Field note · 12 March 2026
When DAU stops helping
Daily active users is a respectable metric for a chat app. It is a weak metric for a lending app that people correctly open twice a month, and a dishonest metric for a hotel super-app whose “activity” spikes every Songkran regardless of product quality.
We still see DAU as slide one in Bangkok board packs. The chart is easy to pull. It moves a little every week, which feels like management. It also lets a team avoid the harder sentence: are paying users still paying, and did the thing we shipped last month change that?
What DAU hides
DAU mixes new install noise, reinstalls after an OS update, staff accounts, and genuine habitual use. In marketplace apps it also mixes buyers and sellers. A “healthy” DAU line can sit on top of a rotting paid-retention curve. Executives who only watch the daily line find out about the rot in a quarter-end commercial review, which is the wrong room and the wrong week.
There is a second problem. Once DAU is on the first slide, every other number is framed as a descendant of it. Activation gets defined as “became DAU.” Retention gets defined as “still DAU.” The dictionary collapses into a single vanity noun.
What we put instead
For subscription products we ask for a weekly active payer count with a 28-day window, plus month-two paid retention. For marketplaces we split demand and supply actives and refuse a blended DAU. For lending, we look at completed key jobs (application started, disbursement, repayment) rather than app opens.
None of this is exotic. It is simply slower to explain in a meeting that has been trained on a green tile. The Board-Ready studio spends a full module on retiring DAU without starting a religious war. The trick is not to ban the metric; it is to demote it to a diagnostic the analytics team may still watch, while the appendix leads with jobs and money.
If your current pack opens with DAU, try this experiment next Thursday: move it to page five and put the exception log first. See whether the conversation gets shorter or longer. In our cohorts it almost always gets shorter, which is the point of Executive App Reporting Analytics — fewer slides, more decisions.