Standard retention formulas feel too basic for what I need. Looking at users who return vs users who actually engage meaningfully.
Anyone built retention metrics that factor in usage intensity? Not just did they open the app, but how much they actually used it.
Just multiply retention by average session actions. Simple but catches the difference between real users and zombies.
We track session depth and frequency together. Someone opening the app 5 times for 30 seconds each gets a lower retention score than someone opening twice for 10 minutes.
Our formula weighs time spent plus key actions. A user completing 3 important tasks in one session scores higher than someone just scrolling.
We only count someone as “retained” if they hit our minimum intensity threshold - usually 2+ minutes of active time or completing at least one core flow.
This killed our vanity metrics but gave us way better insights into who actually sticks around.
Yeah, cumulative value is the way to go. It separates passive users from the ones actually doing stuff.
When you flag people whose meaningful actions are dropping, you’re basically predicting churn before it happens. Way better than waiting until they’re completely gone. Focus on metrics that actually matter.
I just count active minutes per week instead of opens.
Don’t just track opens - look at how deep people engage each session. I focus on users who do my core action multiple times per visit. I bucket them: light (1-2 actions), medium (3-7), heavy (8+). Then I calculate retention for each bucket separately. Heavy users coming back? That’s real retention. Light users returning daily but barely doing anything? They’re slowly churning. This shows you which behaviors actually predict stickiness.