How do you define and measure churn?

Churn is an important business metric for all subscription based businesses. Churn is when a customer cancels their subscription and stops using the service. It is important because it means a loss of ongoing revenue.

Subscription SaaS companies (Netflix being a good recent example) rely on having a base of recurring customer revenue that they can keep adding too. This means that as long as new customers are being added and existing customers are not leaving, revenue will grow predictably. However, if the churn rate is higher than the new customer acquisition rate it means revenue overall will fall. Valuations of SaaS companies are based on accurate measurement of both customer acquisition and churn.

Alex Oppenheimer has written a short blog post describing what it is and how to measure it. Different definitions include:

• Customer churn = customers lost in a given period / total customers at the beginning of that period
• Revenue churn = The ARR dollars lost associated with churned customers in a given period / ARR at the beginning of that period
• Non-Renewals = customers lost in a given period / customers who were up for renewal during that period

Often data scientists are asked to build models that can predict churn. Understanding these definitions is a useful starting point prior to doing that. Another important step is working out what you will do with the model predictions! But that's another topic...