We’ve recently released unbounded retention in our retention reports. Along with N-Day retention, we now have two different ways of measuring retention. So, what are the main differences and use cases for both?
As can be seen in the above example, unbounded retention captures the users who come back on a specific time unit (e.g. 1 month) or afterward. Whereas, N-Day retention looks at only the users who come back on that specific time unit.
A key difference between the two is that because unbounded retention looks at the users who come back on a specific time unit or after, this metric will be at least that of N Day retention, and generally come in higher, since it takes into account the users’ visits on subsequent days.
Consequently, note that unbounded retention can change if a user, who has not returned to your app in a month, suddenly returns to the app. In this instance, retention for all the previous days in the report will increase. This is David, in our example above, who uses the app in January and doesn’t come back again until April.
When to use unbounded retention
If your users don’t use your product at a regular cadence, you may prefer to use this metric. This is generally more true of eCommerce platforms where you may not expect the average user to come back on a daily / weekly / monthly basis. Telcos may also prefer to use this metric, since prepaid users generally tend to come back into the app to reload only when their data runs out.
If you don’t have a time-bound KPI to look at user churn / focus on cadence, unbounded retention is also a good metric to use, since it’s the inverse of the overall churn of your user base.
When to use N Day retention
If you have a stricter measure of retention rate, counting only those users who return to your product within a set period of time, you can use the N Day retention metric. Using a definition of users who use the app at least 1x a month, would only include David in our example above, and exclude Lisa and Chitra.
In addition, if you do care about cadence and rely on users coming in each and every time unit, we’d recommend using N Day retention. In general, this would pertain to gaming and social companies where you’d want users to use your product regularly. This also applies to SaaS products like Mixpanel, and in fact, our key metric looks at users who get day-to-day self-serve answers from Mixpanel to inform their decisions.
For shorter term analyses, like user onboarding or measuring effectiveness of acquisition / marketing campaigns, N Day retention might also be more appropriate.