Customer churn rate: formula, worked example and calculator

What the churn rate is

The churn rate answers a simple question: of the customers you had at the start, how many are gone at the end? It always refers to a period, which should be stated with the number: 4.5% churn per quarter and 4.5% per month describe very different businesses.

It matters because the customer base only grows when new customers outnumber those lost: a company losing 54 customers a quarter has to win at least 54 just to stand still. That is why churn is best read alongside the sales pipeline: the pipeline tells you how many customers come in, churn how many go out.

When a customer is lost: subscriptions and repeat sales

With a subscription, a lost customer is one who cancels or does not renew. In a B2B company that sells repeatedly, a distributor or a services firm, usually nobody cancels anything: the customer just stops ordering. You need a rule for how long without an order makes a customer lost, chosen from the real buying cycle.

• Measure the typical interval between two orders for each customer over the last one or two years: in distribution it may be three or four weeks, in a maintenance service six months.
• Set the threshold well beyond that interval: two to three times the typical interval is a common starting point. With a mixed customer base, a threshold per segment or per customer works better.
• Check the threshold against returns: if many customers marked as lost order again, the threshold is too short.
• Define an active customer the same way, for example at least one order in the last twelve months: a lost customer is an active one who crosses the threshold during the period.

With a six-month threshold, a customer lost this quarter placed their last order six months ago: in repeat sales churn is a lagging indicator, which is why the early signals matter.

The churn rate and retention rate formulas

The denominator is the customers active at the start, and only those: customers won during the period stay out of the calculation, so the rate measures how much of the existing base you keep. Net growth combines arrivals and departures and belongs next to churn, not in its place, because strong acquisition can hide rising churn.

Monthly and annual churn: compound conversion

Churn is not converted by multiplying or dividing by the months. Each month you lose a share of the customers who remain, not of those you started with, so the rates compound like interest, and the conversion goes through retention:

Multiplying by twelve overstates annual churn, and the more so the higher the rate: 10% monthly churn does not become 120% a year, but 71.8% (1 − 0.9^12).

Example: one quarter at a B2B distributor

A wholesale distributor, whose customers order every three to four weeks on average, treats a customer as lost after six months without an order. At the start of the quarter it has 1,200 active customers; over the three months 54 cross the threshold and 70 place their first order. The figures are illustrative.

The shortcut would give 4.5% × 4 = 18.0% a year, more than a point above the correct 16.8%. Net growth of 1.3% says the base is growing, but slowly: nearly eight in ten new customers only replace those lost.

Customer churn and revenue churn

The churn rate counts customers, and each weighs the same: if mostly small ones leave it overstates the loss, if a few large ones leave it hides it. That is why revenue churn is measured too, and in repeat sales it includes customers who stay but buy less.

In the example, in thousands of euros: the 1,200 customers had bought 3,000 in the previous quarter, the 54 lost accounted for 60, and among those who stayed some bought 150 less and others 90 more. Gross revenue churn is (60 + 150) ÷ 3,000 = 7.0%, above the 4.5% customer churn, and net revenue retention (3,000 − 60 − 150 + 90) ÷ 3,000 = 96.0%. It is a common picture in B2B: more revenue is lost to customers who shrink than to customers who leave. If sales are seasonal, compare the quarter with the same quarter of the previous year.

Average customer lifetime and customer value

With a constant monthly churn of 1.52%, the average customer lifetime is 1 ÷ 0.01523 = 65.7 months, roughly five and a half years; with an average margin of 300 euros a month, a customer's value over time is about 19,700 euros. It is a rough estimate that does not discount future value and starts from contribution margin rather than sales, but it helps compare scenarios: bringing monthly churn down to 1.2% extends the average lifetime to 83.3 months.

Reading churn by cohort

A single rate mixes customers with ten years of history and customers who arrived last month, who usually churn more in their first months. A cohort analysis groups customers by the quarter they were won and follows how many are still active after 3, 6, 12 and 24 months: it helps show whether customers from a campaign stay less than others, or whether a price list change weighed on those who came after it. The same works by channel, region or sales agent.

Common mistakes

• Putting new customers in the denominator: 54 ÷ (1,200 + 70) gives 4.25% instead of 4.5%, and the more the company acquires, the lower churn looks.
• Netting lost against new: (54 − 70) ÷ 1,200 gives a negative churn rate, which is just net growth with the sign flipped.
• Mixing periods: comparing monthly churn with quarterly churn, or converting by multiplying by the months.
• Ignoring seasonality: with a fixed threshold, customers who order only in certain seasons look lost out of season and then come back.
• Changing the definition of a lost customer: shortening the threshold from six to four months raises churn without any customer behaving differently.

Early signals and a churn risk score

Churn measures what has already happened. To act earlier, look at signals that typically come before a customer leaves, compared with each customer's own history rather than the average:

• Less frequent orders: the gap between one order and the next grows.
• A shrinking basket: fewer lines or less value per order.
• Complaints and support requests on the rise.
• Growing payment delays, read with accounts receivable aging.

Combined into a score, they help decide whom to call first; they are not a certain forecast, and the weights are tuned on the customers who actually left in the past. In Muvia a Python function works out every Monday each customer's risk, comparing recent order frequency and spend in the management system with that customer's last year; each sales agent receives the list of whom to call in order of value, and an alarm flags a customer who moves into the high band. This is the use case B2B customers at risk of churning; to bring together the orders, support and payments of each customer, see customer 360.

Calculator

Enter the customers at the start of the period, those lost, those gained and the length of the period: the calculator returns churn and retention, monthly and annual churn, net growth and average customer lifetime. The starting values are those of the example.