Dunning and involuntary churn

Not all churn is a decision. A large share of cancelled subscriptions are involuntary: the card expired, the payment was declined, the funds weren’t there. The customer didn’t choose to leave - they still want the product, they just stopped paying without realising, and a subscription quietly lapsed. In a lot of recurring businesses this involuntary churn is 20-40% of total churn, which makes it one of the biggest leaks in the whole model and the one nobody’s emotionally invested in fixing.

Dunning is the process of recovering those payments - the retries, reminders and card updates that turn a failed charge back into a paying customer. It’s the closest thing to free money in recurring revenue, because you’re not acquiring anyone or convincing anyone. The customer already decided they want the product. You’re just fixing a broken transaction, at zero acquisition cost.

What dunning actually involves

• Smart retries. A failed charge retried on the right schedule recovers a meaningful share on its own - funds arrive, a temporary decline clears. Retrying blindly the next day wastes attempts; retrying on the patterns that actually recover, spaced out and timed around paydays, does far better.
• Pre-dunning. Catching the problem before it happens by emailing customers whose cards are about to expire to update them ahead of the failed charge. The cheapest failed payment is the one that never fails.
• Card-account updater services. The card networks can supply updated details when a customer’s card is reissued, so the subscription bills the new card automatically without the customer doing anything.
• The recovery sequence. Email and SMS telling the customer the payment failed and making it one click to fix. This is a conversion funnel like any other - the “convert” action is just updating a card - and it responds to the same optimisation: clear message, minimal friction, good timing.

Why it’s so underworked

It’s unglamorous and it belongs to nobody. It’s not acquisition, so marketing ignores it. It’s not product, so product ignores it. It looks like billing, so it gets treated as finance plumbing rather than the retention surface it is. Meanwhile a few points of recovered involuntary churn flow straight to the bottom line with no offsetting cost, a better return than almost any acquisition test.

Where it goes wrong

• Treating it as a billing problem, not a CRO one. The recovery sequence is a conversion flow and rewards the same testing - subject lines, timing, friction, the number and spacing of touches. Left on a default setting, it under-recovers.
• Over-aggressive retries. Hammering a card with daily retries annoys customers, racks up processor fees, and can flag the account with the card network. Recovery is about the right attempts, not the most.
• Dunning emails that read as spam. A “your payment failed” message that looks like a phishing attempt gets ignored. It has to be trustworthy and obviously from you, or the customer won’t act.
• Ignoring the voluntary cases hiding inside. Some failed payments are a soft cancellation - the customer let the card lapse on purpose. Pushing those too hard is better handled as a cancellation conversation than a payment retry.