Everyone talks about retention like it's one number. It's not. And treating it like one is how companies waste years solving the wrong problem.

Here's what usually happens. A SaaS company looks at its monthly churn rate, sees it's 5%, and declares "we have a retention problem." Then they hire a customer success team, build an onboarding flow, and send a bunch of "we miss you" emails. Six months later, churn is still 5%. They're confused.

The reason nothing changed is that they never bothered to understand when people leave. And when matters more than how many.

There are three kinds of churn, and they have completely different causes and completely different solutions.

The first is early churn. People sign up and leave within the first week or two. They never got to the "aha moment." They didn't understand the product, or the product didn't match what the marketing promised. This isn't a retention problem. It's a positioning problem, or an onboarding problem, or an acquisition problem. You're bringing in the wrong people.

The second is mid-term churn. People use the product for a few months and then drift away. They got value for a while, but something changed. Maybe their needs evolved. Maybe a competitor did something better. Maybe they just forgot why they were paying. This is where customer success teams actually help. Check-ins, feature education, integration support. The standard playbook works here because these are people who could stay — they just need a reason to.

The third is late-stage churn. People who've been with you for a year or more and finally leave. This is almost always structural. Their company changed direction. They got acquired. They switched to a platform that bundles your functionality. You can't prevent most of this, and trying to will burn enormous effort for almost no return.

The illusion is treating all three as one problem. When someone says "our churn rate is 5%," they're averaging together three completely different phenomena. It's like a doctor saying "your body temperature is 98.6" when your left hand is frozen and your right hand is on fire. The average is meaningless.

I've seen this play out dozens of times. A company with terrible early churn invests in customer success for existing users. A company with structural late-stage churn rebuilds their onboarding. They're working hard. They're just working on the wrong thing.

The fix is simple but requires discipline. Break your cohort data into time periods. What percentage of users leave in week one? Week two through eight? Month three through twelve? After a year? Plot these separately. You'll almost always find that one period dominates, and that period tells you exactly what kind of problem you have.

If your week-one drop-off is 60%, stop worrying about customer success. Fix your signup-to-value pipeline. Make sure people understand what your product does before they sign up — not after. Tighten your marketing so you attract people who actually need what you sell. The best retention strategy in the world won't help if you're pouring the wrong people into the top of the bucket.

If your mid-term churn is the issue, look at usage patterns. People who churn at month three usually stopped using a key feature at month two. Find that feature. Figure out why they stopped. Usually it's because they never knew it existed, or it was too hard to set up, or they didn't see how it connected to their workflow. This is fixable.

If late-stage churn is your biggest leak, honestly, stop trying to fix it. Focus your energy elsewhere. A certain amount of customer turnover is just the cost of being in business. Trying to prevent all churn is like trying to prevent all employee turnover. Some of it is natural and healthy.

The retention metric that actually matters isn't your overall churn rate. It's your churn rate for users who make it past the first month. Those are your real customers. Those are the people who understood your product, found value in it, and chose to pay. If those people are leaving at a high rate, you have a real problem. If most of your churn is people who never got started, you have a different problem — and it's probably an easier one to solve.

Stop staring at one number. Start asking when. (And if you're unsure which number to stare at instead, read The One Metric.)