Churn rate

Churn rate is the percentage of customers who cancel or stop paying in a given period. Start the month with 100 customers, lose 5, and your monthly churn is 5%. It's just retention flipped around: 95% retention is the same as 5% churn. Simple to calculate, brutal to ignore.

The reason churn decides whether a business lives or dies is that it compounds. 5% monthly churn quietly bleeds 46% of your customers over a year; 10% monthly churn loses you 72%. So you can win 100 new customers a month and still shrink, because you're pouring them into a leaky bucket. That's why fixing churn often moves revenue more than doubling your ad spend, and why investors stare at it harder than they stare at growth.

To actually do something about it, you have to break churn apart rather than watch one blended number. Two cuts matter most.

First, split it by cohort, the customers grouped by when they signed up. Say you're running a product-analytics tool like Amplitude and you chart retention curves per signup month: you'll often spot a cliff in month two or three, right when people decide whether you're worth keeping. If January's cohort retains far worse than March's, you go find what changed (usually onboarding) instead of guessing.

Second, separate voluntary churn (people choosing to leave) from involuntary churn (failed card payments, expired cards). They have completely different fixes. Involuntary churn is often just a dunning problem, recoverable with a few well-timed emails. Say you wire up Customer.io to fire a payment-retry sequence and a win-back flow for cancellations, you can claw back a chunk of churn that had nothing to do with how good your product is.

Then track the number religiously. Pick monthly churn as a north star, set a target (say from 5% down to 3% inside a year), and put it on a dashboard the whole team sees. Pipe it into something like Databox so it updates weekly without anyone rebuilding a spreadsheet. Churn reduction compounds the same way the loss does, so every single point you claw back keeps paying you for years.

An analytics platform noticed customers typically churned in months 4-6 after purchase. Exit interviews revealed customers hit the feature limits of their tier and didn't want to upgrade. Instead of losing customers, the company bundled 'commonly requested' features into lower tiers without raising price. Voluntary churn dropped 3%, and the expansion revenue from upgrades more than offset the value given away.

Calculate churn by cohort

Don't just track overall churn. Track by customer cohort: customers acquired in month 1, month 2, etc. This reveals whether churn is improving (are you better at onboarding now?) or worsening (was last month's cohort lower quality?). Cohort churn also reveals the riskiest period: many companies see high churn in month 2-3 when customers evaluate alternatives.

Separate voluntary and involuntary churn

Involuntary churn (failed payments, deprecated features) has different solutions than voluntary churn (customers choose to leave). Track them separately so you can address the right root cause.

Analyse why customers leave

When customers cancel, conduct exit interviews. What was their original use case? Did they achieve it? Why are they leaving? Are competitors involved? After 20-30 interviews, you'll see patterns. 'Not enough features for enterprise' is different from 'product too complex for what we need'. Different patterns require different solutions.

Set churn targets and track them religiously

Your north star should be monthly churn rate. If you're at 5% monthly (46% annually), your goal might be 3% (34% annually) within 12 months. Set targets, track weekly, and tie team incentives to hitting them. Churn reduction compounds; every 1% improvement is significant.

A project management tool tracked churn by customer segment. Enterprise customers had 2% monthly churn; SMB had 8%. Investigation found SMB customers were power users who quickly outgrew the product and switched to enterprise solutions. The company created a mid-market tier with additional features, reducing SMB churn to 4% and capturing revenue growth they were losing to competitors.

A SaaS company tracking monthly churn by cohort found customers acquired in January had 12% churn by month 3, February had 15% by month 3, but March had only 8% by month 3. What changed in March? They redesigned onboarding. The insight prompted them to rollback February's onboarding experiment and double down on March's approach. Cohort analysis turned a vague problem ('our churn is high') into a specific solution ('fix onboarding').

Determines long-term business sustainability

A business with high acquisition but high churn is unsustainable. You're constantly replacing customers instead of growing. Lowering churn even 2-3 percentage points often has more impact on revenue growth than doubling acquisition spending. Retention is cheaper than acquisition.

Reveals product-market fit problems

High churn often signals customers don't see ongoing value. They tried your product, weren't convinced, and left. This is different from the product being bad; it's misalignment between what customers need and what you deliver. Churn cohorts show which customer segments or use cases aren't working.

Impacts valuation and investor confidence

Investors obsess over churn. A startup with high churn rates and fast growth is a red flag; they're burning money acquiring customers they'll lose. A startup with low churn and slower growth is attractive because it's building sustainably. Churn is more important to investors than raw growth rates.