Retention is one of the most important components of the user lifecycle, and should hold a central spot in any mobile marketer’s dashboard. On top of providing an invaluable indication of the quality of an app, retention metrics have a direct impact on the calculation of customer lifetime value (LTV) for freemium apps.
User retention however suffers from an existential problem: the lack of consensus over a common definition of what it actually means or how it’s computed. Retention benchmarks are already extremely different across app categories, or the type of service provided. If, on top of this, different definitions are used across the board, the metric stops having any meaning at all.
There is unfortunately no single, canonical way to calculate retention: different methods might be relevant for different analytical purposes. It is up to you as an app publisher to figure out which one(s) work out the best for your own goals.
Below is a simple typology with five definitions of retention. Out of these, two are the most widely used (and get mixed up the most). The other three are less common but can still become relevant for certain types of apps or publishers.
Disclaimer: the names given here are far from official, but rather attempt to describe the specific use case as well as possible. If you can think of a better name, please let me know
Here is the general formula to calculate retention across all definitions:
Regarding the legend, “tick” signs mark the days on which users need to open the app at minima in order to be considered as retained for a given definition. “The X” signs are only there for the sake of clarity as they mark days which do not contribute to considering the users retained for a particular definition.
Common Types of User Retention
1. Classic Retention
- How to calculate it: how many users come back to the app on Day+N.
- Where and why to use it: classic retention is the easiest to put together from an analytics point of view and also the most widely used. It gives a good idea of overall retention levels in the app.
- Day 28 example: only users who came back on Day 28 exactly are considered retained. If and how often they came back before that has no impact.
2. Rolling Retention
- How to calculate it: how many users come back to the app on Day+N or any day after that.
- Where and why to use it: rolling retention has long been used by Flurry as their default definition of retention, as it enables publishers to understand the churn rate of their app (users who have not dropped out yet), as it is per definition equal to 100% minus churn.
- Day 28 example: users who came back on Day 28 or any day after that (such as on Day 43) are considered retained.
Additional Types of Retention
3. Full Retention
- How to calculate it: how many users open the app every single day until D+N.
- Where and why to use it: full retention is extremely restrictive and not so widespread as a metric, but it is a good minimum benchmark and gives an idea of the level of engagement in the app.
- Day 28 example: only users who open the app every single day from Day 1 all the way to Day 28 are considered retained.
4. Return Retention
- How to calculate it: which proportion of users come back to the app at least once within N days.
- Where and why to use it: return retention is traditionally used in gambling as it shows how many people failed to drop out after the first open (called retainers), which then also allows to retarget the ones that have dropped out.
- Day 28 example: all users who came back at least once before Day 28 (for ex. on Day 5 or 16) are considered retained.
5. Bracket-Dependent Return Retention
- How to calculate it: it is a specific and restrictive case of return retention, where brackets between typical retention marks such as Day 1/3/7/28/60 are defined. Day M is then designated as the bracket mark just below the target retention mark N (ex: if N is 28, then M is 7). The metric will measure the proportion of users who come back to the app at least once during one of these bracket. Users returning at least once between Day M and N will be considered retained.
- Where and why to use it: This method is useful to get an idea of user behavior and usage patterns in your app depending on the service provided and your own goals.
- Day 28 example: users opening the app on day 10 will be considered retained for the purpose of this metric. However, users coming back on Day 5 but not any time between Day 7 and 28 will not be considered as retained.
Which retention do you use? Do you think the typology should be improved? Let me know in the comments or get in touch directly! This article was originally published on the AppLift blog.
About the author
Thomas heads up content marketing at app marketing platform AppLift. As such he’s in charge of sourcing, curating, creating and distributing insightful content to increase visibility and thought leadership for the company. Thomas loves to scrutinize the relentless and trilling developments of the mobile industry. He can be reached at firstname.lastname@example.org Also see: Applift Blog | Insights | Twitter | Facebook