Cohort analysis groups customers by when they shared a starting event, such as the month of their first purchase, and follows each group over time so you can compare them fairly. A cohort is one such group.
Category
Metrics
Also known as
cohort; customer cohort; acquisition cohort; cohort report
It separates "we have more customers" from "our customers stay longer", which a single total can't do. It also shows whether this quarter's new customers behave differently from last quarter's, which a blended average hides.
In practice
Read the first-to-second purchase rate by acquisition cohort, monthly, with the holdout group in the same view. A change in the welcome series shows up as a break between cohorts. A change in the market shows up in all of them at once. A rising count of active customers can hide that each new cohort retains worse than the one before.
Something out of date?