How can I analyze users who did NOT do an event?

We commonly think about the actions that users perform, although patterns in the actions that users do not perform can also surface meaningful insights. Here are a couple ways that you can investigate which users do not perform particular events:

Explore

The behavioral filters in the Explore report allow you to filter for people profiles of users who have not performed an event. Simply select the Event header of the filter menu and select the event in question and apply the “was not performed” option:

Explore was not performed

This will return a list of users who have people profiles who did not perform a particular event within a date range.

Cohorts

The Cohorts builder allows you to define and save definitions of custom user groups that you can use when filtering for users in other reports. You can create a cohort made up of users who did not do an event by selecting the event in your filter criteria, and selecting the “did not” option above the event.

Cohorts Did Not

Alternatively, when using a cohort in your reports, you can toggle between analyzing users who are “in” or “not in” the cohort.

Cohorts Users Not In

Note: Cohorts is available on Enterprise plans. Visit the pricing page for more details about Mixpanel’s billing plans.

JQL

For advanced questions that involve directly querying raw data, Mixpanel’s Javascript-based query language (JQL) can also be used to return users who have not performed an event. Standard Javascript comparison operators (!=, !==, etc.) can be applied, in filter() and reduce() transformations to return only users who do not satisfy a certain set of criteria.

To get started using JQL, check out the JQL developer guide.

When in doubt, consider the events that users do

Often questions that involve investigating the actions that users do not perform can be also researched by examining the actions that users do perform. For example, when investigating the behavior of users who do not log in, traits of users who do log in can lend some insight into what encourages a user to log in, and highlight what the users who do not log in may be missing.