DAU, MAU, Retention, ARPU: Game KPIs Explained

Every game studio tracks numbers. Far fewer agree on what those numbers mean. Someone says retention is fine, someone else says the game is leaking players, and both are looking at the same dashboard. The problem is rarely the data. It is that four core metrics get thrown around loosely, mixed up, and quoted without context.

This post breaks down the four you will hear in almost every meeting: DAU, MAU, retention, and ARPU. What each one measures, how it lies to you, and how to read them together instead of in isolation.

DAU and MAU: how many people actually show up

DAU is Daily Active Users. The count of unique players who opened your game on a given day. MAU is the same idea over a rolling 30 day window. Two players who log in five times each still count as two, not ten. That is the point. These metrics count people, not sessions.

On their own they tell you scale and not much else. A game with 50,000 DAU sounds healthy until you learn it had 500,000 last month. The raw number is a vanity trap. What matters is the trend and the ratio between the two.

That ratio has a name: the stickiness ratio, DAU divided by MAU. It tells you how often your monthly players come back within a month.

• A stickiness of 0.5 means the average monthly player shows up roughly 15 days out of 30. That is a daily habit game. Think competitive multiplayer or anything with daily rewards.
• A stickiness of 0.1 means they show up about 3 days a month. That can still be fine for a session based or narrative game people binge and put down.

There is no universal good number. A puzzle game and a live service shooter should not have the same stickiness, and forcing one to look like the other usually means bolting on daily login mechanics that annoy the people who already liked the game.

Retention: the metric that predicts everything else

If you only get to keep one metric, keep retention. DAU and MAU are outcomes. Retention is the cause.

Retention measures the percentage of players who come back N days after they first installed. The three checkpoints almost everyone uses:

• Day 1 retention. Of the players who installed yesterday, how many came back today. This is mostly a measure of first impression and onboarding.
• Day 7 retention. Did the game survive the first week. This is where the core loop either hooked people or did not.
• Day 30 retention. Long term fit. The players still here after a month are your real audience.

Rough benchmarks for mobile games, knowing they swing hard by genre:

Two warnings about these numbers.

First, always read retention as a curve, not three disconnected dots. A game with great Day 1 and terrible Day 7 has an onboarding that oversells and a core loop that underdelivers. A game with mediocre Day 1 but a flat curve after Day 7 has found a small loyal audience. The shape tells the story, not any single value.

Second, retention is cohort based or it is meaningless. “Day 7 retention is 12%” only makes sense for a specific group of players who installed in the same window. Blend cohorts together and a marketing spike of low quality installs will tank a number that has nothing to do with your actual game quality.

ARPU: turning players into revenue

ARPU is Average Revenue Per User. Total revenue over a period divided by the number of active users in that period. It answers one question: on average, what is a player worth to you.

The catch is the word average. In most free to play games, the vast majority of players spend nothing. Revenue comes from a thin slice of payers, sometimes a fraction of a percent. So ARPU is really two hidden numbers smashed together: how many people pay at all, and how much the payers spend.

That is why serious studios also watch ARPPU, Average Revenue Per Paying User. The gap between ARPU and ARPPU tells you which lever to pull.

• Low ARPU, high ARPPU: your payers spend well, but almost nobody converts to paying. Fix the conversion step, the first purchase, the entry price point.
• Low ARPU, low ARPPU: people convert but spend little. Your monetization depth is shallow, there is nothing compelling to spend on past the first few dollars.

Chasing ARPU without splitting it is how teams end up shoving aggressive offers at everyone and burning retention to juice a quarter.

Why no single metric is safe alone

Each of these four can be gamed, and optimizing one in isolation usually breaks another.

Push daily login rewards and DAU climbs while engagement quality drops. Run a cheap user acquisition campaign and MAU jumps while retention craters because the new installs were never a fit. Hammer players with purchase prompts and ARPU ticks up for a month while Day 30 retention quietly bleeds out.

The metrics only mean something in relation to each other. Rising DAU with flat retention means you are buying growth, not earning it. Strong retention with low ARPU means you have an audience that loves the game and a monetization model that is leaving money on the table. Flat MAU with rising ARPU might mean a shrinking but increasingly loyal and spending core, which can be healthy or a slow death depending on whether new players still arrive.

Read them as a system. One number is an anecdote. The relationships between them are the actual diagnosis.

Where to start

If you are setting this up from scratch, do it in this order.

1. Get clean cohort based retention first. It is the hardest to fake and the most predictive.
2. Add DAU, MAU, and the stickiness ratio to watch the shape of your active base over time.
3. Split monetization into ARPU and ARPPU so you can see conversion and depth separately.
4. Put all four on one screen, filtered by the dimensions that matter for your game, version, country, platform, game mode, so a change in one is visible against the others.

That last point is where most off the shelf analytics tools fall down. They give you the metric but not the context, and a KPI without the ability to slice it is just a bigger number to argue about. The goal is not a prettier dashboard. It is being able to answer “did that patch help or hurt” before the next one ships.