Cohort Analysis

A method of grouping customers by a shared characteristic — typically the month of their first purchase or signup — and tracking their behavior over time. Cohort analysis reveals patterns that are invisible in aggregate metrics. It shows, for instance, that customers acquired in January repurchase at a different rate than those acquired in June, or that a specific acquisition channel produces higher long-term value. In upsell architecture, cohort analysis is used to identify the prediction window: the timeframe in which a customer is statistically most likely to make a repeat purchase, and thus most receptive to a related offer.

Example: A pet food brand analyzes cohorts to find that customers with medium-sized dogs reorder a 15-pound bag approximately every 28 days, then triggers a reminder email at day 26 with a complementary treat offer.