Most churn taxonomies are organised by what the customer felt: too expensive, not using it, found an alternative, life changed. Those categories are true and almost useless, because they do not tell you whose desk the fix sits on. A more practical cut — the one behind why subscribers churn, sorted by who can fix it — divides the exits into three piles: the ones your product team owns, the ones your operations own, and the ones nobody owns because the customer simply left the category. Each pile gets a different response, a different budget, and a different level of worry.
The product pile is the one everyone reaches for first: features missing, quality slipping, a competitor's launch. It is real, and it is also the slowest to move and the easiest to hide behind. Before assigning an exit to this pile, check the billing record. A striking share of subscribers who cite "not using it enough" on an exit survey had a failed payment in the previous cycle that nobody worked properly. The survey answer describes the feeling at the exit door; the ledger often describes the actual cause, which was a card, a retry, and a message that never landed.
The operations pile is where the fixable churn lives: failed payments mishandled, cancellation flows that offer nothing between "stay" and "gone", pause options that exist but are never presented, price rises announced like ambushes. These exits share a trait — the customer did not decide against the product so much as the machinery decided for them. A renewal that fails on an expired card, retried blindly for a week, then cancelled automatically, is recorded as churn. Nobody chose it. It is the operational equivalent of a shop locking its door during opening hours and logging the lost sales as lack of demand.
The nobody pile deserves honesty rather than effort. People move, budgets close, needs genuinely end. The correct response to this pile is grace: a clean exit, a door left open, a win-back note timed to when the circumstance might change. Businesses that refuse to accept this pile waste their save budget annoying people who were never saveable, and train their teams to distrust the save data entirely.
Sorting this way also fixes your metrics conversation. Voluntary and involuntary churn are not two flavours of one problem; they have different owners, different remedies, and different ceilings. Involuntary churn — the payment-failure slice — can plausibly be cut by half or more with disciplined recovery work, because the customer never voted to leave. Voluntary churn moves in points, slowly, with product work. Reporting them blended produces a number that flatters nobody's work and directs nobody's week.
The sorting also changes how you staff the problem. Product-pile churn belongs in roadmap discussions, reviewed monthly against cohorts. Operations-pile churn belongs in a weekly queue with names and dates on it, because its remedies are procedural and its losses compound silently. Nobody-pile churn belongs in neither — it belongs in a clean exit experience and a light-touch win-back calendar. When all three piles share one meeting and one metric, the urgent procedural losses wait politely behind the interesting product debates, month after month, while the retry calendar stays exactly as broken as it was.
Try the exercise on last month's exits. Take fifty cancellations, read the billing history before the survey answer, and sort them into the three piles. Most teams find the operations pile is larger than they assumed and the product pile smaller — which is good news, because operations is the pile you can fix this quarter, with checklists instead of roadmaps.
That operations pile, specifically its failed-payment slice, is the whole of what Swipe Again does: retries sorted by failure family, card updates made one click deep, outcomes counted on the original dollars over ninety days, flat $179 a month. Sort your reasons by owner first. Then spend on the pile that can actually move.
Top comments (0)