Tracking Retention in Your Windows 10 Apps. Part 1.
Windows Dev Center gives you a pretty good picture of daily acquisitions of your apps, channels users come from, and their demographics. HockeyApp and Application Insights can give more details on the daily usage of your app. Unfortunately, as far as I know, no service available to Windows 10 (UWP) developers can help you track retention of your users.
(Note: I could be wrong and I’d be happy to be wrong about this. Let me know if you are aware of a UWP-compatible service to track this)
In the part 2 I’ll try to lay out an easy, while most likely not the most efficient, approach to DIY retention tracking for UWP apps. But first…
Why would you want to track retention?
Unless you are selling a paid app with no trial, the total number of downloads is a typical vanity metric. It gives you a number to look at, but it doesn't tell you anything about how your app is used and liked (or not) by the people who download it. Understanding how many of your users continue to use your app after a certain number of days could help you evaluate value the app provides, predict your future user base, LCV, make improvements, track the effect of those improvements, evaluate quality and ROI of different user acquisition channels, and much more.
Different types of retention
There’s no standard definition for retention tracking. The way you track it mostly depends on your goals. You can find a good breakdown of different methodologies here. We track two types of daily retention for our app AppRaisin:
- Rolling retention. This is defined as percentage of users who installed the app on Date-X and used it Y or more days later. We use it to understand the overall use base and churn (number of users who stopped using the app completely). It also gives you a good benchmark to compare to the overall industry since this is the method Flurry uses and publishes on their blog.
- Classic (or strict) retention. Percentage of users who installed the app on Date-X and used it exactly Y days later. These numbers help us predict the number of daily active users in the future (a metric that we consider one of the most important to us). It also reflects the effect of retention-oriented changes we make much more vividly.
Depending on the type of your app you may want to pay more attention to your monthly retention numbers. Or you may care a lot about users coming back every single day.
Some AppRaisin retention numbers and how they help us
All-time rolling retention:
By comparing our Day-28 numbers to the iOS/Android retention matrix provided by Flurry we got some peace of mind from seeing that we do better than average apps in our category on Android and better than any average app on iOS.
Our average Day-28 retention in March was 36% while in May it has grown to 41%. So we know that we are on the right track overall.
All-time classic retention:
As you can see classic retention numbers are much more modest (as they should be).
On June 9th we have launched push notifications for “Hot news” (with more than 100 votes). That moved Day-7 classic retention from 24.8% in the last 10 days of May (right before the effect of the feature should have kicked in) to an average of 30.8% since the feature was launched.
That is almost 20% improvement and we were able to measure it by tracking retention in a “classic” way. Rolling retention is unable to help us with this since the new feature affected Day-7 rolling retention for all the older users as well.
So, how do you track it in UWP Windows 10 apps?
I hope I was able to wet your interest for keeping track of user retention numbers for your apps. Unless someone points me to a service with Universal Windows Platform support, I will post a simple DIY solution for retention tracking in the Part 2. Subscribe (follow) above and stay tuned…
