Monthly active users (MAU) is the count of unique users who engage with a product within a 30-day (or calendar month) window. It's the most commonly cited engagement metric in investor reports and product dashboards because it captures the broadest view of a product's active user base. Spotify reports 675 million MAU as of early 2025. That number includes everyone who opened the app at least once in 30 days, regardless of how deeply they engaged.
MAU matters for business context. For experimentation, it's one of the least useful engagement metrics to put on an individual A/B test.
Why is MAU difficult to use as an experiment metric?
The 30-day window creates two problems for experiments.
Low sensitivity. MAU changes slowly. A feature change that affects daily engagement takes weeks to register in MAU, because MAU counts users across an entire month and a single active day is enough to be counted. A user could have a meaningfully worse experience for 29 days and still count as a monthly active user if they opened the app once. This makes MAU nearly useless for detecting the impact of individual experiments within typical 2-4 week run times.
Long observation windows. To measure MAU properly, you need at least 30 days of post-exposure data per user. Many experiments can't run that long without blocking other tests from using the same traffic. Confidence's analysis framework supports shorter-window metrics that are more practical for experiment-level decisions, while MAU can be tracked as a longitudinal metric across many experiments.
When does MAU belong in an experiment?
MAU works as a long-term guardrail across a program of experiments rather than within a single one. If a team ships ten features in a quarter, each validated by its own success metrics, MAU tracked over the same quarter tells you whether the cumulative effect of those changes grew or shrank the user base. Spotify uses holdout groups and cumulative holdback evaluations to measure exactly this kind of aggregate impact.
MAU can also serve as a secondary metric in experiments that specifically target re-engagement or reactivation. If you're testing a feature designed to bring back lapsed users (users who haven't been active in 14+ days), MAU is one of the few metrics that will capture whether those users came back and stayed within the measurement window. But even in this case, a shorter metric like WAU or 14-day retention is typically more sensitive and more practical.
How does MAU relate to DAU and WAU?
The ratio between these metrics tells you something about engagement depth. A product with 30 million MAU and 10 million DAU has a DAU/MAU ratio of 0.33, meaning roughly a third of monthly users come back on any given day. This ratio, sometimes called "stickiness," is itself a useful product health metric.
In experimentation, the three metrics form a hierarchy of sensitivity. DAU detects changes fastest but is the noisiest. WAU smooths out daily variation while remaining responsive within a typical experiment window. MAU is the most stable but the least sensitive to individual changes. Confidence supports all three as metric definitions in your warehouse, and the right choice depends on the hypothesis and the observation window your experiment can support.