Cohort Overview

The Cohorts tool allows you to understand how often your customers return and engage with your website or product. 

  • Analyze drivers of user retention the factors related to churn
  • Create user cohorts based on repeated behaviors and attributes
  • Identify customers with high lifetime value (LTV)

Every analysis in Cohorts begins with an event. To Define a Cohort you will identify users who complete an event at least once within the established data range. Like in Segmentation or Funnel, you may narrow your search by using a Filter or you may create a combination of events by using +Did [Not] Perform.

Next, you must Select a Breakout. Breakouts may include an Event Property, a User Property, or a User Segment. You may also add a Generation breakout. Examples of generations are: Hour, Day, Week, or Month. For example, if you select Month, then the cohorts will represent all users who performed the first action within the month. 

Finally, you must Select a Target Behavior. A target behavior is an event and it represents the subsequent behavior that you would like to analyze. 

All queries within Indicative a fully customizable using Settings like date range and interval. The Recurring setting shows the retention of users who perform the same target behavior multiple times. The First-Time setting shows the distribution of time it took for users to first perform the target behavior. Results may be displayed as a Non-Cumulative Percent or a Non-Cumulative Count. When conducting a first-time analysis, you may also select Cumulative Percent or Cumulative Count. There are three different chart types: table, comparison chart, and area chart. You may also include “Annotations,” which mark milestone events.

What are Time Generations?


Time Generations are an automatically generated property type available in Cohort. They group users based on the hour, day, week, or month that they completed the prerequisite cohorting action (set in Define a Cohort). These groups are then charted and compared against each other, exposing differences in behavior over time.


For example, if a cohort is defined as those who completed the event Create Profile, the target action is Purchase Product, and the time generation is Hour of, then users are broken out into cohort based on the hour that they completed Create Profile. These groups may then be compared against each other for completion of the target action, Purchase Product.

"By" Clause in Cohort

In Cohort, "By" clauses are applied differently than in other tools. When building a cohort analysis, the "By" clause applies across all query rows.

2 users found this helpful