Customer churn rate:
the share of customers you lose each period

What churn actually counts

Churn rate answers one narrow question: out of the customers you had when a period began, what fraction had stopped buying or cancelled by the time it ended? A subscription app that starts March with 600 paying users and sees 24 cancel has a March churn rate of 4%.

The word 'customer' needs a definition before the number means anything. For a subscription business it is usually an account with an active paid plan. For a shop it might be someone who bought at least once in the last 90 days and then did not return in the next 90. Two companies using different definitions will report numbers that cannot be compared, so write the rule down before you start tracking.

The period matters just as much. Monthly churn suits software and memberships. Annual churn suits contracts, insurance and anything renewed once a year. Quote the period every time; '4% churn' on its own is incomplete.

The churn formula and its mirror image

Take the customers lost during the period, divide by the customers you had at the very start, and multiply by 100. New customers who joined during the period stay out of both numbers, because they were never part of the starting group.

The mirror image is retention: 100% minus churn. With 24 lost out of 600, churn is 4% and retention is 96%. They carry the same information, but teams tend to act on churn because a falling number feels like a target.

Why a small monthly rate is a big yearly loss

Churn repeats every month on whatever base is left, so it compounds. Surviving customers after twelve months are the starting group multiplied by (1 − monthly churn) twelve times. A 4% monthly rate leaves 0.96 to the power 12, which is about 61% of the original group, so around 39% have gone.

The shortcut of multiplying by twelve (4% × 12 = 48%) overstates the loss because later months have fewer customers to lose. Use the compounding version when you convert a monthly rate into a yearly one.

The last column uses the rough rule that average lifetime is one divided by the monthly churn rate. At 4% that is 25 months, a useful figure when you estimate how much revenue one customer brings before leaving.

Customer churn, revenue churn and cohorts

A single blended rate can hide the real story. Three splits are worth making once you have the basics:

• Logo churn counts accounts. Revenue churn weighs each account by what it pays, so losing one large client shows up properly.
• Voluntary churn is a customer choosing to leave; involuntary churn is a failed card payment or expired mandate. The fixes are completely different.
• Cohort churn groups customers by the month they joined, so you can see whether newer customers stay longer than older ones.

Most subscription businesses find churn is highest in the first one to three months and then flattens. A blended 4% might be 9% for new customers and 2% for those past the six-month mark. Improving onboarding then beats discounting for long-standing customers.

Mistakes that distort the figure

• Dividing by the end-of-period count instead of the starting count, which flatters the result when the base is growing.
• Counting a paused account as churned in one report and active in another.
• Mixing periods, such as lost customers from a quarter divided by a starting count from the first month.
• Including trial users in the starting group when your lost count only covers paying users, or the other way round.
• Reading a one-month spike as a trend; seasonal businesses should compare the same month year on year.

Turning the number into decisions

A churn figure only earns its keep when it triggers a question. Start by splitting the 24 lost customers by reason: price, a competitor, a missing feature, a poor first month, or a payment that simply failed. Even a rough tag in your cancellation form gives you a ranked list of problems.

Next, look at timing. If most of the 24 left within their first 30 days, the issue sits in sign-up and onboarding. If they leave near the renewal date, value is not being noticed during the term. If they leave right after a price rise, the pricing change is the likely cause.

Finally, put a rupee value on the loss. Losing 24 customers who each pay ₹500 a month removes ₹12,000 of monthly recurring revenue, which is ₹1,44,000 over a year if none are replaced. Set that against the cost of a retention offer or a better onboarding sequence and the priorities usually become obvious.

What counts as an acceptable rate

There is no universal good number. Consumer subscriptions with monthly billing commonly see churn that would alarm a business selling annual contracts to large firms. Compare yourself with businesses that bill the same way, serve the same kind of customer and charge a similar price.

A more practical test is whether acquisition outpaces loss. If you add 40 customers a month and lose 24, the base is still growing, but the cost of replacing customers is eating into profit. Lowering churn from 4% to 3% on a base of 600 saves six customers a month without spending a rupee on advertising.

A conversion-style tool can help on the front end, since keeping churn low and winning new customers efficiently work together. Use the related calculator to check how many visitors turn into buyers, then compare that with how many leave each month.