Retaining customers is often cheaper than acquiring new ones. Analysing churn shows where and why customers leave.
Defining Churn
- Contractual businesses: churn when a subscription ends.
- Non-contractual businesses: churn must be inferred, such as no purchase in 90 days.
Choose definitions carefully and apply them consistently.
Key Metrics
- Churn rate and retention rate per period.
- Revenue retention, including expansion and contraction.
- Customer lifetime.
Retention Curves
Plot the percentage of each cohort still active over time. Flattening curves indicate a loyal core.
Finding Drivers
- Compare churned and retained customers' behaviour.
- Survival analysis to account for customers still active.
- Exit surveys and support tickets.
- Predictive models to flag customers at risk.
From Prediction to Action
Knowing who might churn matters only if interventions work. Test retention actions with controlled experiments — and target people likely to be persuaded, not those who'd stay anyway.
Segment
Churn often differs sharply by channel, plan and customer type.