What counts as an active user?
Three decisions define the number: who counts as a person, what counts as activity, and which 30 days you measure. Every team answers these differently, which is why one company's MAU is rarely comparable to another's.
A person is usually a unique account or device identifier. If the same customer uses a laptop and a phone without logging in, some tools count them twice. Identity resolution decides how accurate the count is.
Activity should mean a meaningful action, not just a page load. A login is a weak signal because it can come from a password manager or a background token refresh. A better signal is the core action your product exists for: sending a message, creating a record, running a report, or viewing a dashboard.
The window is either a calendar month or the trailing 30 days. Calendar months line up with billing and board reporting. Rolling windows give a smoother daily trend line. Pick one and keep it fixed so the series stays consistent over time.
Because each person is counted once, MAU says nothing about how much they did. Someone who opened the app once and someone who lived in it every day both add one to the total. That is why MAU is a measure of reach and is usually read together with user engagement metrics.
Why does MAU matter for SaaS teams?
MAU is the denominator for almost everything else. Adoption rates, feature usage, support load per user, and revenue per active user all depend on it. If the base is wrong, every ratio built on it is wrong too.
For product managers, a flat MAU while signups grow means new users are not sticking. That points to an onboarding or user activation problem rather than a marketing problem.
For customer success leads, MAU inside a single account is an early warning system. When active users in a paying account fall month over month, renewal risk rises long before anyone opens a cancellation form. Tracking this per account is one of the most practical ways to get ahead of churn rate.
For product marketers, MAU shows whether launches reach people. A feature announcement that lifts feature usage without lifting MAU reached only existing active users. One that lifts both pulled dormant users back in.
Boards ask for MAU because it shows whether a product is used, not just sold. Seats purchased is a sales metric. Seats active is a product metric.
Examples of how MAU is defined in practice
The right qualifying action depends on what the product is for. A few generic patterns:
- Team messaging tool: a user counts if they send or read at least one message in the month. Simply having the desktop app running in the background does not count.
- Project management software: a user counts when they open a board, update a task, or comment. A user whose only activity was receiving notification emails does not count.
- Mobile banking app: a user counts on any foreground app session, since checking a balance is the core job. A silent push notification delivered to the device is not activity.
- B2B analytics platform: a user counts when they view a dashboard or run a query. Admins who only manage billing are often excluded because they never touch the product's value.
Consumer platforms such as social networks report MAU in public filings, and each one publishes its own definition in the footnotes.
How to define and track MAU well
- Write the definition down. State the identifier, the qualifying events, and the window in one shared document. Change it only with a dated note.
- Choose a value action, not a login. Ask what a user must do to get value, and count that. Logins overcount; deep feature use undercounts. Pick the middle.
- Exclude internal and automated traffic. Filter employee accounts, test accounts, bots, and API service users. These inflate the base and hide real trends.
- Segment the number. MAU by plan, by company size, by signup cohort, and by platform tells you far more than one total. User segmentation is what turns MAU from a vanity metric into a working one.
- Pair MAU with retention. Rising MAU can hide rising churn if acquisition is strong. Track user retention by cohort alongside it, as described in this guide to user retention rate.
- Close the loop with communication. When a segment goes quiet, tell them what changed. Product updates delivered in-app or by email are the cheapest lever for waking dormant users, and the impact shows up in next month's count. See how to increase SaaS user engagement for tactics.
MAU vs DAU, WAU and common mistakes
DAU and WAU apply the same logic to a day or a week. The ratio of DAU to MAU is often called stickiness: it estimates how many days per month a typical active user shows up. A month-end reporting tool can have a low ratio and still be healthy.
Retention asks whether the same people come back. MAU can rise while retention falls, as long as new users arrive faster than old ones leave. The two metrics answer different questions and should be reported side by side. This list of customer engagement metrics shows where each one fits.
Feature adoption is usually expressed as a share of MAU. If the base is inflated, feature adoption looks worse than it is.
Common mistakes:
- Counting seats or licenses instead of people who used the product.
- Comparing your MAU with a competitor's without checking that the definitions match.
- Changing the qualifying event mid-year and reporting the series as if nothing changed.
- Treating a single month's dip as a trend. Holidays, fiscal year ends, and outages all cause one-off drops.
- Reporting MAU without cohort context, which hides whether growth comes from new signups or from returning users.
How AnnounceKit handles monthly active users (MAU)
AnnounceKit does not replace your MAU dashboard. It works on the other side of the number: getting quiet users back and keeping active ones informed.
- A changelog page on your own domain and 10+ in-app widget display modes put product updates where users already are.
- Segmentation lets you target a release note or survey at a specific group, such as accounts whose usage dropped last month.
- Email digests and Slack delivery reach users who have not opened the product recently.
- NPS surveys and feature requests with voting and Jira sync capture why users disengage, not just that they did.
- AI post generation and an official MCP server for AI agents keep the publishing cadence up without extra headcount.
Pricing is flat per project from $79/month and is not tied to your MAU, so a growing user base does not raise the bill. A 15-day free trial is available.