Cohort Retention model

What is cohort retention?

Cohort retention analysis groups customers by the month they made their first purchase and tracks what percentage of each group is still buying at each subsequent month.

This reveals patterns invisible in aggregate metrics — for example, customers acquired during a sale event may have high first-month numbers but terrible 3-month retention compared to customers acquired organically.

Reading retention curves

In Explore → Cohort Retention, each row represents a cohort (e.g., "Jan 2024 cohort") and each column shows the percentage still active at that time interval (Month 1, Month 2, Month 3…).

A healthy store typically shows:

• Month 1: 100% (everyone makes at least one purchase by definition)
• Month 2: 25–40% repeat purchase rate
• Month 6: 15–25% still active
• Month 12: 10–20% still active (your loyal core)

What to look for

• Cohorts with unusually high early retention — What was different about how those customers were acquired? Replicate it.
• Cohorts with a sharp drop at a specific month — Was there a product issue, price change, or service failure at that time?
• Recent cohorts vs. older ones — Is retention improving or declining over time?