Daily active users is a lazy proxy
Daily active users is easy to plot and easy to love. It rises when a notification is loud, when a bug forces a retry, when staff log in to “just check.” It is a count of attendance. Attendance is not health, any more than footfall is the quality of a library.
The laziness is not in the arithmetic. It is in the sentence we allow DAU to finish: “the product is doing well.” A banking app can have a calm, infrequent DAU and be doing its job. A puzzle game can have a frantic DAU and be exhausting the people it claims to delight. App Analytics that stops at the headline has chosen not to look at purpose.
In the studio we ask a ruder question: active at what? Opened the app? Completed the job the app exists to do? Paid? Invited a colleague? Each answer produces a different curve. The honest organisation publishes the one that matches the promise on the homepage, and keeps DAU in an appendix for the people who still want a pulse.
There is a second laziness: mixing beings. Support accounts, demo logins, and the founder’s mother will inflate a small product faster than any growth loop. We have watched a United Kingdom education company discover that a third of “students” were teachers in staff mode. Their DAU did not fall when they excluded those accounts. Their self-respect did, briefly, and then the remaining chart became usable.
If you must keep DAU on a wall, pair it with a kernel action and a note on who is excluded. If the pair ever diverge, believe the kernel. The wall can wait.