queues it is the largest single variable.
Then comes communication, and here the order matters. An email sent after the final retry is an obituary. An email sent early, while retries are still working in the background, gives the subscriber a chance to fix the problem themselves — a new card entered once is worth more than five retries against the old one. The best-performing message in most queues is also the plainest: your payment failed, here is the secure link, here is what happens if it is not fixed.
The top layer is measurement, and it is the one most often skipped. Recovery rate means little without a denominator. Count only the failures that were actually recoverable — hard declines and cancelled cards were never coming back — and suddenly the honest number is both smaller and more useful, because it shows what the stack is really doing. A fuller breakdown lives in a guide to subscription dunning software — worth reading before the next tool comparison, because it separates the layers most vendors blur together.
None of this requires exotic technology. It requires treating each layer as its own job: keep credentials current, retry on a schedule that respects how money actually moves, ask the customer early and plainly, and count honestly. Businesses that do this rarely describe dunning as a feature. They describe it as an operation — one that runs every day, whether anyone is watching the dashboard or not.
That is also the honest test for any tool in this category. Do not ask whether it "does dunning." Ask which layers it covers, which it leaves to your billing platform, and which it leaves to you. The gaps between layers are where renewals go to die quietly, one plausible-looking decline at a time. More on how one team approaches the whole stack is at Swipe Again's failed payment recovery overview.
Top comments (0)