Churn is not one problem.
That's the mistake every team makes at least once. They see the number going the wrong direction, build a win-back email campaign, add an exit survey, maybe throw a discount at canceling users — and then wonder why the rate barely moves. It barely moves because they're treating four different problems as one.
The moment I join a subscription business, the first thing I do is break churn into two axes. The first is timing: early churn versus late churn. The second is intent: voluntary versus involuntary. That gives you a 2×2. Each quadrant has a different cause and a completely different fix.
Early involuntary is payment failure on new accounts. Someone just subscribed, their card didn't process correctly, and they never technically started. This isn't a retention problem — it's a billing problem. Fix it with retry logic and dunning sequences before you do anything else, because every email campaign in the world won't recover a subscriber whose card details you never captured correctly.
Early voluntary is the expensive one. Users who subscribed, tried the product, and left before getting anything out of it. This is almost never about the product being bad — it's about the experience between signup and first value being too slow, too confusing, or too silent. Joey Coleman's research across hundreds of subscription businesses puts 20 to 70% of new customers lost in the first 100 days. Not because the product failed. Because the company assumed that once someone paid, the selling was done.
It wasn't. The period between "they gave us their credit card" and "they felt the product working" is the most fragile stretch of the entire relationship. HubSpot went from an LTV:CAC ratio of 1.67 to 3.5 in a single year — not by cutting acquisition costs, not by raising prices. By cutting early churn in half through better onboarding. That one change more than doubled the lifetime value of every new customer they brought in.
Late involuntary is card expiry and failed renewals on long-tenured accounts. These users aren't unhappy — their Visa expired and nobody told them. The fix is proactive: account updater services, payment alerts before the renewal date, a human call for high-value accounts. Most teams ignore this until it compounds.
Late voluntary is intentional cancellation from users who stuck around for a while. This is the hardest one because it means perceived value has been eroding for some time before they acted on it. By the time someone cancels a subscription they've held for two years, the decision was made months ago. An exit survey won't save them. The useful signal is in the data that preceded the cancellation — when did engagement drop? What feature did they stop using? Which cohort are they from, and did that cohort always churn at this rate or did something change?
The mistake is treating all four of these with the same tool. A discount offer at cancellation helps late voluntary users who are price-sensitive and helps no one else. An onboarding sequence helps early voluntary users and arrives too late for everyone else. You need four separate responses to four separate problems.
Here's how I prioritize. Plot your churn on the 2×2. Find the quadrant where the most revenue is leaving. Fix that first and only that first. Churn is non-linear — 4% at $10K MRR means replacing four customers a month. 4% at $100K means replacing forty. The faster you're growing, the faster the treadmill moves. You cannot outgrow a churn rate. You can only fix it.
Once the biggest quadrant is under control, move to the next. Don't run four initiatives simultaneously. The clarity of knowing exactly which problem you're solving is what makes the fix stick.