Skip to main content
A cohort is a group of unique users who meet a set of conditions — or who were saved from a chart, or uploaded as a list. Once saved, a cohort is reusable across Klaritics:
Cohorts solve a common problem: you discover an important group of users mid-analysis — people who added to cart but never purchased, or who retained on Day 7 — and then have to rebuild that logic every time you need it. A cohort captures the group once.
Membership is always evaluated per unique user_id. A user who performs an event many times is still one member.

​
Three ways to create a cohort

Behavioral

Define rules from events, properties, sequences, and time. Stores the rules.

From a chart

Save the users behind a funnel step, retention cell, or chart segment. Stores the users.

CSV upload

Upload a list of user IDs. Stores the users.
The key distinction: a behavioral cohort turns conditions into user IDs, and is re-evaluated against its rules. Chart and CSV cohorts store a fixed list of user IDs, with no rules behind them.

​
Build a behavioral cohort

A behavioral cohort is built from event conditions, user-property conditions, groups of conditions, logical operators, and time conditions.
1

Open the builder

Go to Cohorts → Create Cohort and choose Add Event.
2

Add event conditions

Each condition is an event, optional event properties, and a time condition. For example: purchase_completed where category = Electronics and price ≥ 5000, in the last 30 days.
3

Choose how conditions combine

Pick the group’s operator: OR, AND, or AND THEN. Add more groups if needed.
4

Add user-property conditions

Optionally add a property group, such as plan = Premium AND country = India.
5

Check the count

Use Show Count to see how many users qualify, as a trend over the last 30, 60, or 90 days.
6

Name and save

Name the cohort and save it. It’s then available across your analyses.

​
Operators

A group uses one operator throughout. A AND B AND C is valid; A AND B OR C and A AND B AND THEN C are not. The builder prevents mixing them.

​
Groups

You can add several event groups, each with its own operator, and join them with a single AND or OR:
User properties go in a separate property group, which supports AND and OR but not AND THEN — properties have no order. Event groups and property groups combine at the top level.

​
Property operators

​
Time conditions

Each condition has its own time setting, and different conditions can use different ranges — for example, event A in the last 7 days AND event B between Jan 1 and Jan 31.
Relative ranges are calendar-based and include today. If today is Oct 7, “Last 5 days” means Oct 3 – Oct 7, not the last 120 hours.

​
Sequences with AND THEN

  • Events must happen in order, but other events in between are allowed: a user who does A → C → D → B still qualifies for A AND THEN B.
  • Each step can have an optional time bound, measured from the step immediately before it:
  • Time bounds can be set in minutes, hours, days, weeks, or months.
  • DIDN’T works the same way, and its time bound is also optional — A AND THEN DIDN'T B is valid.

​
Dynamic and static cohorts

A cohort mixing fixed and relative ranges is dynamic. A dynamic cohort whose range has a definite end date stops adding users once that date passes. Membership in a dynamic cohort can change repeatedly. A user who adds to cart on June 1 and purchases on June 2 leaves “Cart, didn’t purchase” — and re-qualifies if they add to cart again on June 5 without purchasing.

​
Naming and saving

Every cohort needs a name. Defaults are Cohort 001, Cohort 002, and so on, and you can change them. On save, Klaritics validates the configuration and name, stores the definition, calculates the current members, and makes the cohort available across your analyses.

​
Save a cohort from a chart

Inside a supported chart, use Save as Cohort to save the users behind a specific population:
A chart cohort stores user IDs only — not the chart’s query, filters, or date range, and it doesn’t update later. If the chart is filtered to the last 7 days, India, and Android, the cohort holds exactly the users in that filtered view. Saving “Purchase Conversion Users” from a funnel does not save the funnel’s steps.

​
Upload a CSV

Go to Cohorts → Create Cohort → Upload CSV.

​
Use cohorts in analyses

  • Apply a saved cohort as a filter in Chart Analysis, Funnels, Retention, and other supported charts.
  • One cohort can be used in many charts.
  • A chart can include several cohorts side by side for comparison — for example, High Intent Users vs. Indian Users on the same funnel.
The chart’s date range always wins. Applying a cohort never overrides it — both are evaluated together. A cohort defined over the last 30 days, applied to a funnel showing the last 7 days, contributes its members within the funnel’s 7-day window.

​
Manage cohorts

The cohort list shows Name, Type, User Count, Created By, Created At, and Last Updated. You can search by name, filter by type (Behavioral, CSV Upload, Funnel, Retention, Chart Analysis), and sort by created date, last updated, or user count. Duplicate to build on an existing cohort. For example, duplicate “High Intent Users” and add country = India to make “High Intent Users: India”. This is how you extend a cohort — cohorts can’t be nested inside each other.
A cohort in use can’t be deleted. You’ll see: “Cannot delete cohort. This cohort is currently being used by other analytics. Remove the cohort from those charts before deleting it.”

​
Permissions

Anyone with access can view, apply, analyze, and export cohorts. Editing or duplicating a cohort’s definition requires editor permission.

​
If an event stops reporting

A behavioral cohort that uses an event which stops arriving stays intact and usable. Its count drops to 0 if nobody qualifies, but the condition is never silently removed.

​
Examples

​
Next steps

Sync cohorts to other tools

Send cohorts to MoEngage or any webhook, once or on a schedule.

Retention

Measure how a cohort returns over time.