A cohort is a subset of objects that share some common characteristic. In data analytics (and more specifically, in our case) we will be referring to users who are grouped by shared characteristics. Usually, that characteristic is a certain type of date, for example, an acquisition date, and the first login date.

Users that were acquired on the same day are in the same cohort.

A cohort analysis allows us to compare the behavior and metrics of cohorts over time.

In the following cohort example, we can see the retention of users over 5 days (numbers from 1-5). Each cohort is determined by Time. Users from the first cohort are users that registered on the system on January 30th, and so on: