EDUCATION CENTER
Other products
Get to know devtodev`s other products: an analytics platform for game developers, an analytics consulting, and outsourced customer support in games.
Main Metrics. Active Users (DAU, WAU, MAU)
Let’s figure out how we can calculate active users.
Published
11.11.2019
|
devtodev

Let’s figure out how we can calculate active users, what indicators we should use to segment them, and what statistical anomalies exist when you assess the audience. 

A project audience is filled up with new users every day. Some of them lose their interest very quickly, one or another sometimes come back to an app, and others use it regularly. And the representatives of all these segments definitely open an app daily. Today we will talk about them - Active users.

Active users are those who had at least one session during a tested time period. These intervals can be different, but mostly the following three groups are tested: daily, weekly and monthly audiences. And they have well-established names:

  • DAU - the number of unique users per day (Daily Active Users);
  • WAU - the number of unique users per week (Weekly Active Users);
  • MAU - the number of unique users per month (Monthly Active Users).

At the same time, it is possible to make similar calculations for any other periods, if they meet KPIs of a company. For example, looking back on the year, we can count a yearly project audience and compare it with previous years to evaluate the dynamics.

It is worth paying attention that WAU for a certain week is not the sum of DAU for 7 days, since we are talking about unique users. For example, one of them can open an app on Monday and Tuesday, and they will get into DAU of Monday and DAU of Tuesday. But within a week (from Monday till Sunday) they will be counted only once.

The same goes for MAU. It is not the sum of 4 WAUs and 30 DAUs. From the point of view of calculation, these indicators are not interconnected and are calculated separately.

To understand these indicators better, let's calculate them using an example.

Let’s suppose we have information about various users’ app sessions for 2 weeks. No matter how many times a day a user logs into a project, this user will still be one unique visitor.

The days when users opened an app are marked with blue:

devtodev active users

So, let’s calculate the DAU for the 1st, 2nd, 5th and 10th days first. To do this, we need to know how many unique users opened an app during these days:

  • DAU of the 1st day = 2 (users 1 and 4);
  • DAU of the 2nd day = 3 (users 2,4,5);
  • DAU of the 3rd day = 3 (users 2,3,4);
  • DAU of the 10th day = 0 (no one opened an app during these days).

And now let’s calculate WAU:

  • It is 5 during the first week (from day 1 to day 7) – all the users opened an app;
  • in the second week (from day 8 to day14) this indicator is already 3 – the first and second users did not have any sessions.

We can choose any random week as well. For example, if we take a period from day 3 to day 9, WAU will be equal to 4.

In the same way, we can count Active Users for a month, for half a year, and for other time intervals.

Only 5 were represented in our example, and in a real project there will be thousands, hundreds of thousands, millions of users who open an app daily. And the way they open the app indicates its stability, quality and scale.

In addition, it is more reasonable to track Active users indicator in real time. If something goes wrong inside an app or on a server and users aren’t be able to use a product, it will immediately affect this metric. To keep it under control, we can group users not by days, but by hours or even 10-minute intervals.

By the way, active users who are currently inside an app represent a separate metric that has its own name. Most often they are called Users online, but we can also find such abbreviations as CCU (concurrent users) - users who are in an app at a certain moment, and PCCU (peak concurrent users) - the maximum number of users which are simultaneously inside an app.

The average CCU reflects the scale of a project well, and the PCCU is very important when planning server load.

Active users dynamics can change not only within a day, it can also gradually rise or fall month by month. It is rather important to control it. Segmentation helps to simplify the analysis of changes in the number of Active users. Thanks to it, you can quickly understand due to which segment of users the indicator is changing.

Here are some variants of active audience segmentation.

By payments:

  • paying / not paying
  • made only 1 payment / made repeated payments

By the date of installation:

  • 1 day / 2-7 days / 8-14 days / 15-30 days / 30- 60 days / 60+ days

By the frequency of visits: 

  • Every day / 4-6 times a week / 1-2 times a week / once a month and less

We can also divide them by countries, devices, operating systems, or by a custom event (which means that the audience is divided into users who performed or did not perform one or another action).

The last segmentation option can be used if an app has any key event that is important for completing a game experience or for creating the right first impression of a product (for example, completing a tutorial, N game levels or visiting a store).

It will be much easier to look for a possible cause of the problem, when you identify the segment where a decrease of active users occurs.

This is what may happen:

devtodev active users

First, the number of active users in Russia begins to decrease, while the number of visitors from Japan increases, which compensates for the fall in another country. If we look only at a general DAU chart, we are unlikely to notice any changes in dynamics. Only when the number of active users in Russia falls even more, we’ll be able to see it on the general chart. Meanwhile, quite a lot of time may already pass, which could be used to find and eliminate the cause of this fall.

Simpson’s paradox is one more statistical anomaly which confirms the importance of segmentation. It is better viewed with an example.

Let’s take 4 countries from the previous example and assume that they have the following paying conversion:

Country New Users Paying Users Conversion
Russia 3300 160 4.85%
Japan 2250 100 4.44%
United Kindom 2400 170 7.08%
China 8600 600 6.98%

The next step is to calculate the total conversion of European and Asian countries:

Region New Users Paying Users Conversion
Europe 5700 330 5.8%
Asia 10850 700 6.5%

Ultimately we get this:

  • The conversion in Russia (4.85%) is higher than the conversion in Japan (4.44%);
  • The conversion in the UK (7.08%) is higher than the conversion in China (6.98%);
  • The overall conversion in European countries (5.8%) is lower than the conversion in Asian countries (6.5%).

All this once again suggests that segmentation may give completely different from general indicator statistic results.

By the way, sometimes, it impossible to determine this tendency clearly by looking at the DAU chart. But when you make groups by weeks or months (converting the chart to WAU and MAU) it becomes more evident.

active users devtodev

Of course, Active users metric is important for the project itself. But in addition, it is also associated with other financial and behavioral metrics.

First of all, the number of new users affects Active users. The more there are new users, and the faster and more stable they come into a project, the faster grows an audience.

The second equally important indicator is Retention (user retention). It shows how users return to a project. If you bring there the new users who will not come back later, then they will not replenish the audience, and such acquisition won’t be effective at all. It is important to engage users in a product so that they will wish to return. And the more the number of this users is, the larger will be an active audience.

Here is a small example:

User Type Variant 1 Variant 2 Variant 3
New users 5000 1000 5000
30 Day Retention 10% 10% 3%
How many users left 500 100 150

We can have good in-app Retention scorecard, but if there is a small number of new users the audience will grow very slowly. And vice versa, if there is a good flow of new users and low Retention, then most of them will leave a project. It also won’t increase an audience.

The larger a project audience is, the more there are potential payers among it. Exactly in this order users become paying:

New users → Active users → Paying users

Besides it is important that a user remains active in a product after making the first payment, because this will increase the chances that he will make repeated purchases.

Thus, Active users affect revenue in direct ratio:

Revenue = Active users * Paying share * ARPPU

The number of active users is one of the most important indicators of a product, which is indirectly responsible for its success. It combines the quality of new users acquisition campaign and retention metrics, directly affecting revenue. Therefore, when analyzing active users, we should also pay attention to the rate of audience growth, because this metric is one of the most positive signs of active product development.

Read more