Churn Rate
The churn rate indicates what proportion of a company's customers leave within a given period of time. It is one of the most important metrics for companies that rely on recurring payments, such as streaming services or software providers.
Many companies don’t make their money from a one-time sale, but from monthly payments. Netflix, Spotify, or a mobile phone plan work this way. For such companies, one number is especially important: How many customers cancel again? That’s exactly what the churn rate measures. The English word “churn” means something like turnover or agitation. If a service loses 3 out of 100 customers in a month, the churn rate is 3 percent.
Why attrition is more expensive than growth
Acquiring new customers costs money. A company pays for advertising, discounts, and free months. It only recoups this money after several months. If a customer leaves beforehand, the advertising was a losing proposition. That’s why people often say: keeping existing customers is cheaper than acquiring new ones.
A high churn rate also acts like a hole in a bucket. The company can keep pouring in new customers at the top and still not grow. At 5 percent monthly attrition, more than half of the customer base is mathematically gone after a year. Marketing then only has to compensate for the loss instead of generating real growth.
For investors, this metric therefore acts as an early warning system. Rising revenues look good, but say little if attrition is increasing at the same time. A low churn rate, on the other hand, is considered a sign that a product is genuinely needed. For business software, values below one percent per month are considered very good, while for private streaming subscriptions a few percent are normal.
How the figure is calculated
The basic calculation is simple. You take the number of customers who canceled within a period. You divide this by the number of customers at the beginning of the period. Out of 2,000 customers at the start of the month, 60 are lost, so the churn rate is 3 percent.
It gets tricky in the details. Do you count customers or revenue? If a provider loses ten small customers, that’s less severe than losing one large one. That’s why, alongside customer churn rate, there’s also revenue churn rate. Sometimes it even turns out negative: the remaining customers add so much additional business that revenue rises despite the cancellations.
A common misconception is equating churn with dissatisfaction. People also cancel because they move, need to save money, or only needed the product briefly. A language course subscription loses users as soon as the exam has been taken. Such attrition can hardly be prevented through better service.
Churn prediction through AI
Companies don’t want to notice cancellations only after they’ve happened. That’s why they use programs that learn patterns from past data. Such systems observe, for example, that users typically log in less often about three weeks beforehand. The model then estimates a probability for each customer of canceling soon. Anyone deemed at risk receives an offer or a phone call.
These predictions are a classic application field for machine learning in everyday business. They are less spectacular than chatbots, but very widespread. Almost every major mobile carrier and every bank works with such models. Even AI companies themselves now pay close attention to their own churn rate, since many users sign up for a subscription just to try it out.
The term regularly appears in news and quarterly reports. When a streaming service reports “increased churn after the price increase,” it simply means: more people canceled than usual. This single number often determines whether the stock rises or falls that day.