Churn is the rate at which customers stop using or paying for your product. It is one of the most important numbers in any subscription or recurring revenue business and one of the most commonly misunderstood.
How to think about churn
The basic churn calculation is straightforward: the number of customers who cancelled or stopped using the product in a given period, divided by the number at the start of that period. A business that starts the month with 100 customers and loses 5 has a monthly churn rate of 5%.
The implication of that number is less intuitive. A 5% monthly churn rate means the average customer stays for 20 months. A 3% monthly churn rate means the average customer stays for 33 months. A 2% monthly churn rate means 50 months. Small improvements in churn rate have large effects on customer lifetime and therefore on LTV.
It also matters how churn is defined. Customer churn, meaning the number of accounts that cancel, is different from revenue churn, meaning the revenue lost from cancelled accounts. A business that loses its smallest customers but retains its largest ones might have high customer churn and low or negative revenue churn. Both numbers tell you something different.
The signals that predict churn before it happens
Churn almost always follows a period of declining engagement. A customer who used to log in daily and now logs in weekly is not churned yet, but the trajectory is meaningful. A team that has stopped inviting new members, a user who has not completed the workflow they started three weeks ago, an account that has stopped generating the output that signals active use: these are predictive signals.
For most products, there is a handful of engagement behaviours that are strongly correlated with retention. Identifying those behaviours, the ones that your retained customers consistently do and your churned customers consistently stopped doing, gives you the early warning signals you need to intervene before the cancellation happens.
What to do with an at-risk customer
The most effective interventions for at-risk customers are ones that address the specific drop-off in engagement rather than generic retention communications. A customer who has stopped using a specific feature gets a message about that feature. A customer who has not logged in for ten days gets a check-in from a real person, not an automated email. A customer approaching renewal who has not engaged with a key part of the product gets a session with someone from the team.
The personalisation required for this is not as technically complex as it sounds. Basic engagement scoring, a threshold for what constitutes at-risk behaviour, and a simple response playbook for each category of at-risk customer can be implemented without sophisticated technology.
Churn is predictable. The customers who are about to leave almost always show the signals in advance. The question is whether you are looking.