“For the first couple of years I was relying on vanity metrics — things like monthly active users. That wasn't really moving the needle. It was only when I moved to event-based analytics that growth really happened.”

Kill vanity metrics — only event analytics let you run the experiments that actually move revenue

MAU and download counts let John feel good while the app barely converted. The switch to Mixpanel event analytics is what enabled the paywall A/B tests that took conversion from 0.5% to 8%. Decision rule: if a metric can't be tied to a specific experiment you could run next week, it's a vanity metric — stop checking it.