Most "churn rate" answers stop at one formula: customers lost divided by customers at the start. That formula is fine until your customer count is changing fast, you have multiple cohorts, or your customers don't all pay the same amount. Which is to say, it's fine if you have one customer.
Real businesses need to know which method to use, when, and what each one tells them. Here are the 4 methods, ranked by sophistication.
For a faster path to insight: take the Churn Health Check and it'll tell you not just your churn rate but whether you're measuring it correctly in the first place.
Method 1: Simple churn rate (the one everyone uses)
Formula: Customers Lost / Customers at Start of Period
Example: You started March with 1,000 customers. Lost 50 during March. Simple churn rate: 50 / 1,000 = 5%.
When to use it: Quick board updates, casual benchmarking, situations where your customer count is stable. It's the easiest to calculate and explain.
Where it fails: When your customer count is growing fast. If you started March with 1,000, lost 50, but added 300 new customers, the 50 lost happened among an effective base much larger than 1,000. Simple churn overstates your actual churn rate during high-growth periods.
Method 2: Average churn rate
Formula: Customers Lost / ((Customers at Start + Customers at End) / 2)
Example: Started March with 1,000. Ended with 1,250. Lost 50. Average: 50 / 1,125 = 4.4%.
When to use it: When your customer count is changing meaningfully during the period. Most growing SaaS companies should default to this over the simple method.
Where it fails: Still doesn't account for the fact that newly-acquired customers had less time to churn during the period. The 300 customers acquired on March 30th had only 1 day to churn, but they're in the denominator as if they had a full month.
Method 3: Cohort-based churn rate (the most accurate)
Formula: For each signup cohort, track what % churns over time. Then aggregate or compare cohorts.
Example: Customers who signed up in January 2026 had 950 active by month 6, started at 1,000. Cohort retention at month 6: 95%. Cohort churn over 6 months: 5%.
When to use it: Whenever you need real insight, not just a number for a slide. Cohort analysis reveals patterns that blended numbers hide entirely.
What it shows you:
- Whether retention is improving or declining over time (compare recent cohorts to old ones)
- When in the customer lifecycle churn happens (first 30 days vs month 12)
- Which acquisition channels produce stickier customers
- Whether changes in your product/pricing actually moved retention
The cost: More work to set up. Requires you to track each cohort separately. Tools like Amplitude, Mixpanel, ChartMogul, and Baremetrics can do this automatically.
Full guide: how to build a cohort analysis chart.
Method 4: Revenue churn (and Net Revenue Retention)
Gross Revenue Churn formula: (MRR Lost from Cancels + MRR Lost from Downgrades) / Starting MRR
Net Revenue Retention (NRR) formula: (Starting MRR + Expansion - Contraction - Churn) / Starting MRR
Example: Starting MRR $100K. Lost $5K from cancels. Lost $2K from downgrades. Gained $8K from upgrades. Gross revenue churn: 7%. NRR: ($100K - $5K - $2K + $8K) / $100K = 101%.
When to use it: Always, alongside customer churn. Revenue churn tells you the actual money story. NRR is the single best metric for SaaS health, and what board investors increasingly care about most.
Why it matters: A SaaS company can have 5% customer churn and 105% NRR (existing customers grow faster than churn) or 5% customer churn and 92% NRR (existing customers shrink despite same churn rate). Same logo churn, completely different business trajectories.
Full guide: net revenue retention.
Which one should you actually use?
Use multiple, for different purposes:
| For this purpose | Use this method |
|---|---|
| Quick monthly check-in | Average churn rate |
| Board reporting | NRR + cohort retention curve |
| Diagnosing problems | Cohort analysis |
| Tracking intervention impact | Cohort comparison (before/after) |
| External benchmarking | Annual customer churn (most comparable) |
Common calculation mistakes
- Using simple churn during growth. Overstates your problem when you're acquiring fast.
- Mixing voluntary and involuntary. They're different problems. Always track separately.
- Reporting customer churn without revenue churn. Hides the real impact when enterprise customers leave.
- Annualizing monthly churn linearly. 5% monthly is not 60% annual. Use the formula: 1 - (1 - 0.05)^12 = 46%.
- Comparing to outdated benchmarks. The "5% average" is from 2010 data. See our guide on real 2026 churn averages.
Most importantly: stop trying to find one number that summarizes your retention. The interesting story is always in the segments, cohorts, and trends. The Health Check diagnoses where your blended numbers might be hiding real problems.
Tools to make this easier
Calculating these manually each month is painful. Use:
- For revenue metrics: ChartMogul or Baremetrics (Stripe-native, automatic)
- For cohort retention: Amplitude or Mixpanel
- For quick calculations: Our churn rate calculator
- For benchmarking: Our churn rate benchmarks by industry and size
- For diagnosing your full retention setup: The 60-second Health Check