Churn and retention

Churn means customers leaving. In a subscription business, it is the customers who cancel or do not renew. In an app or a shop, it is customers who stop buying or stop using the product. The churn rate is the number lost in a period divided by the number you had at the start of that period.

If a company begins the month with 2,000 customers and 80 leave, monthly churn is 4%. The same number can be measured over a quarter or a year. Always check the period, because 4% a month is very different from 4% a year. A 4% monthly churn compounds to roughly 39% lost in a year, since the same loss repeats on a shrinking base.

Retention is the flip side: the share of customers who stay. If 80 of 2,000 customers leave, 1,920 remain, and retention is 96%. Companies also talk about a retention curve, which tracks a group of customers who joined at the same time and shows what share is still active after one month, three months, a year and so on.

The shape of this curve is revealing. A curve that falls fast at first and then flattens is a good sign, because it means a core group of customers has found lasting value. A curve that keeps falling towards zero means the product is not sticky. Many consumer apps lose most of their users in the first weeks, so the question is whether the survivors stay.

Customer churn counts heads. Revenue churn counts money. They can be quite different. If a company loses ten small customers paying ₹1,000 a month each and keeps all its large ones, customer churn looks high but revenue churn is small. If it loses one large customer paying ₹10 lakh a month, customer churn is tiny and revenue churn is heavy.

Business-to-business companies therefore watch revenue churn closely, and consumer companies watch customer churn and activity. Revenue churn also has a net version that adds back the extra money from customers who upgrade. When upgrades exceed losses, the result is negative churn, which the next guide, on net revenue retention, explains.

It depends heavily on the customer. Consumer subscriptions, where users can cancel in one tap, often have monthly churn of several percent. Business software sold to small firms is often in a similar range of a few percent a month, while software sold to large firms on annual contracts may lose only a few percent of customers in a whole year.

So compare a company with others that sell to the same kind of customer. As a rough guide, many investors like to see annual revenue churn in the single digits for enterprise software and are more forgiving of higher numbers for small-business or consumer products. These are rules of thumb, not standards, and they shift with the market.

• Consumer apps: expect high early drop-off; look for a flat tail
• Small-business software: a few percent a month is common
• Enterprise software: annual churn in single digits is often seen as strong
• Always compare the same period and the same customer type

Churn decides how long a customer stays, and therefore how much they are worth. In the lifetime value formula, churn sits in the denominator: halving churn doubles the expected life of a customer and doubles LTV. A small improvement in retention has a big effect on the value of the whole business.

It also acts as a brake on growth. A company that loses 3% of customers each month must win more than that just to stand still. With high churn, a large part of the marketing money goes towards replacing lost customers, not growing. This is why investors often look at retention before they look at growth.

The best retention comes from the product itself: it solves a real and repeating need, so customers have a reason to come back. Beyond that, companies improve the first experience so new users reach the useful moment fast (often called activation), send timely reminders without spamming, fix the causes of complaints, and talk to the customers who cancel to learn why.

Indian startups face some particular issues. Many users have low-storage phones and uninstall apps quickly, payment failures can lead to involuntary churn when a recurring charge fails, and customers in smaller cities may need support in local languages. Fixing payment retries or adding language support can reduce churn more than a new feature would. A customer who leaves is not always lost, so some companies run win-back offers, but these work only if the original problem is fixed.

An overall churn rate mixes old and new customers, and new customers usually leave faster. A cohort view separates them by joining month. If recent cohorts keep more customers than older ones did, the product is improving even if the overall number looks flat.

This is why careful investors ask for cohort tables. They show whether retention is getting better, and whether growth is built on customers who stay. The guide on cohort analysis explains how to read one.