Churn Rate Calculator
Customer churn and retention
Use starting customers, ending customers, and new customers acquired to estimate churn rate, retention rate, and customer losses for a period.
Churn rate calculator for customer churn and retention
This churn rate calculator helps you estimate how many customers you are losing over a time period and what that loss means as a percentage of your starting customer base. If you track customers (or subscribers) month to month, churn is one of the fastest ways to see whether you are leaking demand faster than you are filling it. It is also a useful reality check for growth. You can add customers and still have a churn problem if losses are quietly large.
The calculator uses three numbers most teams can get from basic reporting: customers at the start of the period, customers at the end of the period, and new customers acquired during the period. From those, it estimates customer losses and turns that into a churn rate. It also shows retention rate (the inverse of churn), net change, and a quick view of whether you grew or shrank during the period.
If you know the period length (in months), you can also estimate an average monthly churn rate equivalent. This is useful when you calculate churn over multi-month windows (for example, a quarter) but still want a monthly number you can compare to other months or benchmarks. If you leave the months blank, the calculator assumes a 1-month period and still produces a valid churn estimate.
Assumptions and how to use this calculator
- This calculator estimates customer churn, not revenue churn. If your customers upgrade or downgrade plans, customer churn can look stable while revenue changes materially.
- Customer losses are estimated as: losses = start − end + new. This isolates customers who must have left to explain the ending total.
- If the calculation implies negative losses (for example, end is higher than start + new), the calculator treats churn as 0 and flags it as “net growth beyond new customers,” which can happen with reactivations, data timing, or counting changes.
- The churn rate is calculated against the starting customer count for the period. This is the most common “period churn” definition for customer counts.
- If you enter a period length in months greater than 1, the “average monthly churn” is an equivalent rate that would compound to the same total churn over that period.
Common questions
What is churn rate, in plain terms?
Churn rate is the percentage of your starting customers who stop being customers during a time period. If you start the month with 1,000 customers and lose 50 of them (after accounting for new signups), your churn is 5% for that month.
Why do I need “new customers” to calculate churn?
Because ending customers alone cannot tell you how many left. If you start with 1,000 and end with 1,000, churn could be 0, or you could have lost 200 and gained 200. The “new customers acquired” input lets the calculator isolate the customers who must have left to reach your ending total.
What if I do not know new customers acquired exactly?
Use your best estimate. If you have signups but not net new customers, try to use first-time paying customers or first-time active customers, depending on how you define your customer base. The result will be more accurate when your “new” input matches the same definition used for start and end counts.
How should I interpret retention rate?
Retention rate is 100% minus churn rate for the period. If churn is 4%, retention is 96%. It is a simple way to communicate stability: higher retention means customers are sticking around. For subscription businesses, small improvements in retention often have outsized long-term impact.
When does churn rate become misleading?
Churn can mislead when customer definitions change (for example, counting trials vs paid), when there are large reactivations, or when you have a highly seasonal period. It also does not tell you whether your best customers are leaving. If you can, segment churn by cohort, plan tier, or customer type to find where the problem actually is.