Cohort and retention model

A cohort model groups customers by the month they first ordered and follows each group over time. It is the only reliable way to see whether the business is improving or simply getting bigger.

Building it

Rows are acquisition months, columns are months since acquisition, and each cell holds the share of that cohort still ordering, or the revenue they produced. Twelve rows and twelve columns is enough to see the shape.

Reading it

• Down a column: whether newer cohorts behave better than older ones
• Across a row: how quickly a group stops ordering
• The flattening point: where the curve stops falling, which defines the loyal core

A curve that flattens at a decent level means the business has a real customer base. A curve that reaches zero means the company buys every order it gets.

What it changes

Retention determines how much can be spent to acquire a customer. A shop whose customers order four times over two years can pay far more for the first order than one whose customers never return, even with identical margins.

A common mistake

Averaging all customers together. A single blended retention figure mixes a loyal core with one-off discount buyers, and the average describes neither.