DAU, WAU, MAU

Also called: DAU, WAU, MAU, daily active users, weekly active users, monthly active users, active users metric

Daily, weekly and monthly active users: distinct users who did a qualifying action in a day, a 7-day or a 30-day window. What counts as 'active' must be defined.

DAU, WAU and MAU count the distinct users who were active in a day, in 7 days and in 30 days. They are the most quoted product numbers and the easiest to get wrong, because "active" is a choice, not a fact. A usable definition names five things:

The qualifying action: a real sign of value (adding an expense, sending a message), not just any event. App opens, push deliveries and server-created events inflate the count.
The unit: user, account, device or group.
The window: calendar (June 1–30) or rolling (the 30 days ending today). Rolling windows give a smooth daily series; calendar ones are easier to explain in reports.
The time zone in which a "day" starts. Events are usually stored in UTC, so a traveler's evening can fall on the next day.
Exclusions: staff, QA and test accounts, bots.

$\text{MAU}(t) = \left|\ \{\text{users with a qualifying action on days } t-29 \dots t\}\ \right|$

The key property is that these are unique counts, so they don't add up across time: MAU is not the sum of the month's DAU, and WAU is not the sum of seven days. One person active every day adds 30 to the sum of daily counts but only 1 to MAU. The gap between them is exactly what stickiness (DAU/MAU, WAU/MAU) measures.

Which one to headline depends on the product's natural usage rhythm. A messenger lives by DAU; a bill-splitting app used when someone pays for something shared is closer to a weekly product, so WAU is the honest headline and a daily chart mostly shows weekend dips.

Example

Five Halves users over one week; "active" means adding an expense or recording a settle-up yourself:

User	Mon	Tue	Wed	Thu	Fri	Sat	Sun
A	✓	✓	✓	✓	✓
B	✓					✓
C			✓
D	recurring bill created by the server
E						✓	✓
DAU	2	1	2	1	1	2	1

Sum of daily DAU: 2 + 1 + 2 + 1 + 1 + 2 + 1 = 10. Average DAU = 10 / 7 ≈ 1.43.
WAU = distinct active users in the week = {A, B, C, E} = 4, not 10: A alone contributes 5 to the sum.
D has an event but did nothing: counting "users with any event" would give WAU = 5, overstating it by 25%.

At full scale in June Halves reports an average DAU of 6,000, an average WAU of 18,000 and an MAU of 40,000. Adding up 30 daily values would give about 180,000 "users", four and a half times the real monthly audience.

Common mistakes

"Active" = any event. Server-created events, push deliveries and background syncs make people look active who never opened the app.
Adding DAU across days. Unique counts don't sum; compute WAU and MAU directly from distinct users.
Forgetting staff and test accounts. In a young product a handful of QA accounts can be a noticeable share of DAU.
Ignoring time zones. A UTC day and a user's local day can disagree for travelers and late-night users; pick one and state it.
Headlining the wrong grain. DAU for a weekly product turns every weekend into a false alarm.

Learn it in the course

DAU, WAU, MAU and stickiness · Count active users at the rhythm people actually use the product.
Case: a north star for Halves · Pick and defend a north star metric for Halves, with six months of numbers behind it.
Forecasting active users · New users × retention: build a monthly active users forecast from cohorts, with a separate curve for summer cohorts, and price a cut in paid acquisition in users.
Metric definition: one rule, one number · How to write a metric definition so two people get the same number: qualifying event, unit, exclusions, time window and time zone, on a DAU example.
The north star metric and its KPI tree · One north star metric for the value users get, and a KPI tree of input metrics that teams can actually move.