Customer Retention and Churn

What are retention and churn?

Definitions vary by business model:

Always state the time window and event that counts as churn.

Example

Adding new customers to the starting cohort would hide the churn.

Cohort retention table (customers grouped by signup month):

Reading down a column shows whether newer cohorts retain better, which may indicate that product improvements are working.

Segment analysis

Compare retention by:

• Signup month (cohort)
• Plan type (monthly vs annual)
• Acquisition channel (ads vs referrals)
• Tenure
• Usage pattern (heavy vs light users)
• Region or device

Segment differences show where to investigate. Small groups and different customer mixes can make comparisons unstable.

How it works

1. Define the cohort (who's included)
2. Define the qualifying activity (what "active" means)
3. Define the observation window
4. Build a customer-level dataset (one row per customer)
5. Verify each customer is counted once in the relevant group

Common mistakes

A business can lose a few high-value customers and see a large revenue impact while customer churn looks low