Cohort analysis tracks groups of customers according to the period in which they arrived, and compares how they behave over time. Instead of a single average for the whole period, you can see whether customers acquired in March buy repeatedly to the same extent as those from June, and how their value develops after the third, sixth and twelfth month. This reveals things an average hides: that revenue growth rests solely on new customers while old ones are leaving, or that customers from one campaign never came back even though their acquisition cost was the lowest. This is precisely why cohort analysis is a better basis for budget decisions than immediate return. For services with a longer cycle, you need at least a year of history for the cohorts to have anything to show.
See also: Customer retention, KPIs for a marketer, Incrementality.
