Payment-failure churn is a dunning problem. Intent churn is a decisioning problem. How to split them, who owns each, and how to measure both correctly.
Frequently asked
Related resources
Solutions
See it in the product
The same loops this note describes run 24/7 against your customer base. Watch the workspace decide, experiment and execute 1:1.
Vocabulary
Definitions an answer engine can quote, each one a page of its own.
Compare
Neutral reads on the categories buyers evaluate against a decision layer.
All comparisonsVoluntary churn is a customer deciding to leave. Involuntary churn is a customer leaving because a payment failed. They look identical in a churn dashboard and they are almost nothing alike: different root cause, different owner, different fix, and a different way of measuring whether the fix worked.
Most retention programmes are built as if all churn were voluntary. That is why so many of them spend margin on discounts for customers who never intended to cancel, while a quieter stream of cancellations keeps leaving through a failed card that nobody owns.
Across the subscription bases we work with, involuntary churn is consistently the underestimated half. In card-on-file B2C, payment-failure cancellations typically land between 20 and 40 percent of gross churn, and they skew higher in markets with high prepaid or debit penetration, in bases with long tenure (more card reissues), and after a pricing change that pushes charges past a bank's low-value approval threshold.
The reason this matters commercially: involuntary churn is cheaper to fix than voluntary churn. Recovering a failed payment costs a retry and a message. Reversing an intent to cancel usually costs margin.
The levers are operational, and they compound:
None of this is marketing work. It belongs with payments and billing, and it is measured in recovery rate per decline code, not in campaign performance.
Here the question is not how to collect the money, it is which intervention changes the outcome for this customer, at what cost, with what incremental effect. That is uplift, not propensity: a customer with high cancel probability and near-zero uplift to any offer is not worth a discount, and a customer with moderate probability and high uplift is.
The practical shape is the one described in our churn prediction guide: rank eligible actions by uplift-weighted margin, hold out a slice of equally scored customers, and read incremental save rate weekly.
A single churn rate hides both problems. The minimum reporting split:
There is a real grey zone. Some involuntary churn is passive voluntary churn: the customer stops topping up the card on purpose, or lets the payment fail rather than clicking cancel. The way to separate them is behavioural, not billing-based. A customer who was actively using the service in the week before the decline is almost always genuinely involuntary. One whose engagement decayed for six weeks before the decline had already left.
That distinction changes the treatment. The active customer gets a frictionless card-update path. The disengaged one gets a value intervention first, because updating the card only buys one more cycle of the same outcome.
Markin separates intent churn from payment churn at the signal layer, then ranks the interventions that actually move each one. See Retention decisioning or the telco view, where both streams are largest.
