Getting started?

Level up on fundamentals with Key Concepts then Create and Analyze Predictions to maximize your retention efforts.

Key concepts

Churn vs. conversion Predictions

There are two main goals of any app--to keep users coming back and to get those users to convert on specific events. Predictions enable you to forecast both, and to proactively engage users before they churn or to encourage more users to convert on a chosen Event. Generally, churn predictions forecast whether users will not behave in some way and conversion predictions forecast whether users will behave in some way.


The results of each Prediction will group users into buckets of High, Medium, and Low Likelihood for churn or conversion based on the Events you have specified. You should use these buckets as Audience criteria for targeted messaging. For example, send users at high likelihood of churn a 20% off coupon while sending your most loyal users a discount referral code.

Create & analyze Predictions


First, you'll want to define what event(s) to associate with the new churn or conversion prediction you're creating. For churn predictions, select whether you want churn to be defined as not performing any events - to forecast the likelihood that users will not perform any events at all; or the following events - to forecast the likelihood that users will not perform specific events that you select. For churn predictions only, you can define the number of consecutive days that the event(s) will not occur in order for a user to churn. You can do this by changing the number in the 'consecutive days' field and can range from 7 to 90 days. For conversion predictions, you may only forecast whether users will perform specific events.

For both churn and conversion predictions, use the AND/OR operators to combine multiple events for more granularly defined forecasting. Predictions does not currently support event attributes. You can only predict whether users will or will not perform an event, no matter what attributes that event has.

Consider using Predictions to:

  • Identify users that are at a high risk of churn, and offer them an incentive, such as access to a new feature, to continue their relationship with your brand and your app

  • Identify users that have a low likelihood of conversion, and offer them an incentive to convert, such as a discount or reward, all without incurring the cost of offering the incentive to your entire user base

  • Discover the key behaviors that are early indicators of user retention; like Facebook's growth team's realization that users who add 7 friends in their first 10 days are more likely to be retained

  • Discover the key user attributes that are related to purchase behaviors; for example, users on iPhone 6 Plus devices and have AT&T as their network carrier have a higher conversion rate than an average user

To illustrate: