Active Users in a nutshell

The most commonly used metric to capture the performance of an app, product, service or a tech service business is active user. In this article, we cover some of the basics of how to define & compute active users. Active users can be used to derive a number of metrics that provide richer insights into user behavior and its impact on business performance.

Definition

Active Users are the users performing certain intentional actions for a certain duration.

Intentional action — Actions performed by the user that leverages a feature of the application or platform. Excludes Login, software update and opening/closing the application/device. Active users can be categorized into the following groups,

• New users
• Retained users
• Resurrected users
• Churned users

New users

A user who has never used the product in the past and is active in the current time period (t). Conventionally, the trend of new users is studied on a monthly basis.

Retained users

A user who has been actively using the product in the current and previous time period.

Returning users / Resurrected users

A user who was actively using the product in any time frame prior to previous time period (0 — (t-2)), inactive in the previous (t-1) time period and active in the current (t) time period

Churned users

A user who was active in the previous (t-1) time period but inactive in the current (t) time period

User Categorization Chart illustrating categorization of users based on their activity in a month

Cohort

The time period used to group the users. Common cohorts — yearly, Quarter, Monthly, Weekly and daily level of granularity with res

Monthly Active users calculated from the User Categorization Chart

Usage

Active users metric is used to track progress of the business relative to the previous time period.

The metric is influenced by

Seasonality — Things that keep repeating on a regular cadence. For eg: Thanksgiving, Christmas, Weekday/Weekends

Daily Active User trend for Wikipedia Android App showing Weekend spikes. Mobile use of Wikipedia is expected to spike on weekends

Chart explaining daily active users for an e-commerce website. The dip seen during weekends that recur every week and the spikes during Nov-Dec are examples of seasonal components.

For eg: Stock apps have higher active users during weekdays than weekends because the stock market is open on weekdays. News app have no change in active users between weekday and weekends

Cyclical effect — Any impact on active users driven by the product life cycle. Cyclical effects are associated with seasonality. The product life cycle impacts the sale and usage of a product. For example: A release of a new IPhone results in the slowdown of sales of existing models thereby impacting the active users of the existing product.

Derived Metrics

Stickiness Ratio

Average of DAU for a month/ MAU for the same month

Average of WAU for a month/ MAU for the same month

This determines the frequency of the use. The value of the ratio is always between 0 and 1. The higher the value of the ratio, the more frequent the users are active in the product.

The ratio reflects if the majority of active users are frequently using the application. The metric helps the business reflect on the baseline usage of the application and identify the most active users on the platform.

The above chart shows the stickiness ratio (Avg. WAU in a month divided by MAU). The growth in stickiness ratio indicates users are using the product more frequently over time. In addition, the dip is a reflection of seasonal users.

Comparing a Banking application, news app, movie streaming application, investing applications and social media

The frequency of the use of a banking application by one user will likely be lower that other applications. So, the Average DAU will be low compared to MAU. In this case, the baseline usage pattern is towards the lower end of the spectrum.

News application, investing applications and social media applications tend to have users who frequently visit the application to consume information. So the average DAU : MAU will be higher compared to other applications.

Gradient — First Derivative

The metric is a comparison of active users in the current period to the active users in the previous period. Defined as the % change in users compared to the prior time period (Week, Month, Quarter, Year). Typically referred to as Week over week aka WOW (%), Month over month aka MOM (%), Quarter over Quarter aka QOQ(%), Year over Year aka YOY (%). Further insights can be derived by comparing the gradient of WOW across Months, MOM or QOQ across years.

The metric captures

1. Rate of growth — the rate at which the active users are growing/shrinking on a periodic basis. For E.g: A positive WoW metric over time indicates growth active users, A negative WoW metric over time indicates drop in active users.

The above chart shows the percentage change in Weekly Active Users over time. WoW for most of the weeks more than 0% indicates growth in weekly active users over time. The peaks and troughs seen in the holiday season seen in the chart indicate the impact due to holiday.

2. Seasonal trends — changes in user behavior due to recurring events (e.g.: weekday, weekends, holidays etc.). These recurring events cause a jump or drop in user activity. Seasonal trends are typically identified by looking at the rate of growth over the previous time periods. For Example, Comparing MOM trends across two years i.e. comparing MOM2020 (MAU[November, 2020] / MAU[December, 2020]) against MOM2021(MAU[November, 2021] / MAU[December, 2021]) captures the user trends during the USA holiday season. For an e-commerce website, there is an increase in monthly users during the holiday season. To derive the actual user growth, one has to look at MOM over years. An increase/decrease in MOM over years during the USA holiday season reflects the actual growth/shrinkage in users.

3. Business cycle — the impact driven by a product life cycle. Similar to seasonal trends, business cycles create a point in time impact on user activity. The point in time impact is reflected in WOW or MOM as a spike or a drop. The impact of the business cycle vs the growth of users can be captured by comparing the MOM of current years against the previous year. In other words, comparing the rate of growth of the current period over the previous period. Example of impact from business cycle: Active users on v1 of a product decline when v2 launches and users migrate.

This is the first of a series of articles we plan to write on Key Performance Indicators (KPIs). Thanks for being here. Leave your comments and feedback below.