Cohort analysis for ecommerce: a practical setup

Set up the cohort

• Group customers by the month of their first order, not by the first ad click, and write that rule in the report.
• Use one customer key across the store and analytics, such as an account ID or a hashed email, and name which one.
• Use the same new customer definition as your data definitions sheet.
• Build the table: rows are cohorts, columns are months since the first order (month 0, month 1, month 2).
• Each cell holds the share of the cohort that ordered again in that month. Use one measure throughout and label it.
• Show each cohort's size beside its row. A row of 40 customers moves far more than a row of 4,000.

Read the table

• Read down a column to check whether newer cohorts keep pace with older ones at the same age.
• Read across a row to see how one group ages. Mark open cohorts as not yet measured, never as zero.

• Compare at the same age only. Never put one cohort's month 1 beside another cohort's month 6.
• Note what else changed in each cohort's first month: a sale, a price change, a new product or a new campaign type.

What to write in the client report

• The cohort rule, the customer key and the source of both, plus the table for the last six to twelve cohorts with their sizes.
• One line per gap between cohorts, with the event that may explain it and the check planned next.