Frequency Best Practices

Mixpanel’s Frequency report lets you conduct a deeper exploration of your user’s retention.

Retention measures how effective your product or service retains existing customers over a day, week, or month.

You can use the Frequency report to learn how frequently users return to use your product or service.

Using the Frequency Report

To generate Frequency reports, click Analysis, Retention, and select the Frequency analysis icon.


Mixpanel groups unique users in time-incremented buckets when they first complete an action, and then groups those same users in subsequent buckets when they return and perform the same or different actions.

The buckets measure the number of unique hours in a day a user was active. There needs to be at least 60 minutes in between events for the user to count in more than one bucket.  Each time they qualify for the next bucket, 60 more minutes have to elapse before they make it into the third, etc.

In the image below, you can see how often customers return and open an app during a day.

On Jan 3, 2020, 9,842 users performed the event “App Open”, and 3,125 fired the event again over a 2 hr period. 1,566 users returned 3 hrs later, 1,009 returned 4 hrs later, and so on.


Ways to Use Frequency Reports

These use cases can help you explore different ways to leverage the functionality of a Frequency report. ––––These use cases cover how to use Frequency reports to:

  • Maximize your analysis of active users.
  • Build a more engaged user base.
  • Define better ways to monetize pay to play games.

Many app developers, particularly social apps developers, want to calculate the amount of monthly active users (MAU) and daily active users (DAU). However, what does an active user actually mean? For example, is an Amazon active user the same as an active Tinder user?

Instead of counting any user who performs an action once as an active user, you can generate a Frequency report that tracks user activity in a specific timeframe. As a result you can analyze any seasonal changes in your users’ frequency levels, and get a more actionable measurement.

The image below is a Frequency report. It shows that while the number of DAUs didn’t grow too much between December and January, the number of App Opens increased, which is more actionable information.


Build Engaged User Bases

Some apps require a base of highly engaged users, such as dating apps. Dating apps need an engaged base of potential matches to satisfy its users. Mixpanel’s Frequency report can tell you which acquisition source generates your most addicted users to grow your user base. The acquisition source indicates the source of user traffic, such as Twitter or Reddit.

To learn which acquisition source provides users who send the most messages, you could build a Frequency report with the event “Active User” and breakdown by the “App Version” property.

In this case, you can compare the segments, and see that users who have "App Version" 3 are more engaged than those who have "App Version" undefined, 2, and 1.


Optimize Pay-to-Play Games

Suppose there is a medieval-themed mobile game that allows players to create knight characters for free. Your product monetization team suggests introducing an in-app purchase for characters after several free credits are expended. If character creation greatly improves the game, you could estimate the amount of revenue you are going to make. Not only that, but you could also monitor changes in the frequency after the change went live. 

In the image below, you can see that more than one third of the users create multiple characters in a day, which indicates potential for monetization.


Here’s another example to test monetization strategy: suppose you own a game and users have complained about having to watch ads before they play your game. To test to see if friend invitations generate more addicted players, you could invite three friends to play in exchange for an ad-free experience.

You could determine whether to offer this option if you knew that players who were invited were more likely to be addicted than players who were not. The frequency chart below, segmented by an "Invited User" property (true if they are invited; false if they are not) indicates that they are more likely to play the game multiple times day than a user who is not invited.


Did this answer your question?



Please sign in to leave a comment.