The Cohort tool measures your users' behavioral patterns by organizing them into related groups and tracking them against a target action over time. In Indicative, a cohort is created by first identifying a set of users who performed an Event or series of Events. Then, the cohort is segmented by time and/or properties.
Cohort are most commonly used for:
- Identifying trends in the user life cycle.
- Understanding user engagement over time
- Understanding user retention and churn over time
- Grouping users to identify behavioral trends
- Calculating Lifetime Value by identifying repeated user performance
- Understanding long-term results of A/B test groups
Begin an Analysis
The Cohort tool is designed to track a user group's completion of a target behavior over time. To help illustrate this, the query builder in Cohort is broken into three parts.
- Define a Cohort is the group of users who will be included in the cohort analysis. They are created by selecting an event or series of events which a user must have completed to be included in the user group.
- Select a Behavior is a single target action that members of the selected user group should complete.
- Select a Breakout is where you select how your cohort analysis will be broken out. You may use event properties, user properties, or time generations.
To get started, follow the prompts in the query builder to fill each of these three sections.
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 (in the case of the image above, clicking an email link). These groups are then charted and compared against each other, exposing differences in behavior over time.
Time generations appear to the right of the target behavior in the query builder. However, they are based upon completion of the initial action, rather than completion of the target behavior.
For example, if a cohort is defined as those who completed the event "Email Link Clicked", the target action is "Free Trial Start", and the time generation is "Day of", then users are broken out into cohort based on the day that they completed "Email Link Clicked". These groups may then be compared against each other for completion of the target action, "Free Trial Start".
"By" Clause in Cohort
In Cohort, "By" clauses are applied differently than in other tools. When building a cohort analysis, the "By" clause selected for "Segmented by" applies across all query rows.