What Is Customer Churn? And How to Calculate It Properly
Post 11 Sep 2026 Customer ChurnRetentionSaaS MetricsRevenue ChurnCustomer Success
Customer churn is the percentage of customers who stopped doing business with you during a defined period, measured against the number of customers you had at the start of that period. The common mistake is to track it by headcount only, and once a year only — which makes a company look healthy while its recurring revenue quietly shrinks from the inside.
What Churn Actually Is: A Two-Line Definition and One Formula
The base formula:
Customer churn = (customers lost during the period ÷ customers at the start of the period) × 100
Two conditions keep the number honest. First, don't include customers who joined during the period in the denominator. Second, lock down your definition of "lost" — is it a contract that expired without renewal, or a formal cancellation request? A definition that shifts between quarters makes quarter-over-quarter comparison meaningless.
Customer Churn vs. Revenue Churn: Why the Two Numbers Diverge
Revenue churn = (recurring revenue lost ÷ recurring revenue at the start of the period) × 100
The gap matters. If you lose two customers out of a hundred in a quarter, customer churn is 2%. But if one of them was your largest account, revenue churn might be 12%. The reverse is also true: losing five small customers can cost you less than one large customer downgrading.
The rule of thumb: customer churn measures the health of your product and service; revenue churn measures the health of your business. Read them together.
Churn on Annual Contracts: How to Calculate It When There's No "Month"
With annual contracts you don't calculate a monthly rate. You calculate against the renewal cohort: every contract due for renewal in the current quarter is treated as one pool.
Renewal rate = contracts renewed ÷ contracts due for renewal in the quarter.
That gives you a reading every three months instead of waiting for year-end. A customer whose contract comes up in Q4 doesn't belong in your Q2 calculation — they simply haven't had the chance to leave yet.
Contraction and Expansion: The Customer Who Stayed but Cut Seats from 40 to 15
In per-seat subscriptions, the most dangerous kind of loss never appears on the cancellation list. A customer who drops from 40 seats to 15 and renews counts as 100% retained — while you've lost 62% of their revenue.
So track two separate lines every quarter:
| Line item | What it means |
|---|---|
| Churn | Revenue from customers who left entirely |
| Contraction | Lost revenue from customers who stayed (fewer seats, fewer modules) |
| Expansion | Added revenue from existing customers |
Lost revenue = churn + contraction. And when expansion exceeds the sum of the two, you're growing out of your existing customer base without a single new sale.
What Counts as an "Acceptable" Number — and Why an Imported Benchmark Won't Fit Your Market
Don't import a benchmark and hold your team to it. The right yardstick is your own number over the past four quarters, segmented by customer size and industry. The differences are large: a government contract is tied to budget cycles and procurement rules, while a small business may shut down for reasons that have nothing to do with you. Blending both into a single average hides the real problem.
Early Warning Signals: What You Can See Three Months Before a Non-Renewal
Nobody decides not to renew on the renewal date. The signals that usually come first:
- Active users declining relative to licensed seats
- A module the customer once relied on falling out of use
- Repeat support tickets without a satisfying resolution
- The main point of contact changing with no introduction to a replacement
- Two consecutive late payments
- No logged contact with the account in 60 days
What matters most is turning these signals into automated alerts rather than a manual monthly review. In Effistar you can set a rule that automatically creates a task for the account owner the moment any of these conditions is met — instead of waiting for someone to notice.
## Who Owns the Number? Splitting Accountability Across Sales, Operations, and SupportChurn is usually a number nobody owns, which is why nobody fixes it. Here's a clear split:
- Sales: first-year churn — usually a promise made in the sale that delivery never met.
- Operations: onboarding-stage churn — the customer who never actually started using the product.
- Support: post-first-year churn — an accumulation of unresolved issues.
Tomorrow's step: pull the list of contracts due for renewal next quarter and add two columns to each — active users today versus licensed seats, and the date of the last logged contact. The accounts that look weak in both columns are your work list for this week.