Journal · 18 February 2026

The quiet cost of vanity DAU

Team working together at a long table

Daily active users is a wide net. It counts the person who opened the app to dismiss a badge, the person who finished the job the product exists to do, and the person whose token refreshed in the background depending on how generously you defined “active.” When that net rises, rooms cheer. When it falls, rooms panic. Neither reaction is analysis.

I sat in a United Kingdom board meeting where DAU had climbed for eleven weeks. Support tickets about “I cannot find my class” had climbed with it. The extra actives were people returning to a broken schedule view, not people training. Vanity is not that the number is large. Vanity is that the number is unexamined.

Identitycachegrid asks a ruder question: active at what? If you cannot finish the sentence without pointing at a vendor default, the row does not belong in a pack. A funded user, a completed session, a published listing — those are narrower, and they move more slowly, which is why they feel less useful to a launch deck.

There is also a cost in attention. Teams that worship DAU start shipping prompts, badges, and emails whose only job is to create an open. The product becomes a reminder machine. Retention Mapping Studio spends a week separating “opened” from “did the thing.” The second curve is usually uglier. It is also the one that predicts whether anyone still pays next quarter.

I do not ban DAU. I put it in an appendix, next to OS mix and crash-free sessions, as context. The lead chart is a cohort of people who reached first value in a given week and returned to do it again. That chart cannot be juiced by a badge without also juicing the underlying work — and if it can, your first-value definition is still a costume.

Return to the journal