How to Calculate Churn Rate Accurately

• The basic churn formula
• Step 1: Pick a measurement window
• Step 2: Define your denominator carefully
• Step 3: Count cancellations cleanly
• Step 4: Add revenue churn
• Step 5: Choose customer churn vs revenue churn for your context
• Step 6: Handle annual contracts correctly
• Step 7: Validate your numbers against MRR movement
• Common mistakes to avoid
• What to do next

Churn rate is the percentage of customers (or revenue) you lose in a given period. To calculate it correctly, divide the number of customers lost during the period by the number of customers you started the period with, then multiply by 100. The trickier work is deciding which customers count, which period to use, and whether to measure logo churn or revenue churn.

Most SaaS dashboards report churn incorrectly. They mix new signups into the denominator, ignore downgrades, or average monthly numbers in ways that hide seasonal spikes. This guide walks you through the calculation the way a finance team or due diligence analyst would expect to see it.

The basic churn formula

The textbook formula is simple:

Churn rate = (Customers lost during period / Customers at start of period) x 100

If you began January with 500 customers and 25 cancelled by month-end, your monthly customer churn is 5%. Use the Churn Rate Calculator to get an instant number, then read on for the nuances that decide whether that number actually means anything.

The key word in the formula is "lost." A customer is lost when their subscription ends, not when they stop logging in. Inactive paying customers are still revenue. Cancelled customers who finish out a paid term should be counted in the month their access actually ends, not the month they submitted notice.

Step 1: Pick a measurement window

Most SaaS teams report monthly churn. Some report quarterly or annual. The window matters because short windows are noisier and long windows hide trends.

• Monthly churn is the standard for month-to-month SaaS and early-stage companies. It surfaces problems quickly.
• Quarterly churn smooths out monthly noise and is common in board reporting.
• Annual churn is the right window for annual-contract SaaS, where most customers only have a chance to cancel once per year.

A common mistake is multiplying monthly churn by 12 to get annual churn. That overstates the number because it does not compound. A 5% monthly churn compounds to roughly 46% annual churn, not 60%.

Step 2: Define your denominator carefully

The denominator is "customers at the start of the period." It is not the average of start and end, and it is not the end count. New signups during the period do not belong in the denominator because they did not have a chance to churn for the full window.

If you sell per-seat, you can measure seat churn separately, but headline customer churn should treat one logo as one customer.

Step 3: Count cancellations cleanly

A cancellation counts when access actually ends. If a customer paid through March 31 and cancelled on February 10, they churn in March, not February. Counting them in February inflates that month's churn and understates March's.

Downgrades are not customer churn. A customer who moves from $500/month to $100/month is still a customer. Downgrades show up in revenue churn, not logo churn. Pausing accounts that resume within a defined window (often 60-90 days) typically do not count as churn either.

Step 4: Add revenue churn

Customer churn tells you how many logos you lost. Revenue churn tells you how much money walked out the door. The two can move in opposite directions.

Gross revenue churn = (MRR lost from churn + downgrades during period) / MRR at start of period

Net revenue churn subtracts expansion revenue (upgrades, cross-sells, seat additions from existing customers):

Net revenue churn = (Churn MRR + Downgrade MRR - Expansion MRR) / MRR at start of period

A best-in-class SaaS company has negative net revenue churn, meaning expansion from existing customers more than offsets losses. Public SaaS leaders like Snowflake and Datadog regularly report net dollar retention above 120%, the inverse of net revenue churn.

Step 5: Choose customer churn vs revenue churn for your context

Pillar-stage investors will want all three. Report them together rather than picking the most flattering.

Step 6: Handle annual contracts correctly

Annual contracts complicate monthly churn because customers can only churn at renewal. If you simply count cancellations each month, you will see a giant churn spike in months that have heavy renewal cohorts and near-zero churn the rest of the year.

The fix is to measure churn against the at-risk denominator. In any given month, only customers up for renewal could have churned. So:

Renewal-period churn = (Customers who did not renew / Customers up for renewal) x 100

This produces a number that is comparable month over month. You can then translate it to an annualized customer churn rate by taking the average across renewal cohorts.

Step 7: Validate your numbers against MRR movement

A sanity check: your reported churn should reconcile with how MRR moved during the period. If you started with $100,000 MRR, ended with $105,000 MRR, added $12,000 in new business, and reported $3,000 in churn, you implicitly added $4,000 in expansion to make the math work. If that expansion number does not exist in your billing data, your churn is wrong somewhere.

Run this reconciliation every month before you publish numbers to investors. It catches data quality bugs early, especially around mid-cycle plan changes and proration.

Common mistakes to avoid

1. Including new signups in the denominator. This artificially lowers churn during high-growth months.
2. Counting cancellations on the date submitted, not the date access ends. This shifts churn into the wrong month and inflates volatility.
3. Annualizing by multiplying by 12. Compounding makes this wrong; use (1 - monthly retention)^12 or just measure annual churn directly.
4. Ignoring downgrades in revenue churn. A customer cutting their contract in half is a real revenue loss even though they did not cancel.
5. Reporting net revenue churn without gross. Expansion can hide a real retention problem; investors will ask for both.

What to do next

Once your churn calculation is clean, the next step is feeding it into LTV calculations. Lifetime value is roughly ARPU divided by churn rate, so any error in churn flows directly into LTV and any LTV-to-CAC analysis you do.

If your churn number is higher than you expected, segment it. Look at churn by acquisition channel, plan tier, cohort age, and customer size. The aggregate number is usually fine; the segments reveal where you are bleeding.